User Comments, Suggestions, or Complaints | Privacy Policy | Terms of Service | Advertising
Page rendered in 1.3708 seconds
56 querie(s) executed
You are here > Home > Hall of Merit > Discussion
| ||||||||
Hall of Merit — A Look at Baseball's All-Time Best Monday, January 13, 20202021 Hall of Merit Ballot Discussion2021 (December 2020)—elect 3 Top 10 Returning Players Newly eligible players Tim Hudson |
BookmarksYou must be logged in to view your Bookmarks. Hot Topics2023 Hall of Merit Ballot Discussion
(168 - 1:46pm, Jul 13) Last: DL from MN Most Meritorious Player: Trophy Case (71 - 6:53pm, Jul 08) Last: DL from MN Most Meritorious Player: 1900 Results (6 - 6:28pm, Jul 08) Last: DL from MN Most Meritorious Player: 1900 Ballot (11 - 4:03pm, Jul 06) Last: DL from MN Most Meritorious Player: 1918 Results (6 - 8:58pm, Jun 29) Last: kenna Most Meritorious Player: 1920 Ballot (25 - 8:55pm, Jun 29) Last: kenna Most Meritorious Player: 1900 Discussion (9 - 1:14am, Jun 29) Last: Harmon Ripkowski Hall of Merit Book Club (6 - 12:45pm, Jun 28) Last: progrockfan Most Meritorious Player: 1899 Results (6 - 1:51pm, Jun 22) Last: DL from MN Most Meritorious Player: 1899 Ballot (10 - 9:54am, Jun 22) Last: TomH Most Meritorious Player: 1899 Discussion (9 - 9:04am, May 31) Last: DL from MN Most Meritorious Player: 1898 Results (4 - 3:22pm, May 06) Last: DL from MN Most Meritorious Player: 1898 Ballot (8 - 10:07am, May 05) Last: DL from MN Most Meritorious Player: 1898 Discussion (7 - 1:08am, May 02) Last: Harmon Ripkowski Most Meritorious Player: 1897 Results (2 - 4:29pm, Apr 06) Last: DL from MN |
|||||||
About Baseball Think Factory | Write for Us | Copyright © 1996-2021 Baseball Think Factory
User Comments, Suggestions, or Complaints | Privacy Policy | Terms of Service | Advertising
|
| Page rendered in 1.3708 seconds |
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.
I would like the Doc to weigh in on how close to his ballot Cravath is...
https://homemlb.wordpress.com/2020/08/13/special-consideration-deadball-hitters/
Now here’s a guy with a serious case for special consideration. Cravath started out in the PCL in its inaugural season. He pasted the league for five years then finishing among the top five in the important offensive categories every single year. The Coast League sold players to the big leagues from 1904 on. Most were cups of coffee as you’d expect, but by 1910 they faucet had been turned all the way on and never stopped running.
Finally, he went to the AL for 1908, and there posted a 136 OPS+ for the Red Sox that would have been eighth in the league had he enough plate appearances to qualify. In other words, he simply kept doing exactly we he had done in California. That wasn’t good enough for Boston, though. They wanted to get young and hungry, Cravath wasn’t fast and didn’t field his position all that well, and they prized athleticism among their outfielders. Cravath was broad and brawny, not lithe and quick. Can’t argue to much with the results, though, because Speaker-Hooper-Lewis, the so-called Golden Outfield, kicked the AL’s butt for years. Cravath got a couple tiny look-sees by Chicago and Washington in early 1909, didn’t hit well, and that was that. Down to Minneapolis and the American Association for the now twenty-eight-year-old.
He had a typically good year in 1909 for the Millers and then something changed in his game. He suddenly started to use Nicollet Park’s tiny right field to his advantage, and put up big-time numbers. He led the AA in everything important or finished second in 1910 and 1911, and, finally, the big leagues noticed. After some controversy, the Phillies wound up with his rights, and it was a match made in Homer Heaven. The Baker Bowl’s right field was nearly identical to the teeny tiny right field in Minneapolis, and Cravath simply kept doing what he did in Minnie. Six home-run crowns later, he retired in 1920 having earned about 30 WAR in this thirties, an astonishing total for someone who barely played in the majors before age 30.
So what was Cravath anyway? A park illusion? A guy whose glove was so bad no one wanted his bat? Or just an unlucky sun of a gun?
I would argue that the one true thing about Cravath is that he was a, if not the, leading batter in every league he ever played in. In the PCL he annually led or placed among the top four in all the important offensive categories. Then he had the 136 OPS+ in the AL. From 1909–1911, he owned the American Association. In the National League, he took advantage of his home park to create mounds of offense for his team. He always hit, always.
If you look at his minor league stats, his AL stats, and his road stats in the National League, they are of a piece. He never stopped being the player he always was, but he somehow added something to his game at home that went above and beyond what he normally did as a hitter.
So going backwards in time…. It seems to me that, at an absolute minimum, Cravath merits credit for his 1909–1912 years at Minneapolis. He’d already shown he could hit in the majors as a twenty-eight-year-old, after all. But that big-league-ready bat didn’t come out of nowhere. Cravath hit hard from the day he arrived in Los Angeles in 1903. He was a finished or darn-near-finished product in his rookie season in the PCL, which may not be all that surprising because he started at age twenty-three. The numbers we have for the Coast League show consistent excellence from day one. Had he started his career in 1908 or 1913, he would have been sold to the majors within a year or three based on the experiences of other players who went to the east after he did.
Here’s the MLE I put together for him. We have no walk data for any of Cravath’s minor league seasons, and it’s an important part of his game, but we do have information on virtually all of those leagues’ pitchers that includes walks and innings pitched. I found the ratio of his Walks/27 Outs to the ratio of his major-leagues’ Walks/9 innings. It turned out to be nearly double the league rate. Then I applied that ratio to minor leagues he played in and ultimately to his known outs. The 1909 line incorporates his big-league play as well, and I’ve tacked his 1908 and 1912–1914 big-league years on here for comparison.
YEAR AGE G PA BAT RUN DP FIELD POS RAA WAA REP WAR NOTE
===========================================================================
1903 22 153 688 -10 -1 0 -3 -8 -22 -2.5 24 0.2 MLE
1904 23 150 660 0 -1 0 -3 -8 -12 -1.5 23 1.5 MLE
1905 24 147 611 12 -1 0 -3 -7 1 0.1 21 2.7 MLE
1906 25 148 625 4 -1 0 -3 -7 -7 -0.8 21 1.8 MLE
1907 26 148 582 20 0 0 -3 -7 10 1.2 20 3.7 MLE
1908 27 94 340 11 -1 0 -3 -4 3 0.3 12 1.7 MLB
1909 28 148 577 13 0 0 -3 -7 4 0.5 20 2.9 MLE/MLB
1910 29 151 661 33 -1 0 -3 -8 21 2.3 23 5.0 MLE
1911 30 154 656 36 -1 0 -3 -8 24 2.5 22 4.8 MLE
1912 31 130 503 9 -1 0 2 -6 5 0.4 16 2.0 MLB
1913 32 147 594 47 0 0 -4 -7 36 3.8 18 5.7 MLB
1914 33 149 605 38 -1 0 -10 -8 20 2.1 19 4.0 MLB
If this MLE is correct, his career value in my way of looking at things and expressed in 162 notation would rise to about 61 WAR, and his peak to around 37 WAR. DRA likes his defense better than Rfield. That rise would put him on cusp of the Hall of Miller and Eric, about even with Bobby Abreu. Does the fact that Cravath’s MLEs for the PCL and for 1909 are lower than his subsequent seasons add circumstantial evidence to the case for his being a creation of his park? I have four technical answers:
This MLE procedure is not as rigorous as the one I use for Negro Leaguers. If it were, I would use z-scores to shuttle Cravath from the minor leagues to the major leagues. Because we do not have walk totals for batters in either the PCL or the AA, we can’t figure the standard deviation of his originating league. It’s entirely possible, and I think somewhat likely, that Cravath stood out from the league more than this MLE method can show.
According to non-BBREF sources, Cravath stole quite a few more bases in the minors than in the majors. In the PCL, he averaged about thirty-three swipes per 154 games. In the AL, he averaged twelve. In the American Association he averaged twenty-six steals per 154. In the NL he averaged eleven. It is possible that Cravath possessed better speed as a younger player than his major league steals totals suggest. There may be a little less negative value in baserunning than we’re able to credit him with.
The fielding numbers you see here are estimates based on his Rfield totals at Baseball-Reference. However, Defensive Regression Analysis (DRA) sees Cravath as about average in the field, range-wise. If DRA is correct, that would add back significant value to Cravath’s career, to the tune of 40 to 50 runs, depending on how you apply it.
Furthermore, DRA and Rfield capture only his major-league record, so all but 117 games of Cravath’s defensive value are measured when he was an older player. This is potentially skewing the data a little toward negative value as well.
I think caution is warranted before writing off Cravath as a one-trick pony. But it was a good trick….
Doc's comments suggest Cravath is at least a low ballot type of consideration, one I'll be weighing heavily before submitting a final draft.
Anyway, that turned into a long post about why I'm conservative with extra credit. If you add 50% credit to Cravath's career, he's borderline (IMO), but that's a crazy amount of credit to speculate about. Careers fall apart or go into steeper decline than expected for all sorts of reasons. What you do with Cravath's park (if anything) is another question, that I will leave alone.
