Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Hall of Merit > Discussion
Hall of Merit
— A Look at Baseball's All-Time Best

Tuesday, January 23, 2018

2019 Hall of Merit Ballot Discussion

2019 - (December 2018) - elect 3

Top 10 Returning Players

Luis Tiant (240), Sammy Sosa (238), Kenny Lofton (236), Andruw Jones (220), Jeff Kent (207), Ben Taylor (197), Johan Santana (186), Buddy Bell (139), Bobby Bonds (124), Jorge Posada (105)

Newly eligible players

Player Name	bWAR 	WS	WAR7	JAWS	HOFm	HOFs
Roy Halladay	64.7	225.5	50.6	57.6	127	45
Todd Helton	61.2	316.5	46.4	53.8	175	59
Andy Pettitte	60.8	228.7	34.1	47.5	128	44
Mariano Rivera	57.1	272.5	28.9	43	214	30
Lance Berkman	51.7	310.7	38.9	45.3	98	44
Roy Oswalt	50.2	175.3	40.1	45.1	59	34
Miguel Tejada	46.9	278.6	36.5	41.7	149	44
Placido Polanco	41.3	215.4	32.2	36.8	42	26
Freddy Garcia	35.7	136.4	28.3	32	38	23
Derek Lowe	34.5	175.6	28.4	31.4	51	19
Kevin Youkilis	32.7	144.3	31.2	31.9	29	23
Vernon Wells	28.7	186.6	26.2	27.4	52	19
Ted Lilly	27	114.3	24.8	25.9	12	16
Travis Hafner	24.8	142.5	24.6	24.7	31	19
Jason Bay	24.3	162.5	24.5	24.4	47	21
Michael Young	24.2	231.2	21.1	22.7	112	36
Darren Oliver	22.6	119.3	17	19.8	20	9
Jon Garland	22.4	117.5	19.5	21	17	9
Ramon Hernandez	21.6	156.7	18.7	20.2	43	26
Ryan Dempster	19.3	133.7	23.8	21.5	26	12
Juan Pierre	16.9	178.2	16.4	16.7	63	23
Octavio Dotel	15.4	95.5	14	14.7	25	13
Jake Westbrook	13.3	78.4	14.6	13.9	14	3
Jose Contreras	13.2	67.8	13.3	13.3	17	7
DL from MN Posted: January 23, 2018 at 12:35 PM | 534 comment(s) Login to Bookmark
  Related News:

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

Page 4 of 6 pages ‹ First  < 2 3 4 5 6 > 
   301. Howie Menckel Posted: December 11, 2018 at 11:10 AM (#5796580)
bump
   302. DL from MN Posted: December 11, 2018 at 11:46 AM (#5796605)
I plan on ending the 1949 MMP on 1/2/2019.
   303. progrockfan Posted: December 11, 2018 at 02:37 PM (#5796728)
Regarding Mariano Rivera (who I think will sail into the HOM), is WPA for closers a bit problematic due to the selection-bias of when they are utilized? I am not sure.

WPA does indeed have a selection bias towards high-leverage situations, which makes it ideal, in my view, for evaluating the value of relievers in comparison to starters. WAR is a whole-value measure, WPA is a success-in-clutch-situations measure. Both have value - WAR much more so generally, in my view, but WPA more so for assigning a realistic value to relievers with their low innnings counts but frequent appearances with the game on the line. That's my 2¢ anyway.
   304. Howie Menckel Posted: December 11, 2018 at 02:46 PM (#5796736)
frequent appearances with the game on the line.

starting an inning and only needing to get 3 outs before 2 or 3 runs score seems to make "game on the line" do some heavy lifting there.

maybe someone can list Rivera's regular-season save percentage compared to other closers in the last 25 years? that would seem relevant to "game on the line."
   305. progrockfan Posted: December 11, 2018 at 04:40 PM (#5796832)
starting an inning and only needing to get 3 outs before 2 or 3 runs score seems to make "game on the line" do some heavy lifting there.

Causal oversimplification alert. The scenario you describe is not in fact a situation with the "game on the line", and therefore not a situation that would register heavily with WPA. Rivera registers very heavily with WPA because a) he had decent career volume, and b) he was used often in situations with the game actually on the line.

maybe someone can list Rivera's regular-season save percentage compared to other closers in the last 25 years? that would seem relevant to "game on the line."

Now, see, I don't agree with that, because I think the definition of 'save' used by MLB is so broad that it renders the stat functionally useless, except as a crude indicator of career volume.

But you're correct, I think, in saying there's probably a methodology (though it would probably require digging through many thousands of box scores), the purport of which would be, "How did (closer X) perform with the game actually on the line?"
   306. Rob_Wood Posted: December 11, 2018 at 05:07 PM (#5796859)

3. Regarding Phil Rizzuto, I am late to this parade but I am coming around to his case. His case seems to rest on receiving proper "credit" for his war-years. I am one who gives fairly generous credit for a player's military service. So, to me, Rizzuto may come down to how we treat his malaria-affected 1946 season in the calculation of his war-time credit. Leaving 1946 to the side for the moment I wonder how robust the typical way of deriving war-time credit really is.

Could someone undertake the following mini-study (I have no idea how difficult this would be)?
- Select all major leaguers who had 4+ WAR in each of their age-23, age-24, and age-29 seasons and played more than 75 games, say, in each season from age-23 through age-29 (maybe relax this to 3.5 instead of 4.0?)
- Take the average of these three seasons WAR and call it X
- Look at their WAR for their age-25, age-26, age-27, and age-28 seasons
- Take the average of these four seasons WAR and call it Y
- Calculate ratio of Y/X and call it Z
- Find distribution (mean, median, percentiles, etc.) of Z.

Thanks mucho!


Quick & Dirty attempt.

I have the seasonal WAR values for around 600 of the greatest position players of all-time on my PC as part of my CPASR framework. Missing seasons are omitted (i.e., skipped over). Also, I do not have the players' ages and am too lazy to either look them up or download them in some way. So I did the next best thing.

I selected any player having a 7-year sequence in which seasons 1, 2, and 7 were each at least 4.0. If there was more than one such sequence, I chose the first sequence in his career. I then calculated the ratio of the Avg WAR for seasons (3,4,5,6) to the Avg WAR for seasons (1,2,7). Of course, the idea is to investigate the WAR "profile" of players not missing seasons due to military service and see how often they exceed the average of their endpoints.

Here is the distribution of ratios among the 212 players who qualify. Remember a ratio above 1.0 indicates that the player's "missing" (interim) seasons have a higher average than his endpoint seasons.

0.0 0.5  14
0.5 
0.6  12
0.6 
0.7  29
0.7 
0.8  27
0.8 
0.9  35
0.9 
1.0  35
1.0 
1.1  26
1.1 
1.2  20
1.2 
1.3   6
1.3 
1.4   4
1.4 
1.5   4
above 1.5   0 


Let's remove the 14 players who cratered in the middle. That leaves us with a sub-sample of 198 players. Their average ratio is 0.90 and the median ratio is 0.89. 138 had ratios below 1.0 and 60 had ratios above 1.0.

Admittedly this mini-analysis was very Quick & Dirty. But it probably is useful for our purposes. My preliminary conclusion is that we may want to be conservative in awarding credit for missing seasons in military (maybe around 90% of the player's endpoint performance) since down seasons happen to virtually every player, having nothing to do with military service. So taking a straight average may be too optimistic.

On the other hand, many players who returned from military service found it difficult, for good and obvious reasons, to step right back into the major leagues and perform at their previous levels. Thus one end-point may be skewed downward a bit. Phil Rizzuto and his malaria is a prime example, but virtually all returning players struggled when they first came back.
   307. Kiko Sakata Posted: December 11, 2018 at 05:15 PM (#5796864)
Now, see, I don't agree with that, because I think the definition of 'save' used by MLB is so broad that it renders the stat functionally useless, except as a crude indicator of career volume.

But you're correct, I think, in saying there's probably a methodology (though it would probably require digging through many thousands of box scores), the purport of which would be, "How did (closer X) perform with the game actually on the line?"


That's almost precisely what WPA is designed to do. The end probability for the team in a successful save is 100%. The difficulty of getting there is baked into the team's win probability when the closer enters the game. If the team had a 90% chance of winning the game and they do so, that's +.1 WPA for the closer. If the team had only an 80% chance of winning the game and they do so, that's a +.2 WPA. If the team had an 85% chance of winning the game when the closer entered the game and he leaves the game with the score tied and the winning run in scoring position, that'll score as something like a -.55 WPA for the closer. Add them up and there you are.

If you want to convert that into a record of wins and losses that are comparable to records for starting pitchers and position players that splits credit between pitchers and the fielders behind them, you want my Player won-lost records. Where, basically, Mariano Rivera is a worthy probably mid-tier Hall-of-Meriter and no other relievers really are (Trevor Hoffman and Goose Gossage are probably #2 and #3, with the order debatable). (Here's an article that looks at the top relief pitchers in the stat)
   308. progrockfan Posted: December 11, 2018 at 05:29 PM (#5796873)
@Kiko Sakata:

I think what @Howie Menckel is looking for is a percentage tool, i.e., 'Closer X converted Y of Z possible opportunities with the game on the line, for an AA% game-saving percentage.' Personally I think WPA performs that job admirably - but it's easy to see why a voter might like to compare closers on a percentage basis - particularly as Howie's original question was, "[Rivera's] 35-for-38 - is it really any better than the other guy's 35-for-38, just because Rivera left fewer of his own runners on base?"

I'm paraphrasing you, Howie, and please don't hesitate to correct me if I get it wrong - but I believe the analytical heart of the question you're asking is, Did Rivera convert a higher percentage of leveraged situations than his peers? I'm not aware of an existing stat that answers precisely that question - and it's a fair question too, in my opinion.

(Great site, by the way, Kiko, and one I'll be exploring as opportunity allows.)
   309. bachslunch Posted: December 11, 2018 at 06:43 PM (#5796892)
Carl, thanks for your feedback on the ballot thread. It’s not going to change my already submitted ballot, but appreciated anyway. I realize my thinking is not to everyone’s taste, but it is what it is.

Any chance non-ballot comments can be tramsferred over here? Am thinking it will make things much easier for the folks tallying things up. Just a thought, FWIW.
   310. Bleed the Freak Posted: December 11, 2018 at 07:49 PM (#5796912)
I don't really see them as all that close. Rizzuto has a bigger peak, more career, and he lost what are more typically better years (age 25-27 vs. age 24-26) to the war.


From the Hall of Miller and Eric project, the CHEWS+/MAPES+ scores were:
Rizzuto - 78/76
Pesky - 72/70

Pesky closes some gap if we credit him for "clutch" situational value from Baseball-Reference.
Johnny is at 2.7 + wins and Rizzuto at -2.8 wins.

The war years were sandwiched between Pesky's best work, while with Rizzuto it was less peak work, so you could argue to give Johnny peakier war credit.

I understand having Rizzuto ahead of Pesky, but I don't see a gulf between the two bodies of work, where Rizzuto is a clear HOMer balloting or top half of the ballot, while Pesky is mostly an after thought.
   311. Howie Menckel Posted: December 11, 2018 at 09:44 PM (#5796948)
thanks for your responses, progrockfan - and to an earlier point of your concern about tone, I think you have a keen sense of it.

and I say this as the one who is one of those furthest from you on Rivera (well, not on him being on my ballot, but on the true value of his usage in the regular season. spoiler alert, I am well aware of his DIFFERENT usage in the postseason, which I see as exponentially better choices - and of course the performance, to boot - and thus the massive boost to his career value).
   312. Kiko Sakata Posted: December 11, 2018 at 09:45 PM (#5796949)
Okay, here's my preliminary ballot. Barring somebody convincing me otherwise, I'll move this to the Ballot thread some time in the next week (probably late Sunday?).

The basis of my system is my own statistic, Baseball Player won-lost records. These are calculated based on Retrosheet play-by-play data, so I only have them for the (white) major leagues from 1921 onward and I am missing some games (which I estimate based on games I do have) for seasons from 1921 - 1936. My weighting system ends up favoring peak, but I'm kind of open-minded in terms of Hall-of-Fame case. FYI, if you want to try to put together a ballot based on Player won-lost records, you can customize your weights here.

I've discussed various things through this thread, so I'm just going to post a ballot with required disclosures here.

1. Roy Halladay
2. Mariano Rivera - I can see an argument for either order or these two. I tend to like starting pitchers more than relievers, but Mariano was so, so, so good.

3. Cannonball Dick Redding - he makes my ballot on the strength of Dr. Chaleeko's recent MLE's. Redding probably has a case for #1 or #2. I'm going #3 on the basis of the greater uncertainty of his record.

4. Tommy John - anybody who's read anything I've written in the Hall of Merit over the past however many years I've voted (4 now?) knows that my system LOVES Tommy John. My personal Hall of Merit would have quite a few more pitchers than the actual HOM, which is reflected here - although Halladay and Rivera are new candidates, of course. Below here, I tried to tamp down the pitcher-heaviness of my ballot.

