|
|
Hall of Merit— A Look at Baseball's All-Time Best
Tuesday, April 17, 2007
|
Bookmarks
You must be logged in to view your Bookmarks.
Hot Topics
Reranking First Basemen: Discussion Thread (18 - 10:10am, May 27)Last: TomHReranking Shortstops Ballot (10 - 5:16pm, May 25)Last: Chris Cobb2024 Hall of Merit Ballot Discussion (113 - 9:59pm, May 23)Last:  Howie MenckelCal Ripken, Jr. (15 - 12:42am, May 18)Last: The Honorable ArdoNew Eligibles Year by Year (996 - 12:23pm, May 12)Last:  cookiedabookieReranking Shortstops: Discussion Thread (67 - 6:46pm, May 07)Last: cookiedabookieReranking Centerfielders: Results (20 - 10:31am, Apr 28)Last: cookiedabookieReranking Center Fielders Ballot (20 - 9:30am, Apr 06)Last: DL from MNRanking Center Fielders in the Hall of Merit - Discussion Thread (77 - 5:45pm, Apr 05)Last: Esteban RiveraReranking Right Fielders: Results (34 - 2:55am, Mar 30)Last: bjhanke2023 Hall of Merit Ballot Discussion (376 - 10:42am, Mar 07)Last:  Dr. ChaleekoReranking Right Fielders: Ballot (21 - 5:20pm, Mar 01)Last: DL from MNRanking Right Fielders in the Hall of Merit - Discussion thread (71 - 9:47pm, Feb 28)Last: GuapoDobie Moore (239 - 10:40am, Feb 11)Last:  Mike WebberRanking Left Fielders in the Hall of Merit - Discussion thread (96 - 12:21pm, Feb 08)Last: DL from MN
|
|
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.
1. John (You Can Call Me Grandma) Murphy Posted: April 17, 2007 at 02:12 AM (#2336707)1. take one Nellie Fox
2. add a pile of walks
3. bake for a few extra years
4. serve Willie Randolph in first year eligible
"If, then" is not a good way to analyze the situation. The question is whether Randolph is really the next best candidate.
bb-ref translated stats for Randolph (top) and Fox, two gold glove 2Bmen:
.. G .. AB .... R ... H .. 2B . 3B HR RBI . BB . SO .SB .Avg .OBP SLG OPS
2257 8346 1380 2379 342 .69 .55 .757 1345 689 296 .285 .384 .362 .746
2445 9554 1335 2769 370 116 35 .824 ..746 225 ..76 .290 .349 .364 .713
If Nellie were still eligible, Randolph would be higher on my ballot as of this morning. This surprises my gut feel; I never thot of Randolph as all that special. I think some of that is that by missing a few games per year, he never scored 100 runs in a season. But scoring 80 runs/yr while playing only 130 games and not making a ton of outs is pretty durn valuable to your team.
There was a side-by-side shot of him and Franco in the dugout the other day. Franco (who's four years younger) looked at least ten years older than Randolph. Maybe fifteen.
I love bb-ref's new toy. Let me add a third 2nd baseman to the list
(my apologies in advance if this doesn't line up right)
.. G .. AB .... R ... H .. 2B . 3B HR RBI . BB . SO .SB .Avg .OBP SLG OPS
2257 8346 1380 2379 342 .69 55 .757 .1345 689 296 .285 .384 .362 .746
2445 9554 1335 2769 370 116 35 .824 ..746 225 .76 .290 .349 .364 .713
1872 7082 1167 2174 344 140 83 .960 ..716 290 337 .308 .376 .431 .807
The bottom player is one of my best friends, Larry Doyle. Is it possible that some voters mark down Doyle for the runs he gave away in the field without crediting him for the runs he produced at the plate?
If you rate players per game played, this is moot.
If you rate players by WS/WARP/RCAP (actual PA), a player who batted leadoff his entire career should have his offense value lowered by about 5%.
Of course, Randolph was never destined to have all that many RBI, no matter where you batted him. Not with his TB/PA.
Hey, just like SS in the 1970s! ; )
In short, Frank Howard was Al Simmons. Luzinski was Tom Tresh. Other than that they're pretty similar.
Or IOW, Frank Howard is to Greg Luzinski as Willie Randolph is to Horace Clarke or Johnny Ray or Juan Samuel, at least in terms of peak.
I'm still one of Doyle's biggest fans. What's clear to me, though, is that the electorate is particularly dismissive of glove-men who have mediocre gloves. You really need a WS letter grade of B or above to get any traction as a 2B or SS candidate here. It doesn't help that WARP's wacko fielding replacement really hammers down these guys as well. Of course, understanding why they vote that way doesn't mean I don't wish the electorate would change their mind. :)
Just as a simple measure, 2B and 1B in both leagues had the exact same positional EqA of .272 from 1906-15. Using more sophisticated data, the worst three starting 2B in both leagues averaged 1.0 wins below overall league average (by my measure) from 1906-15, while the worst three starting 1B in both leagues averaged 0.7 wins below overall league average. Nate Silver's Freely Available Talent levels are 0.5 wins per season below the worst three regulars average for 1B, and just 0.1 wins per season below the worst three regulars average for 2B. So my calculated replacement level for 1B in 1910 is 0.7 + 0.5 = 1.2 wins below average per season, while my calculated replacement level for 2B in 1910 is 1.0 + 0.1 = 1.1 wins below average per season. 2B and 1B were most definitely equal around 1910, any way you look at it. Corner outfielders were slightly stronger, with a .278 EqA and a replacement level of 0.9 wins below average.
DavidFoss, not true about poor fielders at key positions--I vote for Toby Harrah! Derek Jeter will get in one day, and so will Mike Piazza. I don't think Kent's good enough.
I do. You've made the 2B-OF comparison before. I know you have the numbers, I was just asking to see them. :)
Well, the 1B comparison doesn't surprise me. 2B really is even with 1B in this era (especially between 1909 and 1914). I'm not sure if this is due to 2B being easier or just a down era for 1B. The 1B gap between Anson/Brouthers/Connor and Sisler has been discussed here before. The best MLB 1B of this era were Konetchy & Daubert (versus Collins, Lajoie & Doyle at 2B).
