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

Friday, July 11, 2003

Stovey and Browning

The two best hitters of the AA, how much do we discount them?

I think the AA was kind of like the USFL. The USFL had Jim Kelly, Herschel Walker and a few other legit stars (although they were young), and some decent retreads like Brian Sipe, but it wasn’t anywhere near the NFL, yet it was a major league, better than what we’d call AAA and they had really good players. Look at the AFL in the early years, it took them quite awhile before they were able to be truly competitive with the NFL. Even to this day, the AFL squads have won just 10 of 36 Super Bowls (28%), despite making up 38% of the teams from before the merger (not counting TB’s win, which is the only one from a post-1970 expansion team). Almost all of the really big stars were in the NFL/NL. It’s an interesting debate.

Joey Numbaz (Scruff) Posted: July 11, 2003 at 04:08 PM | 177 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 2 of 2 pages  < 1 2
   101. John (You Can Call Me Grandma) Murphy Posted: August 05, 2007 at 02:12 PM (#2471826)
Sorry, John, but I do see the system as failing us.


With all due respect, Dan, you're overreacting greatly.

Look, I could say the system failed every time it elected someone that I didn't support, like Carey, Roush, Suttles, etc. In fact, I could also say it if Tony Perez were elected. In fact, everyone of us could say it for the same reasons, but for different players. But I don't say it and many of us also don't because we understand that people have different views of value when it comes to what a Hall of Meriter means and we respect that. These views are also 100% arbitrary, since there is no exact measure of greatness.

Again, I'm 100% confident that the average educated baseball fan noticing that Browning is in the HoM will not think that we're a bunch of maroons. Going over some of the discussion, you would think we were potentionally electing Mookie Wilson (one of my favorite players, BTW). The Louisville Slugger was a great player during his career. No, he was certainly not an inner-circle player and I fully understand his long electoral journey to the present time, but he had a greater impact per season than many players that we allowed into the HoM a long time ago. If you value that, it means a lot.

I have spent many years now with the HoM as a voter and as the guy keeping the project rolling along when Joe couldn't do it anymore. While there were days that I felt like saying the hell with it, I have never have felt that way concerning our election mechanism, research and analysis, inductees or the electorate as a whole. I have felt, and will continue to feel, that we are doing something extremely worthwhile and will be, and is already, respected by a great deal of people. Not too bad, I must say.
   102. John (You Can Call Me Grandma) Murphy Posted: August 05, 2007 at 02:22 PM (#2471828)
Anybody who continues to equate Browning with "mistake" is having no affect whatever on my ballot any more than my issues with Randolph affected yours.


Of course, Marc. That argument will get absolutely no traction. What will convert voters, however, is explaining why certain metrics overrate Browning. But again, Dan doesn't agree with the peak-prime viewpoint, which appears to be his biggest beef when it comes to Pete, not necessarily the numbers or analysis that many of us are looking at.

BTW, I just want to say that Dan has been a great member of the HoM and has been a stellar asset to the project. He has personally helped me with many of the threads and has made my job much easier. While I disagree with him on this issue, I'm very glad that he's a part of our group.
   103. Howie Menckel Posted: August 05, 2007 at 02:48 PM (#2471838)
DanG is indeed one of our more knowledgeable and important members, of course.

Nonetheless, his rhetoric about the impact a Browning election would have on the perceived validity of the project is bizarre at best. It's patently false and is very likely to be proven so soon enough, as both in the short and long terms his predicted outcry (or whatever he's predicting) will obviously never materialize.

Here's a player who any outsider either:
- never heard of; or
- knows put up some of the most impressive offensive marks of all time.

The next level of knowledge - recognizing that in some years the AA was far inferior - nonetheless would have any reasonable voter at that level realizing that his election/non-election will hinge in large part on exactly how much to discount the gaudy numbers. And since it's taken him 100 years to get in, clearly we did not fail to recognize the significant and real holes in his candidacy.

I would no more give the slightest credence to someone who calls a Browning election a project-threatening mistake than to someone who called a Browning NON-election a project-threatening mistake.
There's just too much pro and con here for either position to have any - well, merit - in many, if not most, of our minds.

As for the "less than 30 pct" argument, well, that at least is a debatable one.
We could have had a threshold - say, one-third of the pts - and been willing to elect only those players. And if that left us '20 short' of HOM members, so be it.

But if we did that, I suspect someone would be tempted to suggest at the end, "Hey, if we DID elect enough to match the HOF, who are the extras we'd pick?"
In essence, we are doing that simultaneously. It's not that difficult to see which are the reluctant picks. Indeed someone likely will extricate a rough list of same at project's end.

I guess I'd like to see DanG (or JoeD?) clarify why he thinks Browning's election will be so devastating, relative to Randolph's or Beckley's or Pike's or anyone else's. I suspect even other Browning-bashers disagree with him on that point.
   104. sunnyday2 Posted: August 05, 2007 at 03:27 PM (#2471868)
I am actually somewhat incredulous that Browning is actually going to get elected. He is only 14 points ahead of the next candidate, and every year there are a few voters who don't vote, and who they might be is sometimes decisive. And of course many of us doodle with our ballots year to year. So who knows.

OTOH it's also true that I am pretty much incredulous about the rest of the high backlog getting elected. If as has been stated the only mistake is to elect somebody who hasn't been fully vetted, then where are we at?

Browning--fully vetted, of course; the top backlogger; if not Browning, then who?

Bresnahan--hard to see Rajah getting the call this year, but he's been pretty fully vetted too and recently.

C. Jones--not quite as fully vetted as Browning and Rajah but at least the blacklist credit part of his resume has been vetted. To me, he and Browning are frick and frack; do his supporters see him as better than Browning or are they the same people who support Pete?

B. Johnson--not particularly well vetted; his non-supporters seem to be taking it for granted that we couldn't possibly elect another hitter from his era. Somebody should make a concerted effort to blow up Bob IMO--not because I don't support him, though I don't, but just so that the case against is as fresh in people'sminds as Pete's and Rajah's and Charley's.

Dawson--ah, the Hawk; if not Browning, then maybe Dawson? Of course with the more recent newbies it's been elect the best, forget the rest, but who knows. Not as well vetted as could be, especially the case FOR, the case against is more clear.

Redding--pretty well vetted but it was a while ago, anybody remember the pros and cons?

Perez--the case that he's the best of the contemporary hitters hasn't really been made IMO but here he is. Somebody should probably undertake to remind us why NOT to vote for Doggie and who's better at doing the things that are his strength.

Duffy--never really was very well vetted IMO. Yeah yeah, there's the glove and the peak, but the peak was one year. His supporters could remind us why they support him.

Puckett--the case against has been made though somewhat half-heartedly. If voters like Dawson better then it's probably a moot point, but his supporters probably need to revisit his case.

Leach--who? Not vetted well at all.

Walters--there's been some discussion, but if I thought he had a chance I'd want to revisit the case both way.

Cravath--pretty well vetted.

To me the "mistakes," in that the pros and cons aren't freshly in mind, would be Bob Johnson, Doggie Perez and Tommy Leach. But that's just me and what I remember of the debates over the years.
   105. DanG Posted: August 06, 2007 at 05:03 PM (#2473868)
I suspect even other Browning-bashers disagree with him

Hello, Browning bashers? You've all conceded, have you? Pointless to argue "settled law" I suppose?

Fine, maybe we have a "need for screed" on Charley Jones.
   106. Chris Cobb Posted: August 07, 2007 at 01:16 AM (#2474381)
To carry on with Browning a bit farther, let me comment on the issue of Browning's peak.

Sunnyday2 wrote:

Probably a better (and fairer) view is to just write of '89 as an injury year and define his peak as 1885-88, 90, 5 years with a mean putative OPS+ of 177.5 (adjusted to 166.5) against tougher competition. This is approx. equal to Willie McCovey's 5 year peak and better than everybody else listed above, including Mike Schmidt, if a pennant is a pennant. If you want to apply a "timeline" then who knows. Your mileage may vary. But note that I'm already applying the competition discount (vs. the contemporaneous NL).

Even if a pennant is a pennant, Browning's batting peak is not as valuable as McCovey's, because it didn't dominate in its context to the same extent.

Let's take two views of Browning's league-adjusted OPS+, Brent's and Sunnyday2's, and see how Browning places among the major-league leaders in OPS+, assuming that all other players outside the toughest league also have their OPS+ totals similarly adjusted. I'll put the scores into 4 categories: #1 in majors, top 5 in majors, top 10 in majors, top 10 in league (but not in majors)

Here's Brent

152 -- Top 10 in Majors
144 -- League top 10
168 -- Top 5 in Majors
139 -- League top 10
169 -- #1 in Majors

Here's Sunnyday2

180 -- Top 5 in Majors
151 -- Top 10 in Majors
178 -- Top 5 in Majors
158 -- Top 10 in Majors
175 -- #1 in Majors

Now here's McCovey, 1966-70
165 -- Top 5 in Majors
160 -- Top 10 in Majors
175 -- #1 in Majors
211 -- #1 in Majors
182 -- #1 in Majors

Then there's the issue of playing time. During his five-year peak, Browning played in about 82% of his team's games. McCovey, who was not greatly durable, played in about 91% of his team's games.

Sunnyday2 was only talking about OPS+, not durability, and I'm coming around to the idea that durability is somewhat overrated in comparision to rate of production when playing,

But there's no way, unless you don't adjust for the weaker AA at all, that Browning's 5-year peak as a hitter is equal to McCovey's.

I'm coming around to the view that Browning, even when properly adjusted for his context, is still the top batting candidate by rate among those eligible. But he's not that far ahead of the competition, and I'm not convinced that when durability and defense are also taken into account, he is all that close to meriting election. So, for Browning supporters who are looking at his peak as having the value of McCovey's, I hope you'll recalculate with a somewhat more modest view of Browning's peak value as a hitter.

I'm still reviewing Browning, comparing Browning's rates to those of other available bat candidates, along with his durability and defense. I'll post again when I have more data to share.
   107. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 07, 2007 at 01:31 AM (#2474400)
Chris Cobb, what you are describing as "dominating in its context" is precisely what I have measured with standard deviations. A 180 OPS+ did not buy as many pennants in the 1880s as it did in the 1960s, because the league standard deviation was higher. The same is true of McCovey's expansionary 1960s compared to, say, the 1980's, when an OPS+ in the 150's was good enough to lead the league in many years. And yes, I have taken both the competition discount and standard deviations into account before supporting Browning. If I didn't discount Browning for playing in high-stdev leagues, he'd probably be near the top of my ballot rather than the low middle.
   108. sunnyday2 Posted: August 07, 2007 at 01:56 AM (#2474426)
Chris, what is your consideration set of players whose position was essentially "hitter"?