5. Ben Taylor - it looks to me like the HOM just missed him when he was first eligible. He looks to me like he's probably the best first baseman of the first two decades of the 20th century.
6. Wally Schang - ditto, except he's a catcher (and white, not that that matters).

7. Tommy Henrich - with WWII credit

8. Dwight Gooden - as I said above, my system likes peaks. Gooden was the best pitcher in baseball his first two seasons and hung around long enough to accumulate some career value.

9. Johnny Evers - this is somewhat speculative. I talked about him back on I think the first page of this thread.

10. Jorge Posada - I love the hitting from a catcher; I'm highly skeptical of modern "pitch-framing" numbers.

11. Andy Pettitte - he's something of a poor man's Tommy John. My system is very appreciative of long-career #2 starter types.

12. Johan Santana - see the Gooden comment. Longer peak than Gooden but shorter career.

13. Vern Stephens - World War II is obviously an issue with his 1944 and 1945 seasons, but he was actually even better in 1949.

14. Jeff Kent - my system loves guys who can hit at fielding positions.

15. Luis Tiant - he's in my pHOM and is deserving here. But by the voting rules of the HOM, if I think he's only the 15th-best guy eligible, I'm supposed to list him 15th.

First 11 out:

16. Don Newcombe - I may be under-selling him. I'm not sure how many extra-credit seasons to give him. He also slips just off-ballot due to a concern that my ballot was too pitcher-heavy.

17. Urban Shocker - his peak mostly pre-dates my system. And, again, I didn't necessarily want to put 9 or 10 pitchers on my 15-man ballot.

18. Orel Hershiser - similar case to Gooden and Santana

19. Darryl Strawberry - my system really likes his peak; low-scoring environment masks his raw stats a little, but he was a big part of a lot of wins by the Mets in the mid/late 1980s.

20. Dave Concepcion - there are three 1970's shortstops who are in my pHOM: Concepcion, Bert Campaneris, and Toby Harrah. I go back and forth on how to rank them against each other.
21. Bert Campaneris
22. Toby Harrah - I'm not necessarily being super-strict about the off-ballot rankings by this point. I tweaked things to group these three.

23. Dizzy Dean - I'm missing some games from his peak, so this is somewhat speculative. And the very short career is a concern.

24. Andruw Jones - in my pHOM (although not necessarily yet, because of the number of strong debut candidates in the last couple of years).

25. Gil Hodges
26. Lance Berkman - my system likes Hodges a little better than Berkman in part because first base was a weaker position in Hodges' day. I can't decide if that's fair to Berkman or not, but since they're both (just) off-ballot, I don't know that I need to decide that just yet anyway.

Other required disclosures:

Sammy Sosa - probably in my top 75 or so. Decent candidate. He looks much better in eWins (which are context-neutral) than pWins (which tie to team wins).

Kenny Lofton - he might sneak into my top 100. Baseball-Reference (and Fangraphs) overweight fielding in their WAR calculations. So, I tend to rate players with large fielding components to their WAR somewhat lower.

Buddy Bell - same story as Kenny Lofton but far more so. Honestly, he's not THAT much better than Harold Baines in my system - Baines was a better hitter, but Bell makes that up by playing third base and playing it very well. I think BB-Ref also has problems with their positional adjustments for 3B and SS in the 1970s (see my #'s 20-22 above).

Bobby Bonds - a step below Lofton (who's a step below Sosa, who's at least a step below Andruw Jones). My system doesn't like his fielding as much as BB-Ref.

Other debuts of note:

Roy Oswalt is probably in my top 30 or so.
Todd Helton ranks somewhere in the 100-125 range for me (between Lofton and Bonds). When you park-adjust his numbers, he just doesn't really distinguish himself from a bunch of other first basemen from this era. John Olerud, for example, shows up just ahead of Helton in my weighting system.

Other guys I considered (for whom I'm missing some data, so could be mis-valuing) (in alphabetical order): Dave Bancroft, Bus Clarkson, Kiki Cuyler, Burleigh Grimes, Dolf Luque, Herb Pennock

I think that's everybody, but please feel free to ask about anybody I might have forgotten.

Oh, since he's being talked about here - Phil Rizzuto. Rizzuto's big problem in my system is that he was much worse when he got back from the war. A player with Phil Rizzuto's 1940-41 and 1949-53 could absolutely make my ballot (although near the bottom). But given that he actually played in MLB in 1946, 1947, and 1948, I think it's too speculative to give him any bonus credit for those seasons and without that he's maybe in my top 50 at best, perhaps a bit lower (in the same general range as Sammy Sosa, probably). Johnny Pesky probably ends up in a similar position, maybe a tick below because his career was a bit shorter than Rizzuto's.
   313. Jaack Posted: December 12, 2018 at 12:26 AM (#5796970)
Before I comment on anyone else’s stuff, I wanted to get my current ballot out there for commentary before I make it official:

1. Roy Halladay
2. Mariano Rivera
I’m pretty confident in this order. I can see the case for Rivera on top, but it takes a generous interpretation of his career to make him better than Halladay in my system.
3. Lance Berkman
Based on everyone else's opinion, I'd better get used to him being at the top of my ballot.
4. Tommy John
5. Babe Adams
Adams moves up few slots - I'm pretty convinced that rWAR is underrating him (along with pitchers in general) and if anyone deserves minor league credit, it's him - if he played in the modern game, he never would have spent a year and a half in the middle of his career dominating in the minors.
6. Kenny Lofton
7. Jeff Kent
8. Mickey Lolich
9. Dick Redding
I was pretty low on him, but the new MLEs from Dr. Chaleeko are overwhelming. I'm being very conservative in placing him 9th, but I think he's going to make it this year anyway.
10. Bob Johnson
I phased out all WWII debiting - philosophically it didn't seem right to me to mark a guy down for playing in what was still the highest level league in the world and practically, Johnson was the only guy anywhere near my ballot it affected in any significant way.
11. Kiki Cuyler
12. Bert Campaneris
13. Ben Taylor
14. Roy Oswalt
15. Robin Ventura
This last slot I'm very open to changing - I'm pretty comfortable with the first 14, but this slot mostly feels like a best of the rest - there's not a whole lot of daylight between Ventura and my number 30 player.

16-30: Todd Helton, Trevor Hoffman, Bobby Bonds, Andy Pettitte, Jim Kaat, Dwight Gooden, Billy Wagner, Dolph Camilli, Joe Tinker, Hugh Duffy, Jim Sundberg, Willie Davis, Thurmon Munson, Luis Tiant, Hack Wilson

   314. Jaack Posted: December 12, 2018 at 12:58 AM (#5796976)
On other subjects:

As far as relievers go - I think that for relievers in particular, have a lengthy, productive career is more important than for any other group of player. While there might not be a huge difference between a good year from Trevor Hoffman and a good year from JJ Putz, I think a reliever who can be relied on year in and year out to be a shutdown option out of the bullpen is a valuable asset indeed. Reliever attrition is so high, that avoiding that downfall seems to me to be a major asset for Rivera or Hoffman compared to a couple years of brilliance from Bruce Sutter or John Hiller.

On Dolf Luque... he'd currently be 177th on my ballot if it went down that far. He seems to me to be a late bloomer more than anything. If you re-arrange the years and interpret him generously he kind of looks a bit like an 1920s Dwight Gooden, who is near ballot territory for me. But I'm not that generous and he just has to few great years - 1923 and 1925 are the only ones that look like real standout seasons to me.

On Phil Rizzuto, I currently rank him 134th. He may deserve more military credit than what I'm giving him, but I still find it hard to see a scenario where he gets close to ballot territory. His glove was quite good, but none of the metrics see it as historically so - I don't think he was an Ozzie Smith or Joe Tinker out there. And the bat was good for a shortstop, but mostly below average - career wRC+ of 96, and only 4 seasons at 100 or above. Johnny Pesky I rank lower, but I actually have an easier time with his case - his bat was a step above Rizzuto's and he probably deserves more war credit than Rizzuto because his years surrounding his service were simply better.
   315. Carl Goetz Posted: December 12, 2018 at 11:16 AM (#5797064)
"Carl, thanks for your feedback on the ballot thread. It’s not going to change my already submitted ballot, but appreciated anyway. I realize my thinking is not to everyone’s taste, but it is what it is.

Any chance non-ballot comments can be tramsferred over here? Am thinking it will make things much easier for the folks tallying things up. Just a thought, FWIW."

No problem. Wanted to share my thoughts on the matter. I don't expect anyone to change their ballot for me. All I can ask is for my thoughts to be considered.
I agree that my comments should have been posted here and not on the ballot. I imagine non-ballots on the ballot thread makes things difficult for the ballot counters so I apologize.
   316. Carl Goetz Posted: December 12, 2018 at 11:31 AM (#5797070)
"Rivera registers very heavily with WPA because a) he had decent career volume, and b) he was used often in situations with the game actually on the line."
While I agree that for closers only, WPA is not completely useless, there are still issues with the stat for closers.
1)Situation: 1 run lead heading to the 8th. Setup man shuts down the 3-4-5 hitters in the 8th and closer shuts down the 6-7-8 hitters for the save in the 9th. WPA will give the closer more credit, but who did more to win the game? The way WPA is set up, the 9th is always going to be a more valuable inning than the 8th if the lead is the same.
2)WPA gives all defensive credit and blame to the pitcher. How good you're fielders do is irrelevant in WPA.
3)From where you start in WPA, all that matters is allowing zero runs to score for the rest of the inning. A closer who walks 3 guys and strikes out 3 guys without a run scoring gets exactly the same WPA as a guy in the same situation who gives a 1-2-3 inning.

#1 & #2 are big problems for me and keep me from using WPA as a measure of player value. I'm honestly not sure how I feel about #3 since technically, allowing zero runs in the inning is the goal. Also, the guy who walks 3 players is more likely to blow a save in other situations which will show up in his WPA too.

I will re-state that in its own description of WPA, Fangraphs warns against using WPA as a measure of player value and suggests that it is more of a storytelling stat.
   317. Kiko Sakata Posted: December 12, 2018 at 11:44 AM (#5797076)
#316

Apologies for the shameless self-promotion. My Player won-lost records use WPA (well, technically, they use win probability) as the guts of the system but makes adjustments which, I think, make them a much better measure of player value.

They very specifically address issue 2): on balls in play, I divide credit/blame between pitchers and fielders. They very specifically do NOT address issue 1): I make no quality-of-opposition adjustments. I would be very surprised if these sorts of things didn't balance out, though, if not over the course of a season, certainly over the course of careers the length of which we're evaluating here.

As for issue 3), by expressing things in terms of wins and losses, my stat will provide some information here. A pitcher who strikes out the side will just get some pWins - the number of which will depend on the exact circumstances of the particular game - and no pLosses. Whereas a pitcher who walks the bases loaded and then gives up two line drives, the first of which is caught and the latter of which is caught and turned into a game-ending line-drive double play will probably actually pick up more pLosses than pWins (the walks will, of course, be entirely blamed on the pitcher; giving up line drives is a negative, with the fielder mostly getting the offsetting positive from catching the ball).

Mariano Rivera looks very good in my system - he's #2 on my ballot and I debated putting him #1 - because he was so good for so long - although even with the "so long", Mariano looks much better compared to average (what I call pWOPA) than compared to replacement level (pWORL). No other relief pitchers really register in my system. Based on Player won-lost records, you could make a case for Trevor Hoffman as the second-best relief pitcher in MLB history (you could also make a case against that; there's maybe 4-5 guys with pretty similar career numbers). Hoffman would probably just make the bottom of my ballot if HOM ballots were 100 players long.
   318. DL from MN Posted: December 12, 2018 at 03:20 PM (#5797173)
86. Dick Redding. My punt is 3200 innings at 114 ERA+ for a record of 207-159, i.e. same quality as Chris but a little shorter. About here looks right – a little below Grimes (longer career) and Maglie (better quality.)


This doesn't square with the most recent MLE at all (more like 4900 innings at that 114 ERA+). Even Alex King's earlier MLEs were 3618IP 112ERA+. I split the difference this year with value equivalent to 4200 IP @ 113 ERA+.

He's clearly better than contemporary Vic Willis who you ranked #33. He looks better than your #16 Tommy John also (more IP at better rate). I would say Ferguson Jenkins is a good comp.


   319. Carl Goetz Posted: December 12, 2018 at 04:01 PM (#5797198)
From Karlmagnus ballot with Addie Joss: "If you assume the rest of his career would have been 1800 IP, 120-90 with an ERA+ of 110 (somewhat conservative, assuming you boost his last sick season, though pitchers didn’t last as long as they did later) then 50% credit would put him at 3227IP, 220-142, with ERA+ of 130. 25% credit puts him at 2777 IP, 190-120, with ERA+ of 136. Substantially better than Koufax. OPS+20. Electorate needs to take him more seriously. 121PP."

