Howie
why were there fewer great OFs in that era/or why have we elected so many fewer?
This one really surprises me.

The 1970s was my decade and I thought then that the dearth of outfielders was notable. As Clemente, Mays, Aaron and Robinson passed from the field who were the active players recognized as all-time greats? Johnny Bench; Joe Morgan and Rod Carew; Pete Rose (in the infield more often than not); soon George Brett and Mike Schmidt. Reggie was the big star in the outfield -- superduperstar, yes, but never more than Al Kaline or Willie Stargell or Lou Brock in his accomplishments, a sure HOFer but not one of the greatest outfielders. Same for Jim Rice, Dave Parker, Dave Winfield somewhat later. By the mid-1980s Bill James was an opinion-maker, at least for me, so I knew that Wade Boggs, Ryne Sandberg, and Cal Ripken were all-time greats. Meanwhile the best outfielders, Henderson and Raines, were recognized only by batting position: maybe the best leadoff men of all time.

You just mentioned Brock in the same sentence with Kaline and Stargell.

Once again, Reggie Smith doesn't get credit for the Replacement Win Shares; the replacement player does. Tell me what Reggie Smith did. Not what his team was able to do when he wasn't playing. The biggest problem that I have with DanR's system is that it consistently overrates players who weren't in the line-up.

Yes, Brock Kaline and Stargell were certainly going to the Hall of Fame but never mentioned as all-time great players as Clemente was, I'm old enough to remember, as well as Mays Aaron and Robinson. Reggie and later outfield stars until Griffey Junior(?) or Barry Bonds were in the same big class.

Kaline was long in the tooth when SI named Reggie superduperstar. Maybe Yaz and Billy Williams.

It's worth noting that when Aaron and Robinson retired there must have been many older baseball fans who recalled thought of Gehrig, Foxx, and about ten outfielders as the greatest players of all time. Mathews and Banks hit 500 home runs, which was a big deal then, but I don't recall that they were considered first tier stars. I never heard that Jackie Robinson and Yogi Berra were in the conversation at second base until the first BJHBA. In historical memory those guys were shaded if not dwarfed by the outfielders of their time and that had been true since Rogers Hornsby, I believe, about three baseball generations.

I never heard that Jackie Robinson and Yogi Berra were in the conversation at second base until the first BJHBA.

Yeah, that Bill James really opened my eyes.

>what happens if four teams happen to have a revolving door at a position in an 8-team league?

Well, that just doesn't happen.

As for Concepcion's '84-'85 and '87, he played in >100 games with ? 75 OPS+. Why not include them. They may not help his rate but the method is designed to help anything. And anyway the total (prime) WS above median is more interesting than the median year differential at least to me. You're pretty fixated on the rate. It's the 71.5 that's especially not impressive. Like I say, Rizzuto is in the same range with WWII credit, and against a great cohort.

Like I say, this is just a quick (not) and dirty reality check used along with career and peak and prime and annual totals.

Tell me what Reggie Smith did. Not what his team was able to do when he wasn't playing.

But if you're the team employing the player, then you absolutely consider the value that the replacement level provides. If, eg, I earn 20 win shares in only 100 games played, then the team has 60 other games where it can earn win shares at the position from a replacement player. If I earn those same 20 win shares over 162 games, then the team only earns 20 wins shares, period.

A player who produces at a higher rate, with less durability, is more valuble than the player who produces the same aggregate amount in a season but over a longer period of time. And the difference is substantial. Since our goal is to enshrine the players most valuable to their teams, we should respect and acknowledge that difference.

With all due respect to Chris, I just think that Dan and 'zop are objectively correct here. If a rating system gives credit for play below replacement value, it will overrate players who are durable and/or have long careers. That is all that Dan's charts demonstrate, but it is a huge point that half our electorate consistently forgets.

You're right, Chris Fluit; Smith shouldn't get credit for the replacement player's Win Shares. But neither should the replacement player--because they're worthless. They are Win Shares that would cost a team nothing more than the minimum salary. Like water, they provide a benefit (we need to drink it to survive; the replacement player does get some hits) but are not scarce--the fact that I turn on my tap doesn't prevent you from doing the same; the fact that Team A is getting replacement Win Shares doesn't prevent Team B from getting them as well. Thus, they have no value: water is free, and so are replacement Win Shares. The proper way to compare those players is to subtract 14 WS per season from the Total Win Shares column, giving you Win Shares above replacement. The overall number would change, but the gaps between the players and their ranking would remain constant.

What is your basis for arguing that I "overrate" players who weren't in the lineup? In your mind, what is the "correct" replacement level to use to avoid such "overrating?" I sure hope you have a better answer than "whatever BP WARP or Win Shares say," since--I repeat for the umpteenth time--when a player goes down, he is NOT replaced by the 1899 Cleveland Spiders; he is replaced by a minor league veteran or bench player, whose performance level has been empirically measured a bajillion times. The fact that Bill James and BP have bigger reputations in the sabermetric community than I do doesn't change the fact that they're just dead wrong about this, and if voters took the time to actually look at any of the available data on the question, they would see so for themselves.

Again, if you have managed to unearth some study showing that replacement level is in fact far lower than roughly 80% of positional average, please do the group the enormous favor of sharing it with us. Otherwise, I'd respectfully request that you familiarize yourself with the easily accessible body of research on this question before asserting that my system is "overrating" anybody.

First off, I'm not giving Reggie Smith any sort of credit for years 18-22 when he wasn't playing (why is he getting 14 WS in '83, his last year was '82?).

Second, from those numbers you posted at the end of the last page, Dan, Perez looks pretty good to me. That '67-'73 run is pretty nice. And, if you want to add years to the other guys, his years at or below replacement are similar to the them.

Also, I like Cash, he's 28th on my ballot. And I just took another look at Smith, I think I'll move him up, not high enough to be on the ballot though.

I have no idea if Dan consistently overrates players who are not durable, but certainly John McGraw and Reggie Smith are examples. There are counter-examples like Lundy and Concepcion, too.

Okay, this set of issues (replacement level, in-season durability) is a little tricky. Here is a stylized example that may illustrate a couple of key ideas.

For simplicity we'll measure everything in terms of Win Shares (WS). Suppose we agree that replacement level is 10 WS in 162 games, or 5 WS in 81 games.

Player A earns 16 WS in 81 games played in Year 1 and 16 WS in 81 games played in Year 2.

Player B earns 16 WS in 162 games played in Year 1 and 16 WS in 162 games played in Year 2.

Player C earns 32 WS in 162 games played in Year 1 and then retired.

Of course, each player earned 32 WS total. Player A has serious in-season durability problems.

Calculating WSAR (win shares above replacement) shows Player A gets 11 (16-5) each season = 22 total. Player B gets 6 (16-10) each season = 12 total. Player C gets 22 (32-10) total.

