Yogi wants us to make his HoM induction a necessary day. ;-)

Apparently nobody posts messages about Yogi anymore. The thread is too crowded.

He would normally a candidate for unanimous induction, but he's got some stiff competition and will almost certainly finish 2nd.

My own personal favorite Yogi-ism:

Some asks him, "What time is it, Yogi?" and Yogi replies, "You mean now?"

My all-time favorite Yogi-ism is when he was reading a comic book, while Bobby Brown was reading Gray's Anatomy. Yogi finishes his comic and turns to the future doctor and says "How did yours come out?"

One of the things Bill James once wrote was something to the effect that the legend of Yogi, funny-looking guy who said funny things, carried the danger of eclipsing the memory of Lawrence Peter Berra, one helluva baseball player.

And how much of that legend is really as much about Garagiola as the storyteller as it is about Yogi, the subject of the stories?

If you've never read "I didn't say all of the things I said", or whatever the title exactly is, go to your library tomorrow and borrow it. :)

legend of Yogi, funny-looking guy who said funny things, carried the danger of eclipsing the memory of Lawrence Peter Berra, one helluva baseball player.

Very true. The stories are indeed great and with his first ballot induction a near metaphysical certitude, there is perhaps no need to do more than just re-tell those amusing tales. :-)

So we don't fall into only doing that, we could turn this into a discussion of where Yogi ranks against the all-time great catchers. Gibson is still #1, but Yogi versus Cochrane-Hartnett-Dickey-Campanella is close enough where Yogi's greatness will come out in full detail. (And without researching fully, my first inclination is to rank him above all those guys).

For one thing, after the Raschi/Reynolds/Lopat were gone, the Yankee pitching staff was in constant flux. There was always Ford, but after that it was a mix of guys like Turley, Kucks, Sturdivant and Larsen moving in and out of the rotation -- and yet they always had solid team ERA+. A lot of credit goes to Stengel & the coaching staffs, sure, but some credit goes to Yogi as well. Now, we don't usually quantify stuff like that here and I'm not saying we should, but it shows a level of patience and intelligence not evident in those wonderful stories.

we could turn this into a discussion of where Yogi ranks against the all-time great catchers.

I have Berra #1 among all major league catchers, with Piazza #2 (so far).

I don't slot a guy into my all-time rankings until he is retired. So:

1. Gibson
2. Yogi
3. Bench
4. Campanella (including NeL, I have him as the #6 ML catcher)
5. Carter
6. Fisk
7. Hartnett
8. Dickey
9. Cochrane
10. Ewing

Nobody else is even close to these 10, though Piazza will move in there somewhere and maybe Pudge.

Yogi versus Cochrane-Hartnett-Dickey-Campanella is close enough where Yogi's greatness will come out in full detail. (And without researching fully, my first inclination is to rank him above all those guys).

I certainly have Berra quite a ways ahead of this group. Josh is in the top 10 all-time so far, Yogi is around 20-25, and Dickey/Hartnett are 50ish.

My favorite Yogi story. (Told as close as I can recollect to having heard it.)

Yogi's on a road trip, and he calls home to the missus. He asks her "what did you do last night?"

She answers, "I saw Dr. Zhivago."

He responds, "What's wrong with you now?"

Bill James ranked Bench ahead of him, which seems right; if you look at the numbers, Yogi's very little ahead of the Schnozz (though Yogi was a better fielding catcher, "calling pitchers" may be closer -- the '39-40 Reds were pretty good.)

Yogi's without question the most fun guy among the close catchers, and probably my favorite Yankee (NOT difficult!)

and maybe Pudge.

Huh? If Pudge, the Irod version, retired today wouldn't he easily make a top 10 all-time? I'd have him in the top 5.

In fact, I'll go out on a relatively stable limb and predict that when Pudge retires he will be considered to be the #1 or 2 ML catcher of all-time.

In fact, I'll go out on a relatively stable limb and predict that when Pudge retires he will be considered to be the #1 or 2 ML catcher of all-time.

Pudge? As in Rodriguez? No way. Guy doesn't hit enough, and his peak is pretty middling for an all-time great. I've got him outside the top ten. Gibson, Hartnett, Berra, Piazza, Bench, Campy, Cochrane, Fisk, there's eight guys higher on the totem pole than Pudge2 without even thinking about it. Carter's better too, there's nine. So at best Pudge is #10.