I'd say that's an extremely liberal estimate of the rest of his career given that he was dead.

Ross Youngs' ghost would like to know why he's not at the top of your ballot. He contracted a kidney disease at age 27, played 2 sub par seasons and then died. His 7 seasons from age 21-27 were roughly half a HOFer by career and about 75% there by peak (at least with BBref War compared to average RF HOFer; he'll be even closer once Baines' numbers get factored into the average). If he played until age 35, he'd easily have added 30 more WAR and probably some to his peak. Except for the small problem that he was dead and didn't do any of that. Nor was he capable of it.

I believe that both Joss and Youngs were good bets to finish off Hall of Merit careers had they not contracted the diseases that ended their lives. But that's not the same thing as them doing it in real life.
   320. karlmagnus Posted: December 12, 2018 at 05:50 PM (#5797244)
First, we have a problem that the HOM 2019 ballot does not appear on the main page. We will have an even worse problem of voters dropping off if we make the ballot that tough to find.

On Joss, he had 121 PP while alive, hence ranks between Rivera and Halladay based on what he actually did. Others are not giving him credit for his extraordinarily high ERA+.

On Redding, I was not aware of the latest estimates of his phantom career. You are however pro-rating up to a ML workload, when we have no evidence of anything like that, at least not a decent competitive level. He may be a little low at 85, I'll take another look next year, but even at #33 he's not on my ballot. I'm not rating him as high as the very solid Tommy John, who I think should be in. At 4200 with 113ERA+, your compromise. he would have 96PP, which is definitively below my ballot.
   321. Al "Battery" Kaline Posted: December 12, 2018 at 05:58 PM (#5797245)
*TANGENT ALERT* Not sure where else to post this, since the various help desk/bug report threads are closed...

Is anyone else getting stuck when trying to access the last page of either the Thibs tracker or Baines/Smith thread? For me, all I can see is an Adidas ad at the top, and the rest of the page never loads.
   322. Howie Menckel Posted: December 12, 2018 at 06:33 PM (#5797256)
sometimes I am logged out of HOM threads lately, though I can get in non-HOM ones
   323. Kiko Sakata Posted: December 12, 2018 at 06:36 PM (#5797258)
You are however pro-rating up to a ML workload, when we have no evidence of anything like that, at least not a decent competitive level.


This is a reasonable argument, which leads me to ask a question. For Negro League position players or 19th-century position players, it seems reasonable to me to just do a straight 1-for-1 schedule-length adjustment. If player A played 65 games in an 81-game schedule (I know, nobody played 81-game schedules; it makes for easy math), I'm fine with treating that the same as playing 130 games in a 162-game schedule. Sure, there's some chance that the guy would break his leg in game 97, but that's pretty rare for position players.

But I tend to view pitchers more as they only have so many innings in their arms. So, a guy who was able to pitch 120 innings per year for 15 years may not have been able to last for 15 years if he was pitching 240 innings per year instead.

But this leads to a question. Are Cannonball Dick Redding's innings which form the basis of his MLE's all of the innings that he was pitching those years? My understanding is that for a lot of the history of the Negro Leagues, teams played as many, if not significantly more, non-league games, barnstorming the country and what not, as they played league games. So are Redding's innings low because he really only pitched that few innings or are the innings used in Redding's MLE's low because they represent only a fraction of the innings that he was actually throwing?
   324. Michael J. Binkley's anxiety closet Posted: December 12, 2018 at 06:50 PM (#5797259)
Al-

I have gotten that trying to access any BBTF pages from my laptop this afternoon. I have not had any issuea using my phone, from which I am obviously posting now.
   325. Carl Goetz Posted: December 12, 2018 at 07:46 PM (#5797268)
#321 I've had the same problem today.
   326. Carl Goetz Posted: December 12, 2018 at 07:51 PM (#5797270)
"On Joss, he had 121 PP while alive"
I'm not sure what a PP is, but he only had around 46 WAR on BBref; 44.3 with my inclusion of FIP and other adjustments I make. I only get 21.6 WAA and 7.0 WAG so he doesn't score well on peak for me either.

The thing with ERA+ is that when the league average ERA is so low, it takes less ERs below that to build up higher ERA+. Also, 1/3rd of his runs allowed were unearned so I don't think ERA+ tells the whole story.
   327. Rob_Wood Posted: December 12, 2018 at 10:57 PM (#5797305)
I have been periodically and frequently unable to post in various HOM threads due to the unnerving "lock out" phenomenon. Some times some threads just don't want to receive my pearls of wisdom. People in the past have given hints, such as posting in a non-HOM thread and then coming back to the "locked out" HOM thread, or posting in another HOM thread and then coming back to the "locked out" HOM thread, clearing cookies/history, trying a different browser or computer (including PC vs. laptop vs. phone), etc. Most of the time one of those hints works in a reasonable number of tries -- but not always.

Having said that, I hereby authorize JoeD, DL, and anyone else the power to move my preliminary ballot posted above (post 263) over to the Ballot Thread if you don't see my ballot therein at the time the voting is about to close.

And I echo what somebody said somewhere, let's try to keep the Ballot Thread as clean as possible with just ballots. Any questions, answers, discussion, etc., should be posted in this the Ballot Discussion Thread. Thanks.
   328. bachslunch Posted: December 13, 2018 at 08:02 AM (#5797319)
Thought I'd expand out my position rankings for the heck of it:

1B. Helton, Taylor, Olerud, McGriff, Cash, Cepeda
2B. Kent, Lazzeri, Evers, Phillips, Myer, Pratt
SS. Stephens, Tinker, Fregosi, Aparicio, Campaneris, Tejada
3B. Bell, Bando, Cey, Ventura, Elliott, Harrah
LF. B. Johnson, J. Cruz, Berkman, O'Rourke, Downing, J. Gonzalez
CF. A. Jones, Lofton, W. Davis, Lemon, Damon, Pinson
RF. Sosa, Bonds, Cravath, Hooper, J. Clark, S. Rice
C. Schang, Lombardi, Munson, Posada, Tenace, Kendall
P. McCormick, Tiant, Halliday, Redding, Willis, John, M. Welch, Shocker, Rivera, [Pettitte], Bridges, Santana, Quinn, Cicotte, Finley, Tanana.

A quick look at next year suggests Derek Jeter (duh) and Bobby Abreu will both be on my ballot next year from the newcomers. They will go to the head of their respective positions, best as I can tell.
   329. Dr. Chaleeko Posted: December 13, 2018 at 08:15 AM (#5797324)
Re IP totals in MLEs

I want to be very careful with Karl’s question. I would be glad to answer in some detail, and you can always look up the process document on my site. Remember, the MLEs are designed to place a major-league caliber player into an MLB context from a context that is less well documented and more fractured by its nature. So the point is that we should end up with seasonal MLEs that look like MLB seasons, not like seasons constrained by the level of documentation we have at a given moment.
   330. progrockfan Posted: December 13, 2018 at 09:59 AM (#5797361)
@Carl Goetz:

The way WPA is set up, the 9th is always going to be a more valuable inning than the 8th if the lead is the same.

That's counterintuitive. I'm not saying you're wrong - but - a 1-run lead is tougher to hold with 6 outs to go than with 3 outs to go.

Presumably the 8th counts for less Win Expectancy in calculating WPA because the closer's team has an extra 3 batting outs, during which they can potentially increase their lead...?

@karlmagnus:

2. Addie Joss...

[...]

Bobby Bonds very short career, at a level that keeps him just off my consideration set

I have a deal of difficulty reconciling your #2 ballot placement of Joss with that last statement. Bonds's career was substantially longer than Joss's.

Also, regarding Addie Joss's dominance: Joss pitched from 1902-1910. Let's bear in mind Bill James's comment on baseball of Joss's era:

As late as 1900-1905, there were at least two amateur players who showed up at the park, bought a ticket, and wound up playing in that day's game. This is not "major league baseball" as we understand the term today.
   331. Carl Goetz Posted: December 13, 2018 at 11:23 AM (#5797393)
"Presumably the 8th counts for less Win Expectancy in calculating WPA because the closer's team has an extra 3 batting outs, during which they can potentially increase their lead...?"

I used this Win expectancy calculator for the following example. https://gregstoll.com/~gregstoll/baseball/stats.html#H.-1.8.0.1

Visiting Team has 1-0 lead going to bottom of the 8th.
70.52% chance of winning. (Leverage 2.13)
Visiting Team has 1-0 lead going to top of the 9th.
84.88% chance of winning.
Visiting Team has 1-0 lead going to the bottom of the 9th.
81.73% chance of winning. (Leverage 2.7)
Visiting Team has 1-0 lead at the end of the 9th
100.00% chance of winning (didn't need the site for this one)

8th inning guy gained .1436 WPA
Offense (in top of 9th) lost .0315 WPA
9th inning guy gained .1827 WPA

So even if both 8th and 9th inning guy went 1-2-3 and even if the 8th inning guy faced tougher hitters, the 9th inning guy earned 27% more WPA than the 8th inning guy.

Note that by the same method (assuming the visiting team scored their run in the top of the 1st) the starting pitcher only earns .1788 WPA for throwing a 6 inning shutout. He has to get the 2nd out with either nobody on or a runner on first in the 7th just to slightly pass the WPA value of the closer's 1 inning of high leverage work. This seems extreme to me.





   332. Rob_Wood Posted: December 13, 2018 at 12:42 PM (#5797449)
Yes, that is a well-known bug/feature of WPA. Many people do not use WPA for that reason. Hopefully, people who do use WPA are aware of this telescoping importance of anything that happens in late innings.

I don't know if this analogy works, but imagine if baseball's rules were changed so that the first team to score 10 runs wins (no matter in what inning). Then the parallel WPA calculations would be greatly "skewed" towards scoring that 10th run.
   333. progrockfan Posted: December 13, 2018 at 12:58 PM (#5797458)
@Carl Goetz:
This seems extreme to me.

Me too. Extreme, and with pitch counts trending the way they are, unattainable by most starters. That's why I can't (and don't) advocate general usage of WPA.

Having said this, WPA explicitly measures success in leveraged situations - not overall game-winning value (what's what WAR is for). The starting pitcher would earn far more WAR, the reliever far more WPA. So, assuming that leverage is at its highest in the 9th inning of a close or tied game, WPA is doing its intended job in your example.

So, the question becomes: How important is late-inning leverage in baseball games?

Here's another way to phrase it. If managers could find the roster spots, would they carry enough pitchers to allow each pitcher to face each batter only once per game, every game? We're seeing sporadic use of 'openers' now to face the first pass through a lineup - which would seem to indicate that for at least some managers, the answer would be Yes: it's quite valuable to have opposing batters see a given pitcher only once.

Well, baseball rosters aren't big enough to permit that. Specialist relievers are therefore almost always used at or towards the end of the game. Further, teams tend to use their ace specialist at the very end, and lesser relievers earlier on, as setup men.

Why the very end? Because leverage is at its highest then, thereby giving the ace's specialist innings their highest possible value. It als provides an extra edge against pinch hitters.

Is that sound strategy? Baseball managers certainly seem to think so. But then, they also used to embrace the sac bunt, the low-percentage steal, and other dubious tactics.

For years Bill James advocated closer by committee. Boston said, What the hey, he's a smart guy, we'll give it a go. It was a disaster. One possible lesson to take from this is that a designated ace reliever should indeed close the game. (At any rate, that's the lesson Boston took, and no other team in MLB seemed to disagree.)

In sum: MLB managers have concluded, for a number of decades now, that end-of-game leverage is sufficiently important to assign an ace pitcher to that and only that. WPA probably does ascribe too much credit for these situations - but at least it provides some form of measure of the (quite possibly exaggerated) value that MLB managers actually do ascribe to their closers.

Of course, we don't have to agree.
   334. Carl Goetz Posted: December 13, 2018 at 02:32 PM (#5797515)
"So, assuming that leverage is at its highest in the 9th inning of a close or tied game, WPA is doing its intended job in your example."
Assuming the 9th is actually 27% more important than the 8th and that fielders had nothing to do with the results. I'd still argue that leverage is flawed because it doesn't take into account who's batting. If the 8th was the other teams' 3-4-5 or 2-3-4 hitters, the 8th is actually the higher leverage inning. But leverage calculations assume all the hitters are average. That's just wrong, period. Its fine if looking at context-neutral numbers like WAR if you assume all opponents are average; on average, they are. (or close to it over the long run anyway) But the entire point of WPA is to look at what the player does IN context. The context for a relief pitcher is not just the base out situation and the score. Key parts of that context are the "Who" 1:who is batting for the other team, 2:who is on base and 3:who is fielding behind you (and 4 are you listening to The Who cause that just makes everything better). Beyond a super complicated black box mixed model, I don't know how you'd account for those things. But if you are going to use leverage or WPA, you need to understand that these flaws are baked in.