In terms of calculating the total value that each team receives (including the replacement value for games the player in question did not play), player A's team receives 16+5 = 21 each season for a total of 42. Player B's team receives 32 of course since he played every game. Player C's team receives 32 for Year 1, but we then must add in the 10 for Year 2 to get the proper total of 42 or else we'll claim that Player A is more valuable than Player C.

From this simple stylized example, I think a few ideas are illustrated. First, according to a WSAR perspective a higher replacement level implies less value to in-season durability. Compare the example above with a replacement level of 0, say. If a team can easily find a very good player (say in Triple-A), then it doesn't matter much whether their regulars can stay in the lineup.

Second, from a team total value perspective it is important to tally over the same number of seasons. Or else as we saw above Player A looks better than Player C and Player B looks the same as Player C.

Third, related to the first point, there are many "levels" of "replacement level". The time a player is not in the lineup can affect who a team can find to replace him. The quality of a player that a team finds with 5-minutes notice is apt to be significantly lower than the quality of a player that a team finds over an off-season. Analysts have argued that the replacement level tends to be very low in the very short term and asymptotes to the league average over time. Thus,we must be careful in assessing replacement level when in-season durability is concerned.

Lastly, from a pennants-added perspective, being linear WSAR undervalues in-season durability. Clearly Player C's team has a better chance of winning a pennant than Player A's team. Yet each has the same WSAR total.

So, I lost all my older HOM stuff, but I managed to rebuild my system with a big emphasis on Dan's numbers, with a bit of my personal biases thrown in. I wanted to ask about Bobby Bonds, which surprised me a lot. He looks a lot like Reggie Smith, and that might get him into my ballot (one of my personal biases is a bigger emphasis on in-season durability, by the way).

AJM--that's my estimate of his Japan year MLE in Win Shares.

Sunnyday2--I repeat: What is the criteria by which you determine who is being "overrated?" According to straight Win Shares, I overrate such players, because straight Win Shares uses a batty replacement level of 52% of league average. I have never heard of a study suggesting that actual replacement level is in that ballpark. I reiterate my request for voters who seem to be asserting that anything that disagrees with Win Shares must be incorrect to actually go and read the literature and research on this. It is Win Shares (and BP WARP) that overvalue playing time and undervalue rate because their replacement level is too low, not the other way around.

Rob Wood--yes, the higher your replacement level, the more rate matters and the less playing time matters. The need to compare over the same number of seasons is precisely why I added in the 0-WS seasons for the players who played less than 22 years in my previous example.

I completely disagree with the increasing-replacement-level-over-time concept. Even if we accept that what would generally be called Freely Available Talent (FAT) may not in fact be freely available at a moment's notice within a season, the guys who actually replace starters who go down don't come from outside the organization; they come from the bench. Keith Woolner has shown that backups' production is the same 80% of positional average that FAT players are. Ergo, the in-season replacement level is the same as the offseason replacement level. At the other end of the spectrum, league average players will never be freely available, by definition (at least outside of Lake Wobegon). Sure, over a period of years it will be easier for a team to find an above-replacement player than it is over a period of days. But that above-replacement player will in turn earn an above-replacement salary. If you are going to compare a player to league average in the long run, you have to factor in the cost of acquiring and retaining league-average talent.

As for the pennants added issue, this is why I use the salary estimator. The market will pay $0 for all performance up to replacement level (e.g., the first 3 BP WARP or 13ish WS per season); after passing the threshold where players accrue value, it pays progressively more for each marginal win than it does for the previous one. My ranking system uses empirical data to apply this pattern to HoM candidates. This is why Doc Gooden is on my ballot.

Papa Bonds was a terrific player, and definitely beats Lil' Reggie on a peak/high prime basis. But Smith has enough extra career to edge him out comfortably for me.

the guys who actually replace starters who go down don't come from outside the organization; they come from the bench. Keith Woolner has shown that backups' production is the same 80% of positional average that FAT players are. Ergo, the in-season replacement level is the same as the offseason replacement level.

Don't you also have to take into consideration the guy who replaces the bench player?

Not if the guy who replaces the bench player sits on the bench...I mean, yes, theoretically, you take a hit from FAT to almost-FAT at your second or third pinch hitter/defensive replacement, but that gap is just not big enough to amount to more than a small fraction of a win, so I don't feel any compunction about ignoring it. This is not at all the case with bullpens, where you absolutely do have to take into account the whole "chaining" process to calculate the proper value of high-leverage relievers. But for position players, I just don't think it moves the needle enough to be worth taking into account.

It matters a very small bit for starting pitchers too, since replacement starters are often long relievers, and then you have to factor in the mopup man getting those long relief innings as well. I count that at .005-.010 of winning percentage on a rate basis. But for hitters it's just too small to matter.

Overrated in this context only means that you rate the player higher than the consensus. I wouldn't claim that anybody overrates or underrates players against an objective standard because I have no clue what the objective standard (a players "real" value as opposed to what somebody or a "system" thinks the value is). Well, OK, I take that back. Somebody here overrates Jake Daubert, Ginger Beaumont and George Kell. Oh, and Sam Leever. But otherwise....

DanR #114
Keith Woolner has shown that backups' production is the same 80% of positional average that FAT players are. Ergo, the in-season replacement level is the same as the offseason replacement level. At the other end of the spectrum, league average players will never be freely available, by definition (at least outside of Lake Wobegon). Sure, over a period of years it will be easier for a team to find an above-replacement player than it is over a period of days. But that above-replacement player will in turn earn an above-replacement salary.

Perhaps Keith Woolner's perspective is management consulting, so monetary compensation is an important element of the analysis. Fantasy league consulting is important to some analysts and in that world, too, a cost of acquiring a player is an important element.

But if we go that way here, we should give Rob Wood's Player C credit for retirement, which saves the club his annual salary --or should debit Player A because his two 81-game seasons cost the club two annual salaries.

from Rob Wood #112, a too-short quotation, for reminder only
. . . Player C earns 32 WS in 162 games played in Year 1 and then retired. . . .
. . . Second, from a team total value perspective it is important to tally over the same number of seasons. Or else as we saw above Player A looks better than Player C and Player B looks the same as Player C. . . .

Sunnyday, that's an important clarification! The word "overrate" implies a normative judgment, that the evaluation is incorrectly favorable. There's no doubt that my system values high-rate, low playing time guys more than the group consensus, but that doesn't make it wrong. I think the group has formed that consensus (which is obviously in my view a misguided one) because of its reliance on WS and BP WARP, both of which share the same flaw of a too-low replacement level. I think if my numbers had been around since the beginning, and if my name had as much sabermetric credibility as Bill James's and BP's do, the consensus might have evolved quite differently.

Paul Wendt, a team always has to pay 25 salaries no matter who its players are. Retirement doesn't save a team money, it just means the team will be paying someone else...