Except his career really isn't short. He was playing since age 22 at the highest league available to him. I give him minor league credit for 1907, 1909, 1910 and 1911. 4 extra seasons at basically his career average production. This is the most minor league credit I give anyone.
Gavy Cravath – Retrosheet now has home road splits for Gavvy for his entire career!
1908-20 G AB HR RBI AVG OBP SLG OPS
Home 616 1898 93 331 .304 .398 .545 .943
Away 610 2057 26 247 .273 .366 .419 .785
That home 943 OPS would be fourth behind Ruth, Cobb and Hornsby for the period, just ahead of Speaker and Shoeless Joe.
The 785 OPS as a road player? Tied for 23rd for the period with Sam Rice, which is still very good. Nestled in between Zach Wheat, Jack Tobin, Braggo Roth and Steve Evans.
As always the truth is somewhere in the middle, but I’d bet he’s closer to Zach Wheat than Shoeless Joe. If he’s “only” Zach Wheat his lack of bulk probably means he is not a HOM candidate for me.
Here's what I can say, extrapolating from the work I have done so far, which has been a thorough study of the 1903 and 1905 PCL seasons.
For those years, it looks like Cravath would have been around a 90-110 OPS+ hitter in the majors, which would not have been enough bat for a average fielding corner outfielder to win a starting job. Top players from the PCL were signed to the majors during those years, but Cravath, while among the 10 best hitters in the PCL, was not at the level of the players signed. His 1906 season does not look to be of that level either. His 1907 breakout season caught the attention scouts, I think, and got him to the majors. He was definitely the calibre of a starting Major-League corner outfielder that year, probably comparable to his 1908 and 1912 ML seasons. I don't know if he deserves MiL credit, though, because he got signed after one season of showing that he would be a good major-leaguer. I don't think the theory that he was overlooked/trapped in the PCL fits the history of signings and evidence of scouting at the time.
I haven't studied his AA years closely yet, but his first year in Minneapolis looks pretty comparable to his last year in the PCL, so it's a 1908/1912 type season. His 1910 and 1911 seasons are monster, blow-out-the-league performances, and probably map to his 1913-1915 peak. The raw numbers might look even better than that (can't know with more certainty until I have league-based conversion factors), but he was park-aided in Minneapolis just as he was in Philadelphia, so I see it as fair in a back-of-the-envelope calculation to assign him around 11 WAR for these two seasons. With another 3 for 1909, that gets him to 47 BWAR for his career. 3 more for 1907 would be 50. That's just getting into borderline territory for an outfielder in the 1910s. Given the magnitude of park assist involved and the amount of MiL credit needed, I see Cravath as a bit short.
If the MLEs suggest that he was a 140 OPS+ hitter for 1907/1909 and a 180-190 OPS+ hitter for 1911 and 1912, that level of performance would be a different story, but I think his being a 120 OPS+ hitter for 1907 & 1909 and a 150-160 OPS+ hitter for 1911-12 is more likely.
I'm a prime/peak/career voter, using Baseball-Reference, Baseball Gauge, Fangraphs, Kikos W-L records, and some of the players and principles from Dan Rosenheck's studies, along with factoring in clutch and post-season performances. I also lean on the MLEs that Doc has done for Negro League players, the studies Doc and Miller have done at their site, and Matthew Cornwells PARC-d at Baseball-Fever.
With ~99, 100, 101 being the fuzzy in-out line.
1. Andy Pettitte - 117
2. Lance Berkman - 114
3. Urban Shocker - 112
4. Kenny Lofton - 111
5. Jason Giambi - 110
6. Tim Hudson - 109.5
7. Tommy John - 108
8. Johan Santana - 107.5
9. Wally Schang - 107
10. Jeff Kent - 106.5
11. Jim Sundberg - 106
12. Vic Willis - 105.5
13. Joe Tinker - 105.0
14. Bert Campaneris - 104.5
15. Hurley McNair - 104
104s:
Bobby Abreu, Kevin Appier, Orel Hershiser, Thurman Munson
103s:
Babe Adams, Ron Cey, Dwight Gooden, Dale Murphy
102s:
Art Fletcher, Dolf Luque, Don Newcombe, Bobby Veach
101s:
Bobby Bonds, Lonny Frey, Brian Giles, Bob Johnson, Tommy Leach, Johnny Pesky, Vern Stephens, Roy Oswalt, Marvin Williams,
100s:
Newt Allen, Charlie Buffinton, Cesar Cedeno, Eddie Cicotte, Bus Clarkson, Willie Davis, Harry Hooper, John Olerud, Bernie Williams,
99s:
Sam Bankhead, Tommy Bond, Jack Clark, Lazaro Salazar, Sammy Sosa, David Wells,
98s:
Sal Bando, Buddy Bell, Kiki Cuyler, George Foster, Fred McGriff, Tony Perez,
97s:
Dave Bancroft, Tommy Bridges, Gavvy Cravath, Toby Harrah, Phil Rizzuto, George Scales
Newbies - Mark Buehrle 90s
The discussion of Wally Schang and Thurman Munson have caught my attention too. Here is a question I wanted to throw out while the board is very active, about Schang.
Depending on what you do with Buck Ewing who had 48 WAR, compared to Schang’s 47.9….
Wally Schang debuted in 1913 and became the all-time leader in WAR for catchers in 1927, passing Roger Bresnahan. He stayed there until 1937 when Mickey Cocharne passed him. Schang was down to 4th all-time by the start of WW2, with Dickey and Hartnett passing him also. Schang would stay locked into the 4 spot, until Yogi passed him in 1958. Johnny Bench passed him in 1975, Simmons in 1982, Fisk in 1983. He’s still 11th, among people that played half their games at catcher.
If you lower the standard from having only played 40% of your games instead of 50%, which is reasonable, Joe Mauer, Joe Torre and Buck Ewing slot in ahead of him too. Ewing played 47% of his defensive innings at catcher. Schang played 85% of his at catcher.
Everyone ahead of Schang, except Mauer, is in the Hall of Merit. King Kelly just behind him is in. Tenace, Munson, Freehan, Posada and Bresnahan are the rest of the top 20.
At third base, conveniently number 14 in total WAR is Schang’s teammate Homerun Baker. He was the 3b with the most WAR until Eddie Mathews passed him. Baker is in the HOM but the guys around him Bell, Ken Boyer, Bando, Darrel Evans and Longoria mostly aren’t.
Number 14 at 2b is Biggio, SS is Pee Wee Reese, and the guys around them are all in.
Framing Schang’s candidacy that way, my two basic questions are:
If Schang had the most WAR of any MLB catcher until 1937, is that enough evidence alone that he should be in the Hall of Merit?
If Schang was one of the top 6 catchers in WAR in the first 100 years of MLB, is that enough evidence alone that he should be in the Hall of Merit?
I think the answer is yes. But I didn’t put him on my ballot this year, which I’m happy to say is a mistake.
Please feel free to throw stones at this and tell me why this isn’t the right way to frame the discussion. Thanks.
Regarding the assertion that Schang was the best catcher before 1937 I would point out that Bill Dickey (1928-1946), Gabby Hartnett (1922-1941), Buck Ewing (1880-1897), and Mickey Cochrane (1925-1937) all had much higher WAR totals and periods of greatness.
Schang never had even one great season. He was good for a long time but never great. The Hall of Merit's objective is to recognize greatness. I really don't see Schang meeting that criteria even for a short time during his career. His WAR7 is 27.7. You have to go all the way down to Russell Martin in 27th place in JAWS before to run into another catcher as low as that. His best season was 4.4 WAR.
Gene Tenace & Thurman Munson both seem to be more qualified and then we have Joe Mauer and Buster Posey on the way who are both clearly better.
Except, we don't have all encompassing measures for catchers defense, so 27.7 may be WAY understating his true value.
I also would like to see another stab done at catcher defense from this era if and when we get more substantial data. Teams were clearly valuing defense incredibly highly at the position in deadball, and right now I don't think our data is much better than an educated guess. This may be a good opportunity for a WOWY-type of thing, as the standard sample size issues would be less substantial with backups playing so much more than nowadays.
Yes, again, except that if we had those, it wouldn't move Schang much relative to the whole catcher population, who would also commensurately gain value. Schang looks like roughly an average to slightly above average defensive catcher, so perhaps he would gain a few ordinal spots in catcher rankings if we had all-encompassing defensive measures, or perhaps stay roughly the same as really good defensive catchers leapfrog his offensive contributions. Or perhaps it would turn out Schang was better defensively than the surface stats indicate. Or perhaps worse.
I will add some more thoughts about Schang, as the discussion has been interesting to me. Ignoring the last two seasons of his career (because they aren't part of his HOM case at all), Schang averaged 366 PAs a season. This is basically a guy who played about 55% of what would have been a full-time job over the course of his career. Modern catchers play 80-90% full time. Schang was a valuable defender (roughly average at the most demanding position) and a good offensive player (120 wRC+ in his career). The question is, as kwarren is alluding to, how do we account for catchers of Schang's time? Only getting 55% of full-time play is a huge barrier to the accumulation of value. There are two basic approaches: the kwarren approach which says you'd have to be monstrously valuable within your playing time to be good enough to get in the HOM if you only have 55% playing time, because your overall value will be depressed by lack of playing time. Or the fairness-to-position approach (if no one objects to me calling it that) which says compared to the catchers of his time, Schang was among the best, so he deserves to go in. You could also take a hybrid approach to the two.