I don't personally buy some of the assumptions used for WPA, but if you do, I strongly suggest you look into Kiko Sakata pW-pL records which is effectively an improved version of WPA that starts with each play's win value change like WPA but instead of giving the whole offensive value to the hitter, it divides between hitter and baserunner(s). Similarly, it divides defensive contributions between pitchers and fielders. I personally only look at it to break ties between otherwise roughly equal players on the theory that the difference between 2 players' pW-pL record was that the player with the better record performed better in the clutch. I don't use it for my base starting values because it still has the same leverage issues I've detailed above. It is however, far superior to straight WPA if you want to factor in leverage.
Kiko, I'm assuming you're ok with me sharing your website since you probably want to sell more books and stuff :). Also, please let me know if I've misstated anything regarding your system.
http://baseball.tomthress.com/Articles/Basics.php

"For years Bill James advocated closer by committee. Boston said, What the hey, he's a smart guy, we'll give it a go. It was a disaster. One possible lesson to take from this is that a designated ace reliever should indeed close the game. (At any rate, that's the lesson Boston took, and no other team in MLB seemed to disagree.)"
I'm pretty sure Bill James advocated for the relief Ace to be used in important spots in the 7th or 8th and not exclusively in the 9th. Boston tried (assuming you are referring to 2003) a closer by committee because they had a bunch of average to slightly above average relievers. They didn't have a true relief ace. And now teams are starting to deploy a relief ace at differing times in the game; its the so-called Andrew Miller role.

"Of course, we don't have to agree."

If everyone agreed, we wouldn't need a Hall of Merit :)
   335. progrockfan Posted: December 13, 2018 at 02:34 PM (#5797517)
@Kiko Sakata:
2): on balls in play, I divide credit/blame between pitchers and fielders.

I assume this means that strikeouts are credited solely to the pitcher... How about less clear-cut situations, like, say, an error followed by 3 hits - or, 3 errors followed by a hit?- are these credited & debited in the same manner as earned & unearned runs?

As for issue 3) [...] A pitcher who strikes out the side will just get some pWins - the number of which will depend on the exact circumstances of the particular game - and no pLosses. Whereas a pitcher who walks the bases loaded and then gives up two line drives, the first of which is caught and the latter of which is caught and turned into a game-ending line-drive double play will probably actually pick up more pLosses than pWins (the walks will, of course, be entirely blamed on the pitcher; giving up line drives is a negative, with the fielder mostly getting the offsetting positive from catching the ball).

This makes a lot of sense to me.

For a lot of the history of the Negro Leagues, teams played as many, if not significantly more, non-league games, barnstorming the country and what not, as they played league games. [Emphasis added.]

Absolutely true; this was an indispensable source of revenue for leagues, owners, teams, and players alike.

So are Redding's innings low because he really only pitched that few innings or are the innings used in Redding's MLE's low because they represent only a fraction of the innings that he was actually throwing?

I find generous MLEs 100% fair for NgL players not only because they represent what the players should have been allowed to do in MLB, but also, in many cases, they compensate for what they actually did in hundreds of un-recorded and under-recorded games per year. It's an inexact science - but a helluva lot more exact than, say, Ritter and Honig, working from complete and verified statistics, claiming that George Sisler is one of the greatest players ever because look at all those hits!
   336. Kiko Sakata Posted: December 13, 2018 at 02:37 PM (#5797522)
So, assuming that leverage is at its highest in the 9th inning of a close or tied game, WPA is doing its intended job in your example.


This is true, but "of a close or tied game" is doing all of the work there. In major-league games in general, the average leverage actually declines per inning and is lowest in the 8th or 9th inning. Why? Because most games aren't "close or tied" late. I did the math for the first decade of this century (2000-09), so it may not be completely correct, but the median point in a game in which the winning team takes the lead which they will never lose is the 4th inning.

Here's an article I wrote on the subject. I apologize if any of the links within the article are broken: it's a few years old now (it's also pp. 43-48 of my first book if anybody has that).
   337. Kiko Sakata Posted: December 13, 2018 at 02:44 PM (#5797529)
How about less clear-cut situations, like, say, an error followed by 3 hits - or, 3 errors followed by a hit?- are these credited & debited in the same manner as earned & unearned runs?


pWins and pLosses are allocated play by play. So, the error and hits would all be debited to both the fielder and the pitcher (*) - depending in part on the hit type (ground ball, fly ball, line drive) where I have that information (which is pretty rare on hits prior to I'd say the 1980s). I don't do anything like with unearned runs - "well, if not for the fielders, the inning would be over, so everything from here out is blamed on the fielder who made the error". That is, if a pitcher gives up a home run, that's 100% on him, even if it's the 8th batter he faced in the inning, the first 7 of whom reached on fielding errors.

(*) - I don't explicitly treat errors and hits differently. Player won-lost records are outcome-based: what I care about is whether an out is recorded and where the baserunners end up. But I find that, as a practical matter, it tends to work out that errors end up being charged more heavily to the fielders who committed them than to the pitchers who allowed the ball in play.
   338. Carl Goetz Posted: December 13, 2018 at 03:15 PM (#5797545)
"It's an inexact science - but a helluva lot more exact than, say, Ritter and Honig, working from complete and verified statistics, claiming that George Sisler is one of the greatest players ever because look at all those hits!"
I like Sisler a lot, though not because of "all those hits". Imagine Mike Trout catches an awful case of sinusitis (sp?) and has double vision and is forced to miss all of 2019 and then plays 6 or 7 below average seasons after that. I think their Hall of Fame cases would be about the same and, frankly both are easily in for me.
   339. progrockfan Posted: December 13, 2018 at 04:03 PM (#5797581)
I like Sisler a lot

Me too, actually. I rate him as the finest first baseman prior to Gehrig. But Ritter and Honig looked at his 1924-30 play and determined it was on the highest levels, based on surface stats like hits and BA and not taking into account even simple metrics like outs made or average vs. league.

Imagine Mike Trout catches an awful case of sinusitis (sp?) and has double vision and is forced to miss all of 2019 and then plays 6 or 7 below average seasons after that. I think their Hall of Fame cases would be about the same and, frankly both are easily in for me.

Yeah, I think that's about fair. Sisler from 1917-22 was a beast.

What I meant to say is that I regard NgL MLEs as a more informative measure of player value than traditional stats were for MLB players prior to the sabermetric revolution - you know, when guys like Luis Aparicio were considered to be big stars because look at all those steals! I think those who question the validity of MLEs sorely underestimate the sheer volume of comparable data points that exist from players who competed in both MLB and the NgL. Speaking metaphorically, I have much greater confidence in the 2010s Hall of Merit evaluation of Ben Taylor than I do in the 1960s Sporting News evaluation of Luis Aparicio.
   340. Carl Goetz Posted: December 13, 2018 at 04:19 PM (#5797595)
"What I meant to say is that I regard NgL MLEs as a more informative measure of player value than traditional stats were for MLB players prior to the sabermetric revolution - you know, when guys like Luis Aparicio were considered to be big stars because look at all those steals! I think those who question the validity of MLEs sorely underestimate the sheer volume of comparable data points that exist from players who competed in both MLB and the NgL. Speaking metaphorically, I have much greater confidence in the 2010s Hall of Merit evaluation of Ben Taylor than I do in the 1960s Sporting News evaluation of Luis Aparicio."

I agree with all of this. I'd add that I usually assume Ngl players peaks are higher than they appear due to smoothing of the numbers. For example, if a Ngl player played 77 of 79 games that year, the numbers get rated up to 154 game season, the extra games are at the player's career average, rather than what they actually did that year. It tends to smooth out the peaks and valleys that occur in most players (other than Harold Baines) careers.

And I knew you weren't ripping Sisler. You brought him up and I like talking about players that I like.
   341. progrockfan Posted: December 13, 2018 at 05:53 PM (#5797627)
I like talking about players that I like.

Me too. That's why George Van Haltren is #20 on my ballot. I don't think he has a realistic chance of making the HoM, but it helps keep his memory alive, 115 years after he played his final game.

As we both enjoy talking about our favorite players, let me talk a bit more about Sisler.

His HoF plaque reads, "Credited with being one of best two fielding first basemen in history of game." The other referent here is obviously Hal Chase.

Now, Mr. James dismisses that view of Sisler's defense. I don't have the book in front of me, but in ranking Sisler #28 at first base (iirc) in the New Historical Abstract, he says something like, "Win Shares fails to see that this is true", and goes on about how Win Shares does document almost all other fine-fielding first baseman.

Well, piffle. And here's why piffle:

* First base was not only a much more important fielding position when the game was dominated by ground balls, but was a wholly different type of fielding position as well. First basemen played with different positioning and depth prior to the live ball, and reacted differently to things like bunts and full counts on the batter. I strongly suspect first was the defensive postion whose nature was most altered by the arrival of the live ball - and this alteration would've taken time; the process of change would've continued into the 1920s, through a goodly chunk of Sisler's active career.

* This means that any modern system of defensive evaluation would, I strongly suspect, struggle to accurately quantify first-base defense prior to 1920. (Hell, even today they haven't got third base right yet!)

* Hal Chase would've sold a pennant-clinching game for a plugged nickel. Everything he did in the field must be discounted accordingly. Therefore,

* Contemporary opinion held that George Sisler was much the finest fielding first baseman in baseball history up to and through his time - in an era when first base was considered a key defensive position.

* In the case of pre-1920 first-base defense, I trust contemporary opinion more than Win Shares - especially as Win Shares sees this "one of best two fielding first basemen in history of game" to be a poor fielder. Whilst often enlighening, the discrepancy between contemporary opinion and modern analysis on this subject is simply too vast to credit.

So, in addition to having the finest offensive run at first base prior to Gehrig, I also credit Sisler with being a fielder to stand with the likes of Power and Hernandez. Further, I credit him with being MLB's top first baseman not only for 1917-22, but also for 1925, well into his decline phase. I realize that's partly in default, as Gehrig hadn't yet become GEHRIG, but there it is nonetheless.

All of this adds up to a pretty formidable historical portrait. I suspect that these days, as his image has taken a hit from the rightly-discounted decline in his second half, people sometimes forget his first, because he's as good a peak candidate as anyone. Add in top-flight defense and you have one of the better first basemen of all time. That's my take anyway. ;)

   342. Dr. Chaleeko Posted: December 13, 2018 at 05:54 PM (#5797628)
For the benefit of posterity, I just want to repeat an idea from upthread. Our job with MLEs is to take a person who didn’t get a complete, equitable, or equal opportunity to play in an MLB context and place what we know about them within an MLB context to better understand their careers. The job is NOT translating their stats directly from their league of origination to the destination league. To do the latter would be perpetuating the injustice done to these ball players. When we retroactively look at a fractured and incomplete statistical record (which is so because of racist policies and attitudes) and say that we shouldn’t prorate to the MLB workload level because we don’t see evidence that the player had the ability to throw that many innings, we are very close to a line of thinking we probably don’t intend nor would want to pursue.
   343. Carl Goetz Posted: December 13, 2018 at 07:44 PM (#5797659)
I rank Sisler #13 on my 1B list. He falls between McCovey and Buck Leonard for me, though if Joey Votto has another Joey Votto season in 2019, he may move ahead of all 3. Of currently eligible 1Bs, I'd put in Helton, Taylor and Chance. Of unelected by the actual Hall, I'd add Hernandez, McGwire, Helton, Palmeiro and Start in that order. I could go either way on Will Clark.
   344. JoeD has the Imperial March Stuck in His Head Posted: December 14, 2018 at 09:22 AM (#5797733)
That's why George Van Haltren is #20 on my ballot. I don't think he has a realistic chance of making the HoM, but it helps keep his memory alive, 115 years after he played his final game.


He was so close. Not sure if you are aware or not, but he was the Rizzuto of many years ago. He was first runner-up more than once I think, but never got over the line. I'll give him another look too.

I've been wanting to do a complete overhaul of my ballot for quite sometime. This should be the year. There was less urgency the last few years because we were only realistically going to elect recent players due to the big glut of them coming through.
   345. JoeD has the Imperial March Stuck in His Head Posted: December 14, 2018 at 10:33 AM (#5797769)
Now, Mr. James dismisses that view of Sisler's defense. I don't have the book in front of me, but in ranking Sisler #28 at first base (iirc) in the New Historical Abstract, he says something like, "Win Shares fails to see that this is true", and goes on about how Win Shares does document almost all other fine-fielding first baseman.


I wonder what Michael Humphries DRS says about Sisler's glove. I agree that contemporary opinion shouldn't be dismissed out of hand when the game was so much different.

There's a two prong argument here though - one is how good was he relative to his peers. The bigger one, IMO is where on the spectrum did 1B fall pre-1920? I think it was certainly more important than RF/LF. It was more important than CF until around 1915 or so. You can see this by comparing how well those players did hitting. Heck even through about 1950 or so 1B and LF/RF were about the same.