And my comment about poor fielders at key positions was directed more to the electorate as a whole and not specifically to you. I like Toby Harrah. Not enough to vote for him, but I like him. But there was a guy here a while back who tried to put down Ezra Sutton by accusing him of being the Toby Harrah of his era. There were other flaws in his argument (he didn't account for the season-length changes back then), but the point was that he was using Toby Harrah to invite ridicule. Who else is getting short shrift? I'd guess Lombardi & Lazzeri. I'm not sure we need more 1930s guys, but both of these guys lower on people's radar screens then their hitting levels would warrant.
Doyle's a big one though. I had him above Childs. He hit better and for longer, but fielding concerns put him 25-30 places below Childs in the balloting.
I know that Doyle has taken a number of knocks for playing the position he did in the league and time he did. I like middle IF defense as much as the next voter--I thought Fox was a worthy candidate, I suspect I'll vote for Randolph. I voted for Bancroft and Rizzuto. I might even vote for Davie C. someday, and I'll give them all the credit in the world for saving more runs in the field than Doyle or the mediocre fielding middle IF of your choice. But, at the same time, guys like Doyle and Lazzeri and Jeff Kent (who is a pretty good comp for Doyle in everything except personality) ought to get credit for being significantly better at generating runs at the plate than the guys mentioned above. Whether you'd rather have Larry Doyle or Nellie Fox, Vern Stephens or Phil Rizzuto, Jeff Kent or Willie Randolph on your team seems more a matter of personal preference than anything else.
I don't know why you keep saying that it's *only* Doyle's poor fielding that holds him back. It's the fielding *plus* the fact that he was playing an offense-first position. If 2B = 1B then, would we elect a stone-gloved 1B with his hitting numbers? No. Thus, we don't elect Doyle either.
Andrew M, I think Rizzuto's fielding advantage was simply worth far more runs and wins than Stephens' hitting advantage.
2B around 1910 were absolutely a dime a dozen.
Hi
can anyone (KOJK?) with Sinin's bb encyc look up Randolph's (& D Concepcion's) RCAP for me? Thanks.
Let's say that all the best-hitting shortstops of the 1990s and early 2000s had all for some strange reason been moved to second base instead. The best-hitting shortstop of the era becomes Royce Clayton, while second basemen routinely lead the league in all kinds of individual and aggregate hitting categories. Does that make Royce Clayton a HOMer?
and the answer depends on how fungible (overlap, easily movable from one to another) these middle infielders are.
we all agree that LF-RF is probably 90% fungible. CF to corner OF, less but still some there, LF to 1B a bit, 2B-SS some, 2B/SS to 3B less, C to anywhere virtually 0%.
A player who is best in MLB at his position is very valuable, since if you're tyring to win a pennant, he is the one you want to trade for. UNLESS there are 4 guys who are playing some other position that could easily be moved over if they had to, to fill in a gap.
AL 2B--Larry Gardner 2-36-.283 in 113 games
AL LF--Patsy Dougherty 1-43-.248
NL 2B--John Hummel 5-74-.244
NL LF--Your choice Fred Clarke or Zack Wheat, probably Wheat who played 150+ games versus Clarke's 120+ at age 37
Let's try 1911.
AL 2B--Amby McConnell was the only non-all-star 2B to play 100 games
AL LF--Nixey Callahan 3-60-.281
NL 2B--Dots Miller 6-78-.268 and Hummel is still around
NL LF--Clarke and Wheat still in the mix though Josh Devore may have been better
This of course doesn't prove anything, just anecdotal, but it seems that journeyman quality LF were at least a bit easier to come by for these couple of years.
TomH, any SS could play 2B, but even elite fielding 2B would be bad SS. Most 2B are failed SS, after all. I know, Nellie Fox already got elected, I know.
sunnyday2, my data are based on a 54-player sample (worst three starters at 2B in each league for 9 years), so I'd like to think I've gone a bit beyond just picking eight middle of the pack players and eyeballing them. That said, I agree that LF were *slightly* deeper than 2B for those years (0.2 wins per full season). 1B were not. If you're interested I'm happy to send you the list of players and their stats.
Ummmm.... how much longer until you guys actually start talking about Willie?
I'll disagree with this. Shoo-in HOM-ers, yes, inner circle? How much better would Collins be than Sam Crawford if not for the positional adjustment? Better Just because there's a couple of stars at a harder position and a shortage of stars at an easier position doesn't necessarily make them equivalent. Does anyone think McInnis & Collins had equal defensive skill? Does anyone think 2B couldn't shift to 1B as they got older if they hit well enough? Does anyone think a 1B could switch to 2B? These "spectrum inversions" happen from time to time but I don't see them as anything more than historical curiosities. When they needed a place in the field to try and hide Jack Fournier, those chose 1B.
I'm on board with 2B and 3B swapping places in the spectrum, that's well documented. But I'm not going to be grouping them in with OF & 1B.
Remember that the number of stars at a position has zero impact on my positional ratings. I look at the bottom of the distribution to determine positional value, not the top. The question you have to ask is, if playing 2B was much harder than playing 1B back then, why were there so many guys like Dick Egan (not counting his All-Star 1909)? He was among the worst regulars at his position in the majors, but still hit close to the league average *and* played excellent defense. At shortstop, that production would be the best at the position in MLB; at 2B it was among the worst.
Your rCAPs, sir.
Randloph 235
Concepcion 126
this is a gross overgeneralization. Most of us, given a hour of spare time, could come up with elite 2Bmen who could also play short. (Pokey Reese is the first one that comes to my mind). And, the particular skill at turning the duece would eliminate some SS from playing 2B (since not all have been tried there, it's kind of hard to assert "any shortstop could..."). When SS who start losing range, are they moved to 2B? Wouldn't that make most sense,, if 2Bmen were fialed SS? But we often see SS move to 3B or OF instead. There's probably a god large study there for someone to make....
I suppose replacement level for 2B also is affected by GM/managers decisions not to let a great hitter play there for fear of injury.
Is the tendency also for 2Bmen to miss more time with injuries? If so, this would put a premium on 2B playing time, or at least would again affect 'replacement level' in that we'd have to make sure we're not merely measuring replacement as worst three starters, if indeed each team's backups get more time at 2B. But it's only a postulation I once more make without the data to back it up - man, some of us laggards oughta get cracking the data instead of posting what-ifs.....