And of course, if Browning is not as good as McCovey, well, it's not like McCovey is his competition today.
   109. Chris Cobb Posted: August 07, 2007 at 02:05 AM (#2474434)
Dan R wrote:

what you are describing as "dominating in its context" is precisely what I have measured with standard deviations.

Yes. What I am trying to find are less highly processed ways of pointing out what complex and thorough metrical analyses show, so that one doesn't have to accept the complex metric in order to accept the analysis.

Could you say a bit more about why you rank Browning as you do? On your 2002 ballot, you say that Browning is the best hitter available, bar none, and I agree that is probably true, and that a lot of other things drag him down, which I also agree is probably true. So why do you place him ahead of, say, your next six: Pesky, Bancroft, Tiant, Nettles, Schang, and Grimes? My 2003 ballot is probably going to have 4 of those 6 on it, similarly placed, and I understand why you support Pesky and Schang, although my view of peak vs. career value and of how to assess catcher playing time leads to both of them being a bit farther down in my rankings. But I don't see Browning ahead of them. Why do you?
   110. Chris Cobb Posted: August 07, 2007 at 02:14 AM (#2474455)
Chris, what is your consideration set of players whose position was essentially "hitter"?

And of course, if Browning is not as good as McCovey, well, it's not like McCovey is his competition today.


True. As I noted above in my comment to Dan R, I am prepared to agree that Browning is the best hitter available, at least in terms of rate and peak.

I'm working on comparing Browning to a set of peak value hitters. My consideration set includes all players who placed in the top 10 in league OPS+ at least 6 times. This set includes

10 times -- Bob Johnson
9 times -- Pete Browning
8 times -- Charley Jones and Chuck Klein
7 times -- Orlando Cepeda, Larry Doyle, Frank Howard, Jimmy Ryan
6 times -- Gavvy Cravath, Norm Cash, Ken Singleton, Rocky Colavito, Dale Murphy, Bob Elliott, Tony Oliva, Jack Clark, Bobby Veach, Hack Wilson

I also checked a number of other hitters, who checked in at 5 times or less. These players may be ballot-worthy, but their argument is a career argument, and what I am trying to determine first is if Browning is the best or not of the set of players whose peak and career size are most similar to his own.

Checked but not kept in consideration set: Reggie Smith, Jim Rice, Dave Parker, Rusty Staub, Don Mattingly, Frank Chance, Al Rosen, Andre Dawson, Tony Perez, Kirby Puckett, John McGraw, Bobby Bonds, Sal Bando, Ron Cey, Hugh Duffy, George Van Haltren.

I intend to check Alejandro Oms by OPS+, to see how many times he would have placed among the league leaders if he had been in the majors, but I haven't gotten to that yet.
   111. Chris Cobb Posted: August 07, 2007 at 02:20 AM (#2474466)
P.S. I know this sorting method isn't entirely fair to expansion era players, who face a larger pool of competition to make top 10 spots. It's possible than in 8-team leagues, some of the players who missed the cut would have made it into the set. But they would still be behind some of their contemporaries in the set, and for those players who are closer on the face of it to Browning I will do my best to make acccurate cross-period comparisons. The goal is not to order all of these players, but to see if Browning is at or near the top of the group when a thorough assessment is made.
   112. sunnyday2 Posted: August 07, 2007 at 02:33 AM (#2474507)
Chris, one name sure jumps out of your consideration set of "hitters" and that of course would be the one guy who also carried a glove. Not a great glove but a glove. That of course would be Leapin' Larry Doyle. May I recommend him to you?!
   113. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 07, 2007 at 03:22 AM (#2474692)
Chris Cobb--the simple answer is that Browning comes out a nudge higher in my salary estimator. I am using the AA-to-NL translation rates implicit in BP's all-time adjustments, and then make a further adjustment for standard deviation based on the actual stdev of OPS+ in the AA for those years (I didn't try to extend the regression to the AA, because the coefficients would likely change in weaker leagues). So in the 1882 AA, I have Browning at 74 runs above average, which gets reduced to 66 runs above average after correcting for the fairly high scoring environment in that league, to 36 runs above average after the AA discount, and 26 runs above average after the standard deviation adjustment, or 2.7 wins above average in 69 games. That straight-line adjusts to 6.3 wins above average per 162 games. Using the 1890's replacement levels (since I don't have them for the 1880's), I have replacement 2B at 2.5 standard deviation-adjusted wins below average per 162 games, which makes Browning 8.8 wins above replacement per 162 games. ((8.8^2)*212730) + (8.8*402530) comes to just over $20 million in a full season. Since Browning played in 87% of his team's games, I multiply that by .87 to get a salary of $17.5 million for Browning in 1882. Repeating that for every year of his through 1892, and then using my actual WARP numbers for the end of his career, brings him to a hair over $100 million. That's juust enough to edge out Bancroft ($99.2M) and Pesky ($99.0M with war credit), and more comfortably ahead of Nettles ($95.6M). The pitchers and catchers I just sort of place by feel for now.

Browning's placement is *tremendously* sensitive to my estimates for the AA discount and standard deviation. Even a 5% change in either one would send him either soaring to the top of my ballot or falling off it entirely. (With no standard deviation adjustment or league quality discount, Browning would be an inner-circle Hall of Famer in my system!). But I think that my methodology for determining both of them is reasonable, and the fact that it has resulted in a down-ballot placement of Browning, similar to the electorate's consensus view, has given me extra confidence in my analysis. If you want, I can send you the spreadsheet with my season-by-season estimates of Browning's value, which you can then compare at ease to my WARP numbers for post-1893 players available in the Yahoo group.

Do note that my standard deviation adjustments are available in the .zip archive in the Yahoo group, along with the regression equation used to calculate them. So you can easily pluck the ease-of-domination (what you call "dominating in context") adjustment out, tweak it to taste, and scrap the rest if you'd like.

Thanks very much for your interest in my thought process.

As a staunch enemy of Larry Doyle, I'll just tip my hat into the ring to dissuade Chris from voting for him with my standard schtick--2B replacement level was insanely high when Doyle played, no different than first base, making good-hitting 2B's a dime a dozen. If you wouldn't vote for a poor-fielding 1B with his offensive numbers, you shouldn't vote for him. The fact that he hurt his teams in the field probably more than most 1B ever could, and that he played in leagues that were rather easy to dominate, should only hammer the nails into his coffin.
   114. Howie Menckel Posted: August 07, 2007 at 04:04 AM (#2474840)
"My consideration set includes all players who placed in the top 10 in league OPS+ at least 6 times. This set includes

10 times -- Bob Johnson
9 times -- Pete Browning
8 times -- Charley Jones and Chuck Klein
7 times -- Orlando Cepeda, Larry Doyle, Frank Howard, Jimmy Ryan
6 times -- Gavvy Cravath, Norm Cash, Ken Singleton, Rocky Colavito, Dale Murphy, Bob Elliott, Tony Oliva, Jack Clark, Bobby Veach, Hack Wilson"
..............

Interesting list.
I vote for Johnson in mid-ballot, some war issues.
I have Browning a little above that.
No on Jones for even earlier era and on Klein for so many rivals.
Cepeda on my ballot. Doyle no on awful fielding in a hitting-2B era, Howard I like and is close, Ryan not quite.
Cravath on and off my ballot, Cash no on playing time for his era, Singleton on occasion, Murhpy in the discussion, Elliott yes as 3B, Oliva nothing else, JClark playing time issues, etc
   115. sunnyday2 Posted: August 07, 2007 at 04:23 AM (#2474901)
I would just suggest looking at the 2B offensive records of the '10s. I use Neft and Cohen which lists the starting lineups for each team each year as a quick eyeball check. Granted nobody it much in those days (except, of course, Larry Doyle) but the 2B offense was pretty pathetic as a rule.
   116. sunnyday2 Posted: August 07, 2007 at 04:25 AM (#2474905)
Make that "nobody HIT much."
   117. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 07, 2007 at 04:43 AM (#2474963)
sunnyday--Huh? 2B had a .272 EqA on average from 1907 to 1915 (not counting the Federal League), which was actually a *higher* figure than first basemen in the period (.271) and not far off the production of corner outfielders (.278). By contrast, third basemen and shortstops lagged far behind at .263 and .257, and of course catchers and pitchers held up the rear at .245 and .202. By contrast, from 1930 to '37 (not selective endpointing here--those are the last years I have positional EqA data for), 2B managed just a .257 EqA. That's an *enormous* difference.

And no, it's not just Collins and Lajoie--according to my WARP, the worst six starting 2B in the majors averaged 1.1 wins (summing offense and defense) below average per 162 games from 1907 to 1915. They averaged around 2.4 wins below average per 162 games from 1920 to 1980. There's no way to slice it that doesn't show 2B as a major offensive position in the years surrounding 1910, with a bottom and middle of the distribution comparable to that of 1B and a top (Collins, Lajoie, Doyle) that was actually much higher. Sunnyday, are you just looking at raw offensive numbers? It was the deadball era, after all. Any context-adjusted stat--OPS+, BP WARP, WS, my WARP, etc.--will reflect the strength and depth of 2B in this period.
   118. John (You Can Call Me Grandma) Murphy Posted: August 07, 2007 at 01:14 PM (#2475135)
Hello, Browning bashers? You've all conceded, have you? Pointless to argue "settled law" I suppose?


There is no such thing as settled law here. If that were the case, Dickey Pearce would have never been elected decades ago, Dan.

Fine, maybe we have a "need for screed" on Charley Jones.


If it's more a screed than pointing out something that we didn't know already about Jones in a negative light (other than peak/prime/career considerations, which you will never win), I would respectfully state that you might think about doing something else (like touting Bresnahan? :-).
   119. Chris Cobb Posted: August 07, 2007 at 01:31 PM (#2475153)
Howie ran down who he has been voting for from my "Browning Consideration Set." Myself, the only one I have been voting for is Charley Jones, who has just squeezed on to the bottom of my ballot in recent elections. Once I have finished this analysis, the next step for me will be to decide if _any_ of this group, even the top 2 or 3, are ballot-worthy when compared to the career candidates who have come to dominate my ballot in recent years as we have gotten deeper into the backlog.

