I completely disagree. Jeter's peers (ARod, Nomah, Tejada, Larkin) are simply better than Sewell's. Jeter was 48% better than much better players than Sewell was 50% better than. A star glut (or drought) should have absolutely no impact on a player's ranking. Zero. Stars aren't evenly distributed by position in the short term - the sample size isn't big enough.

It's like penalizing a player for having bad teammates - it's something the player cannot control.

It's like penalizing a player for having bad teammates - it's something the player cannot control.

I'd argue that's totally wrong. Teammates play under the same conditions in the same era.

Sewell was able to play SS in an era where others with similar batting skills could not play SS but had to play 1B or RF. That made him more valuable IN HIS ERA.

What your "method" would do is penalize SS's in the 60's or Catchers in the 1900's by comparing them to eras where a Mike Piazza could be a catcher or an ARod could be a SS. That's penalizing a player for something he has no control over.....

Stars aren't evenly distributed by position in the short term - the sample size isn't big enough.

In the interest of full disclose, I have to admit I was taking my argument a little bit to the extreme to make a point, but in reality I use a "20 year" floating window for my comparisons to mitigate the sample size issue as Joe is correct in that talent will not be PERFECTLY distributed across positions in the short term.

Even using 20 years, Sewell compares as one of the best SS over a TWENTY year period, and that's nothing to sneez at...

I support Joe's case here.

While I would agree that Mike Piazza might not have made it as a catcher in the 1960s or 1970s as he did in the 1990s, that's because I can readily identify ways in which the defensive demands of the position changed between eras. I don't see any evidence here that the defensive demands on shortstops rose substantially around 1920. Is there any evidence aside the decline in shorstop hitting?

If the only reason that "Sewell was able to play SS in an era where others with similar batting skills could not play SS but had to play 1B or RF," then it seems that Sewell's most valuable skill was arranging to be born in 1898 rather than in 1885 or 1908. I agree that his value was increased because of the weakness of the shortstop pool, but I have to regard that value as purely circumstantial and therefore not contributing to the merit of the player.

If there was a reason having to do with the requirements of the position that screened out heavier hitters (as, for instance, playing on Astroturf surfaces that call for tremendous range, quickness, and arm for shortstops), that's another matter entirely, as the players who make it at shortstop receive credit for meeting higher demands. But there's nothing about the 1920s game that placed such demands on shortstops, as far as I know.

It doesn't make it hard to compare players across eras. If Joe Sewell was 50% better than his peers, and Derek Jeter(?) was 48% better, then Sewell was more valuable, and he should rank higher.

I don't necessarily disagree with this, Kevin, because there is a hidden value that superiority at a position creates for a team, but Sewell's dominance would have been impacted if Dobie Moore, Dick Lundy and John Beckwith had been allowed to play in the majors.

WS has some things built in to give more fielding WS in eras with low strikeout totals. This factor isn't as strong as it should be, but it is there.

The range of this factor (fielding/pitching split) is as follows (for an average team):

.317 (Min, 1997+; 6.6 K/9)
.324 (1884, lowest seen to date in 1940)
.325 (Average split as claimed by James; 4.5 K/9)
.327 (1911-14, lowest post 1886)
.333 (Mid 1920's)
.335 (1893-4, highest post 1877; 2.2 K/9)
.342 (1872, highest ever, 0.6 K/9)

Average full-time SS in 2001: 7.2 Fielding Win Shares
Average full-time SS in 1907: 7.5 Fielding Win Shares (adjusted to 162 games)
Average full-time SS in 1872: 7.8 Fielding Win Shares (adjusted to 162 games)

It doesn't make a big difference.

<i>but Sewell's dominance would have been impacted if Dobie Moore, Dick Lundy and John Beckwith had been allowed to play in the majors.

John,

I just don't understand this. Almost every players dominance would be affected if certain banned players were allowed to play. Dazzy Vance's dominance would have been impacted if Smokey Joe Williams, Dick Redding and Joe Rogan had been allowed to play. Roger Bresnahan's dominance would have been impacted if Louis Santop had been allowed to play. Max Carey's and Edd Roush's value would have been impacted if Cristobal Torriente had been allowed to play.

In short, we're not supposed to vote on what could have happened, but on what did. Joe Sewell was the best shortstop in the ML during his playing career.

Joe Sewell was the best shortstop in the ML during his playing career.

I totally agree, but he only dominates by default. We just can't pretend that the Negro Leaguers played in the Phantom Zone. They played at the same time on this planet as Sewell and have the stats to back them up.

We just can't pretend that the Negro Leaguers played in the Phantom Zone. They played at the same time on this planet as Sewell and have the stats to back them up.

John,

I don't pretend that they played in the Phantom Zone. I realize that they played at the same time as other (white) players. However, slotting in the Negro Leaguers isn't an exact science. What's wrong with being the best in ML baseball at your position? Should Sewell (or any other player) be downgraded because there were several other candidates who weren't allowed to play(at preliminary glance they don't seem as good as Sewell to me, except for Lundy, who I haven't looked at yet.) Don't forget that Sewell played against tougher competition than Moore, Lundy or Beckwith. He also played in the harder league in ML baseball. His competition wasn't too bad either. Bancroft and Jackson were both pretty good.

I have Sewell higher than Moore, Lundy and Beckwith and intend to keep him that way. Not all Negro League players were the black Honus Wagner -- Lloyd, who may have been, was a distinctly earlier generation. Sewell, like Beckley and Schang, were the best at their positions for an extended period and deserve a moderate extra credit for this (well OK, Santop may have been better than Schang, but I don't believe Petway was.)

We are in severe danger of resolving too many of the huge uncertainties about Negro League players in their favor, and electing far more of them than they rationally deserve (no more than 12-15, demographically.)