Personally, I do take a hybrid approach. If Schang had gotten 75 or 80% playing time, would his offense (or defense) have suffered from the wear and tear? Probably, IMO. Also, in general I am strict when it comes to hypotheticals. We see what Schang's value actually was.
I think there is a case to be made out there for Schang. But I also think that case would have to thread a lot of needles in order to convince me. I'm only speaking for myself, many others are already convinced. If modern catchers only got 55% playing time, I doubt we'd be stumping to elect any of them. In that scenario, one of them would still stand out as being among the iron men of catchers and among the most valuable. But now I am getting into speculation myself :)
..............
"Modern catchers play 80-90% full time."
they do?
-I see it important to the "positional fairness" argument that two NeL contemporaries of Schang -- Louis Santop and Biz Mackey -- have been elected. I'd like to see their presence more fully acknowledged in arguments re Schang's case.
-If catcher defense is significantly more valuable relative to other positions in Schang's time than WAR allows, then Schang's case in my system would be considerably strengthened.
-Schang's peak is hurt not only by lack of playing but also by lack of peak production. His peak WAR/162 games is lower than that of any of his elected major-league contemporaries and lower than any elected catcher. At 4.56 BWAR/162 games, Schang's best play was barely of "all-star" quality. The HoM electorate, in the aggregate, has generally wanted to see evidence of peak performance that rises to a higher level than Schang showed. Of course, if his defense is being underrated, then adding defensive value would bring up his peak rate also.
Considering these points, I would find analyses from Schang advocates that dig in on the topic of catcher defensive value to be the most useful. One thing that comes to my is that, given usage patterns, WOWY-style analysis of the catchers in Schang's period ought to be quite feasible, for those with the statistical power to do it, and if the game-level data via Retrosheet is sufficient to support it.
Do modern catchers play 80-90% of full time? If you consider 650 PAs as the benchmark of full time, then the ones we might ever consider for HOM do. Jason Kendall averaged 90%, Lance Parrish averaged 80% (removing cup of coffee and hanger-on seasons), Thurman Munson averaged 89%, Sundberg averaged 76% until age 36 season, Darrell Porter didn't, only about 65%, Javy Lopez 74% removing a season mostly lost to injury, Brad Ausmus 74%. It looks like it may be fairer to have said 75%, but then nobody really considers Lopez or Ausmus for HOM either.
For current guys: Yadier Molina 77%, Brian McCann 78%, Russell Martin 78% (including injury season), Salvador Perez currently 84% (still young), Joe Mauer only played 6 seasons as an honest-to-goodness catcher (do people realize it was that short??), but he played 89% full time in those seasons. Buster Posey going strong at 85% removing his injury season.
Look at HOMers and it's clearer: Gary Carter almost 90%, Fisk 77% without adjusting for strike or 2 of his 4 injury years, Bench 90%, Freehan 81%, Yogi Berra 86%, Piazza 87% removing injury year, Rodriguez at 80% without adjusting for any strikes or injuries. Campanella 77%, Bill Dickey 75% and almost from the same time frame as Schang, Ted Simmons 82% without strike adjustment, Torre 85% (through 1968 when he was primarily a catcher) removing injury season.
For guys who've gotten in HOM or got consideration, roughly 75% of full-time playing load basically looks like a lower bound. Bresnahan might the closest comp, IMO, in terms of playing load, except that teams played about 138 games a year in his years of full-time catching, compared to 154 in Schang's. Through his 20s, Bresnahan was at 82%. Unless you cherry pick away a large portion of Schang's catching career, it's hard to even get him into the 60s. Using age 25-34 (1915-1924) is the most favorable stretch I can hand to Schang and he's at 63.5% playing time load there, but it's also accounting for less than 2/3 of his career value in that stretch. FWIW, the cherry picked Bresnahan stretch was worth almost exactly 2/3 of his career value.
There's kind of a lot to unpack. Why was Schang's playing time so low? Hartnett was at 73% and Cochrane at 83%, Dickey 75%, Rick Ferrell 72% and their careers were just later than Schang. Buck Ewing was at 94% for the years he was a primary catcher (1881-1890 excluding 1887)...I just don't entirely know what to do with all this in relation to Schang. Was Schang at the exact nadir of MLB catcher usage norms? Why couldn't/didn't he play a few games at another position when he wasn't catching, as many of these other guys did? His bat was good enough. He ended up playing a lot of games at catcher over the course of his career because his career lasted at catcher longer than other guys. Was that effectively a trade of bulk in given years for longevity spread over more years? If the light yearly load helped the longevity of his career, should he get credit for low yearly usage? If not, then you are stuck with a marked lack of peak.
Biz Mackey's peak stats can't be great. He's kind of Bob Boone if Bob Boone had managed 4 more seasons at his peak offensive output.
I didn't try to compare Schang's peak to those of the NeL catchers, because the peaks of Negro-Leaguers are so hard to convert to MLEs with exactness.
Reviewing Mackey's current stats at Seamheads, it does look to me like his offensive peak would have been quite a bit higher than Boone's. Boone topped out with a four-season OPS run of 101, 104, 115, 113, before he dropped down to a sub-100 OPS+ for almost all the rest of his career.
Mackey's NeL OPS+ numbers from 1920-31 were 130, 149, 168, 183, 131, 156, 149, 126, 148, 122, 156, 154. The 120-130 seasons might have been around Major-league average, but the 140-150 seasons would have been in the 120-130 range, and the 168 & 183 peak would have been up to 140-150. That's way beyond anything that Boone did offensively. I'd say, in fact, that for the full course of Mackey's prime, Schang would probably be a pretty good offensive comp.
It seems likely, though, that Mackey's peak was considerably better than Schang's, however, because Shang's defense was around average, and his best offensive seasons and best defensive seasons never lined up--that's one reason he lacks a strong peak but has such solid career value. Boone actually has a similar "problem." His best years on offense came early in his career, when his defense was pretty ordinary. By the time he became an elite defensive catcher later in his career, his offensive output had dropped down to 70 OPS+ seasons. As a result, he was a productive major-league catcher for a long time, but never great.
Another kind of arbitrary factor that hurts Schang in my system a bit is that I measure peak rate over five consecutive seasons. Schang put up pretty strong sets of four-season runs in his career, but these two runs were divided by a very weak season, so that he doesn't have a typically five-season peak. I should look at his four-season peaks in comparison to his contemporaries. I think they will still be low, because other players are likely to have higher peaks over four years than five, but it is worth looking at.
Even smaller factor, but Schang (a) was a switch hitter, and (b) spent a decent amount of time early in his career (through 1920) at positions other than catcher, which gave his team the option of running multi-position platoons, sort of how Ben Zobrist did 90 years later.
Just to let you know, the only pertinent sentence is the last one. The Hall of Merit specifically prohibits penalizing players for alleged "character" issues outside of a first-year boycott. Since Sosa has been on the ballot for more than one year, any character issues are immaterial to his case. He should be considered for on-field contributions only.
Yeah, I caught that after I posted it but not in time to edit it.
As to Moore, I was a big proponent of Moore's when he was eligible. Moore and Taylor are not direct comps. Moore was a peak-heavy case from a defensive position while Taylor is a career-argument from a bat position. Just pointing out that Taylor is HoM company on that list. Suttles is probably a better comparison, though again, I wasn't going for a comprehensive argument.
even so....
let's dismiss his rookie cup of coffee season, his injury season, and of course his 2020 non-season.
that leaves Posey with nine seasons - and total games started of 916 at C, 199 at 1B, and 20 at DH. so average of 102 at C, 22 at 1B, 2 at DH.
his career high in games started at C is 122 (and his most complete games is 102 in that same season).
Molina - much more of a bulldog - does have 9 seasons of 120+ games started at C. but even there, he has never had 120 CG at C in a season.
If Koufax had pitched in the 90s & 00s in a hitters park like Pedro, he would have far less. Pitching in a pitchers park in the 60s means it takes a lot less pitches to get through 9 IPs.
Schang played a good number of games at catcher, technically, because very often he only played partial games. He actually only had 1 season of 120+ GC (and 1 of 119), and did play other positions quite often, in fact in 1915 and 1916 he was primarily an outfielder. In 1914 Schang played 107 games, but in nearly half of them he got only 3 PAs or fewer. In 20% of his games he got 2 PAs or fewer. In 17 of his 100 games as catcher, he played only a partial game. I don't want to do this for every season of his career, but many of his seasons look like this. It's a very unusual usage pattern, and I'm certain it must be unprecedented among HOMers. Which you can tell by the fact that no HOMer has an average PA/game close to as low as Schang.
Catchers used to bat 8th, so Schang batted 8th. And then they didn't, just like pitchers always batted 9th, until they didn't. Cocharne was probably the catcher that shattered the mold. It was that next generation of catcher after Schang anyway - Cocharne, Hartnett and Dickey.
Hartnett started 411 games batting 8th, 582 batting 7th. Dickey only hit 8th 81 times, and had 1500 starts divided equally hitting 5,6 and 7. Dickey had 50 starts hitting cleanup, and 3 hitting in the top three of the order. Cocharne though, he had 1,000 starts batting 2nd or 3rd, only 130 batting 7th or 8th.
Prior to that the tradition that catchers hit 8th was strong. Even the catchers that could hit who played before 1920 still batted 8th a lot.