Right now he's fighting with Trouppe, Torre (I've got him at catcher, others may disagree), Simmons, Freehan, Mackey, Brensnahan. I-Rod may emerge from that group, but right now he's "just another guy in the 10-17 range."

Pudge? As in Rodriguez? No way.

Not even close, IMO.

The WARP assessment would suggest the following order for the top 4 major-league catchers of all time:

Johnny Bench, Gary Carter, Pudge Rodriguez, Yogi Berra

In judging Pudge's hitting, it should be remembered that he's played his whole career in the DH league, which shaves quite a bit off of a stat like OPS+. WARP2 EQA has Berra a career .287 hitter, Pudge .285, and WARP sees Pudge as the better defender.

I'm not saying that this is the only reliable analysis, but before you say, "No Way," think about the DH issues for offensive metrics and about the quality of defense.

in a league with a DH would Rodriguez's offensive value go down relative to the average, making his bat less impressive?

Even so, catcher defense can't be important enough in today's game that I-Rod's speeding by all those guys. Not when strikeouts are up and bunts are down.

I don't mean to sound close-minded, but all of his defensive value (about 150 runs or 15 wins of it) is wrapped up in his caught-stealing rate, and offensively, he's not very impressive.

Also is that warp1 or warp3?

in a league with a DH would Rodriguez's offensive value go down relative to the average, making his bat less impressive?

Yes.

Even so, catcher defense can't be important enough in today's game that I-Rod's speeding by all those guys. Not when strikeouts are up and bunts are down.

This is relevant in comparisons to Bench and Carter, but not to Berra, who was catching during an era of low base-stealing.

I don't mean to sound close-minded, but all of his defensive value (about 150 runs or 15 wins of it) is wrapped up in his caught-stealing rate, and offensively, he's not very impressive.

What's the evidence that all of his defensive value is in caught stealing? How do his actual offensive numbers compare to other catchers?

Also is that warp1 or warp3?

It's a combination. Because of the DH, it's my belief that WARP1 does not offer an accurate tool for straight-up comparisons between post-1972 AL position players and players from other eras. WARP3 needs to be consulted for DH era hitting value. However, WARP3 stinks for catcher defense, because catchers who played in eras where catcher defense wasn't that important get their defensive value raised to an all-time standard that I don't think they've earned.

Win shares has Berra and IRod as not close at all, not by a mile. IRod has five years in his career with 20 or more win shares (23,26,27,28, and 25). Berra, in consecutive seasons had 21,32,31,29,28,34,24,31,23,21,and 23). Again, not saying win shares is the only way to look at it, but its my personal favorite.

Of all the statistics used on this site, the one with the least precision, IMO, is catcher defense. Every formula I have ever seen leaves off huge parts of what might be considered an important part of their job.

For career based voters, as Chris points out, there are solid metrics that already put Irod in or close to the top 5 all time of MAJOR LEAGUE catchers. He is only 33 and seems to be quite durable. He won't quite get to 3000 hits (if he played well to 40 he would likely make it) but he will likely retire as the all time hit leader for catchers. He'll have more than 300 homeruns (pretty nice for a catcher). Plus his defense is, or at least was for a decade, A+.

Does anyone have a favourite toy calculation for his chance at 3000 hits. I'd bet it is over 10%, maybe quite a bit over. I'd certainly take 9 to 1 odds.

You might be right, Daryn, now that I look at his numbers (he's only 33?!!) It all depends on what he does in the future.

I just did Irod's favourite toy calculation for 3000 hits. If I did it correctly, he has a 29% shot. That probably should be reduced because he is a catcher, but nevertheless.

Three things

1) A reductive view of catcher defense in today's game.
-A lot of catcher assists (non-SB ones that is) are discretionary, throwing guys out on bunts or on dinkers. The catcher waves off the pitcher and takes the throw.
-Catcher putouts have very little meaning. Mostly they are strikeouts or high popups around the plate.
-Plate blocking is important, but it comes up about once a month.
-The variation of pitcher-handling and game-calling a game has been pretty thoroughly debunked by Keith Woolner (and probably others too), and catcher ERA is a poor measure of working with pitchers.
-WP and PB are much rarer today than ever and are extremely dependent on the repertoire of the pitching staff.