We just can't pretend that the Negro Leaguers played in the Phantom Zone. They played at the same time on this planet as Sewell and have the stats to back them up.

John,

Doesn't this imply that we should discount all Major League pre-1947 players, since their impact would have been affected by Negro League players?

Yes, EVERY player we're considering or have elected is impacted by not competing with the Negro League players, but we're "compensating" for that by using some of the HOM slots for Negro League players, so I don't see how it's really more relevant for Sewell or SS than for any other players or positions...

I find it hard to believe how much people take into account direct contemporaries. If Corey Koskie (to pick a random slightly above average player) played in a decade where all the other 3b or AL 3b were terrible, that doesn't make him better than he is. His RCAP might be the best of all-time, or close to it, but if his era adjusted stats left him as the 70th best 3b of all-time (to again pick a number) he should not be in the HoM. KJOK, am I right to think you disagree with this?

I don't know about big discounts, but I thought what was happening was that Sewell was getting a "special bonus" for playing in a weak-SS era. This is what RCAP is doing. Add in the fact that I think RCAP is only looking at the AL SS's and not the NL ones so next-tier guys like Bancroft, Jackson and even Maranville are not even included in Sewell's RCAP numbers. Basically its the "Ruth Effect" in reverse.

The sanity check to the NeL was just to show that good shortstops were out there and it is perhaps just a coincidence that Sewell was the only good white one of the 1920s. So, no "discount" is being called for. I'm just wary of "special bonuses" for positional shortages.

I would use RCAA when comparing Sewell to other SS's. He's a solid candidate and I wouldn't want to underrate him.

Overall US population expectation is that MLB should be 10% black during this time. However, given the demographic bias in favor of southern white players, and the geographic location of the black population in the south at this time, it could be expected that MLB would be 15% black during the 20's/30's. It also was pointed out before that the composition of HOFers during the 60's was 30% black, so there is "precedent" for the all-star population to be higher than 15% black.

This means that the following expectation is not unreasonable during the 20's/30's: that the bottom 1-5 MLB everyday starting players at each fielding position would be replaced by the top 1-5 Negro League players (possibly more and possibly none depending on the relative depth at each position). Pitchers are harder to figure, but numbers like 5-15 cracking the top-4 on the 16 MLB teams seem reasonable.

If Corey Koskie (to pick a random slightly above average player) played in a decade where all the other 3b or AL 3b were terrible, that doesn't make him better than he is. His RCAP might be the best of all-time, or close to it, but if his era adjusted stats left him as the 70th best 3b of all-time (to again pick a number) he should not be in the HoM. KJOK, am I right to think you disagree with this?

Yes, I disagree. it may not make him 'better than he is' but it makes him MORE VALUABLE TO HIS TEAM TO ENABLE HIS TEAM TO WIN GAMES AGAINST THE TEAMS THAT HIS TEAM IS PLAYING.

I found a couple of quick real life Corey Koskie examples -- Walker Cooper and Sherm Lollar, 8-time and 7-time all stars respectively, mostly due to weak depth at catcher in their playing time period in their league. Unless I'm mistaken, these guys will get no more support from our electorate than Ray Schalk, and nor should they -- even though their performance above their contemporaries was quite strong.

Let's put it another way. To NOT consider era, where the HOM is headed is that we'll have NO 1st basemen from the '00 or '10 time period, but 6 or 8 from the 1990's.

How can you be more "MERITORIOUS" than to be the best at your position during the time you played?

And what's so "MERITORIOUS" about being the 8th best player at your positon?

Walker Cooper and Sherm Lollar, 8-time and 7-time all stars respectively

Not sure they're good examples. Cooper played thru WWII, where there's NO DEBATE that the competition was less than the surrounding years.

And Sherm Lollar gets compared to Yogi Berra, and I don't think he's going to stand out all that much there...

However, slotting in the Negro Leaguers isn't an exact science.

Absolutely correct.

What's wrong with being the best in ML baseball at your position?

Nothing, but we know there is a difference between being the best NLer or ALer than being the best major leaguer. Ossee Schreckengost was the best catcher of the AL during the early 1900s, but only because the Duke of Tralee and Noisy weren't in that league. The same goes for Sewell in the majors during the twenties, though not anywhere near to the same degree, of course.

Don't forget that Sewell played against tougher competition than Moore, Lundy or Beckwith. He also played in the harder league in ML baseball. His competition wasn't too bad either. Bancroft and Jackson were both pretty good.

I agree that Sewell played against better competition than the Negro Leaguers, but it could have been better. If the best African-Americans had played during the twenties or thirties, either Sewell's stats would have been affected by the better pitching or he would have looked smaller at short and third with Moore, Lloyd and Beckwith. I don't think this can be disputed.

BTW, I do find it amusing that I have Sewell placed much higher than you, Rick. :-D

Doesn't this imply that we should discount all Major League pre-1947 players, since their impact would have been affected by Negro League players?

Well, we do that anyway when we take into account competition (like the Davenport translations). As for dominance, I've been looking at that since we were able to view the Negro League stats.

We are in severe danger of resolving too many of the huge uncertainties about Negro League players in their favor, and electing far more of them than they rationally deserve (no more than 12-15, demographically.)

What we're really in danger of is maybe electing a very good, good, good, good, good player instead of a great player. When we start electing Highpockets Kelly or Jesse Haines, then I'll start really worrying. :-)

Yes, I disagree. it may not make him 'better than he is' but it makes him MORE VALUABLE TO HIS TEAM TO ENABLE HIS TEAM TO WIN GAMES AGAINST THE TEAMS THAT HIS TEAM IS PLAYING.