Here are some links to a chart that shows the fluidity of the spectrum over time. It's much more fluid than most realize. Certainly more than just "flip 2B/3B around 1930".

Numbers Version
Excel Version
Dan's original (Excel)

Dan Rosenheck created this, I found a spreadsheet from 2009. I tweaked it to use moving averages and smooth the lines out a little more around 2015-16. It's based on the worst regulars at the position. I forget the exact number, around 10-15% IIRC. This averts 'star gluts' throwing off the average.
   346. bachslunch Posted: December 14, 2018 at 11:03 AM (#5797792)
Sorry, didn't realize Jim O'Rourke was enshrined already as a CF. Revised:

1B. Helton, Taylor, Olerud, McGriff, Cash, Cepeda
2B. Kent, Lazzeri, Evers, Phillips, Myer, Pratt
SS. Stephens, Tinker, Fregosi, Aparicio, Campaneris, Tejada
3B. Bell, Bando, Cey, Ventura, Elliott, Harrah
LF. B. Johnson, J. Cruz, Berkman, Downing, J. Gonzalez, Veach
CF. A. Jones, Lofton, W. Davis, Lemon, Damon, Pinson
RF. Sosa, Bonds, Cravath, Hooper, J. Clark, S. Rice
C. Schang, Lombardi, Munson, Posada, Tenace, Kendall
P. McCormick, Tiant, Halliday, Redding, Willis, John, M. Welch, Shocker, Rivera, [Pettitte], Bridges, Santana, Quinn, Cicotte, Finley, Tanana.
   347. Kiko Sakata Posted: December 14, 2018 at 11:34 AM (#5797811)
I wonder what Michael Humphries DR says about Sisler's glove.


In his book, Humphreys doesn't go through first basemen in as much detail as at most other positions. For most positions, he shows numbers for everybody who played a significant number of innings, broken down by era. For first basemen, he just shows a "Top Forty First Basemen of All-Time" which does not include Sisler. George Sisler's name is also not in the index of the book. I think DRA numbers are available online - maybe at The Baseball Gauge?

For what it's worth, my system doesn't think much of Sisler's fielding. Although that comes with two pretty big caveats. I use Retrosheet data. They haven't released anything prior to 1921 (i.e., for Sisler's first six seasons) and they're missing some games from 1921 - 1936 (although we just finished deducing 1936 last night - but that won't be released until June - and that's six years after Sisler retired anyway). Anyway, from 1921 - 1930, Sisler actually played 1,288 games, of which I have play-by-play for only 944 of them. So, my system mostly doesn't think much of Sisler's fielding after his eye injury.
   348. progrockfan Posted: December 14, 2018 at 12:20 PM (#5797832)
Off-topic (sorry for the intrusion, but JoeD says this is the place to ask atm):

Does anyone here have access to, or know where to get, Ichiro's stats for:

* The 1995 and 1996 Japan Series
* Tne 1993-2000 Japanese All-Star Games

Nippon Sports used to have these numbers online, but they're gone now. Neither MLB, the Mariners, Nippon Sports or Orix has responded to my email requests for this data. Jim Albright was kind enough to respond, but doesn't hold the data.

And, less likely, does anyone have info for:

* His 1996 at-bats against MLB pitchers; I know he went 7-11, .636, but don't have any more info than that

Many thanks! :)
   349. Mike Webber Posted: December 14, 2018 at 04:25 PM (#5797950)
@346

I really like this, it frames up the results so you can make sure you aren't forgetting right fielders (ok, catchers really).
   350. JoeD has the Imperial March Stuck in His Head Posted: December 15, 2018 at 11:31 AM (#5798089)
Re: my comment earlier about catchers - I forgot Posada entirely. I think he's the best eligible catcher we haven't enshrined yet. I would definitely support him for election.
   351. Dr. Chaleeko Posted: December 15, 2018 at 03:30 PM (#5798126)
I’m going to disagree with our esteemed leader twice in one post. (Looks down sheepishly.)

On GVH: Nothing wrong with taking another look at a guy, but we’ve combed every nit out of the 1890s. IMO and by my own calculations, I have us six or seven careers worth of players above expectation for the 1890s and 1900s. Why look back at GVH when Jones and Lofton are better CF candidates from underserved eras?

On Posada: I originally thought like you, Joe, about Posada. Max Ma chi’s handling data shows Posada as one of the worst pitcher handlers. I use Max’s numbers at half strength, and that’s enough to push down from the borderline to also-ran. On the other hand, it pushes Sundberg and Pena up from also-ran to just under the line among Cs.

Also, generally, I’ve just posted my ballot, and it includes Dick Redding in the top spot and no other Negro Leaguers. Since I’ve just run a bunch of MLEs that I hope have been helpful, I want to point out that Redding’s MLEs have a pretty robust dataset behind them with near 2,000 innings of data that I pull from. There are other MLEs that look darned impressive: Hurley McNair, Heavy Johnson and Burnis Wright for example. McNair and Johnson have significantly less data to work with than the average player I ran an MLE for. Wright has about as much data relative to batters that Redding does. However, he is showing right now as a +55 fielder. That’s not impossible, but we have just 25-33% of his fielding data. There are wrinkles like that with a lot of MLEs. And given my analysis of the ballot in front of us, Redding’s was the one MLE that stood out to me as having high-quality performance, robust data, and a matching historical reputation to corroborate them. Ben Taylor has a little less data but what we know matches his reputation well. I would feel fine casting a ballot for him had I chosen too. Ray Dandridge actually has more data available than Redding, and I would have no problem casting a ballot for him had I thought he was among the top 15 outside the HOM.

At my site I have posted an updated file that contains all the MLE career lines. It also now includes a stat called DATA+. I cooked up a junk stat to measure how much data a player has to support his case, then I indexed that for each player against hitters or pitchers, whichever was apt. I’d recommend having a look at it before casting a ballot.
   352. Howie Menckel Posted: December 15, 2018 at 09:58 PM (#5798155)
I was stuck off the site all day thanks to that annoying adidas ad at the top of the site (that for now has cycled out).

I oppose the death penalty in most instances but......
(ok, internet, I'm kidding. life without possibility of parole would be sufficient.)

I suspect it's going to impact our voting if no one is able to fix this fockup.
   353. Carl Goetz Posted: December 16, 2018 at 12:47 PM (#5798211)
Adidas ad is still a problem as of a few minutes ago.
   354. Howie Menckel Posted: December 16, 2018 at 09:53 PM (#5798296)
it's true about those evil adidas ads - if you refresh 5 or 6 times, you can find a new one.

meanwhile, the problem about how sometimes it's impossible to be logged into an HOM thread persists. happened twice again today

at least I'm getting this post higher up the right side
   355. DL from MN Posted: December 17, 2018 at 10:13 AM (#5798334)
I really like this website - I've had great conversations here, learned a lot, fun people - but I really hate this website - unsecure, logins suck, editing is very basic, no graphics, sucks on mobile devices, people can't even comment (which is supposed to be the whole point of the website) without difficulty.
   356. Carl Goetz Posted: December 17, 2018 at 10:43 AM (#5798358)
From Bleed the Freak in Ballot Thread.
"Buddy Bell is getting WAR that should be allotted to shortstops from that era, Kiko has convincing data to avoid Buddy."
From Kiko Sakata in Ballot Thread.
"Buddy Bell - same story as Kenny Lofton but far more so. Honestly, he's not THAT much better than Harold Baines in my system - Baines was a better hitter, but Bell makes that up by playing third base and playing it very well. I think BB-Ref also has problems with their positional adjustments for 3B and SS in the 1970s (see my #'s 20-22 above)."

This isn't one I'm familiar with. Can you guys elaborate on the issue? Is it on the dWAR side or Positional Adjustment side? Bell comes out extremely good defensively on both TZ and DRA.
   357. bachslunch Posted: December 17, 2018 at 10:51 AM (#5798362)
Just a comment to progrockfan: nice ballot, a well-considered and thoughtfully-defended submission from first time voter.
   358. Kiko Sakata Posted: December 17, 2018 at 11:02 AM (#5798369)
Can you guys elaborate on the issue? Is it on the dWAR side or Positional Adjustment side?


It's the positional adjustment. If you add up the WAA of all third basemen in the 1970s, it's positive - i.e., the average third baseman is above average. If you add up the WAA of all shortstops in the 1970s, it's negative - i.e., the average shortstop is below average. Which, in my opinion, is non-sensical.

Here's BB-Ref's table of WAA by position for 1977, which is a fairly typical year for this time period. Bottom line gives MLB totals. Third base is +0.7; shortstop is -0.9. For the entire 1970s, 3B is +6.2 and SS is -6.6. This continues into the 1980s as well. Buddy Bell was a regular from 1972 - 1987. Over those years, third basemen are +8.8 WAA - about +0.6 per season.

Edit to add: I looked at 1970 - 1987. The total WAA for third basemen was positive in every one of those 16 years. The total WAA for shortstop was negative in 15 of the 16 years; the exception being 1983, when shortstop WAA was exactly zero.
   359. Carl Goetz Posted: December 17, 2018 at 11:43 AM (#5798381)
But is that a function of an error on BBref's part or of errors in team building during that period? You have an era with a lot of good 3B (both on the offensive and defensive side) and a lot of guys like Jim Mason, Fred Stanley, Bucky Dent, Freddie Patek, Frank Duffy, Larry Bowa etc playing SS. Teams were putting a premium on SS who could steal a base (even with crappy %), lay down a bunt, hit for empty Avg (ie low OBP and SLG), avoid errors (rather than having good range), etc. Looking back with the information we have now telling us what actually wins games, I think its reasonable to conclude that in terms of winning games for their teams, the average SS of that era was in fact a below average player. Sure, guys like Campy, Harrah and Concepcion stood out during this period, but at best, those guys are borderline considerations. There's not much in the way of inner circle SSs between the 50s and the late 70s/early 80s emergence of Yount and Ripken. As one example, there were a lot of players in the 70s attempting a lot of steals at rates that were actually hurting their teams. A lot of them were SSs and very few were 3B.
I guess my point is; is there any reason to believe that every position should equal 0.0 WAA every year with just a small amount of random noise thrown in? Maybe in the current era, with all the data we have access to, that's a reasonable assumption. I'm not so sure looking back at prior eras that its a reasonable assumption.
   360. progrockfan Posted: December 17, 2018 at 11:55 AM (#5798386)
@bachslunch:

:)

A pleasure to be here.
   361. Kiko Sakata Posted: December 17, 2018 at 12:09 PM (#5798393)
You have an era with a lot of good 3B (both on the offensive and defensive side)


Which limits the value of those guys. If everybody has a good 3B, then merely having a good 3B isn't helping your team win games.

is there any reason to believe that every position should equal 0.0 WAA every year with just a small amount of random noise thrown in?


I think if you're trying to measure actual wins, then I think that every position should be average on average by definition. The way you beat the other team is to put better players at positions. If everybody else is doing it wrong, there's real value to being the only one to do it right. I also think it's a little presumptious to presume that everybody was doing it wrong.

At Baseball-Reference, third base WAA is +0.1 or higher every season from 1959 - 1989. Thirty years is a long time to have a random positional glut.

As one example, there were a lot of players in the 70s attempting a lot of steals at rates that were actually hurting their teams. A lot of them were SSs and very few were 3B.


This isn't the strongest argument to use to defend Buddy Bell as a Hall-of-Merit player. Buddy Bell was 55 of 134 (41.0%) in stolen base attempts in his career. According to Bill James in his New Historical Abstract, that's the worst stolen-base percentage in major-league history (for years w/ CS data, of course).
   362. bachslunch Posted: December 17, 2018 at 12:59 PM (#5798401)
Which limits the value of those guys. If everybody has a good 3B, then merely having a good 3B isn't helping your team win games.


Huh. Maybe I'm grossly misunderstanding things, but doesn't having a good 3B, even if everyone else does at the time, help you win games? It might not differentiate between third basemen in the sense that one is better than another, but there are holes in positions by decade in several instances (catchers in the first couple decades of the 20th century, third basemen in Pie Traynor's time, full-time NL first basemen in the 50s, starting pitchers in the 80s). If there happen to be a load of great 3B who played all at once, why can't they all go in? We have period adjusted WAR and similar things for this purpose, no?
   363. Kiko Sakata Posted: December 17, 2018 at 01:08 PM (#5798409)
Incidentally, it's a little harder for me to figure out exactly what they're doing, but Fangraphs looks to be doing a better job with positional adjustments, at least for third base for Buddy Bell's career. Basically, since World War II / integration, third basemen have tended to hit at right about league-average (excluding pitching). For example, from 1947 - 2018, in my stat, I get third basemen with an aggregate winning percentage (adjusted for pitcher hitting) of 0.503. So, basically, the positional adjustment for third base should be right around zero.