One item we ought to remember about Randolph, and guys like him, is that leadoff hitters get many more PA per game, but each PA is worth slightly less in run-scoring opportunity (fewer men on).
If you rate players per game played, this is moot.
If you rate players by WS/WARP/RCAP (actual PA), a player who batted leadoff his entire career should have his offense value lowered by about 5%.
Do you mean that users should deflate his offense value measured by those statistics?
quoting David Foss twice:
I'm still one of Doyle's biggest fans. What's clear to me, though, is that the electorate is particularly dismissive of glove-men who have mediocre gloves. You really need a WS letter grade of B or above to get any traction as a 2B or SS candidate here. It doesn't help that WARP's wacko fielding replacement really hammers down these guys as well.
WARP and Win Shares frequently disagree sharply about fielding value, so we consider them two independent judgments ;-)
Who else is getting short shrift? I'd guess Lombardi & Lazzeri. I'm not sure we need more 1930s guys, but both of these guys lower on people's radar screens then their hitting levels would warrant.
Er, they should be lower than their batting would warrant.
TomH
it's only a postulation I once more make without the data to back it up - man, some of us laggards oughta get cracking the data instead of posting what-ifs.....
need stable computing platform. Don't fix rotten planks; use them as excuses.
SS are usually not moved to 2B because they lack "range" in the sense of moving left or right, it's usually because they don't have the arm to make an effective throw to 1B after moving to their right.
Many players are drafted at SS- if they have the range but not the arm they move to 2b, the arm but not the range to 3B, if they are not cut out for the infield to the OF. The vast majority of these moves are made at the minor league level not in the majors.
veteran SS are moved to 3b or the OF if they can hit enough to play those positions but can no longer field SS effectively.
Do you mean that users should deflate his offense value measured by those statistics?
yes
example:
from one previous stretch of MLB data, leadoff batters had .505 runners on base on avg when they came to bat.
cleanup hitters averaged .74; it is the most prodcutive spot per PA.
At least one previous study published by SABR showed that the relative importance of an atbat varies almost exactly proportionately with the number of runners on plus the batter (bases juiced is 4 times as crucial as no one on).
So, typical cleanup hitter's PAs are (batter=1.00 + .74 runners) 1.74/1.505 = 16% more important than #1 hitters.
Of course, leadoff men come to the plate about 8% more often.
WS, WARP, RCAP, take into account more often, but not the 'less runners on' part.
There probably are other nuances, such as the increased importance of getting on when the guy following you can slug which this does not measure.
Someone with a lineup simulator (gee, me again if I got my home computer fixed..) should be able to estimate the amount of this effect.
This doesn't mean of course that the only value given to a hit should be the bases reached. Obviously a single is more important than a walk because it is more likely to advance and/or score a baserunner, etc.. But I don't like to give credit to players for having the privilege of hitting with more runners on base. The only people I like to leverage are those that are leverage by their managers relief pitchers, and I guess pinch hitters (though none of the latter are HOM candidates).
Seriously, isn't seperating leadoff hitters and cleanup hitters because of this the same as seperating hitter from good and bad teams or hitters from good and bad run environments?
1. leadoff hitters get more PAs.
2. leadoff hitters' PAs have less leverage.
you can account for both, or neither, but don't give one without the other.
WS/WARP/RCAP, I believe, are guilty of giving credit for (1) and not penalizing for (2).
Replacement level is a rate--playing time has no part in it. If your guy goes down, you plug in a replacement. If your replacement goes down, you plug in another replacement. It's the freely available level, and while not *infinite*, given the shape of the talent distribution there are presumably at least 10-20 freely available guys capable of playing around rep level at any given time at a position.
I have no idea if you're right about 2B being hurt more often. But it's certainly true that catchers have less PT, and that doesn't decrease the replacement level at all, at least in my system--remember that, at heart, my replacement levels are Nate Silver's Freely Available Talent levels, which is the average production of players over age 27 making less than twice the league minimum salary. It just means that if you don't subjectively boost catchers' value over replacement--by multiplying (increasing their playing time) rather than by adding (lowering their replacement level)--you will wind up with a HoM nearly entirely devoid of catchers.
Pesky/Stephens
Wright/Barnes
Williamson/Burns
teammates, teammates, teammates, supporting DanR's general point
Au contraire,
One view of Earl Weaver and Cal Ripken (is it only Weaver's own view?) is that numerous big men were capable of playing shortstop, and didn't because everyone followed a social norm. Compare African-American men playing quarterback until about the same time.
One view of Honus Wagner through 1902 is similar. This may be Bill James view: Wagner was a superior shortstop bud he didn't look like a shortstop, so years passed before he became a regular there.
Lassus, try the search feature to find some discussion of Willie Randolph.
I agree with you on the big point that it feels wrong to discount linear weights and other measures of run production by batting position, with a heavy discount for batting first because those batters' plate appearances have especially low relative value or leverage.
But batting position is the result of a management decision analogous to that which uses some relief pitchers in high leverage and others in low leverage situations. and at the same time analogous to that which uses starting pitchers for 200 innings, some relief pitchers for 100 innings, and some for 50 innings. Management knows that the fourth batter will appear with more runners on base, and will appear fewer times, than will the first batter.
Because Joe McCarthy was getting pretty senile has always been my theory.
I have no problem with the scenario (adjusting for # of AB's without adjusting for leverage), though I could also be convinced of adjusting for neither. I also always thought that WS did the latter. While it splits up WS by games, doens't it divide within the team by RC which would be sensitive to PA's? How does WARP adjust for the extra PA effect?
Also, I am not necessarily sold on the leverage argument that Paul makes. If I read Tom correct, and maybe I didn't, leadoff hitters have th least amount of leverage because ideally 7/8/9 are the worst hitters. However, a manager does not choose his worst player to hit in thsi position for obvious reasons. In fact, he can't because the guy batting after the worst hitters will always have the lowest leverage no matter the batting position no? So there are only a few position where a player can be properly leveraged.
Also, if you are getting leverage do leadoff hitters get any bonus points for the role in getting on base in order to leverage the #3 or 4 hitter? Shouldn't some of that leverage go to the guy who actually creates the situation by getting on base?