It's hard to argue with this. All things equal, if Koskie was playing teams that had John Murphy talent manning the hot corner, Koskie's teams would gain a huge advantage by this. Of course, why those other teams would want to use sore armed 39 year old players that can't hit a curve ball needs to be questioned. :-D

This strikes me as the old debate about *value* versus *skills.*

Typically the value position has been about respect for all eras--that players in the 1870s or '80s had value and should be recognized for that, regardless of whether they could step off a time machine and play in the MLs 100 years later.

Meanwhile, timelining represents the skill position. Players today are better in absolute terms and so we have to discount their predecessors. But the same question has been applied to, say, Harry Stovey and Sherry Magee. Were they playing against the best competition or not? Do the deserve to be credited with 100 percent of their apparent value, or not? Were they really as skilled as their apparent value would suggest, or not?

This is all complicated a bunch when you add in peak vs. career, but frankly value vs. skill (in the form of timelining) is probably even a greater determinant of our voting than peak vs. career. Or maybe not, I guess if that were true then Dobie Moore would be ahead of Hughie Jennings.

But rating Joe Sewell is not unlike rating and ranking Stovey or Magee, or Moore or Beckwith. You ask yourself how good his competition was, how his value is best understood. In other words, his value tempered by a sense of what his skills were, not so much from a timelining perspective but compared to all of the skills that were out there, whether directly in his competitive universe or not.

The difference between Sewell's WS and, say, Dobie Moore's is that we know that Moore's is wholly theoretical. But in a sense so is Sewel's. It is based on the assumption that value is value. Some people like that assumption, some don't, but it's still an assumption.

I agree with whoever said that exactly the same considerations as we've been applying here to Joe Sewell must equally apply to every player ever.

It isn't simply an old debate.
KJ wrote in #103,
I was taking my argument a little bit to the extreme to make a point, but in reality I use a "20 year" floating window for my comparisons to mitigate the sample size issue as Joe is correct in that talent will not be PERFECTLY distributed across positions in the short term.

Even using 20 years, Sewell compares as one of the best SS over a TWENTY year period, and that's nothing to sneez at...

Make that 21 years, so that "average at position" in 1921 is the all-MLB average at position, 1911-1931. That is a one giant step from "value" in sunnyday's sense; one giant step from [VALUE] TO ENABLE HIS TEAM TO WIN GAMES AGAINST THE TEAMS THAT HIS TEAM IS PLAYING, per KJ #116.

Even using 20 years, Sewell compares as one of the best SS over a TWENTY year period, and that's nothing to sneez at...

True. At his retirement in 1933, Sewell has 124 RCAA. Cronin has almost caught him in RCAA at that point with 117. It takes Cronin until 1937 to catch him due to a mid-career slump. Vaughn passes Sewell in 1935. Appling catches him in 1942.

Sewell's a good candidate, but I just think RCAP overrates him... the baseline for mid-twenties AL shortstops is just too low. Using RCAA instead of RCAP... and comparing him just against SS's, he still stands out ahead of Bancroft & Jackson... but its not as inflated a number as RCAP.

All this is just Lee Sinins's batting stats, of course. No fielding.

Visit
Games Played by Position, 1901-1919
and see that shortstops played more career games at their position than did counterparts at 3B, C, 1B, and 2B. That is true at rank #1, #5, or #20 on the Games Played list.

Just for comparison here are the leading active SS on WS:

1. ARod 338 if you slot him at SS
2. Jeter 275
3. Vizquel 249 and I saw a column somewhere arguing him for the HoF
4. Tejada 208 and I haven't seen such a column for Miggie just yet
5. Garciaparra 204 has quickly fallen behind in the triumvirate
6. Valentin 181 if you call him a SS
7. Renteria 179
8. Aurilia 159
9. Clayton 153 if you call him a SS :-)
10. Furcal 144 at what age? He could become the guy they want Vizquel to be

So on the all-time list now (ML only)

1.Wagner 655
2. Young 423
3. Ripken 419
4. G. Davis 398
5. Dahlen 393
6. Appling 378
7. Vaughan 356
8. Wallace 345
9. ARod 338!
10. Cronin 333

11. Banks 332
12. Ozzie 326
13. Trammell 318
14. Larkin and Reese 314
16. Maranville 302

17. Aparicio 293
18. Campaneris 280
19. T Fernandez 278
20. Boudreau and Sewell 277

22. Jeter 275
23. Concepcion and Bancroft 269
25. Stephens and Long 265

I would say that ARod is already in, Jeter maybe yes due to peak and timeline considerations but maybe not yet, Vizquel comes up short, Tejada is making progress, Garciaparra probably no longer really is. (Vizquel comes up #32.)

And Sewell? Just eyeballing it to a 162 game schedule (adding 5 percent) gets him to 291, still below 2 guys who are not going to make it, and in a dead heat with one guy (is that "one guy"? or "only one guy?") who already has. If you get into seasonal and other adjustments for everybody else, he trails Rizzuto as well.

If we were going to elect 25 SSs--and we might though 25 per each position on average would be too many--I'm still not sure I see Sewell as one of the top 25, much less one of the top 20 ML SSs. I know we're not supposed to do it, but I would vastly more prefer to save a couple of SS spots for Trammell (whom the HoF has clearly mis-evaluated) and Larkin (just in case), not to mention ARod and Jeter and maybe Miggie (sorry Nomar, it ain't gonna happen).

Anyway, just some thoughts upon eye-balling the career WS totals for active players.

Robin Yount is number two in case anyone (like me) was puzzled.

some thoughts upon eye-balling sunny's list:

Maranville accumulated 25 more win shares than Sewell. In 9 more seasons. Let's think about who comes out ahead in that deal.

Aparicio played 18. Campaneris 20. You get the point. How about a win shares above baseline analysis.. :)

But I'm OK with saving deserving spots for Trammell Larkin ARod Jeter etc

Aparicio played 18. Campaneris 20. You get the point. How about a win shares above baseline analysis.. :)

How about a list of WS/162 games? Sewell would do much better there.