There are two things that today's catcher must be good at: gunning down runners and catching a lot of innings despite getting dinged up a bit.

I mean no disrespect to I-Rod or anybody, but the real difference in catchers is in their arms and their footwork. He's the best at both, no doubt. But that's like Mazeroski and the deuce. I-Rod's a better hitter than Maz, but that only gets him into the top 15 in my book.

2. Chris asked for a little show-me about his offense, so here's the top 10 OPS+ seasons of several popular catchers and their career total (with any luck the coding will make a table):



3. Daryn, I'd take your 9-1 bet and probably your money on Pudge getting 3000 hits. 33 year old catchers who don't play in good hitters parks aren't good bets to be racking up the hits at age 37, let alone age 40. ; )

While Berra played in a low stolen base era, I would still have to think that catcher defense would be more important than it has been over Pudge's career. more bunts, more balls in play. Of course there are probably those that don't think that should count against Pudge. It is only luck that he didn't play in the 1980's or dead ball era, where he would have been a monster.

I would still put pudge behind Piazza, whose defense I think is underrated. I know that a lot of pitchers have complained about pudge, more than I have heard about Piazza. And if Piazza had even a halfway decent throwing arm, he would have been much better.

Let's just say that Mike played at the absolute perfect time to maximize his value.

I would think that catcher putouts would be valuable if you subtract the amount of K's. A lot of athletic plays getmade behidn the plate.

I don't know how to deal with the discretionary assists.

Schnozz: 161/153/147/140/139/138/130/130/129/123 Av: 125. Why no love for Schnozz?

The basic line of argument here is, "We can't measure components of catcher's defense other than SB/CS. Therefore other components are either unimportant or don't exist."

I respectfully disagree. I don't believe that a catcher has no impact on the game other than his ability to prevent stolen bases.

The problem is that a lot of it is very difficult to measure. So what happens is that when people try, they fail, and then assume its becuase the effect doesn't exist. People then think it would be okay to put virtually anyone back there, any it wouldn't make any difference. I just don't think that's the case.

Daryn, I'd take your 9-1 bet and probably your money on Pudge getting 3000 hits. 33 year old catchers who don't play in good hitters parks aren't good bets to be racking up the hits at age 37, let alone age 40. ; )

I'lll decline the kind offer but point out that Comerica is not a terrible park for getting hits (as opposed to HRs), particularly for the type of hitter Irod is. He certainly won't get 3000 hits as a catcher.

I have two long term bets in the works at $100 a pop. Both were made in October 2001. Arod will retire with the career HR lead and Ichiro will make the Hall of Fame. They are both interesting bets with no clear leader as of yet, in my view.

jschmeagol,

I agree about I-Rod that he shouldn't be punished for his birth certificate. What I have trouble with is someone who wants to say he's a top-three catcher when:
a) he's the worst hitter of the candidates for the honor by far, including having the lowest batting peak
b) most of the other guys were plus-defenders themselves---and those that aren't could really mash
c) evaluating the value of catcher defense is a dubious task at any level, so boosting a guy upward for it into the realm of hitters he clearly does not belong with is a decision that begs questioning.

Which is to say, we ought to have fun with Freehan and Simmons!

I agree about I-Rod that he shouldn't be punished for his birth certificate. What I have trouble with is someone who wants to say he's a top-three catcher when:
a) he's the worst hitter of the candidates for the honor by far, including having the lowest batting peak
b) most of the other guys were plus-defenders themselves---and those that aren't could really mash
c) evaluating the value of catcher defense is a dubious task at any level, so boosting a guy upward for it into the realm of hitters he clearly does not belong with is a decision that begs questioning.

I agree with all of this, and would add that, if I remember correctly, Texas pitchers performed worse when IRod was catching than when he wasn't.

Someone who wants to say he's a top-three catcher

To be clear, I am not actually Ivan Rodriguez. Rather, I am a proponent of his. :)


Let's revisit the debate in 2013, when we have all of the picture. Either way, Berra and Bench are going to be difficult to pass if you give any credit to peak value. And Piazza might sneak in there too.