Dan R -- Thanks for your explanation. It is useful to see what assumptions you are making on the 1880s and the AA in particular. Don't worry about my voting for Doyle. A more granular analysis of OPS+, which I haven't gotten far enough with to post yet but have gotten far enough with to reach some preliminary conclusions, shifts him towards the bottom of this group, and the fielding metrics suggest that his defense was so poor at second that his real defensive value was no higher than that of a good corner outfielder.
   120. Bring Me the Head of Alfredo Griffin (Vlad) Posted: August 07, 2007 at 02:41 PM (#2475192)
"The only anecdote that I've found about Browning's baserunning is in Wikipedia, which states that 'Browning's baserunning was also considered sub-par, exacerbated by his refusal to slide.'"

Just for reference: I'm the one who added that to Wikipedia, and I got it from here.
   121. andrew siegel Posted: August 07, 2007 at 03:39 PM (#2475240)
For me, durability is a big issue for Browning. He has only three seasons that project out to more than 142 games when converted to 162 games; his whole career is only nine years and change when figured fractionally. As a direct result of his lack of durability, his numbers sag a bit when you look at runs created (6 times in the top 10 in league) rather than adjusted OPS+ (9 times). 9-plus seasons of 144ish OPS+ spread out over 13 seasons with poor fielding puts him one cut below for me.

I am a Browning-skeptic (he's about 55th on my list of eligibles), but I see the arguments in his favor. If you discount the AA a bit less than I do, if you give him a bit more credit for playing tough defensive positions (albeit badly), if you value in season rate over in season durability (a la Dan R), he can easily rise to the top of the ballot. Everybody is very close at this point. Put differently, he's not that dissimilar from guys I have supported such as Keller, Kiner, and Cravath, but he's also similar to guys like Tiernan, Frank Howard, Fournier, Klein, and Albert Belle who are nowhere near my ballot.
   122. Dr. Chaleeko Posted: August 07, 2007 at 06:41 PM (#2475436)
Wow. From 15 posts in several years to 120 posts in a couple days. Not bad, guys.
   123. EricC Posted: August 07, 2007 at 11:06 PM (#2475698)
Pardon the delay in responding. I was without internet service the last few days.

Brief replies to the comments to my Browning baserunning post:

Stovey The success of Stovey's teams in actual runs scored to expected runs scored was average to good. The most consistent team successes in the AA were by STL and CIN. Arlie Latham is one individual whose reputation might be much higher if game logs were available.

Browning's SB in 1887 He was 4th in the league in "SB", the one year in his career that he was in the top ten. Note that SB were not defined the same back then, and there is no caught stealing to determine net baserunning value.

Browning's defense & his 2B season A modern equivalent might be Danny Tartabull starting out as a SS. In both cases, the entirety of their careers leads to the conclusion that they were subpar defenders overall.

Effect of baserunning: small? Maybe, I don't know. If Browning's teams consistently scored 3-4 percent runs fewer than expected and it had nothing to do with him then the effect is zero. If he was entirely to blame for the reduction of scoring at the team level, then his individual effect would be 9 times greater than 3.5%, that is, about 30%.

Park factors and runs scored Imagine an 8-team league with idential teams playing their games in identical parks. They all end the season 81-81 with the same OBP and Slug. All of the teams score 1000 runs, except for one team where the players occasional try to advance by crawling on their hands and knees, getting thrown out and leading to the team only scoring 930 runs. Then the average runs scored per 162 games in their park is (1000 + 930)/2 = 965, and the average runs per 162 games in the other parks is (13 1000 + 930)/14 = 995. This is how poor baserunning can appear as a park factor that suppresses runs. In practice, this effect is unlikely to be large.
   124. EricC Posted: August 07, 2007 at 11:27 PM (#2475736)
Comparison of actual team runs scored, BPro equivalent runs, and team batting runs above average for the 1890 PL.
Team followed by actual runs followed by EQR, then BRAA:

NYI 1018 603 -17
BOS 992 655 +59
BWW 964 585 -39
PHQ 941 605 -11
PBB 835 655 +58
CLE 849 699 +96
CHI 886 566 -80
BUF 793 612 -9

Actual runs scored and BRAA appear almost completely uncorrelated. In particular, note Browning's team CLE was 6th in actual runs scored, more than 150 behind the leader, and comes out top in BRAA!?! Individual BRAA sum to team BRAA. If team BRAAs doesn't make sense, then individual BRAAs don't make sense.

Anybody know how BroP derives EQR, and what explains the above discrepancies? The only things that I can think of are (1) WARP is using a runs created formula that is not appropriate for the era; (2) gross error. Any thoughts?
   125. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 07, 2007 at 11:39 PM (#2475764)
If I'm not mistaken, one-year PF is just equal to the ratio of total runs scored in away games to total runs scored in home games, plus 1 divided by 2 times 100. So in your hypothetical case, the team in question would score 930/2 = 465 runs at home, while its opponents would score 500 runs in its home ballpark, for a total of 965 runs at home. The exact same would be true for its away games. 965 divided by 965 is 1, plus 1 is 2, divided by 2 is 1, times 100 is 100.

EQR are standardized to a 4.5 R/G, Pythagorean exponent of 2 league. BP's raw run estimation is UEQR.
   126. EricC Posted: August 07, 2007 at 11:53 PM (#2475790)
Dan- you're explanation of park factors makes sense, thanks for explaining.

EQR are standardized to a 4.5 R/G, Pythagorean exponent of 2 league. BP's raw run estimation is UEQR.

None of this explains how EQR are uncorrelated to actual team runs in the 1890 PL, does it?
   127. EricC Posted: August 07, 2007 at 11:54 PM (#2475795)
Also, raw UEQR in the 1890 PL are closely correlated to team runs scored, so whatever is happening is happening in the conversion from UEQR to EQR.
   128. EricC Posted: August 07, 2007 at 11:55 PM (#2475801)
you're -> your in #126. I do speak English.
   129. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 08, 2007 at 02:18 PM (#2476792)
I can't make the numbers add up either, I'm sorry to say. According to BP's About EqA file, EQR is just easily derived from EqA, EqA is ((winpct/(1-winpct))^.2)*.26, and winpct is calculated using the Pythagorean theorem from team UEQR and league average runs allowed. Well, the 1890 Players' League New York Giants had 958 UEQR in 3597 outs, and the league scored 7278 runs in 28560 outs (if I am right that BP is using a flat estimated 75% SB success rate). 7278*3597/28560 is 917 runs for an average team in New York's outs, 958 UEQR and 917 RA is a Pythagenport exponent of 2.13 and a winning percentage of .523, and ((.523/(1-.523))^.2)*.26 is an EqA of .265. New York's season-adjusted EqA is listed at .257, which is a gigantic difference.

This is why you should all use my WARP...because at least my methods are transparent and replicable! ;)
   130. Paul Wendt Posted: August 08, 2007 at 11:33 PM (#2477392)
Park factors and runs scored Imagine an 8-team league with idential teams playing their games in identical parks. They all end the season 81-81 with the same OBP and Slug. All of the teams score 1000 runs, except for one team where the players occasional try to advance by crawling on their hands and knees, getting thrown out and leading to the team only scoring 930 runs. Then the average runs scored per 162 games in their park is (1000 + 930)/2 = 965, and the average runs per 162 games in the other parks is (13 1000 + 930)/14 = 995. This is how poor baserunning can appear as a park factor that suppresses runs.

Some people calculate park factors by this naive method, but does any source of adjusted statistics use this for the run-scoring park factors that are part of the adjustment?

One important source of differences among park factors and adjusted stats derived by their use is the multi-season averaging: three-year? equally weighted? Another is the accounting for signficant changes in the ballparks, which are breakpoints where the normal multi-season averaging is not used.
   131. Chris Cobb Posted: August 09, 2007 at 12:29 AM (#2477529)
Here is a first step at ranking Pete Browning against the eligible peak-prime hitters. I am starting with this group because a) these are players whose type of value and whose career shape is similar to Browning’s, making the grounds for comparison easier to establish and b) this is the group that plays to Browning’s strengths. Career voters and voters who highly value pitching and defense are unlikely to be supporting Browning. It’s the peak/prime case, centered around hitting, that needs examination. I hope to address the other kinds of comparisons before Monday, but this is a start.

Let me stress that I am not at this point arguing for a particular ranking. I have not been a Browning supporter, but I am trying to re-examine the evidence in a thorough and open-minded way. As we all learn more, our positions may change.

I began with a comparison set of 20 players, all of whom are eligible candidates with at least six appearances among the top 10 in OPS+ in their league. I then ranked these players in three ways: by peak batting, by durability, and by fielding, with the aim of placing them in three groups: top, middle, and bottom. I’m not trying to make fine distinctions here, just to get the lie of the land.

For batting, I ranked players by their placement among league leaders in OPS+, but I favored higher placements as follows. For each season that a player was best in the majors, I awarded four points. For each season that a player was in the top 5 in the majors, I awarded three points. For top ten in the majors, two points. For top 10 in a league but not the majors, 1 point. For these determinations, I made three alterations to player records. For Pete Browning and Charley Jones, I used adjusted OPS+ scores for their AA years. I used Brent’s numbers for Browning (sunnyday2’s numbers led to similar results, but I found it helpful to use Brent’s OBP+/SLG+ breakouts), and I applied the conversion factors deduced from Browning to Jones’ AA years. The third alteration to these records is that for Gavvy Cravath, I used Brent’s American Association MLEs for 1909-11.

Here are the results of this ranking

Score – Player (seasons in lg top 10)
Top Group
19 – Gavvy Cravath (9)
17 – Pete Browning (8)
16 – Mike Tiernan (7)
15 – Chuck Klein (8)
15 – Jack Fournier (7)
14 – Charley Jones (5)
14 – Frank Howard (7)
14 – Ken Singleton (6)
Middle Group
13 – Jack Clark (7)
12 – Bob Johnson (10)
12 – Hack Wilson (6)
11 – Orlando Cepeda (7)
11 – Jimmy Ryan (7)
11 – Dale Murphy (6)
Bottom Group
10 – Norm Cash (6)
9 – Rocky Colavito (6)
9 – Bob Elliott (6)
9 – Bobby Veach (6)
8 – Larry Doyle (7)
7 – Tony Oliva (6)
7 – Jake Daubert (6)

I let the top group be a little large because Jones is missing two blacklist years when he would likely have been a top 10 hitter, while Howard and Singleton were both post-integration/post-expansion players, competing against a larger pool.