PS. since this thread is also dedicated to Dave Bancroft.

WS

Bancroft 269/31-27-26-22-21-20-20-20-18-17-11-11-10 (13 years ? 10)
Sewell 277/29-29-26-24-23-22-21-21-21-17-16-15 (12 years ? 10)

Sewell is better for 9 of the 12-13 years, though Bancroft has the best year either ever had and is even two more times and has the 13th year.

For Bancroft it's 61 percent offense and 39 percent D, for Sewell it's 68-32. Considering they're within 8 WS overall, that is surprisingly close. It speaks highly of Sewell's D but also reflects well on Bancroft's offense. Bancroft is a A glove, Sewell an A-.

OPS+

Bancroft 98/122-22-19-10-9-9-4-(6 yrs<100)
Sewell 109/147-16-16-11-11-9-9-8-2-1-(3 yrs<100)

Take away the 147 and they're pretty comparable, but of course why would you take away the 147? This (Sewell +11 points OPS+) I am surprised that their offensive WS are so close.

Playing Time

Bancroft 1913 G (1873 at SS), ~7700 AB+BB, 6 years ? 140 games
Sewell 1903 G (1216 at SS), ~8200 AB+BB, 9 years ? 140 games but all of them are also ? 150, ultra-durable)

They seem to me to be extremely comparable. The biggest differences

1. # of games at SS, advantage Bancroft
2. Sewell's value is packed into fewer (fuller) seasons, advantage Sewell
3. Sewell had the one big year
4. Bancroft a slightly better defender
5. Sewell a slightly better hitter

I'm not going to obsess about two comparable players being 75 spots apart in the voting, we all know how that goes. If I didn't know better, I'd say if Sewell, then Bancroft.

I've got Sewell 10th and Bancroft 16th so I guess I agree, and I'm using WARP numbers, not WS or OPS+ like you used. I'd say Bancroft easily put up more defensive value since he never moved to 3B.

Interesting to me here - Bancroft played 10 more games, but Sewell had 500 more PA's (-HBP). That tells me that Bancroft had a good number of apperances as a defensive replacement or pinch hitter. A Relief fielder.

>Interesting to me here - Bancroft played 10 more games, but Sewell had 500 more PA's (-HBP). That tells me that Bancroft had a good number of apperances as a defensive replacement or pinch hitter. A Relief fielder.

I don't think you can conclude that. It's one hypothesis (or two: and the defensive replacement hypothesis is much more likely than the PH hypothesis). But again, I think there are better, more likely hypotheses:

Another is that Sewell batted at the top of the order and Bancroft near the bottom for 13 years. Over the course of 1900 games, how often would a #1 or #2 hitter get an extra PA vs. a #7 or #8 hitter? 500 would be a pretty reasonable guess, I would think.

And then there's a third--that Sewell played on better offensive teams that went through the lineup more times than Bancroft's did. Combine that with different places in the order and I'd be pretty sure you've covered it.

On Bancroft vs. Sewell:

I have Bancroft slightly higher. He was a better defensive shortstop than Sewell was in a context in which shortstop defense was even more important.

On the PH hypothesis: My old Macmillan Encyclopedia provides the following PH data for Bancroft and Sewell.

Bancroft as PH, career: 3-17
Sewell as PH, career: 5-25

And here's a fourth hypothesis on the PA differences: Bancroft became a regular in 1915, so the first third of his career took place before the 1920-21 offensive spike.

Well, first of all there's the shortened seasons in 1918-19. Looking at their careers, Sewell's leagues averaged about 1000 more AB/Season, which is 125 per team each year. As the regular leadoff hitter, my off-the-cuff guess is Joe's getting 15 of those. Over 13 seasons, that'd be about 300 AB. Throw in the difference in lineup positions, and that should just about cover it.

Another is that Sewell batted at the top of the order and Bancroft near the bottom for 13 years. Over the course of 1900 games, how often would a #1 or #2 hitter get an extra PA vs. a #7 or #8 hitter? 500 would be a pretty reasonable guess, I would think.

On the same team, a difference of six batting positions generates about six PA every nine games --more than 1200 pa in 1900 g.

Who but a catcher plays 15 years batting "#7 or #8"?
SABR Deadball Era Committee has
Bancroft in batting positions 2-7-2-1-6 for Philadelphia 1915-1919.

I’ve moved this discussion of Ozzie Smith and Rabbit Maranville from the 1987 election results thread, as it pertains significantly to Maranville’s case.

59. Chris Cobb Posted: October 17, 2006 at 09:51 PM (#2215736)
It's nice to see Maranville getting some discussion :-).

On the subject of Maranville: does anyone have data on his RCAP? It would be interesting to know in the context of this discussion. If possible, it would also be handy to see the first half of his career: say, through 1922, vs. the second half of his career. He was a bit above average as a hitter most seasons prior to 1923. He was continually below-average after. I think the advent of the lively ball, from which Maranville did not noticeably benefit, had something to do with that, though age and alcohol abuse surely played their roles as well.

On the subject of Ozzie Smith: I agree that he will be elected easily. BP sees him as an average hitter for his career, pitchers included. Without pitchers, he'd be, what, about 5% below average? I'm sure he was an above average hitter for his position during his career.

60. Mike Webber Posted: October 17, 2006 at 10:23 PM (#2215850)
Here is the data from the SBE for Maranville -
I dunno Chirs, I guess you are technically correct that "He was a bit above average as a hitter most seasons prior to 1923" but 1924 and 1925 put him back into red figures.

He was 20 in 1912. So his age 27 season was 1919.