I think a retirement often saves a team money, as an over-the-hill veteran is replaced by an inexpensive farmhand.

By some of the theory I see here, an established and beloved star who is now fading possibly could get 'credit' for not putting the team in a tough spot?

That is only the case if said over-the-hill veteran was overpaid to begin with (which is often the case).

I certainly don't care how much surplus value (above his salary) a player adds or subtracts to his teams from a Merit perspective. I just use the overall market compensation pattern to determine a) where value begins to accrue (around 80% of positional average) and b) the shape of the curve for compensation above this zero point (x^1.5).

3B

This is close.

1. Ed Williamson +11.5 WS in his median year versus the position median (noting that he played SS for 2 years) and +109.5 WS versus the position median for his prime years totalled up

2. John McGraw +13 and +105.5. The career total is better than expected though I did fudge in his favor by counting 2 years that would not usually qualify by my definition of "prime" but which seemed to have sufficient value to count

3. Tommy Leach +13 and +102

(gap)

4. (Ken Boyer) +9 and +91. A mistake in my book, only 9 prime years
5. Al Rosen +11 and +83
6. Bob Elliott +9.5 and +81

I also calculate the differential versus the "best" at the position each year and Leach and Rosen come out on top there, not McGraw, which might surprise somebody. McGraw, Elliott and Williamson are pretty close. Boyer of course was pitted against Eddie Mathews and got demolished. Leach had a fairly weak cohort whereas Rosen just demolishes the competition (AL only).

Wow, I'm stunned to see that McGraw does that well in a Win Shares-based system--I would think WS would just clobber him for the missed games. Perhaps the median players you're comparing him to also had limited playing time?

I agree that Boyer was a bad choice.

Not really fair to compare a guy playing in a 12-team league (with no competing league) to one in an 8-team league, especially when the latter is in the weaker league during one of the greatest league strength disparities in the game's history.

>Not really fair to compare a guy playing in a 12-team league (with no competing league) to one in an 8-team league, especially when the latter is in the weaker league during one of the greatest league strength disparities in the game's history.

Meaning McGraw and Leach?

The rank order here is not nec. how they end up on my ballot, just the rank order that this particular analysis would suggest. It's one column on the spreadsheet. I do think Leach was a very valuable player, but the comp might be Tony Phillips. I can't quite decide whether we under-rate them because they don't have an obvious space in a matrix (a position) or not. I think Leach was more valuable than Bresnahan. ie. about equal year-to-year but therefore more valuable for the career. I'm a peak voter, but their peaks are not that different in this case.

McGraw's medians were Hallman (2B), Corcoran (SS), Shindle, Everitt, Cross, Irwin, Casey and Cross again. Adj to 162 games the medians are 15, 16, 13, 11, 14, 13, 11 so, yes, those are pretty low (well, other than the 15-16 which are 2B and SS), especially considering that they are straight-line adjusted to 162 games which usually inflates the adjusted total pretty significantly.

Actually what is shocking is that McGraw was the #1 3B only 1 time. #1 were G. Davis 25, J. Collins 26, G. Davis 21, Collins 34--in 1899 McGraw was #1 at 34 versus Honus Wagner at 26.

Leach's are 12, 8, 20, 13, 15 at 3B, and 16, 13, 13, 13 in CF. You could argue this undervalues Leach because I define prime as ? 100 OPS+ for 3B (versus 75 for C, 2B, SS). Back before about 1920 I should probably evaluate 3B like a "glove," i.e. ? 75 OPS+. But unlike McGraw, Leach was #1 at his position 4.5 times--1.5X at 3B, 2X at CF.

Williamson was #1 2.5X, one was at SS in 1888. #1 were McVey, Kelly, Bradley, Williamson and O'Rourke, Sutton, Williamson, Sutton, Sutton, J. Ward (SS), Williamson (SS). The medians were Hankinson, Warner, Whitney, Bradley, Farrell, White, Hankinson, Esterbrook, Kuehne, Irwin.

In sum, Williamson's medians were 26 WS versus a median of 14.5 (and a position best of 31) for 10 years.
McGraw was 26.5 WS versus a median of 16 (and a position best of 30.5) for 8 years.
Leach was 26 WS versus a median of 13 (and position best of 24) for 9 prime years, but he also had several "non-prime" years.

Leach BTW was 24-32-13 at 3B and 25-24.5-13 in CF. The best CF were not as good as the best 3B, but the medians were the same.

OK, that's not *so* shocking. Davis and Collins are both eaaaaasy HoM'ers--no shame in finishing behind them. That said, I have McGraw as the best shortstop in the game in 1893 (Herman Long's WS would be inflated by the team's outperforming its stats which I don't credit), and the best 3B in the game in 1895 (over Davis), 1898 (over Collins), 1899 (over Jimmy Williams, who was a heckuva ballplayer for a little while), and 1900 (over Collins).

I was referring to McGraw and Rosen. Being the best 3B in a very weak 8-team league (with the stars in the NL) isn't exactly the same as being the best 3B in the majors.

Rob Wood, Andrew M, DanG, Mike Webber, jimd, Chris Cobb: Your thoughts on Dick Lundy?

DanG, ...: Your thoughts on Dick Lundy?

I work for the church; likewise, I move in strange and mysterious ways.

Lundy is clearly a legitimate candidate, not to be dismissed casually. I have a couple open ballot spots, maybe he'll get one. Or maybe not.

Unlike some here, I feel no urgency to make drastic ballot revisions before the 2008 election. A wholesale reassessment of the borderliners is planned by the 2009 election.

Well, you asked.

I guess the worry is that the group will start to fall apart by '09, so this may be the "last shot" with the full electorate necessary for legitimacy. Anyways, Lundy appreciates your consideration. :)

Hey, it's equally Everybody's last shot in that sense. So let me ax this question:

Make a short list of the most dominating teams of all time. Let's say, the three best.
Not one-year wonders, but 3 or more seasons of true success.

Okay, now name the best player on each of those teams.

You got three player's names? Good.

Are they all our HoM, and in your PHoM? Wouldntcha think so? I mean, does it not make a boatload of sense that the BEST player on the BEST dynasty team ever MIGHT BE WORTHY of the HoM?!?

But for many voters among us, the answer would be "well, two of them are obviouos choices, but I don't think the third guy deserves to be on my ballot".

Which to me sounds kinda strange. Howzit sound to you?

Next week I am, AGAIN, putting the best player on possibly the greatest team ever assembled on my ballot. I wish more than 4 voters would join me this time.

Presuming you're talking about the Cubbies, according to BP the best players on those teams were:

1903: Kling
1904: Evers
1905: Tinker
1906: Chance
1907: Evers
1908: Tinker
1909: Brown
1910: Brown
1911: Sheckard

Brute force here, but Hughie Jennings he was not.