I am pretty sure that MLB contracts are for one year, but where do you find players mid-season for a short contract? Most of the plaeyrs are already signed and playing at a AAA club.
Paul,
I will have you know that I have earned that surname through 25 years of laziness and procrastination!
There were starting to toy with the idea of shifting Pesky to 3B already in 1947 -- swapping him with Pellagrini for a couple of dozen games. Pesky's fielding numbers look to be down that year. Baseballlibrary.com says that Pesky got married in the offseason and gained 30 lbs. I'm guessing that's one of the reasons they went out and got Stephens in the first place... they had already decided to move Pesky.
According to BP, Stephens' fielding numbers in 1946 and especially 1947 are fairly solid. It could be that Stephens was the better option in the spring of 1948.
This issue is famous. Both David Halberstram and Bill James have weighed in on it.
Yes he has 312, more than any other eligible 2B. And he had that year with 31. But here's the problem.
31-23-23.
I mean, even 31 is not that much. Doyle, Myer and Bobby Avila had better years and, hey, Steve Sax had a 31, and Lazzeri and Stanky had 30s. Still, ok, 31 is a nice year.
But 23 as your 2nd and 3rd best? And 54 for your top 2? Here are some of the guys who did better on all 3 of those numbers. Doyle, McAuliffe, Myer, Evers, Gilliam, Avila, Sax.
If Willie Randolph, why not Bobby Avila.
Randolph 31-23-23/114
Avila 34-24-24/124
Randolph .276/.373/.351/105
Avila .281/.359/.388/104
Randolph B+
Avila A-
Yeah, I know that Avila only had about 5200 ML PAs but he probably played as much ball as Randolph did over the course of his whole career. I understand you'd want to see the MLEs, but my gut is that they are pretty comparable for career while Avila was clearly better at his peak.
YEAR LG AGE PO AVG OBP SLG G PA AB H TB BB ops+ sfws
----------------------------------------------------------------------
1944 NL 20 2b .276 .353 .348 154 612 547 151 190 65 98 18.5
1945 NL 21 2b .263 .337 .312 114 450 404 106 126 46 81 11.1
1946 NL 22 2b .311 .392 .408 146 589 520 162 212 70 127 24.5
1947 NL 23 2b .302 .382 .403 122 491 435 131 175 56 109 19.6
1948 NL 24 2b .205 .270 .222 59 227 208 43 46 18 34 2.3
1949 AL 25 2b .300 .390 .404 140 563 497 151 201 66 109 22.6
1950 AL 26 2b .299 .387 .383 80 230 201 60 77 29 98 7.0
1951 AL 27 2b .304 .374 .410 141 602 542 165 222 60 112 24.0
1952 AL 28 2b .300 .370 .415 150 664 597 179 248 67 118 24.0
1953 AL 29 2b .286 .353 .379 141 617 559 160 212 58 98 22.0
1954 AL 30 2b .341 .404 .477 143 614 555 189 265 59 142 34.0
1955 AL 31 2b .272 .368 .400 141 619 537 146 215 82 107 20.0
1956 AL 32 2b .224 .317 .318 138 583 513 115 163 70 69 14.0
1957 AL 33 2b .268 .334 .289 129 509 463 124 134 46 73 13.0
1958 AL 34 2b .253 .349 .365 113 430 375 95 137 55 98 12.0
1959 2lg 35 2b .227 .314 .322 93 298 264 60 85 34 74 5.0
======================================================================
.282 .360 .375 2004 8097 7217 2037 2709 881 101 273.6
If you use 162adjWS, his career total is 287.
Avila 287/34-24-24-24-23-22-20-20-18-14-13-12-11
Randolph 312/31-23-23-22-22-21-20-20-20-19-18-17-15-13-11
Avila is as good or better every year through #8, then Randolph gets back the marginal WS in years #9-10-11. Then their last 2 double-digit WS seasons are about equivalent. That leaves Randolph's career advantage as 2 seasons of 17-15, his 12th and 13th best seasons. Those are the only seasons that his career advantage relies upon. Granted, Avila was only "as good" the rest of the way, unless you're a peak voter, in which case his top 6 years are consistently if slightly better than Willie's.
Avila 101*/142-27-18-12-9-9-7 (?100 in 7 of 14 full years)
Randolph 105/133-27-22-12-8-7-5-4-3-2-2-0-0 (? 100 in 13 of 14 years)
*104 in actual ML action
Pretty comp. Avila is better by an average of 1.5 points through year #7, after which Avila had no more above average years. But during Avila's below average years (in order from high to low, not chronologically) Randolph was in the 100-105 range. Avila clearly had some inferior seasons during that time but also had 5 98s. So Randolph's advantage here is pretty slight.
So offensively Randolph has a slight edge, defensively Avila has an edge. During 7 full-time ML seasons Avila averaged about 137.5 games per season, not great, but Randolph was at about 133.5. For their top 5 seasons in games played Avila was at 150-143-141-141-141, Randolph was at 153-147-144-143-142. Very slight advantage for Randolph.
In summary, Avila seems clearly to have been the better player at his peak. Randolph has the better career even looking at Avila's MLEs. But it also seems likely that adjusting to league and culture changes couldn't have helped Avila much. I can see the argument for Randolph to be ahead and I can see a peak voter preferring Avila.
Bottom line: If we are really committed to MLEs, Avila is a 287 WS 2B, the 2nd best total eligible today, and with a better peak than the top career guy. Where does he fit into y'alls ballots?
I don't buy that at all. Whatever Pesky was doing had nothing to do with getting Stephens. Stephens was a stud star, an obvious perfect Fenway fit. The only question was exactly how big of a package did the Red Sox need to present to the Browns. The Red Sox's acquisition of Stephens had nothing to do with any sort of problem with Pesky; whether Pesky was doing great or terrible had no bearing on the deal.
It was closed, for some reason. Fortunately, I have the key. :-)
Couple o' lineup sims I did:
created a typical lineup that scored 706.4 runs per 162 game season.
added to the leadoff hitter 10 singles, 10 walks, 5 doubles, 5 HR, 30 fewer outs per 665 PA (+45 OBA, +73 SLG). Then did the same with the cleanup and 9th place hitters.
new lineup with better hitters scored extra
1st batter: 32.8 runs per year. Ran 10000 seasons (1.62 million games)
4th hitter: 33.3 runs/yr
last batter: 28.0 runs/yr
I estimate the std dev of runs/yr at 0.3 for 10K trials.