61. Mike Webber Posted: October 17, 2006 at 10:28 PM (#2215867)
Chris, assuming that he is negative -61 offensively, how many runs above average with the glove would he have to be to be HOM worthy? 300?

What do you estimate his defense to be worth?

67. Chris Cobb Posted: October 17, 2006 at 11:57 PM (#2216186)
Mike,

Thanks for the data, and the good questions!

Your question about how many runs with the glove Maranville would have to have to be HoM worthy is not one I can easily answer, since I don't ordinarily use RCAP + FRAA as a measure, since I don't have access to RCAP for all players (I ought to get Sinins' encyclopedia . . . ). I also don't have a metric for valuing runs above average across different career lengths, so I don't know how I would compare Maranville's value above average to players with careers of differing length.

I can say this, however, that, if you look at Maranville from 1913 to 1930 and give him war-credit for 1918, he ends up with this line:

2314 g, -2 RCAP, 227 FRAA (according to BP WARP1)

That's a career of about the same length as Bobby Wallace's, or, more to the point, among eligible candidates, to Nellie Fox's (2367 g, with some coming in 162 g seasons). It's enough career for a middle infielder to be considered a career candidate, and he was 225 RAA. During his prime, 1913-22, he was 44 RCAP and 153 FRAA over 10 years, or about 20 runs per year on average. We could make a direct comparison, by RCAP and FRAA, of Maranville to Fox, and be assured that it was a fair comparison, since it would be over similar numbers of games played.

How would Nellie Fox look, for his career, by RCAP and FRAA (it wouldn't be an entirely fair comparison, because WARP doesn't like Fox's fielding, but we can be generous in moving those values upward)? How would Fox look over his 10-season prime, 1951-60?

A comparison to Fox on these terms would begin to answer the question of whether Maranville still looks like a serious candidate if you totally remove WARP1's FRAR levels from the picture (which I think are fuelling much of the skepticism of Maranville's [and Bancroft's] viability as candidates). I am prepared to rethink my position on Maranville based on the evidence of these comparisons (with suitable grains of salt for WARP1 FRAA for Fox), but if they show Maranville in a favorable light, I hope others will start rethinking Maranville.

68. Mike Webber Posted: October 18, 2006 at 12:30 AM (#2216223)
How would Nellie Fox look, for his career, by RCAP and FRAA (it wouldn't be an entirely fair comparison, because WARP doesn't like Fox's fielding, but we can be generous in moving those values upward)? How would Fox look over his 10-season prime, 1951-60?


Fox
Career 2367g, 139 RCAP, 62 FRAA
51-60 1532g, 185 RCAP, 76 FRAA

Maranville
Career 2670g, -61 RCAP, 192 FRAA
14-23 1458g, 46 RCAP, 154 FRAA

I just doubled his 1917 and threw out 1918 for a War adjustment.

Sewell - as a candidate that is in, but not a slam dunk.
Career 1903g, 346 RCAP, 110 FRAA

I suppose there is a context arguement to be made, but I am not going to try it tonight. If Rabbit's prime was 4 years earlier when scoring total were really in the tank it might mean something.

What do you think of that Chris, and we should cut and paste all this into Rabbit's thread.

OK, so the pairings I would want to look at are these:

2300+ games
Fox - 2367g, 139 RCAP, 62 FRAA
Rabbit - 2314 g, -2 RCAP, 227 FRAA
--advantage, Maranville

10-year prime
Fox - 51-60 1532g, 185 RCAP, 76 FRAA
Rabbit - 13-22 1454 g, 46 RCAP, 154 FRAA (2 war-shortened seasons)
--advantage, Nellie

Then Rabbit has another 2 1/2 seasons as, essentially, a replacement level player. These add little to his case, but help his counting stats a bit.

The peak/prime voter would prefer Fox based on these stats, while the prime/career voter would prefer Maranville, I should think.

The trouble I have with Maranville is pretty simple. No matter where he comes out on RAA, when I put a list of SS together he keeps coming out about 10th or 11th. I am partial to SS but even I cannot squeeze 10 or 11 SS into my top 30.

In 1924, Maranville was a second baseman. Although he was certainly capable of playing SS, the Pirates added Glenn Wright, a better all-around player, and shifted the Rab to 2B. This almost certainly kills his RCAP, since one of the 2Bs to whom he is being compared is Hornsby.

-- MWE

I think adding Bancroft to the RCAP/FRAA discussion would be worthwhile. I think it has been useful seeing Fox score ~200 runs while Sewell scored ~450 runs.

Bancroft
Career 1913g, 157 RCAP, 110 FRAA
17-26 1244g, 171 RCAP, 91 FRAA

No war credit for 1918 because he was + in fielding and negative batting, I call it a wash. In 1919 in 92 games he was 3 FRAA and 2 RCAP. So I bumped that up to 5 and 3 - did not adjust the games figure.

2300+ games
Fox - 2367g, 139 RCAP, 62 FRAA
Rabbit - 2314 g, -2 RCAP, 227 FRAA
--advantage, Maranville

10-year prime
Fox - 51-60 1532g, 185 RCAP, 76 FRAA
Rabbit - 13-22 1454 g, 46 RCAP, 154 FRAA (2 war-shortened seasons)

While I'm here I glanced Bert Campaneris the 1970's version of this player, thinking it might be close but he is well behind this group, especially when yo adjus the schedule length.
career - 43 FRAA 149 RCAP
1968-1977 138 RCAP, 89 FRAA

Still one of my fave articles in BRJ, from #15 in 1986. Rating the shortstops by David Neft, based on LWTs.

• He started with a batting factor which for the 109 SSs rated (with 5 years or more as a "regular"), with .375 as the arbitrary league average. The range was Bobby Wine at .289 to Honus Wagner at .484. (Total range of 195 points.)