Chief Meyers had a career OPS+ of 117, and he started in the 8th slot 657 games, 7th 218, 7 everywhere else.
Kling hit 8th 787 of 1124 games started. Bresnahan had 358 starts in the 8th slot but he also hit in the first third of the lineup 520 times too. Bresnahan had speed unlike most catchers. Steve O'Neil hit 8th in 1,300 of his 1,400 starts. Mike Grady who was 11th in the majors in OPS+ from 1901 to 1906 batted in the 8th spot more than any other position, and batted 6-8 almost exactly the same as he batted in one of the first 5 slots.
So we have almost no catchers from that time period, and we have no one he had so few PA/G.
If Santop and Mackey had played MLB, there is a good chance they would have also spent a lot of time at the bottom of the order too. Or they might have moved to first like Chance and Foxx plenty of others.
Schangs Splits by Batting Order slot
Per his splits on BB Ref
Split G GS PA
Batting 1st 30 23 114
Batting 2nd 19 11 63
Batting 3rd 48 30 149
Batting 4th 194 192 801
Batting 5th 161 159 675
Batting 6th 313 285 1187
Batting 7th 228 182 774
Batting 8th 728 627 2542
Batting 9th 120 0 132
Speaking of Sosa, he's short of your ballot, but a serious in the running vote getter, wanting to share the latest on players getting votes, clutch scores from Baseball-Reference. I've noted in before, but Sammy is so woefully bad that I've dropped him from my ballot. On the positive side, Tommy John has started to gain traction, he was also a significant plus in the playoffs.
Bobby Abreu 4.3
Luis Aparicio 0.6
Kevin Appier 4.6
Sal Bando -2.9
Buddy Bell -5.1
Lance Berkman 1.5
Bobby Bonds -0.8
Tommy Bridges 3.7
Mark Buehrle -1.1
Bert Campaneris 6.3
Dave Concepcion 5.8
Kiki Cuyler 4.2
Bob Elliott 1.8
Chuck Finley 3.8
Art Fletcher 3.6
Jason Giambi 3.4
Brian Giles -3.5
Dwight Gooden 3
Toby Harrah 3.4
Tommy Henrich 3.1
Orel Hershiser -3.7
Tim Hudson 0.2
Torii Hunter -9.6
Tommy John 5.3
Bob Johnson -4.6
Jeff Kent -8.9
Kenny Lofton 1.4
Mickey Lolich 0.7
Ernie Lombardi -6.6
Dolf Luque -2.7
Fred McGriff -4.6
Thurman Munson 2.7
Don Newcombe -3.7
John Olerud 0.4
Roy Oswalt 1.4
Andy Pettitte -0.5
Jorge Posada -3.6
Phil Rizzuto -3.2
Johan Santana 3.2
Wally Schang -1.4
Urban Shocker 2.6
Sammy Sosa -17.3
Vern Stephens -0.8
Darryl Strawberry -3.3
Jim Sundberg 1.2
Gene Tenace -5.8
Bobby Veach 1.5
Robin Ventura -4.2
Bucky Walters 2.3
Roy White 5.4
Bernie Williams -3
Hartnett started 411 games batting 8th, 582 batting 7th. Dickey only hit 8th 81 times, and had 1500 starts divided equally hitting 5,6 and 7. Dickey had 50 starts hitting cleanup, and 3 hitting in the top three of the order. Cocharne though, he had 1,000 starts batting 2nd or 3rd, only 130 batting 7th or 8th.
Looking at Schang's splits, he did in fact bat 8th a lot - 728 times, more than twice as often as he occupied any other lineup position (though less than half of his total games). Although, if you adjust for that in terms of plate appearances, it's worth also pointing out that hitting 8th probably inflated his OBP (he posted a .410 OBP in the 8th spot, notably higher than his OBP in any other spot in which he had 50+ games).
Also, for what it's worth, Dickey wasn't hitting 5/6/7 because that's where the catcher always hit; he was down there because Ruth/Gehrig or Dimaggio/Gehrig (or insert other combination from loaded Yankee lineups) were 3/4.
Are these scores in runs, wins, or something else?
https://www.baseball-reference.com/players/s/sosasa01-bat.shtml
In wins, under the win probability area:
Clutch Year
-0.6 1989
0.3x 1990
-0.2 1991
0.xx 1992
-2.3 1993
-1.3 1994
-1.1 1995
-0.3 1996
-1.1 1997
-1.5 1998
-1.1 1999
-0.7 2000
-1.7 2001
-3.0 2002
-1.0 2003
-0.8 2004
-0.6 2005
0.2x 2007
Sosa's context neutral BtWins are 30.5. Pretty close to his actual WAA of 28.3 (good defense and baserunning early in career largely cancel the bad in later career, so his career WAA is not far from his pure offensive WAA).
His WPA wins are 25.0. His situational wins (WPA/LI) are 41.4. Consider BtWins, WPA, and situational wins as basically three different measures of offensive WAA. The difference between WPA wins and situational wins is the clutch metric listed above by Bleed. If you want to ding Sosa for unclutch, it seems more fair to me to ding him roughly 5 wins for the difference between BtWins and WPA wins. Ignoring context, Sosa's batting events should have been worth 30 WAA, in game context they were worth only 25 wins. Alternatively, you have to acknowledge that Sosa's batting events were actually worth 41 WAA, and you can subtract 17 wins of clutch from that number, to arrive at the same value of 25 context-dependent WAA. Still, when you do this, how do you marry it back to actual WAR/WAA which includes defense and baserunning? The easiest way is to note that the difference in context/non-context but non-leveraged WAAs is 5, that is the fair value to dock from total WAA.
Note that Kent is listed as -9 wins, but following the same process as for Sosa I would say you should only dock him 4.5 wins of WAR/WAA for unclutch, not 9.
My theory is that the clutch measure systematically advantages singles hitters over power hitters because leveraged situations make more difference to the win probability of a single. When the bases are empty, there are two outs, and your team is down by two runs, a single isn't going to add much to the team's win probability at that point. But a home run will still add significantly to win probability. Conversely, when there are runners on 1st and 3rd with two out in the bottom of the eighth inning in a tie game, a single will add significantly to win probability. A three-run home run would add a little more, but at that point in the game, one run is what is going to make the difference, and the additional two runs produced will only add a fraction more. As a result, the singles hitters will add a higher percentage of their win probability in their higher-leverage situations than the power hitters, not necessarily because they have raised their game in clutch situations but because their type of hitting means that they need leverage in order for their likely actions to bring about major changes in win probability.
If this analysis is correct, I don't think it "excuses" Sosa's historically low "clutch" value in its entirety, but I do think it means that we shouldn't treat the "clutch" value as a direct measure of "clutchness." That is, I don't think that we should conclude from these numbers that Dave Concepcion was a better clutch hitter than Bobby Abreu just because Concepcion's "clutch" value is 5.8 and Abreu's is 4.3. Nor is it something to apply directly to a player's content-neutral Wins Above Replacement (kcgard's comment above addresses this point better.) Still, even for a power hitter, Sosa's "clutch" value is poor. Compared to his long-career, power-hitting peers, his WPA is very low: of the 32 players with 475 or more home runs, Sosa's WPA of 25 is second-to-last: only Beltre's 18.3 is lower. (The median is 59.3.) For those who place value on contextual performance as measured by WPA, Sosa's 25 WPA is the key value: the -17.3 "clutch" value just contributes to the narrative of how he generated that value.
Maybe everything I've written here was obvious to the folks who've already thought about WPA and Clutch, but I thought I'd lay out my thinking in case it contributes to moving the electorate's analysis forward. Feedback improving this analysis would be very welcome!
In 1923 Mackey led the ECL with a .423 average, .456 OBP, .588 slugging, 1.041 OPS and 190 OPS+.
In 1924 he led with 90 runs scored.
In 1926 he led with 102 hits, 25 doubles, and 163 total bases.
Boone never led any major league in any offensive category.
Mackey's career slash line is .332/.394/.476; Boone's is .254/.315/.346.
Mackey hit .300+ 12 years in a row, .400+ twice; Boone never reached .300 once.
Mackey posted 10 .400+ OBP seasons to Boone's zero.
Mackey posted .500+ slugging 8 times, .600+ once; Boone never once reached .450.
This isn't just some illustion of context. Mackey's OPS+ is 133, Boone's is 82.
The two aren't remotely comparable as offensive players, on peak, prime, or career.
And as Mackey, like Boone, was universally regarded as a defensive master of the highest caliber, a slash line and OPS+ like that puts him in Inner Circle discussions - a realm that Boone doesn't approach.
During Bob Boone's best 5 seasons as a hitter he had OPS+ of 115, 113, 110, 104, 101.
That "first career" is roughly a decade for Biz Mackey. We've been through it in the MMP project and he's never made a ballot. If Bob Boone had been as good as his peak (115 OPS+) for about 4-5 more seasons they would be pretty comparable.
Let me be clear: I couldn't have more respect for the work and words of Dr. Chaleeko and Chris Cobb. The following critique is addressed to methodology, not personality. I have no doubt whatsoever that MLEs work very well indeed for minor-league and Japanese pro stats. It's specifically and solely with Negro League translations that I take issue.