Thinking of your test:

1890s Orioles: Jennings, in
1900s Cubs: Brown, in
1910s Red Sox: Speaker, in
1920s Yankees: Ruth, in
1930 A's: Grove/Foxx, in
1930s Yankees: DiMaggio, in
1950 Yankees: Berra, in
1960 Yankees: Mantle, in
1970s A's: Reggie, in
1970s Reds: Morgan, in
1990s Yankees: Jeter, will be in

Yep, I'd say we've got it covered.

OK, here's a different way to address the question, who was the most valuable player to the Cubs' dynasty? Using my WARP numbers rather than BP's, and weighting each year from 1903-13 by the # of wins above .500 the team was that year (e.g. 1910 gets a weighting of 31 and 1911 gets a weighting of 19), I get:

1. Mordecai Brown, 1174
2. Frank Chance, 917
3. Joe Tinker, 847
4. Johnny Evers, 756
5. Jimmy Sheckard, 713
6. Ed Reulbach, 670
7. Frank Schulte, 586
8. Harry Steinfeldt, 530
9. Orval Ovearll, 528
10. Solly Hofman, 510

Brown was clearly the biggest contributor to the dynasty. Chance was indeed second, but Tinker was not far behind.

Is it not possible that TomH is referring to the NY Giants of the 1910's? Is TomH a Larry Doyle supporter?

I think he's talkin' Cubs. And I assume he's talkin' Chance. But I don't see it. Aside from the fact that Brown and Sheckard are already in, so there's no real gap...the bottom line is the Cubbies had soooo much talent and Chance such playing time issues that it's not nec nor appropriate to make a special case for him. ie. Elect him just because the team was so good. They're already covered but even if they weren't, they just had a whole lotta guys who could play.

But of course Chance was basically done by 1911. Make the cutoff earlier and he was as valuable as Three-Finger using Win Shares, and that's without even discounting Brown for Win Shares inability to give much credit that deserves to go to the defense, as well as Win Shares very low replacement level that obviously hurts Chance more than Brown. Those two are neck-and-neck for best player on the Cubs' dynasty, with Chance being the obvious choice thru 02-08 (when they finished their 3-year consecutive WS run) with Brown better if you look at 07-13.

DanR's post 132 shows Brown the best Cubbie of the 00s, but he was not, not, not. Chance was, and I don't think it's even particularly close. Of the other dynasties, every team had its best player so obviously "in" it aint even questionable, right?

Dominant teams in baseball history tend to have players that have long careers, and/or who who were probably among the top two or three players in baseball at their peak. the 06-08 Cubs were unique among dominant teams, in that they had neither. The Cubs were filled with solid players, playing at their peaks, but without anyone with an unusually long career or particularly dominant play. This is extremely rare - not just among dominant teams, but among virtually all world series winners. The early 70's A's were closest, I guess, but at least they had Reggie.

In 1986, the Mets had a great regular season - 108 wins, won the division by 21.5 games, etc. I remember people trying to make a case that Hernandez or Carter or someone like that should have been the MVP because - well, you win 108 games, and someone on your team has to be the MVP, no matter what Mike Schmidt did. The logic didn't work then, and it doesn't work now (I understand the cases are a little different, but the logic is the same).

The argument Tom makes for Chance was exactly what put him in the Hall of Fame - but its the only argument that can be made on his behalf. His career as a regular player was very, very short, and he was never among the best two or three players in the game. He was very good for a few years, and those few years happened to coincide with the best few years of a whole bunch of other Cubs. That's not a reason to put him in the Hall.

Now that I think about it, Frank Chance was extremely similar in credentials as Keith Hernandez - except that Hernandez had a few extra years as a regular. Similar to Chance, Hernandez was the best everyday player on a really good team, a terrific defensive firstbaseman who was very strong offensively, but never really had an argument to be among the best two or three players in baseball. The Cubs were a little better than the Mets, but they are very close, other than the length of their careers (which wasn't that long for Hernandez, either).

Chance was, and I don't think it's even particularly close.

The Cubs were built around defense; they were much better defensively relative to the league than they were offensively. Johnny Evers was the most important player on that defense. When Evers left the Cubs in 1914, their defense fell apart - and the team that he went to (Boston) suddenly went from near the bottom of the pack to near the top of the pack in defense.

-- MWE

The 1990s Yankee dynasty didn't have any inner circle types. Rivera may be the best closer evar but he's still a closer.

A *very preliminary* finding on relief pitchers which is *subject to revision at a moment's notice*: to adjust for chaining for relievers in a modern bullpen setup, multiply their leverage by .613 and add .065. E.g. a leverage of 2.0 becomes 1.3, and 0.5 becomes 0.4. I got this simply by creating a standard sample bullpen, replacing each pitcher and sliding everyone else up a role, and seeing the effect on the bullpen's leveraged winning percentage (the effective leverage), and then plotting that against the actual leverage and taking the trend line. I used a reliever replacement winning percentage of .470. I can send the data to anyone who is interested.

Evers was the most important defensive player? I think that Joe Tinker would disagree with that. Tinker left after 1912 and the loss of both Tinker and Evers certainly would add up by 1914.

and the team that [Evers] went to (Boston) suddenly went from near the bottom of the pack to near the top of the pack in defense.

That was the 2nd full season for Boston's young SS, Rabbit Maranville.
His RF going up 14% may also have had something to do with that improvement. ;-)
Then again Evers may have been a heck of a fielding coach too.

His career as a regular player was very, very short, and he was never among the best two or three players in the game
Strongly disagree. He was the 2nd or 3rd best player (non-P) in the game for 1902-1906 (Lajoie has an argument). And the only guy clearly ahead of him was merely a fellow named Honus Wagner.

No way.

Most position player WARP2, 1902-06

1. Honus Wagner, 49.8
2. Nap Lajoie, 40.6
3. Bill Bradley, 32.4
4. Elmer Flick, 30.3
5. Frank Chance, 26.9
6. Bobby Wallace, 26.8
7. Roy Thomas, 25.7
8. Cy Seymour, 25.1
9. Sam Crawford, 24.0
10. George Davis, 23.8

WAY
win shares says WARP2 poops its pants

oops, typo- I meant 1903-07. Sorry. Should be obvious upon inspection that 1902 is not a great year for Chance since he had only 240 ABs.

Most position player WARP2, 1903-07:

1. Honus Wagner, 51.9
2. Nap Lajoie, 40.6
3. Elmer Flick, 34.0
4. Frank Chance, 28.8
5. Cy Seymour, 27.9
6. Bill Bradley, 27.2
7. Bobby Wallace, 25.9
8. Roy Thomas, 24.8
9. Art Devlin, 24.0
10. Roger Bresnahan, 23.9
10. Sam Crawford, 23.9

Moreover, Win Shares sees Chance well behind those three as well: Wagner 214, Flick 152, Lajoie 151, Chance 143. Sorry, Tom, but the numbers just don't back you up on this one.