• Then a running factor on a +/- scale from Maury Wills at +.030 to Bill Knickerbocker at -.040 (range 70 points, or more than 1/3 as much as batting, which seems like too much weight).

• And a fielding factor, also on a +/- scale, from Ozzie at +.174 (Bancroft and Wallace and Maranville the others who are over .100) to Jack Barry at -.090 (range of 264 points or significantly more than batting, which seems a bit heavy, but, hey, there you go).

• Then there was a longevity factor of +/- .050 for each year more or less than 10 as a regular.

Results (keep in mind this is as of 1986, through 1985 season)

1. Honus .614
2. Bancroft .538--hey, you knew this was fielding heavy
3. Wallace .527
4. Oz .516--through 1985 only
5. Chapman .498--no decline?
6. Tinker .495
7. Maranville .494
8. Bush .482
9. Bartell .471
10. Cronin .470

Among those eligible (and a couple other highlights)

11 and 14. Ripken and Yount as of 1985
16. Aparicio .453
17. Elberfeld .449
18. Fletcher .446
20. Vaughan .440--hey, you knew it was fielding heavy

21. T. Jackson .439
27. Logan .425
28. Joost .423
29. Doolan .421
30. Peckinpaugh .419

31. McBride .419
32. Wills .416
33. Alley .416
34. Belanger .415
35. Rizutto .415--not much help here
36. Groat .413
37. Hansen .412
38. Fregosi .410--nor here
39. Marion .407--and this is fielding heavy
40. Heinie Wagner .405

Some other highlights

46. Campaneris .391
50. Stephens .385
87. Bowa .314--recent article says he should be a HoFer
92. Kuenn .308--somebody said he should be a HoMer
97. Travis .292
108. Barry .258
109. Chico Fernandez (1957-62) .241

Among the "fielding specialists" discussed recently + one of my favorites:

Bancroft .377 batting +.007 running +.144 fielding
(Wallace) .388 + .009 + .120
Ozzie .329 + .023 + .174
Tinker .370 + .018 + .093
Maranville .358 + .013 + .124
Aparicio .344 + .020 + .064

OK, my fun stat . . .

((OWP - (PosAvgOWP - .073)) * (PA/19)) + (FRAA/9.558)

Essentially gets you WAR.

I use the entire MLB for PosAvgOWP, not just league (8-team league way too small of a sample). The -.073 gives you replacement level. PA/19 adjusts for playing time, FRAA/9.558 is the number of FRAA it takes to swing one game in a 4.50 R/G environment (which is what FRAA is).

Among our suspects (only counting 1913-31 for Maranville)



I really like it as a 'first sort' on the candidates.

While all of the guys outside are below the guys inside, I don't have an issue with drawing the line a little lower - 85.0 is actually plenty higher than some of our inductees at other positions (leaving out 19th Century guys):



Sure seems to me we are a lot easier on our OFs and 1B than our SS/2B/3B/C.

Other careers for "gloves"

Tinker 1803 g, 6 RCAP, 205 FRAA
Jennings 1285 g, 262 RCAP, 128 FRAA
Wallace 2383 g, 195 RCAP, 183 FRAA
Fregosi 1902 g, 203 RCAP, -43 FRAA
Aparicio 2599 g, 51 RCAP, 71 FRAA
Rizzuto 1661 g, 67 RCAP, 97 FRAA

Childs 1456 g, 354 RCAP, 35 FRAA
Doyle 1766 g, 273 RCAP, -144 FRAA
Evers 1784 g, 126 RCAP, 70 FRAA
Lazzeri 1740 g, 325 RCAP, -43 FRAA
Mazeroski 2163 g, -7 RCAP, 224 FRAA

Boyer 2034 g, 122 RCAP, 113 FRAA
Leach 2156 g, 121 RCAP, 115 FRAA
McGraw 1099 g, 459 RCAP, -17 FRAA
Collins 1728 g, 148 RCAP, 174 FRAA
Elliott 1978 g, 241 RCAP, 5 FRAA
Traynor 1941 g, 209 RCAP, 61 FRAA

I know I've left out a lot, but I have to go.

Sure seems to me we are a lot easier on our OFs and 1B than our SS/2B/3B/C.

That, or FRAA is flawed.

Dick Bartell > Duke Snider doesn't smell right to me.

That, or FRAA is flawed.

Yes.

-- MWE

What's the alternative? FRAA seems to jive pretty well with reputation for the old-timers. And why would it be biased in favor of IF over OF? Should we just guess, or better yet, since we can't quantify it (if FRAA are wrong) - just leave all of the glove spots out? That's not a good alternative IMO.

I'm not so sure I'm on board with this idea that "Zone Rating" is the Holy Grail or the The Great Answer Key, and all retrospective measures should jive with it, or they are wrong.

Zone rating is highly dependent on fielder positioning, and screwy things happen when fielders aren't in their traditional positions, which happens a lot more often than you'd think. Not to mention scoring decisions (not OS, Stats, BIS, MLB.com) that aren't exactly the most reliable either.

Sure Zone Rating is a step forward - but to say that WS and FRAA are absolutely useless or shouldn't be used seems ridiculous to me - especially when it's all we have for most of history.

As for Bartell/Snider, Snider is pretty overrated, he had a short career (AB wise, because he was platooned a lot), and wasn't a good fielder. Bartell was an above average hitter (a very good hitter for a SS), a fielding wizard and had a longer career than Snider.

Picking one comparison, where the two are essentially equal (74.8 - 74.3) and throwing it out there as an example of the whole thing kind of loads the question a bit, IMO. It's not that precise. Maybe I should just use whole numbers instead of decimals, which imply and accuracy that doesn't exist.