For durability, I calculated the percentage of team games in which each player appeared during a ten-year prime stretch as a hitter, allowing one year to be skipped for injury (I skipped Jones’ 2 missed years; I used Cravath’s 1909-11 minor league play). Just as a check, I counted the total seasons in which a player appeared in 140 or more season-length adjusted games in his career. In general players with high scores and low scores in these groups were the same.

Here are the results of this ranking

% -- Player (140 g seasons)
Top Group
98 – Dale Murphy (13)
98 – Bobby Veach (11)
97 – Rocky Colavito (10)
96 – Ken Singleton (12)
95 – Orlando Cepeda (13)
95 – Bob Johnson (12)
94 – Bob Elliott (12)
Middle Group
92 – Charley Jones (8)
91 – Frank Howard (8)
91 – Jimmy Ryan (10)
90 – Larry Doyle (8)
90 – Mike Tiernan (7)
90 – Gavvy Cravath (9)
90 – Norm Cash (9)
90 – Tony Oliva (7)
89 – Chuck Klein (6)
Bottom Group
83 – Pete Browning (6)
83 – Jack Fournier (5)
83 – Hack Wilson (6)
82 – Jack Clark (7)

Here the middle group is large, with lots of players clustering around 90% of team games, and then a big gap separating the middle from the bottom group.

For fielding, I wanted to establish a position-independent, all-time sense of good fielding vs. bad fielding, so I ranked players by WARP2 FRAR, career total. (For Browning and Jones, I season-adjusted their totals. For Cravath, I created totals for 1909-11 based on his major-league averages for 1908-9 and 1912-13.)

FRAR2 – Player
Top Group
273 – Bob Elliott
262 – Jack Daubert
252 – Dale Murphy
218 – Bob Johnson
203 – Jimmy Ryan
201 – Bobby Veach
192 – Norm Cash
Middle Group
152 – Charley Jones
147 – Rocky Colavito
135 – Chuck Klein
121 – Larry Doyle
116 – Tony Oliva
113 – Jack Clark
112 – Ken Singleton
Bottom Group
94 – Mike Tiernan
81 – Orlando Cepeda
74 – Gavvy Cravath
67 – Pete Browning
62 – Hack Wilson
41 – Jack Fournier
Sub-bottom Group
8 – Frank Howard

The breakdown is pretty clear here, though Tiernan might belong in the middle group if his total were season-adjusted.

From these lists of comparable types of hitters, the strength of Browning’s case is not obvious on the face of it. He is solidly in the bottom group on playing time and on defense. He is near or at the top of the batting list, however, and the big question, of course, is whether or not his batting value counterbalances his weaknesses in other areas. It is the case that most of the players who are at the top of the batting list are at the bottom of the fielding list. If we draw in EQA2 for a moment, you can see why Joe Dimino has been equating Browning with Hack Wilson. Browning’s EQA2 is .313, Wilson’s .311. Both played in 83% of their team’s games during a ten-season prime. Their all-time defensive values are highly similar (62 FRAA vs. 67 FRAA). Some players appear to have stronger profiles across the board than Browning. Ken Singleton, Dale Murphy, and Bob Johnson are 2, 1, 1 in their group placement, the strongest overall showing.

But how meaningful are these groups, really? I wouldn’t want to use them as a basis for my rankings. But I think we can use them to identify a few players about whom we really need to ask carefully, “If Pete Browning, why not X?”

These players seem to me to be nine: the aforementioned Singleton, Johnson, and Murphy, along with Bob Elliott, Chuck Klein, Charley Jones, Mike Tiernan, Jimmy Ryan, and Gavvy Cravath.

For the rest:

1) Fournier trails Browning in every category, so he can drop (it might seem unfair to do that without MLEs for his minor league seasons, but I’ve checked his thread and they really don’t change his profile in batting peak or durability, so out he goes)
2) Wilson trails or equals Browning in every category also. If Browning, then maybe Wilson, but I don’t think we have to test further the argument that Wilson was better than Browning.
3) Howard’s fielding is so bad, there’s no way his superior durability overcomes it in comparison to Browning.
4) Doyle trails Elliott in every category, so if we are just looking to see if Browning is the best of this group, Elliott is obviously the more significant of this pair.
5) Oliva trails Johnson in every category, even to the point of being in different groups.
6-7) Rocky Colavito and Bobby Veach trail Dale Murphy in all three categories.
8) Orlando Cepeda trails Ken Singleton in all three categories.
9) Norm Cash trails several people in all three categories. He may have a good extended prime or career case, but his profile here in terms of peak/prime is weak.
10) Jake Daubert has a better profile than I expected, but he is, how shall I say it, overrated by OPS+, and he is already about the weakest hitter in the group. His EQA2 is .277, which means (preview of the next study), his BRAR/out rate is 50% of Browning’s. Decent durability and excellent defense at first base just isn’t going to make up that big of a gap in batting value, not in a peak/prime context.

So this study cuts a comparison set of twenty down to ten. Your mileage may vary. The next study will look at this group of ten, focusing in more narrowly on five-year peak and ten-year primes, trying to find reasonable ways to (1) balance the values of batting, fielding, and durability and (2) lining up EQA2 with the OPS+ study to try to find a way to express Browning’s adjusted batting value that can be fairly counted in terms of runs.
   132. Brent Posted: August 09, 2007 at 03:55 AM (#2478119)
Nice study, Chris.

I'll mention three players who could have been included in this group, though I don't blame you for omitting them--Buzz Arlett, Luke Easter, and Julian Castillo. My numbers show Arlett's MLE OPS+ for the ten-year period, 1924-33, as 157. I don't have similar statistics calculated for Easter or Castillo, but I'm sure that they each had ten-year periods with MLE OPS+ above 145. On the other hand, all three had poor defensive reputations and one of the key questions is just how bad were they--Frank Howard bad, or merely Cepeda/Cravath/Browning/Wilson bad? I used to vote for Arlett, and he, Easter and Castillo are still in my top 40. Nowadays, however, the only ones in your set of 10 that I've been voting for are Singleton and Cravath (sometimes).
   133. Paul Wendt Posted: August 09, 2007 at 04:04 AM (#2478136)
Allocating the ten in another way:

never significantly supported here
Klein

supported long ago
Tiernan
Ryan, moreso

now on 9-10 ballots
Elliott
Singleton
Murphy

now on 18-19 ballots
Jones
Browning
Cravath
Johnson

Chuck Klein was a better player than Joe Medwick (elected years ago, rather quickly) but he was an everyday player for only five seasons, Medwick for ten seasons.
   134. Chris Cobb Posted: August 10, 2007 at 01:06 AM (#2479576)
Methodology for looking at consecutive prime. Looking at 10 years total, with one “off year” allowed to be skipped. This is to show Browning’s prime in its best light. Main stats are WARP2, straight-line adjusted for season length. Since EQA is not linear and I can’t readily calculate it, I am using BRAR/400 outs as the prime offensive rate stat, from which I am estimating a 10-year EQA for visual convenience. I am combining that rate state with season-adjusted outs to get adj. BRAR totals. Then I am putting this batting measure together with three fielding measures: WARP2 FRAR, 1/2 WARP2 FRAR (for those who think WARP places too much weight on fielding, and fielding WSx3. Since wins above replacement in WARP are approximately equal to nine equivalent runs, and 3 win shares equals a win, fwsX3 conveniently scales a win shares measure to the WARP system. I am not going to justify my choice of metrics in depth now, but I will say that EQA2 values the early AA more highly than Brent’s MLE conversions do: EQA2 is pretty much in line with sunnyday2’s OPS+ estimates for Browning. For batting value, it is not a harsh measure. WARP’s adjustments to batting value are not all that large, historically, and it never changes batting replacement level, so BRAR are much more consistent across time that FRAR. I have used several different fielding metrics because, esp. for the 1880s, I don’t know which one is better at identifying good and bad fielders, and which one is scaled most appropriately to reflect the actual range of fielding value. I like FRAR2 for this comparison because it provides a clear answer to the question of whether, in an all-time context, a player was a good fielder or a bad fielder.

Without further ado, here are some results.

First, the rates, which suggest that Browning was, indeed the best hitter during his prime, of the players in the consideration set, though not by all that much.

Player BRAR2/400 (est. EQA2)
Pete Browning 64.0 (.318)
Ken Singleton 63.0 (.317)
Gavvy Cravath 59.4 (.314)
Chuck Klein 59.3 (.314)
Bob Johnson 55.4 (.310)
Mike Tiernan 54.8 (.309)
Charley Jones 51.7 (.305)
Bob Elliott 50.1 (.303)
Dale Murphy 46.4 (.297)
Jimmy Ryan 41.7 (.291)

When this rate is converted to a counting stat by multiplying it by adjusted outs, the order shifts somewhat.

Player BRAR2 total
Ken Singleton 620
Pete Browning 602
Charley Jones 574
Mike Tiernan 573
Bob Johnson 563
Chuck Klein 562
Gavvy Cravath 546
Bob Elliott 510
Dale Murphy 507
Jimmy Ryan 446

I am not certain that this metric is not skewed in some ways by using outs as a measure of playing time, since they are not era-normalized, though they may show some platooning or late-inning replacement effects. Most of the 19th-century players have higher outs/game totals than the 20th-century players, so their totals are higher here, despite generally being a bit less durable.

Now, here are three views of fielding value. First, fielding runs above replacement in WARP2, season-adjusted. Second, FRAR2 divided by 2 to reduce its scale. Third, adj. fielding win shares times 3.