For his career, Buddy Bell's positional adjustment is +44. He played the OF his first year and mostly DH his last year, and the positional adjustments for those two years are -5, which is probably reasonable. So, Bell's positional adjustment for his time at third base is +49, which is giving him about 5 wins too many if you think, as I do, that his positional adjustment should be more like zero. Baseball-Reference gives Bell 66.3 bWAR on the strength of batting, baserunning, and fielding numbers (in runs) of +102 (including Rdp), -17, and +174, respectively. Subtract 5 from that and you'd have 61.3.

Fangraphs gives Buddy Bell 61.7 fWAR with components of +93.0, -19.6, and +176.0 for batting, baserunning, and fielding respectively. In other words, Fangraphs sees Bell the same way BB-Ref does - above-average hitter (they're within 9 runs there), lousy baserunner (they're within 3 runs there - in the opposite direction as the batting), elite fielder (they're within two runs there, which is remarkably close) - the only difference, as far as I can tell, is the positional adjustment (Fangraphs and BB-Ref use the same replacement level).

In the specific case of Bell, BB-Ref and Fangraphs both over-weight fielding, by a similar amount, which is why I would argue that even 61 WAR is too many, but that has nothing to do with the direct topic at hand.
   364. Kiko Sakata Posted: December 17, 2018 at 01:12 PM (#5798411)
We have period adjusted WAR and similar things for this purpose, no?


Well, my argument is that in the case of third basemen in the 1960's, 1970's, and 1980's, no, we don't. At least not if we rely on Baseball-Reference.

there are holes in positions by decade in several instances (catchers in the first couple decades of the 20th century, third basemen in Pie Traynor's time, full-time NL first basemen in the 50s, starting pitchers in the 80s)


I think the problem here is that this comes very close to begging the question. There appear to be holes if we assume that the answer is that positional averages are constant over time so that a drop in empirical averages can only be explained by there being holes at those times.
   365. Kiko Sakata Posted: December 17, 2018 at 01:31 PM (#5798416)
Maybe I'm grossly misunderstanding things, but doesn't having a good 3B, even if everyone else does at the time, help you win games?


Suppose we were to start a Strat-o-Matic league (or DMB or whatever). Now, suppose everybody got 1980 Mike Schmidt for their team. Does having 1980 Mike Schmidt on your team give you an advantage over anybody else in the league?
   366. TomH Posted: December 17, 2018 at 01:59 PM (#5798424)
Obviously (#365), in a strict sense, no.

But there is a big difference between that and reality. Some teams would trade for Mike Schmidt and play him at SS or 1B or something, and then a few teams would NOT HAVE the best 3Bman ever.

The question should also be asked, do I baseline my guy to the position average, or to replacement level? What if there are 13 Mike Schmidts/George Bretts and 13 Ken Reitzeses in 1980? Is every Schmidt a +4 and every Reitz a -4, or Schmidt +8 and Reitz 0? But if we had one Schmidt, one Brett, and 26 Retizeses, is it +7.3 and -0.7 respectively, or again +8 and 0? Run a thought experiment and see how you would draft them; I do not believe you will find the 13 Schmidts in the first example are worth 3.3 (4 minus 7.3) games less than those in the second example.
   367. Carl Goetz Posted: December 17, 2018 at 02:06 PM (#5798427)
"At Baseball-Reference, third base WAA is +0.1 or higher every season from 1959 - 1989. Thirty years is a long time to have a random positional glut."
And generally speaking, teams were over stealing, over bunting and valuing Fielding Percentage over Range in SSs for that entire period.

"I think if you're trying to measure actual wins, then I think that every position should be average on average by definition."

By definition, the total player population should have 0.0 WAA. There's nothing definition-wise that should constrain it to zero at each position. I guess I'd ask the follow up question; if every MLB SS was between 0 and 2.0 WAR for a season with a weighted average around 1.2 because there's no stars, should average be defined down to 1.2 for that year?

"Which limits the value of those guys. If everybody has a good 3B, then merely having a good 3B isn't helping your team win games."
Not necessarily compared to replacement. A 5-6 War 3B is still winning you 5-6 games above replacement regardless of where 'Average' is. Keep in mind that the population of major league caliber players represents the extreme right tail of the baseball playing population of the world; and when we look at the superstars, its even more extreme (they're the right tail of the right tail). This is not a normal distribution that we are looking at. Sometimes superstars bunch up at 1 position. That doesn't make them less valuable assuming replacement level is the same. Those superstars can raise the "Average" at a given position.
For example, imagine 2 Scenarios where every CF in the league is the same except for Mike Trout. In scenario A, Trout plays every game and ends with 7.0 War and is the best CF. In scenario B, Trout plays every game and ends with 13.0 War and an alltime great season; obviously still the best CF. The Average WAR of a CF has increased by roughly .2 (6 extra WAR in CF divided by 30 Teams) in Scenario B compared to Scenario A just based on Trout's performance. Lets say in both scenarios, Lorenzo Cain is the 2nd best CF with 6.8 War. Was he really less valuable to the Brewers in scenario B? WAA says yes, but War says no. Replacement level shouldn't have changed in any meaningful way based on Trout's season since he's not a replacement level player. Technically, the way War is calculated, there will be 6 less War over the rest of the baseball population, but in the real world, the value of replacement is the same. If Mike Trout had 700 PAs, the calculated replacement level would increase by .02War/700PAs (6War/184,439 PAs*700) which wouldn't be a noticable difference looking at each other player.

"This isn't the strongest argument to use to defend Buddy Bell as a Hall-of-Merit player. Buddy Bell was 55 of 134 (41.0%) in stolen base attempts in his career. According to Bill James in his New Historical Abstract, that's the worst stolen-base percentage in major-league history (for years w/ CS data, of course)."
Well, 134 attempts isn't exactly a lot for an 18-year career. I estimate roughly 2706 times on 1B for his career (doesn't include reaching on errors) so that's about 134/2706= 4.95% of the time he was on 1B. And that negative is accounted for in his -17 rbaser on BBRef. It was more an argument as to why 3B were generally more valuable than SS during this period. The argument for Bell is that he was a stellar defender and a pretty good hitter. He had a strong career value value and a good sized peak.

   368. Carl Goetz Posted: December 17, 2018 at 02:17 PM (#5798432)
"Suppose we were to start a Strat-o-Matic league (or DMB or whatever). Now, suppose everybody got 1980 Mike Schmidt for their team. Does having 1980 Mike Schmidt on your team give you an advantage over anybody else in the league?"
If every team had Mike Schmidt at 3B and and Larry Bowa at SS, would they be equally valuable? If so, you'd be willing to trade them even up. Would you rather be the guy with Mike Schmidt at both SS and 3B? Or the guy with 2 Larry Bowas?
   369. Michael J. Binkley's anxiety closet Posted: December 17, 2018 at 02:25 PM (#5798437)
I am in the process of finalizing my ballot, but I just wanted to weigh in on this positional value debate.

A quick summary of my version of WAR for position players - I start with an average of bWAR and gWAR. I then adjust this number my a 9-year rolling average of not positional WAA, but Wins Above Positional Median, as to not be skewed by star gluts and droughts.

Thus I agree with a lot of Dan Rosenheck's (and Kiko's) findings when developing his WAR, that shortstops have been generally undervalued throughout most of baseball history, but especially from the late 1960's to the late 1980's, and conversely 3B was overvalued for much of the 1970's and early 1980's. Consequently, Campaneris is mid-ballot for me, and while the positional adjustments do hurt Bell, and as a result he is off-ballot, both TZ and DRA like him enough so he is still in my PHoM.

(On a side note, this is also why DH's do better in my system than by traditional bWAR. For some reason, teams have underutilized the DH. According to my rolling-average research, since the inception of the DH, not once in the entire history of the DH has the median DH had a positive WAA - Ortiz is sliding easily into my PHoM when he becomes eligible).

I then adjust my WAR numbers by a standard deviation adjustment, but that's another issue.
   370. bachslunch Posted: December 17, 2018 at 02:52 PM (#5798453)
Suppose we were to start a Strat-o-Matic league (or DMB or whatever). Now, suppose everybody got 1980 Mike Schmidt for their team. Does having 1980 Mike Schmidt on your team give you an advantage over anybody else in the league?


So, why wouldn't we enshrine all of these 3B if they have an elite career? Especially if they're all better than 3B from other eras?

Sorry if I seem obtuse, but I don't get this.
   371. Kiko Sakata Posted: December 17, 2018 at 03:01 PM (#5798456)
There's nothing definition-wise that should constrain it to zero at each position. I guess I'd ask the follow up question; if every MLB SS was between 0 and 2.0 WAR for a season with a weighted average around 1.2 because there's no stars, should average be defined down to 1.2 for that year?


The constraint is that you have to field a player at every position. If I have a below-average 3B and two above-average SS and you have a below-average SS and two above-average 3B, we can both improve our teams by having me give you one of my above-average SS in exchange for one of your above-average 3B.

If every team had Mike Schmidt at 3B and and Larry Bowa at SS, would they be equally valuable? If so, you'd be willing to trade them even up. Would you rather be the guy with Mike Schmidt at both SS and 3B? Or the guy with 2 Larry Bowas?


If Mike Schmidt could provide more value at SS than an average SS, then sure, you'd rather have two Mike Schmidts, because that would give you an average 3B and an above-average SS.

But in real life, why do we assume that an average 3B would be an above-average SS, which is what BB-Ref is doing in the 1970s?
   372. Kiko Sakata Posted: December 17, 2018 at 03:05 PM (#5798462)
So, why wouldn't we enshrine all of these 3B if they have an elite career? Especially if they're all better than 3B from other eras?


The best OPS from 1964-68 of any player with at least 2,000 PAs was Willie Mays at .933. From 1998 - 2003, there were 26 players with an OPS greater than .933. Were there 26 hitters in the latter time period who were more valuable hitters than Willie Mays?
   373. Carl Goetz Posted: December 17, 2018 at 03:24 PM (#5798474)
"The constraint is that you have to field a player at every position. If I have a below-average 3B and two above-average SS and you have a below-average SS and two above-average 3B, we can both improve our teams by having me give you one of my above-average SS in exchange for one of your above-average 3B."
Right, but that's not always your preferred player. In a year where a lot of 3B get injured, a lot more bad 3B get PT which reduces the average. Yeah, its better for your chances of winning if you have a slightly above average guy who didn't get hurt, but that doesn't suddenly make him a star. It just means 3B is weak that year.

"But in real life, why do we assume that an average 3B would be an above-average SS, which is what BB-Ref is doing in the 1970s?"
He'd be well above average on offense. And BBref is giving 3 runs for 3B and 9 for SS for much of the 70s. The positional adjustment for SSs is still 6 runs more than 3B. And technically all its saying is that the theoretical average 3B was slightly more valuable to his team in terms of wins than the theoretical average SS during that period. Remember, "average" is just a weighted average of everyone who played that position in that year. There may even be some gray area from when a player moved from one position to the other mid game. Value is really about distribution too. In the 13 Schmidts and 13 Reitzs example, the 13 Schmidts are all going to be drafted highly because there's no middle ground at the position and no one wants to be in the 13 team group with Reitz at 3B. In the real world, the distribution is much less drastic but you can still get gluts at various levels of the talent pool. Those gluts are usually bigger at the lower talent levels but that's not always a perfect relationship. Again, using average (mean) is really only a proper baseline if the distribution is normal.
   374. Kiko Sakata Posted: December 17, 2018 at 03:35 PM (#5798483)
And technically all its saying is that the theoretical average 3B was slightly more valuable to his team in terms of wins than the theoretical average SS during that period.


Yes, that's what it's saying. And what I'm asking is, what is the basis for assuming that? What is the empirical evidence that third basemen provided more value to their teams overall than shortstops. To say, "Well, they had more WAR" is to beg the question.

As for the injury hypotheticals, those are arguments for smoothing positional averages in adjacent seasons. That's a perfectly valid thing to do (although I don't do it). I would have no objections to WAA for 3B in the 1970s going -0.1, +0.2, -0.2, +0.1, +0.1, 0.0, -0.1, -0.1, +0.2, +0.3 even granting that adds up to +0.4. But, again, total WAA for MLB third basemen was positive (not even "not negative" it was positive) for 31 consecutive years. Not, the sum of the 31 years was positive. There were no negative numbers from 1959 through 1989. I don't buy it.

We evaluate Willie Mays's hitting by comparing his hitting to the hitting of his direct contemporaries. We compare Willie Mays as a hitter to hitters from other generations by comparing his relative hitting (e.g., OPS+) to the hitting of other players relative to their contemporaries. Why does it not make sense to evaluate third basemen in exactly the same way?
   375. Rob_Wood Posted: December 17, 2018 at 03:57 PM (#5798492)
This has been an on-going discussion for going on 20 years in the sabermetric community. Some people sincerely believe that, essentially, the positional adjustment's main purpose is to ensure that WAA be 0.0 for every position for every season (or whatever the relevant time period is).

Other people do not share that viewpoint.