Now if you'd said Dick Bartell = Duke Snider smells a little funny, then I could see it, but I'd disagree. Was Dick Bartell better than Duke Snider - probably not. Were they pretty close, in terms of career value, probably yes.

Maybe the system weighs fielding a little heavy, but I don't think it does. It's got 19 SS over 70, 16 RF, 20 LF, 14 CF, 9 C, 12 3B, 14 2B, 15 1B.

I think it probably just puts SS defense in it's proper place - and shows that CF is quite overrated defensively before expansion and the huge parks of the 1961-92 era.

So I don't see the Bartell/Snider thing as a problem at all.

Maybe the system weighs fielding a little heavy, but I don't think it does. It's got 19 SS over 70, 16 RF, 20 LF, 14 CF, 9 C, 12 3B, 14 2B, 15 1B.

That's among the eligibles . . .

Joe Tinker (for that matter shortstops during this time) gets less RCAP than would be expected because of the huge outlier that is Honus Wagner.

Esteban . . . good point - but it's not just Honus . . . you've got George Davis in there too. But one guy isn't going to kill the curve if you use major league wide numbers.

Career Pos Avg OWP (using MLB, not just league):

Tinker - .456
Wagner (1901-16) - .453
Sewell (1920-28) - .402
Chapman - .423
Peckinpaugh - .407
Dahlen - .464
Jennings (1891-98) - .464
Glasscock - .457
Wallace (1899-1915) - .455
Ward (1885-91) - .471
Appling - .433
Cronin - .418
Rizzuto - .429
Boudreau - .421
Aparicio - .392
Campaneris - .372
Ozzie Smith - .394
Derek Jeter - .421
Miguel Tejada - .427
Alan Trammell - .392
Cal Ripken (1981-96) - .400
Robin Yount (1974-84) - .376
Dave Concepcion - .375
Rabbit Maranville (1912-31) - .409

SS just hit better until around 1910-15 or so. Not sure why. But it's not just Wagner.

Thanks for the reply Joe, I forgot about Davis's tail end falling in this period. If I understood correctly, using the average for mlb would also work in smoothing out somewhat other outlier situations, like Lajoie, Bonds, Ruth, et al., right?

On a different track, is FRAA calculated with using mlb averages or just the seasonal NL/AL data?

And here's a hypothetical: Would having a ten team league one filled year with 9 Dick Stuart clones's and 1 Jim Thome on defense and a different year the same ten team league filled with 10 Keith Hernandez clones on defense cause the Thome clone to score higher than any of the Hernandez clones on FRAA because of the lower FR average for the Thome clone due to the Stuart clones being his comparison points, even though all 10 Hernandez clones are better than the Thome clone? If I am interpreting FRAA correctly, the Hernandez clones league would have a higher FRA, yielding lower numbers of FRAA for all of them. The Thome clone would clean up in terms of FRAA because of the dreadful Stuart's. However, the Hernandez clones would be shorted on this measure in terms of how good their defense was in comparison to Thome. Am I way off in my understanding of the situation here? Thanks in advance for any feedback that can clear up my confusion.

Looking at the shortstop lists posted in #144. All the HoMers could hit. Only Reese had an OPS+ under 100, at 99, and he would not if he had played 43-45. Also, they all had either long careers or big peaks or dominated when the best at his position were the wrong color for the "Majors".
The Non-Homers usually had both a shorter career and lesser peak than any of the HoMers.
Here are the career and 3 consecutive years "peak" win shares totals for the various players in the list. All totals adjusted to 154 games with war credit.

HoMers:

Wagner 237.0 / 671 / 149
Vaughan 140.7 / 356 / 110
Appling 137.1 (with some war credit) / 408 / 89
Davis 124.4 / 431 / 86
Dahlen 116.6 / 422 / 84
Cronin 116.0 / 333 / 100
Sewell 108.7 / 277 / 76
Boudreau 104.5 (war demerited) / 277 / 86 (no demerits)
Reese 103.1 (war credit) / 384 / 81
Banks 96.4 / 332 / 93
Glasscock 94.0 (not schedule adjusted) / 342 / 87
Wallace 90.6 (not schedule adjusted) / 361 / 75
Jennings 73.6 (not schedule adjusted) OOPS / 239 / 110

Non HoMers
Bancroft 85.7 / 274 / 84
Fregosi 84.6 / 261 / 81 (totals need to be adjusted down for 162 game schedules)
Travis 83.5 (with a ton of war credit) / 269 / 84 (a lot of war credit)
Rizzuto 80.7 (with war credit) / 314 / 80 (war credit)
Stephens 75.8 / 265 / 84
Bartell 74.8 / 252 / 70
Aparicio 69.0 / 293 / 60 (totals need to be adjusted down for 162 game schedules)
Maranville 68.8 / 322 / 71

Looking at each individually:
Bancroft had trouble staying the line-up so his peak is hurt, though he still shows up well and his OPS+ is a good for position 98.
Fregosi: Had the offense, 113 OPS+. Short career, not a great peak.
Travis: Had the offense, 108 OPS+. WWII after age 27 season requires too much speculation for most voters, I think. And short career.
Rizzuto: WWII took peak seasons and OPS+ 93.
Stephens: Offense 119 OPS+, but shortish career and WWII demerits.
Bartell: Lack of peak and OPS+ 96 would be worst among HoMer SS.
Aparicio: Lack of offense, 82 OPS+, lack of peak
Maranville: Lack of offense, 82 OPS+, lack of peak


To my eyes, to be a HoMer SS, generally, you have to hit very well for your position and play a long time or have a giant peak. Boudreau had the peak argument (I guess. I don't remember the for-arguments and he is still not PHOM) and Sewell the positional dominance.

Anybody that wants to see "Joey" Sewell in action (well, drawing a walk) in the 1932 World Series, and some other interesting things, try http://asap.ap.org/data/interactives/_sports/asapretro/.