Leadoff man got a few extra PAs, but it washed out with the 4th hitters' leverage.
Obviously, the 9th hitter comes up less often, and you get less benefit.
-
Somewhat related question:
Pitchers bat 9th, AND they likely hit with fewer men on; because in some crucial (men on, close game) situations, they are pinch-hit for. I wonder if when I adjust for pitchers' hitting, I'm overdoing it somewhat?
long career peakelss wonders, as seen by WARP1:
age 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39
W.R 00 75 78 70 74 87 40 69 59 81 66 55 73 44 60 39 75 24 00 00
J.B. 44 72 77 71 65 81 73 66 43 42 41 80 77 74 57 65 74 53 16 -6
5-yr moving window:
age 22 23 24 25 26 27 .28 .29 .30 31 32 33 34 35 36 37
W.R 59 77 70 68 66 67 .67 .70 .71 68 64 54 58 48 40 28
J.B. 66 73 73 71 66 61 .53 .54 .57 63 66 71 69 65 53 40
diff +7 -4 +3 +3 00 -6 -14 -16 -14 -5 +2 17 11 17 13 12
scary similar, ain't it?
all of the standard WARP1 caveats apply, of course....
Wow that is a pretty good arguement, especially since Jake Beckley might be his biggest competitor - indirectly.
And here is another one. He was primarily a leadoff and #2 man - which if using the anti-Brock arguement means he had lots of opportunities to rack up those win shares.
From Retrosheet below:
Bat Order G AB R H 2B 3B HR RBI BB IBB SO HBP SH SF XI ROE GDP SB CS AVG OBP SLG
1st 879 3437 579 941 136 31 20 248 548 14 258 9 39 26 0 57 78 118 51 .274 .373 .349
2nd 759 2881 435 813 121 19 26 289 429 8 248 17 39 24 0 46 82 88 21 .282 .376 .364
3rd 7 14 2 4 0 0 0 4 1 0 1 0 0 0 0 0 0 0 0 .286 .333 .286
4th 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 .000 .000 .000
5th 7 15 7 7 0 0 0 1 1 0 1 0 0 0 0 0 1 1 0 .467 .500 .467
6th 53 174 12 37 4 0 0 10 29 0 15 1 2 0 0 2 8 5 3 .213 .328 .236
7th 174 565 73 170 21 6 5 59 86 5 53 4 5 4 0 7 21 15 7 .301 .395 .386
8th 244 803 111 204 30 8 3 70 133 10 85 5 8 9 0 19 24 44 11 .254 .360 .323
9th 78 128 19 34 4 1 0 6 16 0 14 1 6 1 0 2 5 0 1 .266 .349 .313
(1) Randolph v. Fox: I agree that they are very similar, but I had Fox 41st last year. I just don't think these guys hit well enough. Further, I think Fox is better because (1) he has a few genuine peaky seasons and (2) he was the best 2B of his era.
(2) Randolph v. Avila/Evers/Lazzeri and a few others: I think Randolph is very comparable to a bunch of guys who we rejected long ago. For the most part, his career totals are higher than these guys but (1) some of that is the longer seasons and (2) I don't see how the difference between first-ballot HoMer and no votes for 2 guys of equal offensive and defensive abilities can possibly be one or two 15 WS seasons. Basically, the if Fox, then Randolph argument is exactly the kind of argument that (by ignoring the other similar excluded players) drove the HoF down its long spiral to mediocrity.
(3) Randolph v. Beckley: Not a bad comp, but (1) Beckley put up substantially more career value when you adjust for season length and (2) Beckley was the only successful long career 1B of his era while Rnadolph had a set of peers.
(4) Randolph v. Nettles: Teammates. Nettles ranks higher all time at his position, was a better hitter, and was an historically great fielder at his only slightly easier position. I have Nettles in the 20s and can't see putting Randolph higher.
(5) Randolph (and Fox) v. Bancroft (and Rizzuto): Let's just say I agree with the camp that has the .265-.275ish good-fielding SS's ahead of the offensively similar 2B.
(6) Randolph v. Doyle and Monroe: The cases for these 3 are so different that your milage may vary, but my methods have these two as the best available 2B by a decent margin. They are in the low 20s and edging closer to the ballot.
I currently have him just above Fox at #27
Carter and Blyleven are #'s 1 and 2 while Stieb is at #8 and could be higher (Walters is third among my backloggers)
How much of a competitor to Randolph will Stieb be?
I agree with the Nettles and Beckley comparisons, they're all in the middle of my ballot.
I agree that these are similar players, but not that Nettles was a better hitter.
By WARP1, Nettles has a .273 career EQA, Randolph .275.
By WARP2, Nettles has a .275 career EQA, Randolph .284. WARP2 is relevant here because it adjusts for the DH.
Nettles is pulled down a little bit by some bad late seasons, but if you peel off his last 700 PA in his last three seasons, he's going to come out similar to Randolph in both PA and WARP1 EQA, I think.
If they are equal offensively, the question comes down to whether a very good to excellent defensive second baseman is more valuable than an excellent to historically great defensive third baseman. WARP and WS have opinions about this: WARP gives the edge to Randolph: 549 FRAR to 435 FRAR. I don't have WS at work, but that's a comparison that also should be made.
While on the subject of Randolph's offense, it might be worth comparing Randolph to the other well-known leadoff hitter we are considering: Lou Brock.
By WARP1, Brock has an offensive edge on Randolph, .282 to .275, but by WARP2, this is reversed, Randolph leading .284 to .281. I think if Lou Brock had been a very good defensive second baseman instead of an average to below average defensive left fielder, we would have elected him easily. Brock obviously has the durability chit that Randolph doesn't, so Randolph is not simply Lou Brock + second base defense, but I think it's worth noting.
As Randolph's lack of durability may be partly due to his being a second baseman rather than a left fielder, it would be good to learn more about Randolph's history: why did he miss so many games? It would also be good to have some comparative statistics on games played per year, IF vs. OF.