If Chance, why not Cy Seymour? Seymour's '03-'07 was 97% as good as Chance's (both according to my WARP2 and Win Shares), and he has 5-6 quality years outside of that stretch (some as a pitcher), while Chance has 2-3.

Bobby Wallace had some nice ERA+ as well before making the switchover as well. I suppose that's another advantage he has on Lundy.

You still need pitchers. Saberhagen had slightly less value than Stieb. John had slightly more than Pud Galvin. Reuschel and Tiant were quite close to Jim Bunning. Depending on peak/prime/career weightings, you've got some good choices there. It would be a shame to choose position players when they are already significantly over-represented (much as I like Reggie Smith).

The interesting pitcher for me is Dick Redding. A fourth of the electors think he's a slam dunk case, a fourth think he's in the top 20, the other half think he's not good enough to make the ballot. It doesn't seem to be a peak/career split either.

Ditto on the pitchers though, any of those guys would be worth inducting. The argument with the least amount of water is John to Pud Galvin. Galvin is not a particularly strong candidate outside of his contribution for his era. I would have no problem with John getting elected but he's hanging around my in-out line and I'm more pitcher heavy than anyone.

I have Tiant/Bridges/Reuschel/Saberhagen at about the 75th percentile among HoM pitchers.

The split on Redding is that it's just so hard to get a handle on what exactly he did or didn't do. Basically, if you regress to the mean when confronted with uncertainty (as I do), the lack of information makes it difficult to ballot him. If you just put complete credence in the scant data we do have, and you're a peak pitcher voter, he can get very high up your ballot.

Redding did have a monstrous peak. He is held down by his lower-quality back end, and I have wavered on him myself. Reading Shades of Glory recently, I was inspired to move him back onto the lower end of the ballot. When doing MLE's, we know of two possible problems. First, peaks are flattened. Second, tails are sometimes longer than they would be in a MLB setting. I believe that Redding had a comparable peak to Dizzy Dean and Bret Saberhagen and more career. Some of that career, though probably would not have occurred in the majors, but enough to give him the bulk to edge the two aforementioned. And he also benefits from WWI credit, which occurred in his prime.

TomH
Of the other dynasties, every team had its best player so obviously "in" it aint even questionable, right?

I suppose that is Frankie Frisch for New York 1921-24 and Kid Nichols for Boston 1891-98?

There was plenty of resistance to Jennings for at least one reason similar to the case against Chance, but Jennings concentrated his value in four full seasons while Chance played close to five equivalent seasons during his six-year prime. By the way, neither Jennings nor Chance matched his best seasons with the team's best seasons, if we trim the player's best to fewer than six.

What if Steib, Galvin and Bunning are all mistakes? What then?

Or, even if not, the "if/then" formula is not very helpful. The question is, "Who are the best backlog pitchers?" Our consensus says Saberhagen, Redding and Walters, none of whom have any commonalities to simplify the analysis. Well, all are peakers but in this case that doesn't make their records comparable.

Which one would be a mistake? I've had all but Sabes on my ballot, but of course there's more backloggers out there now. I've got PHoMers down in the 30s off my ballot, Saberhagen is ahead of some of them. None of them was on my ballot last year. if I could "throw" the election by putting one of them on my ballot, I don't even know which one it would be. They're all close.

But Newcombe and Dean are quite obviously better candidates. If you want to play the "if/then" game--if Ferrell, then Dean. If Pierce, then absolutely positively definitely Newk.

Limited to four-season peak, Jennings matches three of the Orioles four. The Cubs were very good 1903-1905, but not yet at their best. Chance matches only one or two of the Cubs best six. I'll give him two, recognizing 1905-1910. (For more than half the seasons, after Chance replaced Selee the 1905 team was a bigger winner than the 1911. The Matty/McGraw Giants dominated when they won, including 1905 and 1911, sandwiching the Cubs five best seasons.)

"What if Stieb, Galvin and Bunning are all mistakes?"

Then we're looking at about 20% pitching which is _way_ too low.

"If you want to play the "if/then" game"

If Lemon, then Walters? Heck, if Lemon then about 50 pitchers. How in the world did we elect Lemon? It pre-dates me so I can honestly claim ignorance. I started about 2 (real-time) years ago.

Yes, DL, Bridges belongs in that same discussion.

I must admit that I don't quite get the lack of support for John among career voters, assuming that Tango's .370 figure for starting pitcher replacement value is accurate. Does everyone remember to account for his glove? Post-season performance? I see career voters with Saberhagen ahead of John, and that just doesn't make sense to me.

Oh come on--Ferrell has oodles of career and hitting on Dean. Newcombe just depends on your credits, of course (which I obviously give him).

To be clear, Tango's figure is .410 for relievers moved to the rotation, and then you have to factor in the cost of replacing the reliever with a replacement reliever, so it comes out to around .400. I've never seen .370--Tango says replacement pitchers are .380 when they pitch as starters, but that the optimal approach is to promote a reliever, since doing so adds about .020 of winning percentage.

There is certainly no consensus on counting postseason performance. Whatever effect John's glove had is obviously registered in his ERA+. There is pretty strong agreement on the need to adjust for era IP norms, although there are a million different ways to do it. John's career IP do not seem quite so impressive when you factor in that HoM nonentities like Kaat and Koosman were able to last for nearly 4,000 innings in those days.

Agree that Lemon was not our finest moment.

Dizzypaco
Dominant teams in baseball history tend to have players that have long careers, and/or who who were probably among the top two or three players in baseball at their peak. the 06-08 Cubs were unique among dominant teams, in that they had neither. . . . [Chance] was very good for a few years, and those few years happened to coincide with the best few years of a whole bunch of other Cubs.

I'm with Tom that Frank Chance was the best player on the Cubs at his peak. But he was already past his peak in 1908 and I insist (against Tom and Dizzy?) that we include 09-10 among the dynasty seasons for the team. We argue about how much of 1903-13 to include but it starts with 1906-10.

(The Giants have the same 11 seasons, 1903-13. For the Pirates we argue about how much of 1900-12.)

Whoa! It's that rare experience, others are visiting the Hall of Merit at the same time. It means I shouldn't be here.
:-(

Paul Wendt, a team always has to pay 25 salaries no matter who its players are. Retirement doesn't save a team money, it just means the team will be paying someone else...

Yes a team pays at least 25 salaries. When Player A gives only 81 games one year and only 81 games next year, the team pays more than 25 salaries.

Related, how many who play in the majors at FAT level earn only the mlb minimum salary?
On the margin and in expectation, probably the joint focus of Keith Woolner's research, it may be one team can replace that play at cost equal to the minimum salary.