Roy Campanella slugged .500 in his MLB career. This was the all-time MLB record for catchers prior to Mike Piazza, and still ranks second all-time. He still holds three of the top 12 slugging seasons by a catcher. His record of 40 home runs in 1953 stood for 43 years. His 17.33 at bats per home run again set a new MLB record, and again has been surpassed since only by Piazza.
Now look at his MLEs. In 1944 Josh Gibson slugged .588, Campy slugged .582 - but MLEs credit Gibson with .556, Campy (depending on which MLEs you reference) with .411 or .419.
Did Campanella wallow along as a powerless wimp in Negro League ball, then instantly blossom into a record-breaking slugger in the Majors?
The proposition is absurd on its face.
Biz Mackey slugged .4763 in Negro League play - a shade higher than Johnny Bench's .4758 - but his MLEs credit him with .357 slugging, ranking below every dead-ball catcher in the Hall of Fame and the Hall of Merit.
Quincy Trouppe, who slugged .456, is also credited with .357 slugging - again, behind every elected dead-ball catcher.
And again, the proposition is absurd. MLEs downgrade Negro League performance as if it were low minor-league performance.
The Negro Leagues were not minor leagues. Biz Mackey slugged .476 in his career, not 357.
1944 actual slugging: Gibson .588, Campanella .582 - a 6-point gap.
1944 MLE slugging: Gibson .556, Campanella .411 - a 145-point gap.
I challege you to support this absurd conclusion.
My problem isn't with the the conversion factor being high or low. My problem is with the use of a conversion factor that clearly and consistently creates insupportable results.
YEAR_AB__PA__AB__BB_H___TB__AVG_OBP_SLG_AB/G
1922_150_590_556_34_188_292_338_377_525_3.7
1923_150_547_516_31_196_262_379_414_508_3.4
So MLE of .338/.377/.525 and .379/.414/.508. Good enough that Mackey's clearly an all-star catcher. Bubbles Hargrave hit .316/.371/.513 (127 OPS+) in 1922 and .333/.419/.521 (149 OPS+) in 1923 in the National League which was the best batting line in the white majors. Bob O'Farrell hit .324/.439/.441 (127 OPS+) in 1922 and .319/.408/.471 (131 OPS+) in 1923. The MLE is giving Mackey credit as the best hitting catcher in baseball for those two seasons.
Mackey's NGL batting lines those seasons were .364/.418/.588 and .412/.448/.573. A .573 slugging would have been top 10 in white baseball in 1923. Here's the thing though - it was only good enough for 6th among NGL players with 10+ games played. George Scales had a .747 slugging that year and only Ruth topped that in the white leagues.
So, what's your improved conversion factor?
So your suggestion is we take NGL stats at face value for comparing against other leagues across 150 years of baseball history? I don't even do that for the National League.
I compare NgL play to MLB eras other than their direct chronological parallel. When (if!!) you read my book, you'll be in a position to critique my methodology in detail - and I'll welcome the critique.
For now, consider this:
Roy Campanella slugged .486 in the Negro Leagues, .500 in the Majors. His slugging went substantially up, not down, when he made the transition from the NgL to MLB.
Yet MLEs knock 100+ points from Campanella's actual NgL slugging figures, year after year.
This is practically the dictionary definition of reductio ad absurdum.
May I add: I very much enjoy the debates here, when they're civil. There's nothing personal in any of this - at least, not for me. Accusing me of arguing from "emotion," when I've done nothing but cite numbers and you've cited none, is not conducive to reasoned, logical debate. I've never shown you or your opinions anything but respect, DL, and I think I'm entitled to a minimum of courtesy in return.
Insofar as I'm concerned, the jury's out on this debate. If I'm wrong, I'll admit it, as I've done many times on this site. If you're wrong, perhaps you'll consider doing the same.
So far I've been told that:
* I'm arguing from emotion.
* I advocate direct comparison of NgL and MLB stats from the same year.
* I favor 2-3x more NgL HoM players per capita than MLB players.
None of that is true - but that's a side issue. No one has yet to attempt to refute the core of my argument:
1944 actual slugging: Gibson .588, Campanella .582 - a 6-point gap.
1944 MLE slugging: Gibson .556, Campanella .411 - a 145-point gap.
That's absurd. Prove to me that it isn't absurd.
When we look at the all-time MLB leaders for slugging, we see Ruth at .690 and everyone else well below him. Ruth wasn't just a pioneer; he also played in an era that greatly favored his style of play. (An era which he helped to shape, of course, but that's another argument.)
Now, if someone here wants to argue that Ted Williams, in his time and era, was a greater slugger than Ruth in the context of his, and that their disparities in slugging poercentage are a illusion of context and era, that's a strong possibility, I'll listen to that argument. And of course, that argument can be extended to hundreds, probably thousands of player-vs.-player comparisons.
But if someone wants to argue that only Josh Gibson, among all NgL catchers, hit for power - well, that flies in the face of not just the facts, but of history and of common sense.
I assert that it's not possible that all NgL catchers other than Josh Gibson were deadball-type hitters - which is exactly what their MLEs claim. MLEs have Campanella at .397 or .391 slugging. Reductio ad absurdum, ladies and gentlemen, absurd on its face.
As I'm sure you realize, at least one teenager played in Major League Baseball for 115 consecutive seasons, 1871 through 1985. Joe Nuxhall made his debut at the same age as Campy.
Small sample size, sir. Over the course of his NgL career his slugging averages out to 14 points under his MLB career. A conclusion of .397 or .391 career slugging is therefore insupportable.
B) Re Mackey. Those are indeed old MLEs. We have a lot more data now, and Mackey comes out as a markedly better hitter. In terms of the methodology, yes, it’s quite simple in theory and quite complicated in practice. There’s always room for improvement. I think when we compare individual players from any era, it is VERY important that we talk about league R/G contexts and park effects. In Campy’s case, there’s a pretty wide variety of effects going on during both his pre-MLB career and his MLB career. The nice thing about MLEs is that they cut out those effects. Even if you don’t want to deal with a QoP adjustment, those other factors can lead to a very different reading of a player’s SLG.
For what? Not his career, certainly? Again, I'd appreciate if you could link to seasonal MLEs that you are referring to.
1944 was Gibson's age 32 season, Campanella's age 22 season. That is probably at least some of the difference.
If you think there is an error in the math it should be very easy to point out where that error is. Just saying "that's absurd, that's absurd, that's absurd" doesn't show the correct math. Show your work, you're the one arguing against the MLE.
I don't understand that sentence at all.
Circling back to the original discussion - Wally Schang is a 117 OPS+ career hitter. Would an updated version of Mackey's MLEs move his production from a 98 OPS+ hitter to better than Wally Schang? I haven't seen an MLE update for Mackey with the latest data.
Fair enough...
There's no reason why NgL play in 1924 should be compared directly with MLB play for the same year. NgL play might well be analogous to a different era of MLB - and will probably never be exactly the same, as the leagues played different styles of baseball. NgL play featured plenty of home runs (more by far than MLB dead-ball play) and lots of steals and sacrifices (more than most eras of MLB live-ball play). NgL play was an amalgam, a hybrid of the two. Direct comparison between identical calendar years in NgL and MLB is like comparing Rogers Hornsby to Joe Morgan and concluding that Hornsby was vastly superior because "just look at those .400 averages!". Hornsby dominated the second baseman of his time and place, Morgan dominated his. Much more to the point, Mackey dominated the catchers of his time and place, and Boone did not.
To assert that Biz Mackey could not hit for power, as his MLEs do, flies in the face of the actual facts on the ground. May I point out that OPS+ reflects a hitter's prowess in his own time and place. Mackey was a 133 OPS+ hitter within the environment of the NgL of his era; Boone was an 82 OPS+ hitter within the context of MLB of his. Clearly, and by any standard of which I am aware other than MLEs, Mackey was a greater hitter relative to the catchers of his time and place than Boone was to the catchers of his.
Therefore: Again, I think the core methodology is flawed. Evaluate Mackey's production against the catchers of the NgL of his time, and Schang's against the MLB production of his; factor in (as we do for early baseball) the much smaller league size of Mackey's world, and the inevitable tendency for great hitters to dominate a given season; and work from there. As Mackey led in multiple offensive categories and Schang in none, I think the evidence is clear that Mackey was the greater hitter. MLEs offer instead a picture of Biz Mackey removed from his own time and place and shrunk to some imaginary dimension.
I think the evidence is clear that the quality of competition in the ECL of 1923 was inferior to the AL of 1923. I bet Wally Schang would have torn up the International League of that era just like Jack Bentley did. For another example see how much Benny Kauff's slugging dropped when he returned to the National League. He had seasons in the Federal League with OPS+ of 165 and 180 followed by 138, 138 and 142 in the NL.
The reason why MLEs are calculated into year/league context is because they started with the thought exercise of wondering what a player would have batted if they had been in MLB in year X instead of AAA. The conversion factors were based on players who played at both levels in that season. There is less data for NGL and it's typically averaged over several years. There is a good argument for context neutral MLEs rather than adjusting to the offensive era of a particular league or era but I think you can get there by looking at runs created or using AVG+/OBP+/SLG+ batting lines.
I think this is necessary to do comparisons across leagues and eras. Some eras have higher levels of offense others are dominated by pitchers. Seasons have different games played, different levels of competition, even different rules. Even OPS+ is not as good as standard deviation adjusted runs created.
Having studied the evidence at inordinate length over the past year-plus, I don't think there's any year in which NgL and MLB data correspond - within the same calendar year.