Randolph's seasonal win share totals obviously suffer partly because of the durability issue, as do those of others of the infielders in the high or middle backlog. Are we properly assessing them on durability? (It's here, of course, that the issue of replacement level becomes so important -- the lower the replacment level, the more durability is rewarded.)
Nettles A- 93 dWS
Not much to choose.
Nettles 93 dWS in 2504 fielding games
much to choose!
If you want to find "fielding win shares above average" for Randolph and Nettles and you have the win shares book, find the ws/1000 def innings value that is midway between B and C in the letter grade listings for the two positions, and multiply them by the defensive innings listed for the two players in the defensive innings tables, and that will give you the value of an average fielder in those innings.
I expect that Nettles will have more fws above average than Randolph, as he has more FRAA, but you never know. I'd do it myself, but I won't have a chance to look at the book until Friday evening . . .
1. Calculate how many FWS an average fielder at the player's position would have accumulated in his playing time, and subtract. This is FWS above average.
2. Multiply by the league's runs per win (a function of its R/G) and divide by 3. This is WS-FRAA.
3. Multiply this figure by the ratio of the standard deviation of Chris Dial's ZR data to the standard deviation of WS-FRAA for the given position. This expands or contracts the scale of WS-FRAA so it is in line with play-by-play estimates.
By this measure, I have Randolph with 85 career WS-FRAA (outstanding) and Nettles with 150 (among the greatest of all time, particularly since he was below average in the latter part of his career).
By WARP1, Brock has an offensive edge on Randolph, .282 to .275, but by WARP2, this is reversed, Randolph leading .284 to .281. I think if Lou Brock had been a very good defensive second baseman instead of an average to below average defensive left fielder, we would have elected him easily.
Certainly. I suspect that Lou Brock would have been elected if he were Willie Wilson on defense --merely a good CF.
mulder is Kelly?
second base has a higher intrinsic weight than third. in other words, there are more win shares available for second basemen than thirdbasemen
This is not something to correct for unless you disagree radically with Win Shares. That is, the implied correction would be a fundamental revision whereas most corrections people make or imagine are relatively superficial.
(And I am with Win Shares on this. One shouldn't walk down the road toward thinking of "available Win Shares" as a denominator except in the case jimd(?) has noted: the ceiling for fielding value on the team level, at 33% of team defense.)
Mark Shirk
Paul,
I will have you know that I have earned that surname through 25 years of laziness and procrastination!
I had to check your response, having taken so many cracks about my surname over the years, despite the misspelling. Your response is very good to historically great!
My point about the batting order as a management decision encompassing frequency (how many times will he bat) and leverage (with what importance will he bat) should be understood at the highest level --encompassing the frequency and leverage broadly defined and jointly.
I agree with you that the first batter has more leverage than the fourth insofar as higher quality batters follow the first one than the fourth one. But the fourth batter has more leverage, net, I presume. (TomH may mean leverage more narrowly.)
By the way, while its true that the 7-8-9 batters are relatively weak, a major factor that decreases leverage for the first batter is the first inning. The first batter's first time at bat is the only one in the game that is guaranteed to have value only in the reaching base aspect, not at all in the advancing runners aspect. And those appearances constitute almost 25% of the season for an everyday leadoff man.
52. David Concepcion de la Desviacion Estandar (Dan R) Posted: April 18, 2007 at 11:08 PM (#2338929)
Why don't teams just stock their AAA rosters with freely available replacement players in the offseason?
53. yest Posted: April 18, 2007 at 11:18 PM (#2338939)
not that this is where your going but Eddie Gadell and the Tigers replacment had 1 dat contracts
The role of player contract has changed a lot in 100 years not to mention 150. Reading the May 1912 coverage in the Detroit News-Tribune (a while ago) it seems to me that Detroit already needed approval of the League Pres or National Commission to employ the replacement players. I don't recall whether they had written contracts and I'm not sure yest means that; written or spoken, they did have approved agreements to play one day for $50.
My reading of the Burleigh Grimes transaction in spring 1934 (trying to date it precisely a while ago) is that Grimes and manager met on the road and made verbal agreement about his employment. He did not sign until the team returned home to St Louis? er, maybe it was return to New York after a road trip to St Louis.
During Willie Randolph's time the control of player employment and compensation is more advanced, formalized, rule-governed. I am not sure this difference in conditions means that the concept of freely available talent must be used differently for Randolph and, say Buddy Myer and Larry Doyle, if one's focus is on those players (2Bmen) for the Hall of Merit! By the way, what is the timespan covered by FAT in the hands of its author?
You have learned the truth, Paul.
You have learned the truth, Paul.
Are you sure?
Maybe Scully is Kelly?
You have learned the rest of the truth, Jim.
Evers: shorter career, weaker league (NL<AL in 1910)
Couple o' lineup sims I did:
created a typical lineup that scored 706.4 runs per 162 game season.
added to the leadoff hitter 10 singles, 10 walks, 5 doubles, 5 HR, 30 fewer outs per 665 PA (+45 OBA, +73 SLG). Then did the same with the cleanup and 9th place hitters.
new lineup with better hitters scored extra
1st batter: 32.8 runs per year. Ran 10000 seasons (1.62 million games)
4th hitter: 33.3 runs/yr
last batter: 28.0 runs/yr
I estimate the std dev of runs/yr at 0.3 for 10K trials.
Leadoff man got a few extra PAs, but it washed out with the 4th hitters' leverage.
In a sense this result is not surprising -- if 4th hitters PAs weren't leveraged relative to leadoff hitters, it would be optimal to lead off with your best hitter. Since no one seriously considers leading off with a Ruth, Williams, or Pujols, we obviously believe that the extra PAs are generally offset by lack of leverage.
Where I'm hesitant about applying your result, however, is that the best leadoff hitters aren't just better hitters, they're players whose skills are optimized for the leadoff role. Some events, e.g., walks and stolen bases, are worth more leading off an inning and other events, e.g., extra base hits, are worth more with runners on base. Suppose you ran your simulations adding 30 walks, 30 SB, and 3 CS. My guess is that adding those skills may be worth more for the leadoff hitter than for the 4th hitter. And if my hypothesis is correct? -- well, wouldn't that suggest that leverage isn't just a constant, but a variable that depends on the batters particular mix of skills?