Mike Green
I have Tiant/Bridges/Reuschel/Saberhagen at about the 75th percentile among HoM pitchers.

25th percentile?

Evers was the most important defensive player? I think that Joe Tinker would disagree with that.

The Cubs were third in the league in DPs per baserunner in 1912, and fourth in 1913. Cincinnati, the team to which Tinker moved to, was 6th overall in 1912, and 6th as well in 1913.

In 1914, the Cubs and Braves swapped second basemen. Boston, which had been last in the league in DPs per baserunner in 1913, was first in 1914. The Cubs, who had been 4th in 1913, dropped to 7th in 1914. Cincinnati, replacing Tinker at SS with Buck Herzog and keeping the young Heinie Groh at 2B, moved from 6th to 3rd overall.

There's a common (mis)perception that because 2B was primarily a hitter's position in the deadball era, it wasn't as importantly defensively. Nothing could be farther from the truth. Teams tried to place good all-around players at 2B wherever possible - they were expected both to hit and to play defense.

-- MWE

When doing MLE's, we know of two possible problems. First, peaks are flattened.

The pitching MLEs that I have done are not regressed. Peaks in them are, if anything, accentuated. There is no justification for inflating Redding's peak above his MLEs, unless you think the MLEs are flat out wrong.

Much of the early formative discussion of the HoM can be read here:

Since we were earlier talking about the early days. If you read the thread, I'm there too. Unfortunately, I left before the Constitutional drafting because I was tired of waiting. When I next checked in, it was the Crawford/Plank years...then I started voted again around 1930. Just in time for Sheckard and Caruthers and Pearce and them guys.

Wish I could say I'd been a regular all along, but I was there in spirit!!!!

To answer TomH's (I think?) request. I think it would be very helpful to have the MLEs posted in one place for all NgL playres that have been done so far.

However, I think a better venue would be a separate website. Chris, Brent, and I once had preliminary discussions about doing this, but we haven't talked of it in quite some time.

I think the biggest thing that would have to happen would be for a clutch of us to sit down and figure out a set of best practices around MLEs. As in, given what data we have, what's the very best way to do this...what QofP factors should we use...what research needs to be done yet.... Then just go away and do it and post the results on a convenient site.

So in addition to summarizing our findings in the Centennial Report, I see this as another way the HOM can serve the greater baseball community and help raise awareness of bygone players.

Speaking of the Centennial Report.

Would anyone interested in working on the HOM Centennial Report, please email me at

eric[dot]chalek[at]heinemann[dot]com

At this point, I've got a skeleton outline of how it might lay out, drafts of small sections of it, and I'd love to get a small group of folks together to do the writing.

Please let me know by email if you are interested. Eventually, once we've got a few of its sections ready, I think I'll ask Murph to open a thread where the group can review our work for accuracy.

"25th percentile?"

Yeah, that's what I meant.

Any/all calculations for Negro League players (or minor league translations) should be archived on the Yahoo site at a minimum.

It is strange not to account for post-season performance when one is comparing Saberhagen with John, for instance. Similar opportunity, very different results. At a minimum, one ought to simply add in the post-season numbers to the regular season ones.

Through age 38, John's career numbers are right in the Tiant/Bunning range. They are much, much better afterwards. 1200 innings of 94+ ERA is very valuable. Jon Garland should be so lucky!

watch as once again, Hanrahan inserts his foot into his mouth and begins to speak without actually checking what he's saying. Be amazed!

OK, third try: This time I'll look at my data before mis-remembering. Yes, I do have a spreadsheet that shows that Chance was the 2nd or 3rd best player in MLB. But it is for the years 1903 to 1906, as opposed to the previous set of years I had thrown out from poor recall.

player . WS .. AB
Wagner. 170 2066
Chance. 120 1758
Lajoie... 119 1889
Seymour117 2246
Flick...... 115 2326
Crawford105 2250

Is Win Shares perfect? No. Actually it underrates guys who missed time (i.e., Chance). It underrates guys who played on defenses so good that they pegged the meter for team defense (i.e., Chance). Even so, if you DO include 1907, only Lajoie comes out better thean the Peerless Leader; Flick has a few more WS, but in many many more ABs.

Now, does this prove anything? Only my original point; Chance was the best player in the early half of the Cubs dynasty (TF Brown later on), and a truly dominant player in MLB for a brief period. I agree that this does not auto-qualify anyone to be elected. I merely say that for a group of voters diverse as we who give support to both Dizzy Dean and Lou Brock, it is strange that fewer than 5% of us see Chance as ballot-worthy.

To answer TomH's (I think?) request. I think it would be very helpful to have the MLEs posted in one place for all NgL playres that have been done so far.

However, I think a better venue would be a separate website. Chris, Brent, and I once had preliminary discussions about doing this, but we haven't talked of it in quite some time.

I now have a baseball website, and would LOVE to host this!?

>There's a common (mis)perception that because 2B was primarily a hitter's position in the deadball era, it wasn't as importantly defensively. Nothing could be farther from the truth.

Exactly the argument for Larry Doyle. John McGraw, who many say could manage a bit, saw fit to keep putting him out there year after year after year. Like G. Kelly a generation later, he coulda moved him to 1B--e.g. when Fred Tenney went bye-bye after 1909--but didn't. Of course McGraw found a pretty good 1B but no all-star in Fred Merkle. From here it looks as if the Giants might have been better off putting Fletcher or Devlin or Buck Herzog at 2B, if Doyle was so bad. Then when Doyle went to the Cubs in 1916, they put him at 2B and when they decided to move Vic Saier off of 1B they went out and got...Merkle rather than moving Doyle over. And when McGraw brought Doyle back in 1918, where did he put him? Don't even ask. And when 1B Walter Holke went off to war, did he move Doyle over? Nope, he went out and got Hal Chase to play 1B. By the time Chase got blackballed, of course, McGraw had Frisch and G.Kelly teed up...but wait, the 33 year old Doyle played another year in 1920 at 2B. Kelly played 1B and Frisch played 3B--either one of whom would play 2B at one time or another during his career, but not in 1920.

it has been said that 1) Doyle sucked and 2) played a ton of 2B and therefore 3) 2B couldn't really matter as a defensive position. I have always believed what Mike said, 1) 2B couldn't be THAT unimportant and 2) Doyle played there for 15 years so 3) Doyle couldn't just be a glorified RFer.

Of course, even those who've made the former argument now say that Johnny Evers was more important as a glove than Joe Tinker. How is this possible? He was a 2B. You can't have it both ways.

Yes, Doyle was a C+ and Evers an A-. I don't dispute that. But I dispute that a C+ 2B had no defensive value. I don't think John McGraw put guys out there who had no defensive value.

>1200 innings of 94+ ERA is very valuable.

Very valuable? Or, valuable?