How does the 1923 ECL stack up against any team from, say, the American Association? I'd assume the .500+ ECL teams would blow away any FL team. And FL stats are regarded as Major League stats. (Please let me be clear - I'm not advocating this direct comparision either.)
I say, compare each NgL player to his actual peers, compare each MLB player to his, and decide from there.
Is there anyone here who seriously advocates the position that there was only one power-hitting NgL catcher in 29 years of league play?- that in fact, every NgL catcher other than Gibson hit at dead-ball levels, most of them em]below dead-ball levels?
I think no one here would advocate that position explicitly (at least, I hope not). But MLEs advocate it implicitly.
Falsus in uno, falsus in omnibus.
I'd be pleased to link directly to this data, but it's buried amongst 1000+ references for my book and I don't have the raw time to search it out.
Doc, do you perhaps have a link...?
This isn't exactly the same, but between the end of Buck Ewing's career in 1897 and the beginning of Gabby Hartnett's career in 1922 we only have one major league catcher, so it's not exactly impossible.
I don't think it's correct, though, because prior to 1920 (or 1919, depending on how you count) there were no **true** power hitters in the Majors at any position, and even in 1921, the final year of your proposed consideration set, no one other than Ruth hit more than 24.
1922, the year of Gabby's debut, was the first year than anyone apart from the Bambino showed genuine fence-busting power - the first year, in fact, that anyone other than Ruth hit as many as 30.
By contrast, there were plenty of true power hitters in the Negro Leagues throughout the whole of its existence - making it problematic to assert without concrete evidence (as opposed to estimates) that only one catcher was ever among their numbers. Particularly in light of the fact that we have actual NgL numbers showing catchers with plenty of homers.
My impression is the Negro Leagues up until about 1930 was more of a deadball style league - maybe with a few more homers than the MLB norm from 1900 to 1920, but not way more.
Question?
Is that true or just something I half remember from an old article? Doc must have an idea if that is true or not.
Here’s something to chew on as well. I do my MLEs into the NL unless the player debuted in the AL. Very few debuted in the AL, so most everyone goes into an NL environment. In the 1920s, the NL saw a similar surge in offense than the AL did. They didn’t have Ruth and later Gehrig and Foxx in the league but Hornsby and company nonetheless made out nicely. The annual leaders in Rbat in the 1920s NL piled up 60-100 runs (except in 1926). Usually it was the Rajah. Then along came 1931. Offense immediately dropped precipitously, and the league leadership in Rbat from then until integration had a much lower ceiling, usually in the 50s or 60s. Meanwhile over in the AL, scoring in the 1930s remained at high tide with Ruth, Gehrig, Foxx, Greenberg, DiMaggo and many other elites in the league. The AL’s offensive context didn’t come down until much nearer the war. I bring this up as another example of why context matters. A player’s Rbat might shift a little bit if you move him from the NL to the AL. But his slash line will change dramatically, and in some cases make him seem a lot better hitter than the NL version—-even though they are basically the exact same player underneath. This may actually have some bearing on Schang as well because he played in a league that was a little tougher to stand out in because Ruth and Gehrig by themselves tend to distort league averages because they are so far above the mean performance.
The most recently posted updates to my MLEs are here: https://homemlb.wordpress.com/the-negro-leagues/. They do not cover the latest NLDB update.
For Mackey and other top candidates you will find season-by-season MLEs. They do not include any trad stats projections, but you should be able to quickly see that Mackey was a much better hitter than the old MLEs cited in this thread show. Rbat and OPS+ do not match up exactly, point by point/run by run, but they don’t disagree on who the good hitters were and who the lousy ones were. For players near to zero Rbat, OPS+ can swing a wee bit above or below 100, but it’s still telling the same story: he is a roughly average hitter.
Prog, I think MLEs are not so much a flawed methodology but rather may be either answering a different question or answering the same question differently. I’m not entirely sure of your method, but it sounds like you may be taking a more customized approach to each individual player. My MLEs are designed with the idea that a repeatable system that requires as little of my intervention as possible will be fairer in the long run. There are several ways in which the nature of Negro Leagues stats force me to make decisions, primarily about playing time, but on the whole I try to press the red button and see what kind of sausage comes out the other end of the auger. Or maybe I’ve misunderstood what you’ve said above. Sorry if so! Anyway, I suspect we have a central difference in our approach to the context in which a player’s performance occurred. But hey, there’s plenty of room for differing opinions and methods!
How? WAR appears to be badly broken, especially in pitchers.
How? I very heavily use the New Historical Abstract ranking system, because it remains the most robust and satisfactory of all ranking systems
The analytics community has done amazing work over the last 20+ years, I can't understand where you are coming from here.
Johan Santana is a modern/better version of Dizzy Dean, Diz you have #1, Santana is borderline at best.
I don't see this being logically consistent.
You can find Abreu's win shares at Baseball Gauge. In fact you can find Win Shares for any modern player there.
Total OFF DEF
345.5 305.7 39.7
Top 3: 31.7, 28.2, 27
Top 5: 139
Per 162: 23.1
That's got to put him close to Dave Parker on that list. Actually he looks a lot like Bobby Bonds who you have #5 on your ballot. A little lower by rate but more career WS.
There are two black contemporaries of Traynor in the HoM.
This feels like uninformed trolling. Jim McCormick pitched his entire career with a 50 foot distance to the pitcher's box, 8 balls were a walk and players got to call for a high or a low pitch. For 2350 of those innings it was required that the pitcher deliver the ball below his waist. The job of a pitcher then had vague similarities to the job now but was much more like fastpitch softball.
Ugh, no.
Couple things about this. First, Abreu led the league once in doubles, and led the majors in walks in 2006 but was traded midseason so he didn't lead either league.
Second and more notably, black/grey ink and the HOF monitor/standards numbers are NOT measures of value; they were intended as predictions of who WOULD make the Hall, not necessarily assessments of who SHOULD. Which is why players are awarded four points of grey ink for a top-10 finish in RBI (which Abreu did only twice) or batting average (once), but no points for a top-10 finish in OBP (which Abreu did EIGHT times, plus the aforementioned '06 in which he didn't qualify in either league but would have been top-10 in both). Grey ink gives two points for a top-10 finish in walks, which benefits Abreu - but it also gives credit for high finishes in at bats (which Abreu never had because he walked so much) instead of plate appearances (top 10 six times). Basically, black and grey ink, along with the HOF Standards and Monitor, have the biases of the Hall of Fame voters baked into them, and those biases are largely the reason the HoM was created in the first place.
The core value of your and Chris Cobb’s systems is that they’re empirical, and therefore liable to adjustment and improvement. My system is certainly less empirical in a mathematical sense (as I’ve said here many times, there’s no one uber-stat or uber-system I trust as a core for thorough evaluation), and some here may consequently consider my method less valid than yours. Given the current state of MLEs as applied to NgL play and the results they create (e.g., one power-hitting catcher in 29 years of league play), I trust my system (which I’ll outline roughly below) more than I trust MLEs, by a lot. This may change as your empirically-based systems evolve.
What bothers me is not the conclusion that (I’m paraphrasing here) “Biz Mackey would have been a Bob Boone-class hitter in Boone’s offensive context,” a conclusion with which I disagree for a variety of reasons but which is at least plausibly supportable; what bothers me is when that statement is extrapolated to claim, “Biz Mackey was a Bob-Boone-class hitter” – a conclusion defied by 100% of contemporary evidence.
Such extrapolations, drawn by others with all the good will in the world from your perfectly valid work, are in my view not just unwarranted, but serve to distort the true history of baseball. The sun is round. An orange is round. That doesn’t mean that the sun is an orange.
Biz Mackey dominated the NgL catchers of his era. Bob Boone did not dominate the MLB catchers of his. That, to me, is the core fact. Claiming that Biz Mackey’s **real** OPS+ was 98 – because that’s the conclusion begged by MLE extrapolation – biases the history of baseball in favor of MLB at the expense of the NgL. And I think we’ve had enough of that already. That’s the core of my objection to application of MLEs to NgL play – at least as presently employed.
Please understand that what follows is my opinion, and isn’t meant as some sort of Holy Truth Pontification. With that disclaimer duly noted:
An NgL 190 OPS+ and an AL 190 OPS+ can’t be compared directly for at least four reasons: league size disparity, league quality disparity, sample size disparity, and offensive context disparity.
What I do is look at the top 10 in (name your statistical category) in each league, and depending on whether the decision is blindingly obvious (Babe Ruth vs. anyone, 1920) or more subtle (Rogers Hornsby vs, Oscar Charleston, 1924), I consider multiple relationships: the distance (in percentage terms) between #1 and #10, for example, or the spread between #1 and league average. I do this in a range of categories, and determine a rough ‘winner’.
But, as Bill James observes, “Good sabermetrics respects the validity of all types of evidence, including that which is beyond the scope of statistical validation.” For early MLB and NgL play, with their yawning statistical gaps, this is an especially important qualifier.
Consider Hornsby vs. Charleston in 1924: Hornsby was an average glove, Charleston was one of the greatest ever in center. Can I prove this comparison statistically? No I can’t; the data points aren’t there – except that they are, in contemporary reportage especially, and here I can claim the personal advantage of having read many hundreds of contemporary NgL press accounts.