33 year old catchers who don't play in good hitters parks aren't good bets to be racking up the hits at age 37, let alone age 40.

Because I don't give up too easily:

He needs 810 hits to get 3000.

34 2006 160
35 2007 150
36 2008 140
37 2009 120
38 2010 120
39 2011 100
40 2012 20

That would do it. Quite easy, particularly if he starts part-time DHing a bit more once he hits 37.

Daryn,

Really interesting bets on Ichiro and A-Rod. I haven't seen favetoy analysis on the HR record, but since A-Rod's over 400 already and just turning 30 (or is it 29), unless he goes Ken Griffey on us, you've got some decent odds.

The Ichiro one is even more interesting to, however. So Ichiro has been here now five years. He's gotten 200 hits every year. He's 31. Let's say he retires at 36 with about 2000 hits and ten-year career with like a .310-320 average. He's going to really stump some of those pretzel-logic BBWAA guys.

Of course, he'll be an easy HOMer since our Japanese League translations should be done by then! ; )

Daryn, one last item about I-Rod. If he doesn't make it to 3000 as a catcher, he doesn't make it to 3000. He doesn't hit enough to DH or play 1B, and presumably he won't have enough legs left for OF or 3B.

Two reasons Daryn's money wouldn't necessarily be lost:

Carlton Fisk had 1170 hits from age 34 on, playing catcher.
Bob Boone had 881 hits from age 34 on, playing catcher.

Two reasons a straight-up win-share comparison between Berra and Irod may not be accurate:

Irod loses 1 batting win share per 100 games played to the DH
Berra undoubtedly benefits somewhat in win shares by virtue of always playing on outstanding teams

I'm not saying that adjusting win shares for these effects would close the gap win shares sees between them, but it would narrow the gap somewhat.

Berra undoubtedly benefits somewhat in win shares by virtue of always playing on outstanding teams

No he wouldn't. Players don't benefit in win shares by playing on good teams.

No he wouldn't. Players don't benefit in win shares by playing on good teams.

Unless their teams exceeded their Pythags, of course

Unless their teams exceeded their Pythags, of course<i>

If a team exceeds their pythags, the players on that team will be more valuable than would be suggested by their cumulative statistics, by definition. The fact that win shares takes this into account is a strength of the system, not a weakness IMO.

Unless their teams exceeded their Pythags, of course

And good teams, more often than not, to exceed their Pythags, so in reality, despite what Bill James claims, and ignoring if it's correct for them to do so, players DO benefit in win shares by playing on good teams.

If a team exceeds their pythags, the players on that team will be more valuable than would be suggested by their cumulative statistics, by definition. The fact that win shares takes this into account is a strength of the system, not a weakness IMO.

Maybe, but you're ignoring that exceeding their pythags could be totally due to "luck", rather than the players on that team being more valuable.

The fact that win shares takes this into account is a strength of the system, not a weakness IMO.

I'm a Win Shares supporter, so I don't really disagree, Diz.

And good teams, more often than not, to exceed their Pythags, so in reality, despite what Bill James claims, and ignoring if it's correct for them to do so, players DO benefit in win shares by playing on good teams.

Is that a statistical fact, Kevin?

Carlton Fisk had 1170 hits from age 34 on, playing catcher.
Bob Boone had 881 hits from age 34 on, playing catcher.

You know what I'm going to say now, and so I'm just picking, but 2 out of the hundreds of catchers who have played since, say, WW2 doesn't seem like much of a reason why Daryn should be full of hope. Especially when I-Rod's coming off his worst season in a decade.

It is an argument that Sunnyday would appreciate: that HOMers are outliers, but it's not predictive, only qualitative.

As an aside I do recollect that Boone was an absolute fitness fanatic (at least for his day), which helped him retain his durability well into his thirties. I think Fisk might have had similar discipline in a Yankee-chopping-wood-in-the-autumn-and-winter kind of way.

I don't know much about I-Rod's training regimen, but I'd bet the contemporary player (or catcher) has a more consistently applied and helpful one than the player of one or two generations ago (when Brian Downing miraculously discovered the weightroom and became The Incredible Hulk).