HoMer? Or HoVG?

Very valuable. Roughly $60 million worth in today's terms (5 years at $10-$11 and 2 part-years at lesser amounts). Obviously, John is not going to appeal to peak voters.

I now have a baseball website, and would LOVE to host this!?

Which would solve part of the problem!

Mike Green--you either count postseason performance or you don't. I don't, because those opportunities are not granted equally to all players.

I don't know what you mean by a "Tiant/Bunning range"--one is a HoMer, one is a backlogger. That would be like me referring to a "Reese/Rizzuto range" at shortstop--they were contemporaries at the same position, and I support Rizzuto's election, but Pee Wee Reese he was not. I actually have John and Saberhagen as pretty close on career value, with Saberhagen obviously blowing John away on peak and prime. And that's before factoring in that 60s/70s pitchers lasted far longer than 80s pitchers did. I use a .400 winning percentage for replacement starters per Tango's findings, and adjust for era IP norms. Obviously if you use a lower replacement level or don't account for era IP norms, your mileage will vary. But if that's the case, where's your vote for Bobby Mathews? :)

TomH, I would think any voter sympathetic to Chance's profile--high rate, low playing time, short career--would be lining up behind McGraw first, who also played on a dynasty.

sunnyday--I don't know how many times I need to repeat this, but intrinsic positional value and fielding quality are not the same thing, and there is a negligible relationship between them. Defensive value has two components. The first is, how much value did this player create for his team by providing the offense he did at the given position? For that, we need to compare his hitting to his positional peers (average, replacement, whatever). This is what determines intrinsic positional value. The second is, how good was this player's fielding for his position? This is where we look at something like BP FRAA or DRA.

The players with the most defensive value are those who play positions that are associated with weak hitting *and* play them well. This is the case above all with slick-fielding shortstops, everyone from Bill Dahlen to Adam Everett. These players generally have around 2.5 wins of intrinsic positional value (because SS hit about 2.5 wins worse per season than 1B) and another 2 wins of fielding value (because they field their position at 2 wins better than the average SS), so 4.5 wins total.

Players with middling defensive value can go either way--they are either premium gloves at positions with little-to-some intrinsic value, or they are average-to-mediocre fielders at positions with a lot of defensive value. The former is the case with guys like Roberto Clemente or Keith Hernandez; the latter with guys like Toby Harrah or Derek Jeter. These players have about 2 wins of defensive value, coming from the fielding quality side in the first group and the intrinsic positional side in the second. Significantly for Doyle, his contemporaries Eddie Collins and Nap Lajoie fit in this category--although 2B was a hitting position in those days, they still fielded it much better than their peers did.

Players with negligible or negative defensive value are those who are liabilities at positions that are associated with high levels of offense. These players get no credit for playing a weak-hitting position, *and* hurt their teams by fielding poorly. This is the case of Doyle, or of Frank Howard, or of Mo Vaughn. These guys can get as low as about -1.5 wins in total defensive value.

The key point is that intrinsic positional value and fielding quality have nothing to do with one another. In your terms, the difference between an A+ SS and a D- SS is the same as the difference between an A+ LF and a D- LF. In total--summing positional and fielding value--an A+ LF is probably worth about as much as a C/C+ SS, and a D- SS is probably worth about as much as a B LF.

This is why I think Doyle has no place on a HoM ballot. He played a position where freely available talent provided a high level of offense, and he played it quite poorly. The guys we have elected in this class are Harmon Killebrew, Willie McCovey, and Ralph Kiner, and Doyle's hitting is nowhere near their level.

I don't really understand that position on post-season performance. There are all kinds of reasons why players have differing opportunities outside of their control. Pitchers have rotten catchers who cannot frame a pitch if their life depended on it and they lose opportunities for strikeouts. Position players are benched for obviously lesser players, sometimes at the peak of their careers, just because their manager is biased for some reason or another. Anyways, if you're going to compare Schilling and Brown say, to throw out their significant post-season experience on the basis that some other pitcher didn't get the chance to throw 80+ post-season innings, I just don't get. Besides, it's an opportunity that cuts both ways.

The major difference between Tiant and Bunning through age 38 is that Bunning threw 400 more innings. John splits the difference, with almost exactly the same level of effectiveness.

You're not just comparing two players in a vacuum. You're comparing the entire backlog, which has vastly differing amounts of postseason play, to each other. What about the guy with a better regular season record who never threw a postseason inning? Is he better or worse than either Schilling or Brown? If you include the playoff innings in your calculation, you're not comparing apples to apples.

I'd have to check my numbers to see if I agree with that statement on Tiant and Bunning, but just taking it at face value, I'd say two extra years of All-Star performance is a pretty compelling reason to have one guy make the HoM and the other kept out.

Flick and Lajoie are still better, my friend.

Lajoie vs. Chance vs. Flick, no standard deviation adjustment, 1903-06:

Nap


Frank



Elmer



Man, Flick had a TON of PA in 1906. How'd he play in 157 games?

How'd he play in 157 games?

Played every game.

Cleveland apparently had 4 tie games called due to darkness, weather, whatever.

Dan: what do you have for George Stone, 1905-1908, and in particular, Stone, 1906? I never for a moment took him seriously as a candidate, as his entire career was over in an eyeblink, but 1906 looks like quite a year.

Don't know why WS and your WAR sees things so differently. Might be the Cubs' clutchiness and giving Chance partial credit.

Wow, 4 tie games in one year for Flick. That brings to the front a question that could affect, let's say, theoretical MVP voting; MLB counts tie stats, yet this gives some those players more opportunites (PA) that really don't "count" and in a snes an unfair advantage, just as if their team had played a longer schledule. I know, pretty small thing. Another victory for the Win Shares-based theory of allocating value by team wins tho! :)

You think that's a lot, look at Cleveland in 1910. Lajoie played 159 games that year, and still missed 2.

Stone's '06 is a legendary fluke season. That said, he was quite good in the surrounding years before getting malaria and then ruining his ankle. No idea why he started so late.

No standard deviation adjustment



Like all players at left-spectrum positions, I will have Lajoie higher than WS does because I give him more credit for playing 2B and because I use a much higher standard deviation for fielding quality than WS does. Moreover, the '06 Indians were a gigantic Pythagorean underperformer, and one of the all-time great defensive teams; I am sure they ran up against the fielding WS cap. Flick was also on the '06 Indians. Chance, meanwhile, presumably benefited in WS by playing for such a winning team.

I don't really understand that position on post-season performance.

I'm on the record previous to DanR on this one, and I agree with him. Usually it's TomH and I fighting about, though! ; )

To give an extreme example of the potential distoritions, let's look at Derek Jeter. Were we attempting to compare him to other SS for HOM evaluation, we often begin by finding players of comparable career length and performance.