Raw statistical comparison has numerous disadvantages. Charleston has a raw OPS advantage, 1.252 to 1.203, but that’s more than accounted for by the enormous StDev of 234 which applies to 205 at bats versus 536. A StDev of 234 from a mean of 370.5 isn’t a stable basis for any form of comparison. So I turn the exercise instead into series of bullet-pointed comparisons of each player’s excellence within their own league contexts – and in the case of Hornsby vs. Charleston, it’s not necessary to take a deep-dive to see a clear winner:
• Hitting for average: Hornsby.
• Hitting for power: Charleston by a narrow margin.
• Plate discipline: Charleston.
• Baserunning: Charleston by a huge margin.
• Defensive range: Charleston.
• Defensive sure-handedness: Charleston by a huge margin.
• Arm: Charleston by a huge margin.
I’d take 1924 Charleston over 1924 Hornsby in a heartbeat. But run the same comparisons viz. Charleston vs. Ruth, and the results change dramatically:
• Hitting for average:A dead draw. Both players won their respective league batting titles, neither by a decisive margin. (This is an example of how statistical categories can be directly compared; it’s not remotely as “scientific” as MLEs, but neither is it subject to the potential for MLE-imparted illusions.)
• Hitting for power: Ruth by a narrow margin, for 1924 at least. Charleston did lead the ECL in doubles and home runs – but by nothing approaching the gap between Ruth’s home run total (46) and his nearest AL rival (Joe Hauser with 27, a ridiculous 59% gap).
• Plate discipline: Ruth by a huge margin.
• Baserunning: Charleston by a huge margin.
• Defensive range: Charleston - but 1924 Ruth was no wimp in the field.
• Defensive sure-handedness: This may well be a draw.
• Arm: Charleston, but by a narrow margin; Ruth could bring it from right.
The big margins here are plate discipline (Ruth) vs. basepath speed (Charleston), and as all research points to the ability to get on base as the single most important offensive determinant, I’d select Ruth as the 1924 MMOP (Most Meritorious Offensive Player – I don’t have the time to run pitchers - sorry Walter) for all of professional baseball.
This is just a quick ‘n’ dirty comparison, because it’s what I can do without plunging headfirst into serious research for which I sincerely don’t have the time – my apologies for that – but it should at least give an idea of how I make my choices.
I love and respect you all. These are important discussions, and this is the only place on the Web where they take place in real depth. Let the debate continue.
I will co-sign this to a certain extent, but I don't think the NBJHBA system is necessarily better. Looking at Dean in particular, James's system couldn't favor Dean more if he had designed it to do so. The system is career win shares (which Dean won't have a ton of), top three seasons (great for Dean, who was terrific from '34 through '36), and top five consecutive seasons (AMAZING for Dean, who only had five full seasons in his career, but they were all very good and all consecutive.) Dean does well in James's system partly because he was an excellent pitcher, but also partly because James made a couple of arbitrary choices in building his system that happen to work out in Dean's favor.
What I do when using WAR to look at a player's WAR record is try to circumvent the issue that comes up in James's system, and JAWS, and any other system that relies on an arbitrarily-chosen specific number of seasons as a huge portion of its valuation: I take all of the numbers (or at least, all of them from 1-19) and average them in with career value. How this works out is basically peak-weighted WAR: 1*(best season) + .95*(second-best season) + .9*(third-best season) + ... +.05*(seasons 20 and lower). I've been pretty happy with the results I've gotten from this, with the usual caveats about WAR itself.
My CPASR (career pennants added using a sliding replacement level; soon to be renamed CASPER) methodology does not really do it like that but I imagine the results are similar.
FWIW, here is what I like about MLEs: I know exactly what they are attempting to do and exactly (at least in Dr. Chaleeko's case) how they are attempting to do that. Exactly what components do they consider. Exactly how are they handled. Exactly what assumptions are being made, and exactly how that impacts the result that is being generated. If I think the QOP adjustment is too strong, I can take the MLE and basically plug in my own preferred QOP adjustment instead to get an answer I like better. Same for the baserunning assumptions, or any other component that goes into the MLE. And the output allows direct comparison against players from completely separate leagues, because the ultimate output is runs and/or wins. As DL is saying, if something is wrong with the math, find where that is and plug in something better. Another positive trait of Chaleeko's MLEs is that he verifies that the ultimate outputs make logical sense in the aggregate. Do the MLEs suggest a reasonable number of HOM candidates from NgL or do they suggest a number that's 30% too high, or 20% too low? All of that functions as guardrails for the choices and assumptions that go into each component of the MLEs. Personally, I think this is the most critical aspect of the MLEs or any system that is trying to handle non-MLB players: whatever system you use, does it generate a reasonable number of HOM careers based on the pool of players being considered. Any system that doesn't have a check on that outcome is the only system I would dismiss out of hand.
Kaat
Walters
Koosman
Tanana
Harder
Newsom
Dennis Martinez
Leonard
Derringer
Hough
Pettitte
+53 more guys
Dean
Making Dean the 68th best available pitcher by Win Shares. But then, most people have come to the conclusion that Win Shares aren't really very good, and are downright terrible for pitchers. I don't say this as critique of putting Dean on a ballot (I think that is certainly defensible), I say this as a critique of the following statement:
It appears you don't really use Win Shares at all, actually. You use Bill James' player rankings from NBJHA, with some personal fiddling about ballot placement. Win Shares has two live ball position available players rated higher than Brock, neither of which make your ballot. Brock is #3, Tony Perez #4, next:
Abreu
Kent
Parker
Sosa
Pinson
Giambi
Cash
Luis Gonzalez
Baines
Berkman
Cruz
Cepeda
(none of these make your ballot, and look at this list and it's really kind of crazy in my opinion that someone would actively choose Win Shares now that WAR is available, but opinions differ.)
+7 more guys
Bonds
+7 more guys
Aparicio
+ 3 more guys
Murphy
+ 11 more guys
Bando
Rosen and Traynor don't appear to even be in the top 200 available players of the live ball era by Win Shares. I don't suppose there's any way I could convince you to come over to the WAR side of the fence? The downside is it doesn't like Lou Brock close to as much as WS does.
OK, ranting about Win Shares over. Sorry for this sidebar.
It was Win Shares - just not raw career Win Shares with no adjustments. Digging out my copy of the NBJHBA, here were the criteria:
1. Career Win Shares
2. Win Shares in three best seasons
3. Win Shares in five best consecutive seasons
4. Win Shares per season (which is to say, per 162 games for hitters or per 43 starts for pitchers). This is another metric that would favor Dean, because he was either excellent or injured; this doesn't penalize him for missing time.
5. A moderate timeline adjustment
6. Subjective adjustments
This method is probably better than just using raw career Win Shares, because as noted above, the baseline for Win Shares is ridiculously low; the peak and rate measures compensate for that issue to a certain extent. (Not enough that I'd use it, but still.)
Not saying Win Shares are the end-all, be-all; I tend to use WAR instead, due to many of the issues you point out (and also because WAR is on B-R, but B-R selected WAR over Win Shares for many of those same reasons). But it is possible to evaluate players using Win Shares by doing things other than just looking at unadjusted career total. (This is also true of WAR, of course.)
Especially at the top of the lineup, like Lou Brock.
Win Shares was largely panned on delivery. I and several other people wrote articles critiquing it and the rationale underlying the rankings in the NBJHBA.
The various preliminary versions of WAR were developed shortly thereafter and, to my mind, completely obviate Win Shares. I imagine that a case can be made that the developer(s) of WAR learned from the Win Shares "mistakes", but I don't want to speak for Sean or anybody else doing the hard work back then.
I’m a small-Hall guy. I agree completely with your point: That will never become an issue with my approach.
A player who has never appeared on leaderboards is highly unlikely to grab my attention for a ballot unless they have other qualifying points in their favor – a useful initial sieve. Similarly, multiple leaderboard appearances doesn’t guarantee a ballot place. Ozzie Smith has zero offensive leaderboard appearances (well, okay, once in games and at-bats in a strike-shortened year, but that’s not the kind of thing I mean), but he’s in my relatively small PHoM, because he has strong ancillary credits (I trust I don’t have to describe them to a HoM voter). Jim Rice has some significant leaderboard entries, and good defense too, but had lousy basepath speed, lousy plate selectivity, and hit into way too many double plays; plus, half the time he was Mr. Wonderful and the other half Mr. Ordinary. Those are basic examples of the core concept: you have to look at the whole package.
It’s likely that neither Lazaro Salazar nor Carlos Moran will ever sniff my ballot, because I don’t see the point; Salazar’s best stuff came in Mexican League seasons that can’t be counted as top-level NgL play, and Moran’s career was long but undistinguished. They’re interesting in a holistic sense, as part of the history of the NgL, but of no interest to me insofar as this project is concerned. Now multiply those conclusions by 100s of similar players.