Is that a statistical fact, Kevin?

I haven't looked at 'over .500' vs. 'under .500' if that's your measure of a "good" team, but I did look at World Series teams in the past, and I was sure they exceeded their pythag more often than not, but just doing a really quick check, I'm not coming up with that result, so I may have to eat my words..

If a team exceeds their pythag by one or two wins, its really not that big a deal on a player by player basis - it would come out to less than a win share per person. The only way it would be noticable is if the team exceeded its projections by several games, which I believe is relatively rare.

But even if they did, someone is responsible for all those wins. If a team wins 100 games, then the players on that team have to be responsible for those 100 wins - you can't say the players are responsible for 90 wins, and pretend the other ten didn't exist. That's the theory behind win shares, and its one I support.

but just doing a really quick check, I'm not coming up with that result, so I may have to eat my words..

I wasn't trying to put you on the spot, Kevin. I really don't know either way.

The "the players must get credit for those wins" argument is specious because the run and win values of a player's individual contributions is context-sensitive because the relationships between hits & runs and runs & wins is not linear. Berra's individual accomplishments get more traction because he is playing with Mantle and Ford. Berra's performance creates more additional wins for the Yankees than it would for the Browns. These effects are quite small in any single season, but they get more noticeable on extremely good and extremely bad teams or if they add up over the course of a career. For most players, these effects even out over the course of their careers, but for a few players (like Berra or Charlie Keller or Bob Johnson or Ralph Kiner), they don't even out, and over the course of a career they may be worth 10-20 win shares above or below what a comparable player would earn on an average team.

I lost a post.

Irod already has 46 games at DH, he'll probably get at least 46 more. Those extra 90 hits might make it so that he has 3000 hits overall, but not as a catcher.

And last year was a bad year because he was going through a divorce. I think he'll bounce back.

Being the one who first mentioned Pudge...I said that "maybe Pudge" would be a contender for the top 10 catchers someday along with Piazza among catchers who are still active.

Let me clarify my personal opinion and it's only my opinion, of course.

The #10 catcher on my list is Buck Ewing. While Piazza could finish much higher on my list, I really meant that Pudge could maybe come in at #10. Ewing is the ONLY catcher on the list that Pudge will ever pass for the reasons detailed in #25. And of course if Piazza goes on the list, then the best Pudge can do is #11.

Having said all of that, I'm not even saying that he rates higher than Ted Simmons or Bill Freehan or Joe Torre, if you consider Torre as a catcher. It's just that no way can Simmons or Freehan or Torre move into my top 10, they're already done and accounted for. Pudge could move into the top 10 or 11. But he could still finish behind Freehan or Simmons or Torre, too.

So I see Pudge as a pretty damn good ballplayer, vastly superior to a lot of catchers who have been honored at one time or another, including Schalk, Ferrell, Lombardi and many others. But maybe and only maybe top 10 or 11. Maybe #15, which ain't bad and in fact ought to be worth a spot in both the HoF and the HoM. Just not top 9.

Isn't there supposed to be an effect on strength of schedule with WS and winning teams.

We came across this with Bob Johnson vs. Joey Medwick, where because Ducky was always on winning teams and Indian Bob always on losing teams their competition was not equal.

I sort of believe this but think it should have more to do with the pithing staff and defenses on one's team than the voerall records. Plus, it seemed only to effect really strong or weak teams like the 30's A's or the Gashouse gang.

j, I think that is wrong.

It has more to do with how the team does vs. pythag.

jschmeagol wrote:

Isn't there supposed to be an effect on strength of schedule with WS and winning teams.

That's correct: it's another small way in which win shares favors players on better teams by not placing their performance in a neutral context, but the small things add up for extreme teams and over careers.

Sunnyday2 wrote:

j, I think that is wrong.

It has more to do with how the team does vs. pythag.

Yes, it is wrong (or inapplicable) in the case of Medwick: his peak boost in win shares is seems to be influenced by the team exceeding pythagorean expectations.

In the case of Johnson, however, his career win shares are suppressed by the tougher competition effect of playing for bad teams all the time, though he does take it on the chin from Pythagoras also in a couple of seasons.