C---great vids, thanks! You can see Slaughter's mad dash pretty clearly, and it's hard to see if there's actually a double-clutch. What's evident is that Pesky's throw appears to be simply up the line. I don't know if the announcer was overdubbed or not, but he clearly says that Walker doubles.

On a different track, is FRAA calculated with using mlb averages or just the seasonal NL/AL data?

My understanding is that WARP-1 is based on single league data only. WARP-2 adds in both the league quality (relative to each other) and the "timeline" quality (how single league seasons rate relative to each other).

And here's a hypothetical:

You've got it right. This is true of any stat that makes measurements relative to average (not just FRAA but EQA, OPS+, ERA+, Win Shares, etc.) Average can shift significantly over time.

Some here might argue that your example is only a slight exaggeration of a 1B comparison between the 1900's (10 Hernandez') vs the the 1930's or 1880's (10 Thome's; I wouldn't tar them all as Dr. Strangeglove). Whether this was actually required by the way the game was played or was just managerial fad is open to debate.

And here's a hypothetical:

You've got it right. This is true of any stat that makes measurements relative to average (not just FRAA but EQA, OPS+, ERA+, Win Shares, etc.) Average can shift significantly over time.

Yes, and I'd argue IT SHOULD. Value, vs. average or other baseline, is relative to particular season/year's pennant.
And here's a hypothetical:

You've got it right. This is true of any stat that makes measurements relative to average (not just FRAA but EQA, OPS+, ERA+, Win Shares, etc.) Average can shift significantly over time.

I've been looking into evaluating Bancroft in more detail and noticed that right now I'm giving him extra credit. Does Bancroft merit war or minor league credit or was I making a mistake?

Definitely not war, he played in 1918...he did debut a little late, but I'd have to see a strong argument for minor league credit...

This thread is interesting in more ways than one. About 2-1/2 years ago, 123 articles in two weeks; two years later, a revival. Some observations:

Ed Konetchy was on the agenda! He is forgotten and Beckley is still knocking on the door.

Should the HOM honor about the same number of players at every fielding position? Some questions will never go away.

Sewell took some people by surprise. But he didn't quite waltz in, and Stop Sewell set him back a few decades.

Not only Sewell but Childs(2B), Moore, and Jennings have been elected, contrary to expectations expressed here. The title relevant today would be "Long, Tinker, Bancroft, and Maranville". Partly what happened is a shift in favor of high-peak players.

It appears to me, Chris Cobb expected that Bancroft and Maranville would never make his ballot, but he has become one of their supporters.

Philly #5 listed 20 shortstops, providing WARP1 ratings for most of them. Among eight numbered 10-18 and named in bold (eligible but not in the HOM as of Nov 2004), there has been significant change in two ratings: Herman Long down about 5 points, Donie Bush up about 5 points! The other revisions are within +/-2. Only one man knows much about the frequency and magnitude of revisions.

Of course everybody who played in 1918 and 1919 deserve a little extra credit for the short seasons. If a guy had a peak season in 1918 and if you're a peak voter, it could make a difference.

Chris Fluit on walter Rabbit Maranville in "1999 Ballot Discussion"
He did have a pair of good offensive years in 1921 and 1922 but otherwise, he was in the line-up because of his outstanding glove.

He scored more runs in 1921-22. But there is a reason 1920 demarcates the so-called dead ball and lively ball, and he moved from a bad team in a huge ballpark to a good team in a smaller one.

Except for the teammate effect --that is, at the team level-- baseball-reference tries to show the runs inflation directly with AIR, one of its "Special Batting" statistics. For Maranville in Boston and Pittsburgh, AIR is 93-83-78-75-74-76-79-87 and 109-113-109-106)

People may think the Giants dominated the league during their four-year run (Maranville's four years in Pittsburgh), but the Pirates finished 4-8-8-3 games behind, and 11-32-33-37 games ahead of Boston (his team in the Deadball Era). The 1921 Pirates were the highest scoring team of the span and they finished 3-2-3 in scoring thereafter.

As Chris Cobb has observed, Maranville was an average batter, better than shortstop average, and thus an overall player worthy of his stardom in Boston.

By OPS+ he was a weaker batter as a rookie in 1913 than he would be in 1921-22, equal in 1914, stronger in 1915-20 (losing essentially a full season at his peak to the war).

When he scored only 79 runs in 1916 that was 6th in the league. 69 runs in 1917 was only 14th but it was second to Max Carey among the four second-division teams; that is, he led his team and would have led two others.

Breaking his career into four-year spans, except the last is five years,
OPS+
_90 - 1913-16 Boston
107 - 1917-20 Boston (1918 only 11 games)
_85 - 1921-24 Pittsburgh
_63 - 1925-28 Chicago, Brooklyn, St Louis (only 1000 pa)
_70 - 1929-33 Boston

Before and after his batting peak, he ranked 3rd and 2nd MVP as a rookie and sophomore, and 7th in 1924 when a new award was established.

For some reason I checked the list of Sac Hit leaders at baseball-reference. I thought Dave Bancroft might be one of the leaders with more than 200.
He is only 75th. He was too late! Well, the Top 20 includes no one who debuted so late as 1915. The earliest is Keeler 1892. Sixteen of 20 debuted during the ten years 1905-1914, Cobb to Heilmann (I like that pair). With 300 Maranville is 11th, between Speaker and Cobb.

It appears to me that the "200 Sac Club", 82 players, includes 78 players from Keeler 1892 to Mule Haas 1925 plus this quartet.
1947 Nellie Fox (77t)
1978 Ozzie Smith (72t)
1987 Tom Glavine (79t-active)
1989 Omar Vizquel (57-active)
It's making a comeback.