Over what timespan is the average FWS measured? Eg 5- or 11- or 21-year moving average.
When I look at peak, I look at it in two ways. One is total value above average, for which I subtract the value of an average player (in the WARP1 and WS systems) from the player’s value for all the seasons he is above average, then sum the differences remaining. This measure is significantly influenced by seasonal durability.
The other is peak rate. For this, I find the player’s highest rate of performance per 162 games during a period of at least five consecutive seasons. I use five seasons generally, but if the player, due to a single off year in the middle of a longer run of good years, has a better rate over six, seven, or eight consecutive seasons than over five, I use the longer period. I do this for WARP1 and WS also, adjusting for the DH by adding .4 wins in WARP1 or WS to the player’s total per 162 games played. This is a way of looking at quality of play, irrespective of playing time (although I generally do not include seasons before a player became a full time player in the time frames I examine.)
Anyway, I thought others might find it interesting to see how Randolph’s peak rate looks in relation to the “glove” candidates who are his contemporaries, so I compiled peak rate lists in WARP1 and WS, including all of the serious C-2B-3B-SS candidates whose peaks and careers fell mainly in the 1970s and 1980s.
I have put two dividing lines into the lists. One falls at 10 wins/162 games line, which I think nearly marks the line above which the candidates are nearly all obvious HoMers. The other falls below Darrell Evans, who has the lowest rate of any infielder from these decades that we have yet elected.
For ease of comparison, I have denominated the win shares list in wins. I have provided dates for the peaks in the WARP1 table. The WS peaks are usually but not always in the same set of seasons. I did not have the WS peak seasons data handy, so I omitted it. I have included with the WS list the variance of the WS peak rate from the WARP1 rate for each player.
A similar list could be compiled for Dan R.’s WAR, but I haven’t done so yet.
WARP1
Joe Morgan 14.82 (1972-76)
Mike Schmidt 12.52 (1980-84)
Wade Boggs 11.81 (1985-89)
Johnny Bench 11.60 (1972-76)
Gary Carter 11.44 (1979-83)
George Brett 11.30 (1976-80)
Bobby Grich 11.28 (1972-76)
Rod Carew 11.14 (1973-77)
Cal Ripken 10.88 (1982-86)
Ryne Sandberg 10.53 (1988-92)
Robin Yount 10.29 (1980-84)
Ron Cey 10.14 (1975-79)
Carlton Fisk 10.04 (1972-77)
-------------------------------
Alan Trammell 9.92 (1983-87)
Buddy Bell 9.72 (1980-84)
Ozzie Smith 9.66 (1984-88)
Dave Concepcion 9.51 (1974-79)
Ted Simmons 9.29 (1976-80)
Willie Randolph 9.21 (1976-80)
Paul Molitor 9.11 (1979-83)
Gene Tenace 8.95 (1975-79)
Thurman Munson 8.90 (1970-75)
Darrell Porter 8.86 (1979-83)
Darrell Evans 8.65 (1972-76)
--------------------------------
Lance Parrish 8.60 (1982-86)
Bert Campaneris 8.52 (1970-74)
Graig Nettles 8.50 (1971-78)
Lou Whitaker 8.48 (1982-86)
Sal Bando 8.08 (1969-73)
Toby Harrah 7.62 (1974-78)
Win Shares – counted in wins (WS/3)
Joe Morgan 14.34 (-0.48)
Mike Schmidt 13.01 (+0.49)
George Brett 12.31 (+1.01)
Wade Boggs 11.80 (-0.01)
Rod Carew 11.71 (+0.57)
Robin Yount 11.37 (+1.08)
Bobby Grich 10.74 (-0.54)
Paul Molitor 10.73 (+1.62)
Johnny Bench 10.70 (-0.90)
Ryne Sandberg 10.55 (+0.02)
Gary Carter 10.42 (-1.02)
Carlton Fisk 10.33 (+0.29)
Cal Ripken 10.31 (-0.57)
Alan Trammell 10.11 (+0.19)
---------------------
Sal Bando 9.88 (+1.80)
Lou Whitaker 9.66 (+1.16)
Gene Tenace 9.56 (+0.61)
Ted Simmons 9.29 (0.00)
Willie Randolph 9.24 (+0.03)
Toby Harrah 8.99 (+1.37)
Buddy Bell 8.96 (-0.86)
Ozzie Smith 8.88 (-0.78)
Ron Cey 8.82 (-1.32)
Graig Nettles 8.70 (+0.20)
Lance Parrish 8.66 (+0.66)
Darrell Evans 8.61 (-0.04)
---------------------
Thurman Munson 8.49 (-0.41)
Darrell Porter 8.34 (-0.52)
Dave Concepcion 8.12 (-1.39)
Bert Campaneris 8.06 (-0.46)
A few notes. Randolph falls pretty squarely in the middle of what each system defines as basically the borderline group, .75-.80 wins/162 games below the level of the players who will be obvious HoMers.
His peak rate is almost identical in WARP and WS, which suggests that he is a balanced candidate (bat heavy candidates do better in WS, fielding heavy candidates do better in WARP1).
This table doesn’t unearth conclusive evidence that Randolph is a hidden GIANT of his era, but it does suggest that, during his peak, he was a high-impact player, highly comparable in quality to some players whom we have elected.
Willie vs other 2B candidates: tough, since top 2 reutrnees are Doyle and Monroe. Doyle much better bat, but not as good a glove and 2B was a LOT different in 1910. Monroe's case involves reasonable speculation. Compared to Nellie Fox, better stick, almost as good a glove, longer prime, not as durable.
Willie vs other modern infielders: best comps are Nettles and Concepcion - all good gloves, decent bats. Randolph leads in 'batting vs position' if you use RCAP, but Concepcion may be clsoe if you use replacement level, particularly the poorer shortstops of his time. Nettles likely th ebest glove.
--
Randolph's case seems to be that for 16 years, he was above average - typical Willie year is 16 runs better than your JoeAvg 2Bman with the bat&legs;, 7 more with the glove. I take 2 runs away from the batting mark for the leadoff 'more at bats but lower leverage' thing, YMMV. Even so, Willie has enough for me that he climbs to the top of the backlog. I've been voting Beckley at the backlog's top, and Willie looks like him, but league strength makes me choose Randolph.