Player  FRAR2  Player .5*FRAR2   Player  3*fws
Elliott  243   Elliott   122  Elliott   127.8
Murphy   241   Murphy 120  Ryan   127.7
Johnson  194   Johnson 97  Murphy 122.4
Ryan  180   Ryan 90  Browning  113.0
Jones 141   Jones   70  Jones  106.6
Klein 128   Klein   64  Tiernan 94.0
Tiernan  110   Tiernan 55  Johnson 91.5
Singleton 98   Singleton  49  Klein   80.8
Browning  85   Browning   42  Cravath 75.7
Cravath   83   Cravath 41  Singleton  73.1

Range 160 runs
,   Range 81 runsRange54.7 runs
or 17.8 wins   or 9 wins   or 6.1 wins
1.8 fwins
/yr.  .9 fwins/yr.   .6 fwins/yr


Six features of this chart are notable.
1) Win shares views fielding value as approximately 1/2 as valuable, overall, as WARP2.
2) Given that, win shares still compresses fielding value more than WARP, further reducing its relative value.
3) Win shares generally gives more relative credit to 19th-century players, which makes sense, given that it is less rigorously normalized to an all-time context than WARP2.
4) WARP and win shares mostly agree on player order within 19th and 20th century groups. In the 20th century, it is Elliott Murphy Johnson Klein Singleton/Cravath, and for the upper half of the group they even agree fairly closely about value, once FRAR2 is cut in half. In the lower half, the win shares cushion kicks in, rightly or not.
5) WARP and win shares place Ryan, Jones, and Tiernan in relatively similar positions, though fws run about 30 runs higher than 1/2 FRAR2 (about the size of the cushion for low-end 20th century defenders).
6) Browning is the player who receives radically different assessments in the two systems. It would be helpful to know why.

Given that the BRAR2 totals based on rate per out also may favor 19th-century players, esp. pre-1893 players, the combination of BRAR2 and 3*fws may be skewed inappropriately towards 19th-century players. They certainly do better in it. Whether they should or not is an open question.

Now, for a last chart, here are the results of summing BRAR2 and the three fielding metrics, which give three different views of the most valuable 10-year prime among the most meritorious high-prime, short-career players

Player  B+FRAR   Player   B+1/2FRAR  Player  B+3*fws
Johnson   756 Singleton 669 Browning  715
Elliott   753 Johnson   660 Singleton 693
Murphy 747 Browning  645 Jones  680
Singleton 718 Jones  644 Tiernan   667
Jones  715 Elliott   632 Johnson   654
Klein  690 Tiernan   628 Klein  643
Browning  688 Murphy 627 Elliott   638  
Tiernan   683 Klein  626 Murphy 629
Cravath   629 Cravath   588 Cravath   622
Ryan   626 Ryan   536 Ryan   574 


To conclude. it’s clear that choice of metric at this point has a very significant effect on how one evaluates players. It would be interesting to see how using other batting metrics would affect the ordering of this group of 10 players over 10 prime years. I suspect, though, that there would be much less disagreement than there is among the fielding metrics. It’s ironic that ranking Browning, a player whose fame and stature depends on his bat, should turn so much on evaluation of his fielding.

I have the data to slice up these primes into different subsets – 3 yr. consecutive peak, 5-yr. nonconsecutive peak, etc. if anyone wants to see them.

I would welcome discussion of this data!
   135. Chris Cobb Posted: August 10, 2007 at 01:08 AM (#2479579)
Urg. Has the working of the pre-tag changed again?? I used a fixed-width font and spaces, not tabs, but the columns still don't line up. Feh.
   136. sunnyday2 Posted: August 10, 2007 at 01:39 AM (#2479612)
Chris, did you say you used Cravath's MiL MLEs, so that he had the same length prime as everybody else?

Out of this group, I have overlooked Tiernan, who has dropped out of consideration. I need to revisit that, though it might not matter.
   137. jimd Posted: August 10, 2007 at 02:24 AM (#2479667)
6) Browning is the player who receives radically different assessments in the two systems. It would be helpful to know why.

Don't know why. But that won't stop me from speculating about it.

Above, with WS, you point out that there appears to be a "cushion" which holds up the corner fielders. There is most likely a similar "cushion" for each IF position, and it would be scaled proportionately higher when measured in FWS. OTOH, WARP does not hesitate to punish bad fielding stats.

WS does not distinguish between OF positions. In 1886, the PO/G in the Lou OF are estimated by BP to be LF 1.44, CF 1.36, RF 1.47; avg 1.42. In WS, a CF and a RF/LF with a similar range will get the same WS, other things being equal. BP will consider the OF position and decide that that might be exceptional for (1886) RF and terrible for CF, and reward or penalize the FRAA (and FRAR) as appropriate.
   138. Brent Posted: August 10, 2007 at 03:59 AM (#2479739)
I'm sure that each voter will draw different conclusions from Chris's study. Mine is that Singleton, who appears near the top of the list no matter which assumptions are used, deserves more support than he's been getting. Of course, this comment is coming from the BFOKS.
   139. 'zop sympathizes with the wrong ####### people Posted: August 10, 2007 at 04:12 AM (#2479745)
Here's what I don't get Chris; basically, you've chosen a bunch of arbitrary ways to rank players, none of which really capture what really matters:

How many games was he worth to his team?

Any list that purports to rank players by any metric other than that, be it EQA* WinShares over the Outs produced to the square, or some other arbitrary concontion, is basically just a list that ranks players by that metric, and doesn't necessarily, or even probably, measure actual value.
   140. sunnyday2 Posted: August 10, 2007 at 04:44 AM (#2479759)
In other words, it's "poorly calibrated"?
   141. Paul Wendt Posted: August 10, 2007 at 04:55 AM (#2479766)
Six features of this chart are notable.
1) Win shares views fielding value as approximately 1/2 as valuable, overall, as WARP2.


for outfielders. Right?

4) WARP and win shares mostly agree on player order within 19th and 20th century groups. In the 20th century, it is Elliott Murphy Johnson Klein Singleton/Cravath,

Cravath is alone between 1900 and 1930, roughly. Otherwise these are two widely separated groups --only two if it's true that the measures have stable strengths and weaknesses covering the 1930s/40s to the 1970s/80s at once.

Now, for a last chart, here are the results of summing BRAR2 and the three fielding metrics, which give three different views of the most valuable 10-year prime among the most meritorious high-prime, short-career players

how about "the most meritorious players selected for useful contrast with Browning"?

Ryan is a long-career player with a 5-year peak but no prime short of his career.
Murphy is a long-career player who was remarkably unproductive outside his prime years.

Because he finished with four part-seasons, and missed significant playing time along the way, and fielded CF in some seasons, Reggie Smith is closer to the Browning type than are the ultra-durable short-career Ken Singleton and Bob Johnson (or Ryan or Murphy). He ranked in the top ten by OPS+ only five times.
   142. Paul Wendt Posted: August 10, 2007 at 05:11 AM (#2479771)
Chris Cobb:
Methodology for looking at consecutive prime. Looking at 10 years total, with one “off year” allowed to be skipped. This is to show Browning’s prime in its best light.

Alex:
Chris; basically, you've chosen a bunch of arbitrary ways to rank players, none of which really capture what really matters:

How many games was he worth to his team?

Any list that purports to rank players by any metric other than that, be it EQA* WinShares over the Outs produced to the square, or some other arbitrary concontion, is basically just a list that ranks players by that metric, and doesn't necessarily, or even probably, measure actual value.


Arbitrary?
Alternatively, relying on the 10 wins-per-game approximation, they are all calibrated in games per standard-length season.

On a good day one makes assumptions to show Browning in favorable light, tries a few different rankings in some debated family, and finds that he is not top of any one. Voila! that puts an effective ceiling on his merit.
Today was not a good day in this sense, because Browning is at the top of one list.
   143. Paul Wendt Posted: August 10, 2007 at 05:38 AM (#2479778)
136. sunnyday2 Posted: August 09, 2007 at 09:39 PM (#2479612)
Chris, did you say you used Cravath's MiL MLEs, so that he had the same length prime as everybody else?

I don't know the shape of Cravath's career including MLEs.
But they don't all have the same length prime. First, that is a posit for present purpose to show Browning in best light. Second, it is 11 years for some and 10 for others (12 for Jones, who gets no "blacklist credit" here).

Out of this group, I have overlooked Tiernan, who has dropped out of consideration. I need to revisit that, though it might not matter.

Tiernan with Browning is the one who desperately needs to be excused one season of poor play. Elliott and Ryan can put that "best 10 of 11 seasons" to good use. On the other hand, it has no value for Jones or Klein and little for Johnson (quick judgment by OPS+).
I'll wrap it up here by restating that it is a Browning-oriented study.
   144. Chris Cobb Posted: August 10, 2007 at 05:57 AM (#2479788)
sunnyday2 wrote:

Chris, did you say you used Cravath's MiL MLEs, so that he had the same length prime as everybody else?

Yes.

I suspect the deadball context may be depressing his numbers a bit, but I can't put my finger on how exactly. I haven't played around much with outs made as a measure of playing time, so I'm not sure what the pitfalls are yet.

Alex [etc.] wrote:

How many games was he worth to his team?

It would greatly advance the cause of sabermetrics if anyone were to develop a convincingly accurate way of answering this question for pre-1893 ballplayers like Pete Browning. Not having the key to 1880s value in my hip pocket, I thought it might be helpful to develop some heuristic measures derived from the available comprehensive metrics that take a stab at producing a plausible measure of player value.

As to your claim that my choices were arbitrary and unrelated to player value, I direct you to Paul Wendt's answer to that question n 142 above, which discerns my intentions admirably. I think today is a good day, regardless of Browning's topping a list, because we have a clearer idea of the source of his success in one widely used metric that can now be given more careful consideration.

Anyway, aside from scaling fielding win shares to FRAR, all I was doing in the second study was processing numbers that WARP2 offers as a measure of a player's value above replacement in an all time context.

Paul Wendt wrote:

1) Win shares views fielding value as approximately 1/2 as valuable, overall, as WARP2.

for outfielders. Right?


Well, Elliott was a majority third-baseman, so the data isn't limited to the outfield, but certainly the large majority of the data was for outfielders.

how about "the most meritorious players selected for useful contrast with Browning"?

OK, though the most meritorious high-batting-prime, short-career players might also fit the bill. Murphy and Ryan don't fit the set as well as the others because they had more career, but their batting peaks were of a height that put them into the comparison set, unlike, say, Andre Dawson or Sam Rice.

Because he finished with four part-seasons, and missed significant playing time along the way, and fielded CF in some seasons, Reggie Smith is closer to the Browning type than are the ultra-durable short-career Ken Singleton and Bob Johnson (or Ryan or Murphy). He ranked in the top ten by OPS+ only five times.

A more rigorous study would certainly have included Reggie Smith in the comparison set. With more time, I would have found a more consistent way of comparing leaderboard appearances in leagues of differing size. But as the point of this study was to see if Browning's prime, carefully examined, would stand up to the primes of other characters whose case depends on their best 10 years, Smith's profile seemed unlikely to make him particularly relevant to the case. Ryan ended up being a weak fit also, it turns out.
   145. Chris Cobb Posted: August 10, 2007 at 06:06 AM (#2479791)
Tiernan with Browning is the one who desperately needs to be excused one season of poor play. Elliott and Ryan can put that "best 10 of 11 seasons" to good use. On the other hand, it has no value for Jones or Klein and little for Johnson (quick judgment by OPS+).
I'll wrap it up here by restating that it is a Browning-oriented study.