I've heard it bandied about that October PAs are worth something like 3 to 8 times what regular season PAs are worth. Let's call it three. Jeter has 562 October PAs. Multiply times three = 1686 PAs. Add to his 8431 career 162-addjusted PAs, that's 10117. How does that level of opportunity compare to other HOMers (who presumably have comparable performance levels)?



So taking post-season into account, Jeter gains about 20% more PA, and he vaults four places up the PA ladder above one HOM SS with a career currently as long and three others (assuming we can all agree Larkin will be a HOMer) with longer careers. Given that what we're talking about here is opportunity, his opportunity alone is pretty hefty. The next biggest opportunity-getter is Reese, whose 573 3OCT PAs bump him up two slots. Setting aside the question of war credit for the moment, would those postseason PAs alone make Reese better than Glasscock and Wallace?

You may quibble with my 3OCTPA formulation, but if you don't weight October PA more heavily, then they are vanishingly insignificant, not even 5 or 10% of most players' career regular-season PA. So you have to weight them for them to make any real difference in your evaluation.

I don't believe this is fair to the guys with zero or three October PAs. And fairness is one of our constiutional values. If Vaughan had Jeter's October PAs, he be no better a player than we've already pegged him. He'd simply have more chances in October. Jeter is no better a player than we think he is just because he's batted in a couple hundred post-season games. He's great, and October doesn't change that. Just as Dusty Rhodes wasn't great, and 1954 doesn't change that.

Imagine you are comparing Reese, say, to Jeter. Imagine you think them equal (for this purpose). The post-season had one team from each league, not two or four. Reese's post-season opportunity is restricted by his era. If you must use October performance, then there's got to be some way to adjust for the playoff structure. But does that mean you now have to multiply the PAs by 6 or 9 instead of 3? Now PW Reese has the weighted equivalent of 1700 PA added to his resume! Or do you divide Jeter's PA by three instead of multiplying? Now he's around 185. But that's not really crediting him with all the series he played an important role in.

I guess maybe it's just me, but I think postseason credit is a logistical annoyance that tells you little more about the player than you knew before and isn't fair to all eras and players. Unlike war credit it doesn't help you determine the shape and length of a player's career, nor give you more insight into his comparability with other players.

But that's just my opinion.

Eric, if you wanted to weight post season PA's, it would make more sense to weight them more heavily in the one playoff series era than for the 3 series era. Maybe by 6 in the olden days (WS only), by 4 in the intermediate playoff era (LCS & WS) and by 2 in the modern age (LDCS, LCS, WS).

You don't have to leverage the post-season stats for them to be significant. It is a significant part of the load on some modern ballplayers. Jeter is 33, and he has played the equivalent of almost a full season over his career in the post-season. Or perhaps more to the point, Jorge Posada has played 96 games in the post-season, and many people believe that catcher performance varies more with total load rather than with age. If you don't take into account those 96 games, you are not going to get a fair appraisal of his career when it is done, IMO.

and one of the all-time great defensive teams; I am sure they ran up against the fielding WS cap.

Just about all of the good deadball era teams ran into the fielding WS cap.
(As did most good teams before then also.) The players we care about tend to be concentrated on the good teams in this era, and most FWS are capped, and PWS are correspondingly inflated.

I have a new and serious question. When we talk about distributions of talent, replacement levels, freely available talent, etc., we are usually talking about players. But, I wonder, is it possible that the similar effects apply to management? I think there is strong anecdotal evidence that player personnel decisions get loonier in periods after expansion. Is it possible that there aren't enough good GM's to go around and that some of the folks who get to run ballclubs in the period after expansion are not up to snuff? If this hypothesis holds water, then there are lots of consequences for player evaluation. For example, if part of the reason Dave Concepcion towers over the bottom rung of 1970s SS's is that expansion gave ballclubs to a bunch of incompetent old hacks who put the wrong players (no-hit glove men) on the field, then we might view his dominance somewhat differently. Any thoughts?

It makes sense that the expansion clubs might get replacement level front office personnel. Or if they pay to get experienced help, that leaves someone else with the replacement level GM. But it wouldn't explain a majority of the GM's suddenly deciding that no-hit glove men were where it was at. Must have been something in the water. ;-)

I have a new and serious question. When we talk about distributions of talent, replacement levels, freely available talent, etc., we are usually talking about players. But, I wonder, is it possible that the similar effects apply to management? I think there is strong anecdotal evidence that player personnel decisions get loonier in periods after expansion. Is it possible that there aren't enough good GM's to go around and that some of the folks who get to run ballclubs in the period after expansion are not up to snuff? ... Any thoughts?

That is a great ####### idea.

I dunno, I think the supply of competent GM's is infinitely greater than the supply of competent MLB players...it could definitely be that teams are slow to react to the changes brought about by expansion, leading to suboptimal decisions, but I would think that would be equally true of all clubs, and not concentrated in the newly added GM's.

Flick's 0.4 BRWAA for 1906 are probably high, because the fluke combination of his playing CF that year and his super-high PA total is fooling my non-SB baserunning estimator into thinking he was a blindingly fast leadoff man and giving him +2 runs. Then again, he led the league in triples and was third in steals, so maybe he really was fast.

Man, Flick had a TON of PA in 1906. How'd he play in 157 games?

DanR,
Jimmy Barrett set the season record playing with one team, Detroit AL 1904. Barrett played in all 162 team games. There were 10 ties but only 8 replays: 154 + 8 = 162.

Brooks Robinson broke that record in 1961, the first 162-game schedule, and then Maury Wills in 1962, the only 3-game playoff following a 162-game schedule. Robinson and the 1961 Orioles played one tie game, Wills and the 1962 Dodgers none.

When did extra innings become standard practice?

Extra innings were played, depending on how long the sunlight lasted.

The difference between then and now is there was no concept of a "suspended game" where teams would pick up the next day after 18 innings had been played. It was just declared a tie and replayed as if it had been a rainout, but all stats counted.

Yes. There 11-inning ties, wins and losses. And there were 7-inning ties, and wins and losses, probably more in some cities than others by location in time zone, latitude, and typical start time. Probably in all cities there were more shortened game in doubleheaders and in September-October (the season with shortest daylight hours and most doubleheaders).

Yes. The rainouts/ties piled up and September was often loaded with double-headers from them. For an example that I happen to have on hand, both Boston and NY (AL) played 14 double-headers each between 9/01 and 10/10 (end of season), including 5 HTH (with two tie games) in a very tight two team pennant race with a dozen lead changes over that span. The season ended with what was essentially a best-of-5 playoff series played Fri-Mon, with no games on Sunday; one game in NY, DH in Boston, off day, DH in NY. Home team won 4 of 5, with Boston breaking service by winning game 4 and the pennant on Chesbro's famous WP in the top of the ninth.

Forgot to mention that this was 1904 (if you didn't figure it out already).