I am firmly and decidedly in the second camp.
   376. Carl Goetz Posted: December 17, 2018 at 04:30 PM (#5798501)
"Yes, that's what it's saying. And what I'm asking is, what is the basis for assuming that? What is the empirical evidence that third basemen provided more value to their teams overall than shortstops. To say, "Well, they had more WAR" is to beg the question."
Agreed, but there's also no basis for saying they should be equal.

"As for the injury hypotheticals, those are arguments for smoothing positional averages in adjacent seasons. That's a perfectly valid thing to do (although I don't do it). I would have no objections to WAA for 3B in the 1970s going -0.1, +0.2, -0.2, +0.1, +0.1, 0.0, -0.1, -0.1, +0.2, +0.3 even granting that adds up to +0.4. But, again, total WAA for MLB third basemen was positive (not even "not negative" it was positive) for 31 consecutive years. Not, the sum of the 31 years was positive. There were no negative numbers from 1959 through 1989. I don't buy it."
It does seem odd, but it was also a period of a lot of great 3B and a relative dearth of great SSs. That will skew the averages some.

"We evaluate Willie Mays's hitting by comparing his hitting to the hitting of his direct contemporaries. We compare Willie Mays as a hitter to hitters from other generations by comparing his relative hitting (e.g., OPS+) to the hitting of other players relative to their contemporaries. Why does it not make sense to evaluate third basemen in exactly the same way?"
I compare both Willie Mays and all 3B (and everyone else) to replacement level. OPS+ is an improvement over traditional stats, but I don't use it at all for Hall of Merit research. Obviously averages are needed to calculate War, but that in no way means that all positions will average out all of the time. Do you have an expectation that OPS+ will average 100 at each position? Or do you expect to average out for the league as a whole? In a given season, I expect the total War to be 1000 and total WAA to be 0 because that's how the system is set up. I also expect each position's total innings to be equal to each other since you have to be able to field a full team each inning. There's no reason for the offensive numbers to be close though; some positions are more likely to be removed for a PH for example. Beyond that, there's an almost limitless number of ways those War and WAA amounts can be distributed to individual players. I personally feel replacement level captures that distribution better than average.

I'm not even saying you're wrong about the positional levels BBref uses. I'm just saying that there's no reason to believe WAA should be equal across positions, so I need more evidence than that to base an adjustment on. Average in this case is the mean of an extremely skewed distribution with a great deal of random variance thrown in for spice. Its really hard to force fit that into a sterilized mold.
   377. Kiko Sakata Posted: December 17, 2018 at 05:08 PM (#5798521)
I personally feel replacement level captures that distribution better than average.


Despite its name, WAR builds everything off of average. The shift from WAA to WAR is just a final step that shifts the baseline from average to replacement level proportional to playing time. In BB-Ref that shift's accomplished by what they call Rrep. From what I can tell, Rrep is just directly proportional to plate appearances. Here's the 2018 Cubs. If you sort them by Rrep, Anthony Rizzo (1B) and Javier Baez (2B/SS) are at the top with 21 Rrep; they were 1-2 on the Cubs in plate appearances (665 and 645). Ian Happ and Addison Russell had 462 and 465 PA, respectively, but played very different positions (Happ was mostly an OF with a smattering of 2B/3B; Russell only played SS); they both have 15 Rrep.

I'm not saying this is wrong or right or bad or good (as far as I can tell, it's fine and does exactly what it's supposed to). I'm just saying WAA and WAR are built from the exact same underlying assumptions and will, therefore, have the same basic distribution by construction.
   378. Esteban Rivera Posted: December 17, 2018 at 05:22 PM (#5798527)
Hey guys, I’m back after 4 years since my last ballot. It's been a long while, so I’m reworking things from 'scratch' so to speak.

I’m building my consideration set again using the fantastic info at Baseball Gauge. For my consideration set sweep, I’m using a combination of wRAA base for the offense (or Baseball-Reference WAR offensive component) and DRA base for the fielding component (or Baseball Gauge WAR fielding component). For this sweep of position players (or to be more precise, players who are classified as a position player due to value generated playing a position other than pitcher), I’m splitting the pitching component as 50% B-R and 50% BG. I’m also splitting the positional adjustment 50%/50% (differences are only decimal points). No regression done to the fielding component. Finally, for this sweep, the WAA number will be the total of all WAA with negative seasons zeroed out, meaning that if the total sum of that season’s value (after combining the offensive, fielding positional adjustment and pitching if applicable) is negative, then that becomes zero in the sum. I’m not zeroing out the individual components if they are negative, only the total sum. So a season with positive fielding value that is wiped out by negative offense and/or the positional adjustment would be zeroed out in the total and that particular fielding value would not show up in the fielding column total for this exercise. Same applies to positive offensive value wiped out by fielding and/or positional adjustment, or even positive offense and fielding wiped out by positional adjustment value or even pitching.

I’m using a benchmark of approximately 25 WAA at the positions to build a consideration set, as well as adding some notable names that may be in the 20-25 WAA range for evaluation as well. Note that this is to build a consideration set to analyze further, not to decide who I’m supporting or making the final ballot. I’m also going to add players that are around this benchmark on offense only or defense only to the consideration set. One thing I do need to adjust for is for differing schedule lengths and other assorted credits (such as missing seasons due to military/war service, integration/NGL credit, cases where trapped in minors such as PCL can be a difference maker, etc.). Hope to be able to do a ballot for this year, but the previously described process is mainly to build a consideration set. There's a lot of interesting discussion going on, so I'll chime in later on some thoughts I've been mulling over just to see if I'm off base, need to rethink or am on the right path. It's nice to get back on this horse.
   379. Carl Goetz Posted: December 17, 2018 at 06:40 PM (#5798542)
"Despite its name, WAR builds everything off of average."
True, but I think its a little more intuitive presented from Replacement. Particularly when you are adding multiple years together when some of those years were slightly below average but still useful. You can tell more from a War career total than a WAA career total.
   380. Bleed the Freak Posted: December 17, 2018 at 09:53 PM (#5798563)
Welcome back Esteban, and great discussion all
   381. bachslunch Posted: December 18, 2018 at 08:41 AM (#5798601)
@372: maybe I'm not expressing myself well.

The best starting pitchers from 1890 to 1910 would include folks like Cy Young, Kid Nichols, Amos Rusie, Ed Walsh, Christy Mathewson, and Eddie Plank. The best starters from the 1980s would include folks like Jack Morris, Bert Blyleven, early Roger Clemens, Nolan Ryan, and Dave Stieb. I'd sooner choose the former group over the latter, even though they played earlier.

For third basemen from 1900-1950, your best choices would be folks like Home Run Baker, Pie Traynor, Stan Hack, Bob Elliott, Heinie Groh, Larry Gardner, Tommy Leach, Lave Cross, Bill Bradley, and Harlond Clift. From 1951-2000, you've got Eddie Mathews, Ron Santo, Brooks Robinson, Mike Schmidt, Wade Boggs, George Brett, Graig Nettles, Ken Boyer, Dick Allen, and Buddy Bell. It's no contest for me comparing these two groups which I'd choose.

Some eras are blessed with riches at certain positions, while others aren't. Why not induct accordingly?
   382. progrockfan Posted: December 18, 2018 at 09:11 AM (#5798607)
@bachslunch -

If I might be allowed to paraphrase, I think the point you're making is this:

Certain eras feature unusual concentrations of great players at certain positions.

The HoM should acknowlege that not all eras will feature the same number of superstars at each position.

If the 1950s feature a concentration of superstars in center field, the HoM should honor these stars without attempting to enshrine a similar number of players at the same position from each competing era. Just because Mays, Mantle, Snider, Doby and Ashburn happened to play at the same time doesn't automatically mean that the 1990s, say, will also feature five HoM-worthy center fielders.

...Have I got that right? If so, I agree 100%.

Yet another reason why reliance on any one stat or set of stats must, in my view, be tempered with other forms of evidence.
   383. Kiko Sakata Posted: December 18, 2018 at 09:34 AM (#5798619)
Some eras are blessed with riches at certain positions, while others aren't. Why not induct accordingly?


I think that one needs to be very careful that one doesn't beg the question.

One of the earliest and most universally accepted tenets of sabermetrics is that in order to compare hitters from different eras, one has to place a hitter's performance within the context in which it was made - controlling for the park and league in which the player played. We do these adjustments on a year-by-year basis and nobody would consider doing otherwise. Another generally accepted tenet of sabermetrics (and, really, of baseball analysis generally) is that one has to balance the offensive and defensive requirements of a position.

We can't simply compare the hitting value of first baseman and a shortstop and call it a day - Paul Konerko and Derek Jeter were of similar value as hitters and both were lousy defenders at their position. Nobody thinks that the only difference in value between the two players is Jeter's baserunning and longer career. He was also more valuable because the level of offense these two players provided is simply more valuable coming from a shortstop than coming from a first baseman.

Everybody understands the need to include positional adjustments. The question is what positional adjustments to use. And my contention is that if you have a set of positional adjustments which conclude that the position of third base was above average, in the aggregate, for 31 consecutive years. Again, I need to emphasize: not above-average over a 31-year period; above average each and every one of those 31 years. If you have a set of positional adjustments which conclude that the position of third base was above average for 31 consecutive years, I think there's probably an error in your positional adjustment calculation.

Ernie Banks's last season as a shortstop was 1961. Robin Yount's first season that really adds to his Hall-of-Merit case was 1978. From 1962 - 1977, there are no Hall-of-Merit players having Hall-of-Merit caliber seasons while playing shortstop. That's 16 seasons, 0 shortstops.

Meanwhile, within those 16 seasons, we have the following Hall-of-Merit players having Hall-of-Merit caliber seasons at third base: Ken Boyer, Brooks Robinson, Ron Santo, Dick Allen, Mike Schmidt, George Brett, Darrell Evans, Graig Nettles. That's 8 players of this era that I think most people would classify as third basemen. In addition, Joe Torre won an MVP award playing third base. Harmon Killebrew played 105 games at third base in his MVP season. Pete Rose had several seasons that add to his Hall-of-Merit case where he played third base. That brings us up to 11 Hall-of-Merit players who had Hall-of-Merit caliber seasons playing third base. Eleven third basemen, zero shortstops.

Personally, I'd vote for Sal Bando and Ron Cey before Buddy Bell, but that's irrelevant to this discussion. By voting for Buddy Bell over Dave Concepcion or Bert Campaneris, you're saying that you think that, at worst, the 9th-best third baseman of this era (ignoring Torre, Killebrew, and Rose) is better than ANY shortstop from this time period. I think that's unlikely.

Setting positional adjustments that change over time doesn't mean that the best shortstop will always be equal in value to the best third baseman of an era. He won't be. Mike Schmidt, George Brett, Ron Santo, and (depending on how you value peak vs. career) Dick Allen and/or Brooks Robinson are clearly more valuable than any shortstop between Banks and Yount. I just think that if you can't find ANYBODY at a position over a 15-20 year period who's worthy of the Hall of Merit, there's probably something wrong with your positional adjustments.
   384. bachslunch Posted: December 18, 2018 at 09:57 AM (#5798637)
@383: you're comparing SS and 3B from the same era, right? I was comparing 3B from two different time periods. I'm not advocating that everyone at every position who hit better than the best SS should be in. That's not the same thing. Or what poster progrock said above. Still confused here.
   385. JoeD has the Imperial March Stuck in His Head Posted: December 18, 2018 at 10:15 AM (#5798648)
Re positional adjustment ... I do not think WAA at positions should be equal.

But what should be roughly equal (assuming equal playing time) is the WAR for the bottom 10-15% of regulars. Not every year of course, there are always fluctuations. But if you use a rolling 3 or 5 year average or something similar this should be equal. Maybe not in 1870. But the game reached an equilibrium on this surely by 1900 or so. Probably earlier. That's why we typically think they overvalued all of these medicore hitting 1B from before 1920. We were vastly underrating the defensive requirements. Catching throws with barehands, fielding a game that was centered in the infield, not the outfield, etc.

WAA is prone to star gluts. WAR for the worst players is not.

It isn't like there isn't an explanation for why in the 1970s and early 1980s SS/2B is underrated by most non-DanR WAR methodologies. Teams didn't all of the sudden all get dumb about how to value players.

Artificial turf is the reason. You could no longer hide a mediocre fielder as SS/2B/CF. The ball gets there too fast.

As the required defensive skill goes up, the bats will naturally go down. You cannot sacrifice as much fielding for hitting anymore.