All true, except that Klein does benefit by getting to drop his weak 1938 season (I judged Klein would show best 1929-39, less 1938, rather than including the brilliant 1928 1/3 season and forcing the inclusion of the poor 1938 campaign). Johnson benefits a bit by going 1934-44 less 1943.

If one thinks through the careers of these players, Jones actually shows quite well, in that he is not far off the pace for a ten-year prime while not getting to drop any seasons and while losing 2 1/3 prime seasons in the middle of his career.

The question of what these players have _outside_ the "ten-year prime" is a meaningful one, of course, for any voters interested in a player's total value.
   146. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 10, 2007 at 02:42 PM (#2479926)
Here's what I don't get Chris; basically, you've chosen a bunch of arbitrary ways to rank players, none of which really capture what really matters:

How many games was he worth to his team?


'zop, have you been possessed by aliens or something? Aren't you a peak voter? You've said time and again that you could care less how many total games a player was worth to his team--what matters to you is whether he was ever one of the best players in baseball (and if so, how great, and for how long).
   147. 'zop sympathizes with the wrong ####### people Posted: August 10, 2007 at 02:48 PM (#2479930)
'zop, have you been possessed by aliens or something? Aren't you a peak voter? You've said time and again that you could care less how many total games a player was worth to his team--what matters to you is whether he was ever one of the best players in baseball (and if so, how great, and for how long).

Indeed. But I measure greatness by "# of wins earned by player in a season".
   148. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 10, 2007 at 03:00 PM (#2479937)
Then you'd still be a career voter. Would I be correct in synthesizing that you vote based on something like WARP minus X per season? Value above average? Value above 20% above average? Etc.?
   149. DavidFoss Posted: August 10, 2007 at 03:13 PM (#2479945)
Indeed. But I measure greatness by "# of wins earned by player in a season".

HappyJackChesbro must have lost his internet connection, or he'd be chiming in and saying "Vote for me!" :-)
   150. 'zop sympathizes with the wrong ####### people Posted: August 10, 2007 at 03:45 PM (#2479976)
Then you'd still be a career voter. Would I be correct in synthesizing that you vote based on something like WARP minus X per season? Value above average? Value above 20% above average? Etc.?

Dan, I do consider myself a career voter. I just believe that each win provided by a player is exponentially more valuble than the last. My belief is supported by the way the actual free-agency market values players.

This leads me to weigh peak years more heavily than "ordinary" seasons.

In addition, I believe that there is a line that all HoMers should cross in their best season...somewhere around 9.5 BP WARP, I guess. If you don't reach that at your peak, you were never really great. It is, however, highly unusual for a player that accrues sufficient career value to be worth in my system not to reach a reasonably high peak at his best.

Furthermore, I believe pitcher injuries are essentially random, so, rewarding pitchers for longevity at the expense of peak is like rewarding the pitchers who survived bombing runs in the B-17; maybe the guns just missed their plane? For that reason, I am highly sympathetic to high peak, short career pitchers if the peak was clearly great (Dean, Koufax).
   151. DanG Posted: August 10, 2007 at 05:34 PM (#2480078)
In other words, it's "poorly calibrated"?

How about: "Quantitated so as to be askew from plumb"?
   152. sunnyday2 Posted: August 10, 2007 at 07:38 PM (#2480168)
And I of course consider myself to be a career voter, too. It's just that I only look at that part of the career that has the most peakiness.
   153. Chris Cobb Posted: August 10, 2007 at 09:00 PM (#2480222)
Here’s an addition to the Browning study: Browning contemporaries Fred Dunlap and Ned Williamson, both of whom had 10-year major-league primes in the 1880s. First, a look at fielding evaluations, just of the nineteenth-century players in the consideration set, and then a full listing of the 10 high-peak hitters plus Dunlap and Williamson, combining hitting and fielding.

Player  FRAR2  Player .5*FRAR2   Player  3*fws
Dunlap  364 Dunlap  182 Williamson 218.1
Williamson 353 Williamson 177 Dunlap  174.1
Ryan 180 Ryan  90 Ryan 127.7
Jones   141 Jones 70 Browning   113.0
Tiernan 110 Tiernan  55 Jones   106.6
Browning 85 Browning 42 Tiernan  94.0

Range 279 runs
,   Range 140 runs,   Range 122.1 runs
31 wins, or 15.6 wins, or  13.6 wins, or 
3 wins/year 1.6 wins/year  1.4 wins/year

Player  B
+FRAR Player  B+1/2FRAR Player  B+3*fws
Dunlap  812 Singleton  669 Browning   715
Johnson 756 Johnson 660 Singleton  693
Elliott 753 Browning   645 Jones   680
Murphy  747 Jones   644 Tiernan 667
Singleton  718 Elliott 632 Johnson 654
Jones   715 Dunlap  629 Klein   643
Williamson 715 Tiernan 628 Elliott 638
Klein   690 Murphy  627 Murphy  629
Browning   688 Klein   626 Dunlap  629
Tiernan 683 Cravath 588 Cravath 622
Cravath 629 Williamson 579 Williamson 579
Ryan 626 Ryan 536 Ryan 574 


In the fielding numbers, I’ll note that fielding win shares scales its infielder numbers much closer to 1/2 FRAR2 than to FRAR2. Williamson is rated ahead of Ryan, Jones, and Tiernan by almost exactly the same amount in 1/2 FRAR and fws, meaning each gets a very similar upward bump. Dunlap, however, drops down. This may be an effect of win-shares’ altered formula for pre-1930 second base. Browning remains by far the largest anomaly, however.

Dunlap’s WARP totals sure make one wonder why Browning is on the cusp of election and Dunlap is receiving almost no support (yes, jimd has been asking that question for years!) It may be that WARP2 overrates nineteenth-century second basemen by projecting them into an all-time context where they have a different spot on the fielding spectrum than they did in the nineteenth century game, but Dunlap could be docked quite a bit and still rank with players who are getting a lot more support than he is.
   154. jimd Posted: August 10, 2007 at 10:41 PM (#2480289)
6) Browning is the player who receives radically different assessments in the two systems. It would be helpful to know why.

Above I gave two possible reasons for this. I've thought of a third.

Win Shares does not discount for league quality; FRAR2 does.
Chris, has any AA discount been applied to FWS from 82-84?

I looked at the season-by-season discrepancies between FRAR2/2 and 3*FWS.
They are concentrated between 1882 and 1887.
82 10
83 6
84 11
85 11
86 10
87 17
88-94 11 (no more than 3 in any season).

IF cushion would only apply in 1884 (Browning appears to have been a good IF in 1882, terrible in 1883-84, though he got more playing time there in 1884). CF evaluation issues would apply from 1885-88, and would peak in 1887 (when he was both bad and a full-time CF). League quality issues would be in 1882-84, maybe 85.
   155. Paul Wendt Posted: August 10, 2007 at 11:00 PM (#2480302)
Dunlap’s WARP totals sure make one wonder why Browning is on the cusp of election and Dunlap is receiving almost no support (yes, jimd has been asking that question for years!) It may be that WARP2 overrates nineteenth-century second basemen by projecting them into an all-time context where they have a different spot on the fielding spectrum than they did in the nineteenth century game, but Dunlap could be docked quite a bit and still rank with players who are getting a lot more support than he is.

This study does not show Dunlap in best light, not even in good light. Because he played only ten years, he does not benefit from the "10 seasons in 11 years" option.
As Browning and Tiernan (who did play 11+ years), Dunlap needs it "desperately", burdened with OPS+ 75 in year ten. He missed 40% of years eight and nine, and the best light for him would be 7-year or maybe 5-year.
   156. jimd Posted: August 10, 2007 at 11:14 PM (#2480321)
Dunlap’s WARP totals sure make one wonder why Browning is on the cusp of election and Dunlap is receiving almost no support

Easy. Browning's a bat. Dunlap's an ambidextrous, gloveless fielding wizard.
So what if Fred has 3 seasons better than Browning's best. (WARP3)

Thanks for the good words on Dunlap. I'd like to welcome Juan V aboard the Dunlap bandwagon, and I do miss Michael Bass, who has supported Fred even more strongly than I do.
   157. Chris Cobb Posted: August 10, 2007 at 11:47 PM (#2480400)
Win Shares does not discount for league quality; FRAR2 does.
Chris, has any AA discount been applied to FWS from 82-84?


Whoops!

No, not at all. Thanks for catching this. For consistency, what I think I should do is find the competition discounts in FRAR2 for all the relevant player seasons and apply them across the board. That will take some time, so I think I'll do a quick fix first, which would be to adjust for AA/NL quality differences only in Browning's record (and I'll do the UA year for Dunlap.) That I can have ready tomorrow morning.

Above I gave two possible reasons for this. I've thought of a third.

I should mention that I think your reasons are plausible. I intend to put up a year-by-year breakdown of Browning's fielding value by WARP and win shares to see if that will help confirm or rule out these reasons.

Thanks for the good words on Dunlap.

You're welcome. They seemed justified by the numbers.

Paul Wendt wrote:

This study does not show Dunlap in best light, not even in good light. Because he played only ten years, he does not benefit from the "10 seasons in 11 years" option.
As Browning and Tiernan (who did play 11+ years), Dunlap needs it "desperately", burdened with OPS+ 75 in year ten. He missed 40% of years eight and nine, and the best light for him would be 7-year or maybe 5-year.


What Dunlap "desperately" needs is MLEs for his play in International Association! Then we could look at 1879-88. (If only we had the data to do it . . . ) Since the study shows 10 consecutive years, it indicates that there is no counterargument to an assertion of value by this measure. He is far ahead of Browning in season-adjusted WARP2, even when Browning gets the benefit of dropping a bad year.

I certainly have no objection, of course, to your reminding everyone that this particular "slice-of-career" perspective is designed to allow Browning to appear at his best. Since that is the case, I hope supporters of Browning will think carefully about how well Dunlap does when placed into the same frame, even though it is not the most favorable presentation for him.
   158. sunnyday2 Posted: August 11, 2007 at 01:11 AM (#2480626)
Of course there is a discount for Dunlap's monstrous year in 1884 in a league that was probably about half as good as the IA. Dunlap remains in my consideration set, BTW, but just in the interest of full disclosure. I discount the UA by .65, yes, that's the discount, the actual values are .35 of the NL.
   159. Paul Wendt Posted: August 11, 2007 at 02:32 AM (#2480812)
.35 calibrated in runs, about equivalent to OPS+?
calibrated in wins, about equivalent to win shares?