TomH - you pretty much summed up exactly what I considered when evaluating Randolph and you came to more or less the same result. Randolph makes my ballot and I only need to decide whether I want to PHOM him or Ken Boyer, who are practically equivalent for me.
WS most definitely does not define replacement level as zero. Replacement level in WS is around .200 WPct.
WS defines replacement level as the 'marginal' level, which is .5 of league runs scored or 1.5x league runs allowed.
In a league where an average team scores 750 runs in 162 games, that's a .111 overall WPct if a team has replacement level offense and defense. A 'replacement' level offense with an 'average' pitching staff/defense would post a .212 WPct.
WS most definitely does not define replacement level as zero
WS DOES define replacement level as zero. It's creator, Bill James, says so.
Read "Win Shares and Replacement Level (Or ... Pandora, shut your trap)" in the Win Shares book.
Quote from page 107: "There is no replacement level contemplated within this system."
Then again, just because he says it's so doesn't make it so. The definition of WS explicitly ignores the Pythagorean (i.e. non-linear) effects of runs on W-L. Taking those effects into account causes there to be an "effective replacement level", which is what Joe is talking about. The effective offensive replacement level is at .068 (11-151, assuming a defense that is just as execrable). The effective defensive replacement level is at .158 (25-136). A team with both components at "effective replacement level" would be at .105 (17-145); maybe if I used PythagenPat instead of an exponent of 2, it would be .111? Close enough.
In a league where an average team scores 750 runs in 162 games
The "effective offense replacement" team would score 52% of that (390 runs) and the corresponding bad defense would give up 1442 runs (1/.52 or 192% of league average).
The "effective defense replacement" team would give up 152% of that (1140 runs) and the corresponding bad offense would score 493 runs (1/1.52 or 66% of league average).
Win Shares has different effective replacement levels for offense and defense (due to its assumption of a linear conversion from runs to wins). The effective replacement level for offense is considerably lower than that for defense which is why Win Shares is so highly biased in favor of offense. Hitters get more "free" Win Shares that pitchers and fielders do not get ("free" meaning no value content).
When I wrote above that in win shares replacment level equals zero, I was describing how to interpret the meaning of a player's fielding win shares, not the nature of the win shares zero point. If you are trying to compare the defensive values of two players at different positions as the win shares system sees it, all you have to do is look at their total fielding win shares: that is equivalent to looking at FRAR in BP's WARP, which is adjusted for position.
If you want to know how far a player was above average as a fielder at his position in win shares, to find an equivalent in the win share system for FRAA in BP's WARP, that requires some more complex calculations.
That is all I was claiming in the post above. The Randolph thread is probably the wrong place for analysis of the zero-point in the win shares system, although jimd's point about the discrepancy in replacement level between offense and defense in win shares cannot be reiterated too often, at least until more voters start to notice. . . .
The effective offensive replacement level is at .068 (11-151, assuming a defense that is just as execrable). The effective defensive replacement level is at .158 (25-136). A team with both components at "effective replacement level" would be at .105 (17-145); maybe if I used PythagenPat instead of an exponent of 2, it would be .111? Close enough.
> In a league where an average team scores 750 runs in 162 games
The "effective offense replacement" team would score 52% of that (390 runs) and the corresponding bad defense would give up 1442 runs (1/.52 or 192% of league average).
The "effective defense replacement" team would give up 152% of that (1140 runs) and the corresponding bad offense would score 493 runs (1/1.52 or 66% of league average).
What is the reasoning that specifies the "just as execrable" or "corresponding bad" defense (or offense, given the defense)? "Corresponding" in the sense that when paired with average offense (or defense) the team win would win the same number of games? Is that the criterion --founded on the Pythagen theory that runs ratio determines wins ratio, ie (W/L) a function of (RS/RA)?
Over his career, Willie Randolph was about 20 wins better than an average offensive player, and about 47 wins better than 75% of average.
Over his career, Jim Gantner was about 8 wins worse than an average offensive player, and about 13 wins better than 75% of average.
Over his career, Frank White was about 17 wins worse than an average offensive player, and about 9 wins better than 75% of average.
What does that mean for Randolph? You can see how much better on offense he was than two near-contemporaries who kept their jobs forever.
Some people who rank a little ahead of Randolph, offense only, in my system: Childs, Evers, Lazzeri, Herman, Gordon.
Some people who rank very close to Randolph, offense only, in my system: Doerr, Lopes.
Some people who rank a little behind Randolph, offense only, in my system: Myer, McAuliffe, Huggins.
It's similar to the "equivalence" of 125 ERA+ to 125 OPS+. A team that scored 25% more runs than league average would be 125 RS+. A team that gave up 80% of the league average would be 125 RA+ (1.00/.80).
"Corresponding" in the sense that the other half of the team is just as bad. The .52 RA+ offense is paired with a .52 RS+ defense. Or the .66 RS+ defense is paired with a .66 RA+ offense. When someone says "a team of replacement players", I don't think of a replacement level offense that can field at an average level for average pitchers (though maybe others do). I think of a team ALL of whose players are of little value to any other team.
Pairing a "replacement level" offense with an average defense (or vice-versa) to determine the location of the WS effective replacement level doesn't capture how low it really is.
"Corresponding" in the sense that the other half of the team is just as bad. The .52 RA+ offense is paired with a .52 RS+ defense.
So I guessed. That is one analytical sense. We don't know that such equally bad offense and defense measured in runs is equally common, or equally easy to acquire on short notice, etc. Your point is a troubling asymmetry, and we may presume Bill James didn't ground in empirical work the "unequally bad" offense and defense that play analogous roles in Win Share (I rely on jimd #90). But it may be that replacement level offense and defense are unequally bad measured in runs, if defined in another way such as by reference to the market for players.
The main point of #90, I think, is to simply to quantify the badness of the Bill James replacement level offense and defense, not to argue that they should be equally bad.
True. But they are equally nasty in their impact on W-L, if you take Pythagoras' word for it.
Wow. He has been dead for 2,500 years, yet he has become a leader in the sabermetric movement. ;-)
You must be Registered and Logged In to post comments.
<< Back to main