I think I know NgL play well enough to be able to identify the outlier players, and I’m pretty sure I read every past ballot thread at the HoM prior to submitting my first. My ballot consists of everyone I think has a realistic place on a ballot – but let me be clear: On this year’s ballot submission there are only two (2) players, Mr. Easter and Mr. Taylor, who are in my PHoM, and in Mr. Easter’s case that’s only because I conducted extensive additional research to establish the legitimacy of his candidacy in my mind. Mr. Taylor misses an elect-me slot because catchers are under-represented, which vaults Schang over him; even though NgL first basemen are under-represented in my view, catchers are more so, and representative proportionality was raised (legitimately in my opinion) as an issue with my preliminary ballot. Apart from new entrant Abreu, who I think clearly meets HoM standards, the others are players who voters here have singled out, now or in the past, as names of interest, and when I looked into them I agreed with their cases sufficiently to place them on my ballot. My top 8 I consider electable by the standards of the HoM, which are much broader than mine; everyone from the 9th spot on down (apart from Van Haltren, who’s just an old favorite and also reasonably meritorious in his own right) is hitching a ride because I’m required to nominate 15 names, and they did show significant merit in their careers.
My approach takes a big initial investment of time and reading, and if a player grabs my interest there’s additional time spent in researching their case. For me, on this ballot, that’s Luke Easter and Hugh Duffy; I put probably 20-25 hours into Duffy, and more than 100 into Easter. For the rest, I look at the historical consensus here and agree or disagree.
I hope that answers your questions…?
Eric, thanks for outlining how James uses Win Shares in his rankings. You touched on it even in a previous comment about how James' rankings were tailor made to suit Dean, but I couldn't piece together just from that how the rankings could give certain results. Also, Brock, I didn't intend to sound combative about your use of Win Shares or Bill James' rankings, but yeah, admittedly I don't much love Win Shares. However, no one can deny that Win Shares represent a logically consistent way to compare players, and so they're a perfectly legitimate way to do so for HOM.
Rob Wood, I have much the same impression you do. The WARs that came shortly after Win Shares probably owe a great deal of their existence and ideological direction to Win Shares, what it did well and what (subjectively) could/should have been done better. I also agree they obviate Win Shares, but that is a question of ideology, largely. Bill James often had brilliant insights, which he got 80% of the way to the final working solution, and which then got adopted and "perfected" by people who understood the math but would never have had the concept itself. As a mathematician, I've often remarked that when it comes to successful projects, working out the math is easy, the hard part is formulating the idea correctly. At which James really was brilliant often times.
Koufax pitched 2324 innings, with an ERA+ 131. Was he also "borderline at best". I really don't get how these two pitchers are viewed so differently by so many voters.
In HoM year 1972, when Koufax was elected on the first ballot, Al Peterson ranked Koufax 9th on his ballot, which he submitted on March 15, 2006, noting that "As a voter who tries to balance peak and career he places here with the excellent peak making up for the career shortcomings."
Here's the players on that ballot:
1. Robin Roberts
2. Dick Redding
3. Bobby Doerr
4. Edd Roush
5. Bob Johnson
6. Jimmy Ryan
7. Rube Waddell
8. Tommy Leach
9. Koufax
10. Biz Mackey
11. Hugh Duffy
12. Cool Papa Bell
13. Frank Chance
14. Billy Pierce
15. Dobie Moore
16-20: Oms, Minoso, Mullane, Sewell, Mendez
21-25: Childs, F Jones, Van Haltren, Browing, Berger
26-30: Poles, Bridges, Sisler, Easter, Kiner
31-35: Byrd, Gordon, Willard Brown, Keller, Lundy
Quite a different set of players from 2021, and a different evaluative context. A lot of these players have been elected, of course. Some who haven't been have disappeared from Mr. Peterson's ballot & near ballot -- Jimmy Ryan, Hugh Duffy, Frank Chance, George Van Haltren, Fielder Jones, Wally Berger, Spotswood Poles, Tommy Bridges, Luke Easter, Bill Byrd. The early ML hitters and the centerfielders all looked pretty good in Win Shares, which was the most influential comprehensive metric at the time (Koufax's peak looks very strong in WS, btw, in part because the low replacement level rewards lots of IP). The NeL players are generally ones with strong reputations that didn't turn out to be backed up by the rigorously reconstructed historical record that eventually became available to the electorate -- along with the much improved versions of WAR that reduced the cachet of some players who had looked stronger in Win Shares. (The WS bias for centerfielders and innings eaters is what got Edd Roush and Eppa Rixey elected, btw: if BWAR had existed and been the influence then that it is now, neither of them would have made the cut. The knowledge base of the electorate was very different in 2006!)
A few players, however, are still around, neither elected nor buried deep in the backlog.
Tommy Leach was 8th on Mr. Peterson's 1972 ballot, and he is 4th on his 2021 ballot (that would suggest that 5th place would be the absolute upper bound for Santana). Bob Johnson was 5th, ahead of Koufax; now he is 9th. Tony Mullane was 18th, the second pitcher down from Koufax; now he is 6th. A lot has changed between 1972/2006 and 2021, but it doesn't look as if Mr. Peterson's voting has changed capriciously or that his assessment of Santana is radically inconsistent with his assessment of Koufax.
For a voter who values IP in peak and career, Koufax is marginally ahead of Santana, and a marginal difference in the current context can easily result in a different of 30 places in the rankings. I'd be inclined to suggest that Mr. Peterson's ranking is not giving post-1990 pitchers enough credit, with none making the ballot. Pettitte and Hudson, both with substantially more career weight than Santana, are very close to the ballot, and their ranking ahead of Santana seems consistent with the premises of his system. In 1972, based on the evidence available, he viewed Koufax as a bit above the borderline, but not a top-of-the ballot player. I'd never accuse anyone who ranked Tommy Leach ahead of Sandy Koufax as treating Koufax as an all-time great! If Sandy Koufax ranks there, viewing Santana as "borderline at best" seems pretty consistent to me.
For what it is worth, on my two key stats Koufax has .78 CASPER, 31 WVA, and 49 WVR. Santana has .81 CASPER, 28 WVA, and 44 WVR. So the two do look similar when using my regular season "value" stats. Of course, Koufax also had a stellar post-season record which is not reflected in any of those stats.
The importance of park effects to Sandy's otherworldly 1962-66 run would knock him well down from a #1 placement for me, except perhaps on an exceptionally weak ballot - but his post-season dominance, at home and on the road, would probably give him an elect-me slot. As both pitchers had substantial post-season innings and can therefore be compared head-to-head, post-season performance is for me the biggest difference between Sandy and Johan, and why, on this year's relatively weak ballot, Sandy would rank, say, #3, while Johan ranks #5.
Onto Santana - am I being a bit harsh on Santana in 2021...maybe. From the early days of HOM balloting I tried to avoid the "shiny new toy" vote for new eligibles, just being cautious. Two reasons for taking it slow on the new borderline candidates - the unending eligibility & the contemporary argument. If Santana didn't make it in the first couple years that's fine. If Santana retired early before his pitching cohort was done that's fine. We had time, each "year" election was 3 weeks in real life. We'd have contemporary players soon joining, things work out.
Now that the HOM is current in history is there an urgency I should be following with each new year? For slam dunks I'd say sure but Johan is not that. I've got my pet project players...Tommy Leach & Tony Mullane are still here clogging my ballot after they were vetted 15-20 years ago. I think they're worthy but each time thru the yearly exercise I ask myself: why throw away ballot slots for these no chancers? It's what I do...more than likely I felt the player on the ballot is a better representative for the era/position than someone else we elected prior.
Now to revisit my Fielder Jones interest...that's a name I haven't uttered in a very long time and it brings a little smile to my face...
Chris this is a point I should consider - that I'm going to give the most modern pitchers a tough mountain to climb. OTOH I think that individual pitcher usage decisions made by organizations today are part of that mountain.
He does use Win Shares - I have chunks of that book essentially memorized - but with a massive "subjective component," which I think accounts for the discrepancies you rightly indentify between Win Shares and rankings. I can't knock that too much, as a) I'm massively influenced by James and b) there's a strong subjective element in my own system - stronger than James's I'm sure.
Definitely not true. I'm not defending the system, just pointing out that James uses a lot of non-trad wriggling to come up with his ratings.
And I agree with you, by the way: Hanging around forever, even at a severely compromised level of competence, does have too much of a Win Shares effect.
Brock is an interesting case for me. He's not in my PHoM, I recognize the significant deficiencies in his candidacy, and I know there's a surfeit of outfielders already in - but I do think he meets HoM standards, and comfortably too.
For example, he's the only holder of any significant AL or NL offensive record who's not recognized by the HoM (I think I've pointed that out elsewhere) - and without checking, I'll bet he's also the only 8x leader in a major offensive category (at least, post-19th century) who's not enshrined.
But I acknewledge the overwhelming HoM consensus against him - that's one of the benefits, I suppose, of re-reading old ballot threads; and given the quality of the baseball brains gathered here, I've elected to defer to the consensus.
Found a new winner:
9x:
SB: Luis Aparicio
6x:
BB: Roy Thomas, Eddie Yost
SB: Bert Campaneris, Vince Coleman, Maury Wills
And RIP / almost 1 year anniversary for drumming legend Neil Peart.
One particular Peart lyric was very important to me when I was growing up, and still resonates today:
You can choose a ready guide
in some celestial voice;
if you choose not to decide,
you still have made a choice.
You can choose from phantom fears
and kindness that can kill;
I will choose the path that's clear -
I will choose free will.
Great writer too.
Bleed or Freak is acceptable, alludes to a track from Alice in Chains Facelift.
I'm a HUGE music fan, heavy in nature, metal, hard rock, alternative, blues, jazz, progressive, etc.
Good and bad music in all genres, just not a fan of basic/generic music, give me something with weight and gravity please : )
You must be Registered and Logged In to post comments.
<< Back to main