Then Dunlap's 7-year pattern, representing his full seasons 1880-86, looks like this:
OPS+ 141 155 120 146 87 122 121
Win Shares 17 15 14 16 13 14 17 (1884 slightly his worst season)

--
What Dunlap "desperately" needs is MLEs for his play in International Association! Then we could look at 1879-88. (If only we had the data to do it . . . )

I can send IA 1879 official statistics to a street address. They were reprinted in the Minor League Baseball Research Journal (title?), SABR Minor Leagues Cmte.
Or I can describe their scope.
   160. Paul Wendt Posted: August 11, 2007 at 02:42 AM (#2480833)
Here is Glasscock's 1880-87 counting 1884 at .35
OPS+ 88 92 145 106 92/79 121 151 104
Win Shares 10 10 19 14 13 17 22 17

schedule increasing from 84 to 126 games
   161. Chris Cobb Posted: August 11, 2007 at 02:42 AM (#2480834)
OK. So using FRAR that were competition adjusted but using fws that were not did not make sense. Here is a corrected set of charts. The first one shows fielding win shares for each player’s 10 seasons, season-adjusted and scaled to FRAR by multiplying them by three. Then the competition adjustment is applied. It is 1/2 * FRAA2 – FRAA1. all season-adjusted. WARP2 adjusts for competition level by making a linear adjustment to average value, which can be found by subtracting FRAA in WARP1 from FRAA in WARP2. I then multiplied this number by .5, since win shares places about half the value on fielding as WARP2 (for outfielders, at least) and finally season-adjusted it. Finally, I have placed in parentheses the value of 1/2 FRAR2, season-adjusted, for each player for comparison purposes.

Player 3*fws – comp. adj. = new total (1/2 FRAR2)
Williamson 218.1 – 30 = 188.1 (177)
Dunlap 174.1 – 39 = 135.1 (182)
Elliott 127.8 +2 = 129.8 (122)
Murphy 122.4 + 3 = 125.4 (120)
Ryan 127.7 – 10 = 117.7 (90)
Johnson 91.5 – 5 = 86.5 (97)
Tiernan 94.0 – 9 = 83.0 (55)
Browning 113.0 – 30 = 83 (42)
Klein 80.8 – 2 = 78.8 (64)
Jones 106.6 – 35 = 71.6 (70)
Cravath 75.7 – 7 = 68.7 (41)
Singleton 73.1 – 7 = 67.1 (49)

Now, here is a table showing total value above replacement by this metric, which combines WARP2’s BRAR with fielding win shares, converted to FRAR, and competition adjusted, and everything season-adjusted.

Totals BRAR2 + (3*fws) + (1/2 *[FRAA2-FRAA1]), everything season-adjusted
Singleton 686
Browning 685
Tiernan 658
Johnson 649
Jones 645
Klein 641
Elliott 640
Murphy 632
Cravath 615
Dunlap 583
Ryan 564
Williamson 549

By this measure, Browning remains near the top, but he is nosed out by Ken Singleton, who now leads by both measures that place less weight on fielding. I note that the top 5 players on defense fill five of the bottom six slots on the combined value list. Myself, I have a hard time believing that win shares is scaling fielding value correctly, but this is what the metric shows.

Finally, a re-posting of the season-adjusted WARP2 results, expressed in runs:
Dunlap 812
Johnson 756
Elliott 753
Murphy 747
Singleton 718
Jones 715
Williamson 715
Klein 690
Browning 688
Tiernan 683
Cravath 629
Ryan 626
   162. Paul Wendt Posted: August 11, 2007 at 01:50 PM (#2480957)
Chris Cobb #157
But as the point of this study was to see if Browning's prime, carefully examined, would stand up to the primes of other characters whose case depends on their best 10 years, Smith's profile seemed unlikely to make him particularly relevant to the case. Ryan ended up being a weak fit also, it turns out.

Earlier you (Chris Cobb) mentioned both 10 seasons and 5 seasons. Despite his long career, Ryan must show up better for 5 than for 10, perhaps much better. For some, the career/peak combination is appealing or even required (Alex?).

Despite his fine 10-year prime, Chuck Klein must also show up better for 5 than for 10, because his 5 full seasons in Philadelphia were so dominating.

And Bob Elliott at 5 years has the third-base postwar run, shedding the outfield wartime run. He should be getting a lot of support from peak voters, for he can't be dismissed as another corner outfielder or as one more from the overrepresented thirties. (Oops, black and white baseball were mature systems in parallel, we shouldn't have elected Averill and Medwick. For peak voters, I think that must be the argument against Chuck Klein. Else what's not to like?)

#161
I note that the top 5 players on defense fill five of the bottom six slots on the combined value list. Myself, I have a hard time believing that win shares is scaling fielding value correctly,

I agree, but note that four of the top 5 players on defense are the(?) four of 12 who did not enjoy ten prime seasons so "10 of 11" does not show them in best light.

--
145. Chris Cobb Posted: August 10, 2007 at 02:06 AM (#2479791)
> Tiernan with Browning is the one who desperately needs to be excused one season of poor play. Elliott and Ryan can put that "best 10 of 11 seasons" to good use. On the other hand, it has no value for Jones or Klein and little for Johnson (quick judgment by OPS+).
> I'll wrap it up here by restating that it is a Browning-oriented study.

All true, except that Klein does benefit by getting to drop his weak 1938 season (I judged Klein would show best 1929-39, less 1938, rather than including the brilliant 1928 1/3 season and forcing the inclusion of the poor 1938 campaign). Johnson benefits a bit by going 1934-44 less 1943.


ok

136. sunnyday2 Posted: August 09, 2007 at 09:39 PM (#2479612)
Chris, did you say you used Cravath's MiL MLEs, so that he had the same length prime as everybody else?

[Yes.]
Oh, I suppose Cravath has no worst season to excuse. His ten best seasons are all in a run. Or do you Chris go back to the PCL 1906/07 and drop 1908? Brent gave the pre-1908 a close look. I haven't noticed how many voters "give credit" so early.
   163. sunnyday2 Posted: August 11, 2007 at 03:10 PM (#2480998)
Paul, my recollection though I don't have the numbers in front of me: If you extrapolate Dunlap's WS out to 162 games, then .35 for 1884 looks just like his adjacent seasons.
   164. Paul Wendt Posted: August 11, 2007 at 04:26 PM (#2481023)
By win shares it's close, 1884 is slightly his worst season, 16% below his 1880-1886:

Then Dunlap's 7-year pattern, representing his full seasons 1880-86, looks like this:
OPS+ 141 155 120 146 87 122 121
Win Shares 17 15 14 16 13 14 17 (1884 slightly his worst season)


But I suppose that the NL veteran who enjoyed the best UA 1884 season played somewhat better than in his typical NL season. Down 16% is implausible.

Pete Palmer inferred multiplier 0.76 for OPS+. That happens to give Dunlap 187.5 rather than 87.5
Dunlap's 7-year pattern, OPS+ 1880-86, using multiplier 0.76 for 1884, his full season in the Union Association
OPS+ 141 155 120 146 187 122 121
That is a moderate career year, 20% better than his best.

(0.76 is inferred from the 1883-1885 records of all major league players. For pitchers the finding was 0.90, iirc, and in the printed encyclopedias Palmer used 0.80 as a convenient round number for both batting and pitching. Still uses it, I believe.)
   165. John (You Can Call Me Grandma) Murphy Posted: August 11, 2007 at 07:22 PM (#2481077)
And Bob Elliott at 5 years has the third-base postwar run, shedding the outfield wartime run. He should be getting a lot of support from peak voters, for he can't be dismissed as another corner outfielder or as one more from the overrepresented thirties. (Oops, black and white baseball were mature systems in parallel, we shouldn't have elected Averill and Medwick. For peak voters, I think that must be the argument against Chuck Klein. Else what's not to like?)


The problem with Elliott is that he almost runs into the Eddie Mathews era, where third basemen perceptions start to change. That may be affecting some voters image of him, since I'm surprised he's not much higher at this point. I think the best ML third baseman of the Forties deserves much more love.

As for Averill, I liked him when adding in MLE credit. I wasn't that crazy about Medwick, however.
   166. Esteban Rivera Posted: August 13, 2007 at 01:48 PM (#2483208)
Chris, what a great study. I've been reconsidering Dunlap for the past few weeks and this convinces me that he should be higher on my list. The only question I have is if the BRAR2 and FRAR2 numbers are from Prospectus? If that's the case, then the issue would be how much trust or stock one puts into their all-time league adjustments, which could affect the placement of some of the players if one does not believe the adjustment is enough or if one believes it is too harsh.

Thanks again Chris, this has been very informative.
   167. Chris Cobb Posted: August 13, 2007 at 03:19 PM (#2483273)
The only question I have is if the BRAR2 and FRAR2 numbers are from Prospectus? If that's the case, then the issue would be how much trust or stock one puts into their all-time league adjustments, which could affect the placement of some of the players if one does not believe the adjustment is enough or if one believes it is too harsh.

Yes, the BRAR2 and FRAR2 numbers are from WARP2 on Baseball prospectus. The scale of adjustments is, of course, a big issue. In the case of Browning, we have two sets of adjusted scores to consider for batting. One is the set of adjusted numbers that Brent calculated on the basis of his study of players who moved between the NL and the AA. The other is the BP conversions. I compared the two together, and I found that the BP adjustments reduce Browning's performance less for the early years of the AA than Brent's do, so of the _presented_ conversions, BP's is the milder. If we want to adjust for fielding quality, BP's adjustments are all we have to work with. They are quite steep for early AA, but there's not much difference between the adjustment for later AA and the adjustment for the NL at the same time.