Unless I see evidence to the contrary, I am convinced that SS in the 1970s required more defense quality than any position/era of the modern game. DanR's WAR shows this. Which is part of why he (and those of us that use it) rate Concepcion and Campaneris higher than what you'd get out of traditional WAR.
   386. Carl Goetz Posted: December 18, 2018 at 10:30 AM (#5798653)
"Everybody understands the need to include positional adjustments."
The point of the positional adjustments is to adjust for the fact that an average defensive SS is more valuable on defense than an average defensive 3B who is in turn more valuable than the 1B on defense who is in turn more valuable than the DH (again on defense). If there are a group of players who are offensively making up that defensive difference and more(such as 3B were vs SS from 1959-1989), you can have a period where the average 3B is better than the average SS.
The point of positional adjustments isn't affirmative action for low offense positions. Its to recognize that different defensive positions provide different overall defensive value to the team. This is why dWAR on BBref includes the positional adjustment. What you do at the plate and the on the base paths is completely separate. There's not 1 set of averages for SS and 1 for 3B. All hitters are compared with all other hitters (I believe pitcher hitting may be separated, but can't remember for certain). If a catcher hits a HR on offense, it doesn't count more towards their offensive WAR than if he were a 1B. It shouldn't since it doesn't matter what position he plays in the field while he's batting.

I'll put it another way. What mechanism exists that would make the average 3B more valuable on offense than the average SS by the exact same amount that the SS is better on defense?

Now again, I'm not arguing that you're wrong about problems with the way BBref (or anyone else) calculates the positional adjustment. I'm arguing that I would need to see flaws in the BBref methodology for calculating positional adjustments to believe there are flaws. Telling me that the averages do not equal when there's no logical reason that they should doesn't prove a thing to me. Beyond being an interesting fact that jives with my historical observations of the period, 3B averaging .6 WAA tells me nothing. There's a difference in the numbers. It could be a real difference in the real world. It could be a flaw in their calculations. It could be that the positional adjustment is too great for 3Bs or too small for SSs. It could be that the collection of guys who played 3B during the period were just more valuable than the collection of SSs at the same time. It could also be that the positional adjustments are giving SSs too much credit and that the 3B were even more valuable than War/WAA thinks.
As a more extreme (but very real) example: there aren't a lot of great Negro League RFs and 2Bs. Is that a reason to try and force fit Marv Williams or George Scales into the Hall of Merit? Or do we just elect the great CF, SS, etc that appear to have been more valuable? I'd go with the latter.
   387. Howie Menckel Posted: December 18, 2018 at 10:40 AM (#5798659)
It isn't like there isn't an explanation for why in the 1970s and early 1980s SS/2B is underrated by most non-DanR WAR methodologies. Teams didn't all of the sudden all get dumb about how to value players.

I disagree. Clearly quickness at SS became more valuable in an artificial turf era - I just think teams overcorrected on it.

Larry Bowa got 9100 PA with a 71 OPS+ (and 15 HR). geesh.

his teams won plenty of games - but so did the Yankees with Jeter's defense at SS.
   388. Kiko Sakata Posted: December 18, 2018 at 10:57 AM (#5798667)
you're comparing SS and 3B from the same era, right? I was comparing 3B from two different time periods.


It's the same argument. Over the three decade time period from 1960 - 1989, Buddy Bell was probably somewhere between the, what, 10th and 12th best third baseman in baseball? How many third basemen from the 1920 - 1949 time period has the HOM elected? One? I know they elected Stan Hack but I'm seriously drawing a blank on who else they might have elected. If you're coming to the conclusion that the 12th-best third basemen from one time period is better than the 2nd-best third baseman from another time period, you should probably double-check your assumptions. It's not impossible - and setting positional averages so that all positions are equally valuable every season does not preclude the possibility - but it sure seems pretty damn unlikely.
   389. Kiko Sakata Posted: December 18, 2018 at 11:02 AM (#5798672)
Clearly quickness at SS became more valuable in an artificial turf era - I just think teams overcorrected on it.


Maybe. But that's an untestable assumption. If "quickness at SS became more valuable in an artificial turf era" that argues for a different positional adjustment for SS in the 1970s. Which, right off the bat, means you're acknowledging an error in how BB-Ref calculates WAR. But how do you KNOW that teams over-corrected for it? You're just assuming your answer.
   390. Howie Menckel Posted: December 18, 2018 at 11:02 AM (#5798673)
You're just assuming your answer.

am seeing a lot of that around here
   391. Carl Goetz Posted: December 18, 2018 at 11:22 AM (#5798690)
"If "quickness at SS became more valuable in an artificial turf era" that argues for a different positional adjustment for SS in the 1970s. "
Not necessarily. If BBref is already accounting for it correctly, it wouldn't. BBref's positional adjustments do change over time so just saying SS on artificial turf is harder does nothing to argue that they aren't accounting for it properly. Its again "assuming the answer". So far, Joe has presented a reason that SSing may have been more difficult than we think on artificial turf. But, I have no idea how this DanR calculates his positional adjustment and only a vague one on how BBref calculates theirs. I need more information to decide if the first half of your sentence actually argues for the 2nd half.
   392. Kiko Sakata Posted: December 18, 2018 at 11:33 AM (#5798698)
If BBref is already accounting for it correctly, it wouldn't. BBref's positional adjustments do change over time


Not so you can really notice it (I'm pretty sure they do change somewhat; but the changes are pretty damn tiny). Look at Rpos for Joe Cronin, Bert Campaneris, and Derek Jeter. All three of them have Rpos that maxes out at 10 in seasons where they play 150+ games at shortstop.
   393. bachslunch Posted: December 18, 2018 at 11:34 AM (#5798699)
@388: I've got my best six 3B not in the HoM as Bell, Bando, Cey, Ventura, Elliott, and Harrah in that order. Are you saying I should move Elliott to the head of the line just because his period is underrepresented? If so, what's the objective reason that's not simply position era fairness? If you head down the BBRef WAR list for 3B for the era, you start coming across names like Harlond Clift, Willie Kamm, and Jimmy Dykes. Should we be giving them some kind of preference above guys like Bell, Bando, Cey, and Ventura as well?
   394. Kiko Sakata Posted: December 18, 2018 at 11:39 AM (#5798701)
I have no idea how this DanR calculates his positional adjustment


If I remember correctly, Dan R (Rosenheck) calculates his positional adjustments essentially the same way I do - i.e., they're empirical and allowed to change over time (I don't recall if they tied to empirical average every year or merely across multiple years). Dan also calculated replacement level empirically by position. That is, replacement level for shortstops was calculated by looking at "replacement level" shortstops. I don't remember exactly how he defined who was "replacement level". He had a really nice system. Unfortunately, I think he stopped calculating his numbers at some point and I think that point was long enough ago now that we're missing some data from him for players who are now eligible for the Hall of Merit.
   395. Kiko Sakata Posted: December 18, 2018 at 11:59 AM (#5798715)
Are you saying I should move Elliott to the head of the line just because his period is underrepresented?


No, I'm saying, are you sure that you're not over-rating post-expansion third basemen relative to pre-expansion third basemen? Are you judging Elliott against the correct standards?

Elliott's fielding is judged - in any fielding system - against the other third basemen of his time period. Is Elliott's bat also being judged against the batting standards of third basemen of his time? At BB-Ref, Buddy Bell seems to be getting a positional adjustment of 4 runs per year while Elliott is getting a positional adjustment of 0. Is that a fair adjustment? I don't really see how it is. That implies that third basemen of Bell's time were expected to provide 4 fewer runs of offense per season than third basemen of Elliott's time. That's not consistent with the average offensive performance of third basemen in these two eras.

Even correcting this, I still have several more recent third basemen rated higher than Bob Elliott - Toby Harrah (who I think of as a SS, because it's his SS years that push him to the front of this list), Ron Cey, Matt Williams, Sal Bando, Al Rosen (I like peaks). None of whom made my ballot (Harrah was closest - again, in large part because he was a shortstop who could hit in an era when very few shortstops could hit).

Letting positional averages change over time doesn't preclude the possibility of star gluts. Pie Traynor probably was the greatest third baseman of the 1920s. And Pie Traynor is not on my HOM ballot and is not terribly close to it - he'd probably fall between Rosen and Elliott, although they're all far enough down the list that I haven't thought too long or hard about their exact rankings. I don't vote for Buddy Bell more because I like some of his contemporaries better than him (Bando, Cey) and because I value pitchers more than the overall HOM. But I also think that BB-Ref over-rates Bell (and Bando and Cey), in part, because they over-rate third basemen during his career.
   396. Carl Goetz Posted: December 18, 2018 at 12:13 PM (#5798726)
If I read the old thread correctly (which seems to jive with what you are saying), DanR doesn't have a positional adjustment at all strictly speaking. He just calculates replacement level as the average of the worst 3 regulars at each position which would then "bake" in the positional adjustment with replacement level.
On one hand, I understand what's being done. I see the problems he describes with using averages to calculate replacement level. I do have some issues with this methodology though as well.
1) 3 Regulars is a small sample size.
2) Worst three regulars should be above what we typically consider to be replacement level. If replacement level is truly freely available minor league talent, the average replacement level player should be a backup, not a regular.
3) This assumes efficient markets for players. There are usually some backups at positions who are better than some other teams starters but are blocked from starting on their own team. This partially offsets my objection #2 above since we do see approximately replacement level (or below) players getting regular playing time in every season. I'm not sure if there's 3 for every position though.
4) Replacement level is no longer tethered to a specific replacement team performance level. Not sure if I have a problem with that, but thought it worth noting.
5) This does have the problem of assuming defensive value based on offensive performance. The 3 replacement players are total players (hitters/baserunners/fielders). ie no replacement hitters or replacement fielders (which I believe to be correct). The problem is, we don't know what portion of a replacement SS is replacement level and what portion is the positional adjustment since they are baked together. We are effectively determining the relatively positional values on defense by looking at the total package of the player. This requires the assumption that replacement level players are equal regardless of position. I'm certainly more comfortable with this assumption than the assumption that average players should be equal since the closer one gets to replacement level, the more players there are at that talent level, and the more interchangeable they should be. That said, different types of replacement players exist at different positions. For example, a young SS is more likely to be brought up initially because his "glove is big league ready" (think Orlando Arcia) whereas a 1B/LF is more likely to come up with an iron glove and a big league ready bat (think Kyle Schwarber). I'm not sure how I feel about this in the big picture, but I'm definitely uncomfortable with a few players' bats affecting what is ultimately a defensive adjustment.
   397. Carl Goetz Posted: December 18, 2018 at 12:23 PM (#5798731)
"That implies that third basemen of Bell's time were expected to provide 4 fewer runs of offense per season than third basemen of Elliott's time. That's not consistent with the average offensive performance of third basemen in these two eras."
No, it implies that the defensive value of an average 3B was 4 runs higher (relative to the rest of the defensive spectrum) during Bell's era than Elliott's. Positional adjustment is a purely defensive measurement and has nothing to do with offense. Elliott could make this difference up by being 4 runs better than Bell with offense true; but he could also make it up by being 4 runs better on defense compared with the average of his time than Bell is compared to his time.
   398. Kiko Sakata Posted: December 18, 2018 at 12:32 PM (#5798737)
Let me throw out another way of thinking about this. This is me thinking out loud. Let's suppose that the average shortstop in the 1970s was better defensively than, say, the average shortstop in the 1990s. I don't think this is terribly controversial. In the 1970's, teams felt they needed to emphasize shortstop defense because of the prevalence of Astroturf and the number of speedy players in the game. In the 1990s, teams found they could - and perhaps should - trade defense for offense even at shortstop. Derek Jeter would, of course, be the quintessential example here.

In any fielding system, Dave Concepcion is being measured against the average fielding shortstop of his era. But if the average shortstop is better in the 1970s than in other eras, isn't this under-rating Concepcion's fielding relative to shortstops from other eras? And wouldn't one of the benefits, then, of boosting Concepcion's positional adjustment be to give him this missing credit for his fielding? In other words, if managers are trading hitting for fielding at a particular position in a particular era, whether they're right or wrong to do so, isn't some positional adjustment appropriate to be able to compare the fielding value of players at the same position across eras?
   399. Kiko Sakata Posted: December 18, 2018 at 12:35 PM (#5798743)
the defensive value of an average 3B was 4 runs higher (relative to the rest of the defensive spectrum) during Bell's era than Elliott's.


How would you calculate such a thing? It could well be a lack of imagination, but I can't think how to calculate a positional adjustment based purely on defensive stats.
   400. Jaack Posted: December 18, 2018 at 12:53 PM (#5798757)
BBRef's positional adjustment for 3rd basemen in the 70s is about +4 runs per year.

For 3rd basemen in the 50s it's 0 runs per year.

That is, frankly, absurd. It's clear the position is much stronger in the seventies than in the fifties. Pretty much any standard you use will tell you third base was a stronger position in the seventies. I haven't seen anyone argue why the positional adjustment should be higher for 3rd basemen in the 70s than in the 50s.

We're not lacking 70s third basemen. It seems clear to me that BBRef WAR is systematically overrating 70s third basemen.

Needless to say, I think I might b overrating Buddy Bell. And he doesn't make my top 50.
Page 4 of 6 pages ‹ First  < 2 3 4 5 6 > 

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Dynasty League Baseball

Support BBTF

donate

Thanks to
Francis
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 1.1666 seconds
41 querie(s) executed