Let me add that the adjustment for _quality of competition_ in WARP2 -- the modification of value above average -- is not very large for fielding value except for a few very weak years: it is seldom more than 3 runs, season-adjusted. You can see the scale of these adjustments over 10 years in the first table in post 161 above. The competition adjustment in batting are also usually small. The larger changes in WARP2 have to do with shifts in fielding replacement level, which WARP uses to sort out the relationship between pitching and fielding in defensive value. When these are translated into WARP's all-time context, which is basically post-1970 baseball, the changes can be substantial. But when comparing players from different eras, I think this adjustment is appropriate. The value of fielding relative to pitching was considerably higher in the 1880s than in the 1970s, but the relative value of position players to one another was not so different, so we shouldn't say that 1880s position players as a group were more valuable than 1970s position players as a group because the former had more defensive value. People often think that the WARP2 adjustments are "harsher" than they really are because they don't recognize the meaning of the shift in fielding replacement level, which has nothing to do with the quality of the players but is simply a feature (though not an especially sensible one) of the WARP accounting system.
   168. Howie Menckel Posted: August 13, 2007 at 04:24 PM (#2483325)
"The problem with Elliott is that he almost runs into the Eddie Mathews era, where third basemen perceptions start to change. That may be affecting some voters image of him, since I'm surprised he's not much higher at this point. I think the best ML third baseman of the Forties deserves much more love."

I'm generally the Consensus-meister General around here, and Elliott probably is my most "out of the mainstream" vote, so..... you do the math!

:)
   169. Paul Wendt Posted: August 13, 2007 at 06:36 PM (#2483513)
. . . done

If not for Elliott, Howie Menckel would have the dull grey jersey locked up.
   170. sunnyday2 Posted: August 15, 2007 at 12:07 PM (#2485615)
This is sort of a reaction (not a response per se but a reaction) to the current discussion of Browning vis-a-vis Eddie Murray, but I thought I would move it over here since it's more about Browning than about Murray. What is interesting to me is one of the themes of discussion over the years, a weak theme, but a theme nonetheless. It is the idea that, We've elected enough players from that era. And the comparison really becomes Browning (or C. Jones) and Bob Johnson, the "hitters" who come from eras from which "we've elected enough players."

As a supporter of Pete Browning AND Charley Jones (and Ed Williamson), and NOT of Bob Johnson, I've said it myself. I don't support Bob Johnson because we've elected enough players from that era. The obvious follow up is, well, haven't we also elected enough players from the 19C? My answer is, well, maybe, but they're not the right guys.

So why can't Bob Johnson's supporters (and of course I also support Dizzy Dean) say the same thing? Well, they can.

So, anyway, my insight after all these years is that "we've elected enough players from that era" is in fact never a valid justification for a vote. I promise not to use it against Bob Johnson anymore (if all of you agree not to use it against my 19C guys ;-)
   171. Paul Wendt Posted: February 24, 2008 at 07:17 PM (#2698648)
1.
Harry Stovey scored runs. For example, as a rookie with a rookie team that finished fourth in runs, he ranked second in NL 1880. In 1882 the Worcesters were last in runs but Stovey ranked third. He scored 90 in 84 games, pitcher Lee Richmond scored 50 in 55, and two men tied for third with 33 runs.

During the 19-aughts, project time, OCF made this point emphatically. He says in a Group 3 Ballot comment today, "I'm still convinced that baserunning made a much, much larger impact in his day than in later baseball, and that he was uniquely good at it."

2.
Harry Stovey played first base for about 5-1/2 seasons 1880-86: roughly half time '80, '82, '86, otherwise full time.

A few days ago I mentioned AA 1Bmen Comiskey, Orr, and Reilly in amplifying the point that Stovey played first at the wrong time to get much credit according to any version of runs above position-average.
St Louis 1B Charlie Comiskey hit OPS+ 93 during Stovey's 1B tenure, thereafter 77.
New York 1B Dave Orr arrived in the majors late in 1883 and played 3+ seasons as a 1B colleague.
Cincinnati 1B John Reilly arrived in 1883 when Stovey joined the league.

In 1884, Reilly-Orr-Stovey led the league with OPS+ 190-190-185.

In the National League, ABC were not alone (Anson-Brouthers-Connor).

Firstbasemen at bat, NL 1882 (Stovey's and Worcester's last NL season)
OPS+, rank on team (ordered by team standings)
186 1 Anson (GORE 148)
142 2 Start (HINES 151)
138 2 Morrill (WHITNEY 183)
199 1 Brouthers
113 3 Phillips (GLASSCOCK 145)
_80 6 Powell (BENNETT 151)
185 1 Connor
137 1 Stovey

Parenthetically I have identified the regular players with OPS+ greater than Stovey's. For Boston Jim Whitney played 61 of 85 games including 48 pitcher starts (GS).
So Stovey at OPS+ 137 ranked 11th in the league behind five 1Bmen and five others.

During its maiden 1882 season the AA provides a comical contrast. By good fortune my stab at 1882 has turned up a good (extreme) example for assessment of player hitting both (1) relative to fielding position and (2) relative to own-league or major-leagues average.

Firstbasemen at bat, AA 1882 (first AA season, several startup teams)
OPS+, rank on team (ordered by team standings)
_96 6 Stearns (CARPENTER 155)
123 3 Hecker (BROWNING 222, WOLF 140)
105 3 J.Latham (O'BRIEN 143)
_57 8 Lane (SWARTWOOD 186, TAYLOR 148, MANSELL 142)
_85 5 Comiskey
111 3 Householder (BROWN 144)

Eight regular players hit 140 or better, none at first base. Indeed, the firstbasemen were led by Louisville's Guy Hecker at merely OPS+ 123. Hecker played 78 of 80 games including 11 pitcher starts; next year he started 50 of 98 games at pitcher.

Despite losing Hecker to the pitcher's box, the regular AA firstbasemen were much better batters in 1883, the league's second season.

Firstbasemen at bat, AA 1883
OPS+, rank on team (ordered by team standings)
161 1 Stovey
121 1 Comiskey
150 1 Reilly (JONES 147)
_90 3 Brady
_90 7 J.Latham (BROWNING 177, WHITING 141)
104 5 Field
185 1 Swartwood
101 3 Stearns

Six batters including three 1Bmen hit 140 or better.

Dave Orr became a regular player, 1B for the New York Mets, late in the season (age 24), hitting 195 in 12 games. In 1884, Reilly-Orr-Stovey led all batters with OPS+ 190-190-185.
   172. Paul Wendt Posted: August 07, 2008 at 08:55 PM (#2894540)
> I have Stovey as clearly a left fielder,

For what it's worth he is '1b,p' in the 1879 National Association season statistics. He was a regular player, at least very good; certainly no MVP candidate, contemporary or statistorian.
I suppose that even one NL MVP season as 1B would be short of tipping his career fielding position to first.

Heh-heh,
Harry Stovey and the man he exchanged with, Henry Larkin, did not play long enough at 1B and LF respectively to earn win shares letter grades there (maybe because the seasons were shorter then).
After the position change, Bill James gives Stovey B+ in left, Larkin a rare D- at first.

Skimming the print Win Shares, here are the trailing fielders at firstbase.
F: Nick Etten, Dick Stuart, Frank Thomas
D- Henry Larkin, Al Oliver, Mo Vaughn
D: Eddie Morgan
D+ Dale Long, Gerald Perry

not so unusual as I thought when I was half-way through the alphabetical list.
Al Oliver is B+ in the outfield which is at worst below average for fulltime centerfielder.
   173. GGC don't think it can get longer than a novella Posted: May 01, 2009 at 08:32 PM (#3160754)
99. John (You Can Call Me Grandma) Murphy Posted: August 05, 2007 at 09:46 AM (#2471815)

A thread entitled “Stovey and Browning” should contain more than seven posts.


Dan, they were discussed extensively in many discussion threads.


Those of us reading these a few years later appreciate the centralized discussion.
   174. John (You Can Call Me Grandma) Murphy Posted: May 02, 2009 at 03:00 PM (#3161323)
Those of us reading these a few years later appreciate the centralized discussion.


I understand completely, Jon, but much of the discussion about these two was already hashed out before Joe posted this thread, however.
   175. Bleed the Freak Posted: March 29, 2010 at 11:23 AM (#3487786)
I find Harry Stovey and Pete Browning as two of the most interesting figures to assess for proper value in baseball history.

Whether the following statistics are merely junk stats, they may be worth at least mentioning, and have some merit.

Regarding, Harry Stovey he has a calculated speed score of 8.4, the highest during his time-era, and this does not include his first six seasons. Vince Coleman has the highest total ever at 8.8.

Fangraphs calculation for speed score can be explained as follows:
Speed score (Spd) is a statistic developed by Bill James that attempts to rate how fast players are. Different locations include slightly different components, but the FanGraphs version consists of, “…Stolen Base Percentage, Frequency of Stolen Base Attempts, Percentage of Triples, and Runs Scored Percentage.”

While this statistic colors Stovey as great, it unfortunately cannot utilize stolen base percentage and frequency of stolen base attempts, since these were not tracked during his time era.

Stovey also posted an impressive 451 wRAA, good for 103rd all-time. The only non-HOM player with a higher wRAA is Bob Johnson. It is arguable that, compared to the non-HOM players, Stovey would have the most value, defenseively, before docking of AA seasons, until you reach Jimmy Ryan at 376. Stovey played 1B when the defense was much more valuable for half his career, and LF for the other half.

wRAA is based upon Tom Tango's wOBA, a statistic that is an improvement on OPS, IMO.

wRAA – This is the number of runs above average based off wOBA. It is calculated as (wOBA – lgwOBA) / wOBAScale * PA

wOBAScale is typically around 1.15, but changes slightly from year to year depending on the weights used in wOBA.

Granted, some air should be taken out of these numbers with his AA play, but at least we are talking about a player that was hopefully not on the mistakes side of the HOM, but merely in the lowest quartile. Stovey did leave the AA before it's final rotten quality years and did not play in it during its early formation years, playing in the PL/NL from 1880-1882, 1890-1893, and the AA from 1883-1889.

Pete Browning shows up at 424 wRAA, with his poor defense, does not seem to improve his standing.

Merely some food for thought. The threads have been really quiet. I stumbled across this information last night, and thought others might be interesting in knowing as well.

I will be interested in following the MVP project, but the task seems to be too large of a time commitment to become involved with.
   176. Dag Nabbit is part of the zombie horde Posted: September 03, 2010 at 08:12 PM (#3633183)
   177. Mark Donelson Posted: September 08, 2010 at 08:03 PM (#3636275)
I will be interested in following the MVP project, but the task seems to be too large of a time commitment to become involved with.


Apparently you weren't alone there. :)
Page 2 of 2 pages  < 1 2

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
dirk
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 1.0688 seconds
49 querie(s) executed