Chris Cobb wrote:

The "the players must get credit for those wins" argument is specious because the run and win values of a player's individual contributions is context-sensitive because the relationships between hits & runs and runs & wins is not linear. Berra's individual accomplishments get more traction because he is playing with Mantle and Ford. Berra's performance creates more additional wins for the Yankees than it would for the Browns. These effects are quite small in any single season, but they get more noticeable on extremely good and extremely bad teams or if they add up over the course of a career. For most players, these effects even out over the course of their careers, but for a few players (like Berra or Charlie Keller or Bob Johnson or Ralph Kiner), they don't even out, and over the course of a career they may be worth 10-20 win shares above or below what a comparable player would earn on an average team.

"Specious" is surely too strong:
- The non-linearity is pretty small in the range of .350-.650 winning percentage in which most teams play.
- A bigger problem, IMO, is the way win shares handles negative values, which does tend to hurt good players on poor teams.
- How to treat the difference between actual wins and Pythagorean wins is unclear. Some of the difference reflects the actions that players and teams take to win individual games at the cost of reducing the expected run differential--such as pursuing one-run strategies or employing relief aces. On the other hand, some of the difference is simple luck in timing. I don't think anyone can say how much is one or the other.

My bigger concern, though, is that the argument that players on good teams have an advantage overlooks a huge effect going the other direction, in terms of reduced playing time. We know that Keller was ready to advance to the majors at least a year earlier (he'd won the Minor League Player of the Year award in 1937), but he was held back because the Yankees had 3 outfielders with OPS+ of 168, 157, and 142. Gil McDougald could have had a long career as a shortstop with most other organizations. Whitey Ford was among the top 5 in IP only four times because Stengel's pitching staff was strong enough to allow him to spot Ford against the top teams. Roy Campanella, Don Newcombe, and Junior Gilliam all spent extra time in the high minors. So maybe it's true that a player on a good team gains 0.5-1.0 WS/season from the quirks of the system, but they may also lose that many in any given season from getting an extra half-dozen rest days (due to a strong bench), and also lose 15 to 40 at either end of their careers as they compete for a spot in the lineup.

Elston Howard is another great example - blocked behind Berra and didn't get a chance to be the number one catcher until he was 31.

but just doing a really quick check, I'm not coming up with that result, so I may have to eat my words..

I wasn't trying to put you on the spot, Kevin. I really don't know either way.

John - I made an assertion, and you asked me to back it up. Nothing wrong with that! Unfortunately, my memory may have been faulty.

I just took ALL MLB teams 1871-2005 over .500, and all 1871-2005 MLB teams over .500, split them, and calculated their Pythag using 1.83 exponent.

The "over .500" group had an actual win% of .571, and a pythag of .566.

The "under .500" group had an actual win% of .426, and a pythag of .432.

I stand corrected....

Brent wrote:

My bigger concern, though, is that the argument that players on good teams have an advantage overlooks a huge effect going the other direction, in terms of reduced playing time. We know that Keller was ready to advance to the majors at least a year earlier (he'd won the Minor League Player of the Year award in 1937), but he was held back because the Yankees had 3 outfielders with OPS+ of 168, 157, and 142. Gil McDougald could have had a long career as a shortstop with most other organizations. Whitey Ford was among the top 5 in IP only four times because Stengel's pitching staff was strong enough to allow him to spot Ford against the top teams. Roy Campanella, Don Newcombe, and Junior Gilliam all spent extra time in the high minors. So maybe it's true that a player on a good team gains 0.5-1.0 WS/season from the quirks of the system, but they may also lose that many in any given season from getting an extra half-dozen rest days (due to a strong bench), and also lose 15 to 40 at either end of their careers as they compete for a spot in the lineup.

I like this argument, if for no other reason it leads to upgrading Bob Johnson even further then.

Since the 1929-32 A's were winning 100+ games a year, Indian Bob can stew down in the minors. Let's overlook the fact he's outhitting a couple of the outfielders Connie Mack has. Bing Miller and Mule Haas are hitting .300 albeit with few walks and little power, but you don't change a winning team. Then bring Johnson up come 1933, quickly dismantle the team, and let him suffer on a set of bad teams to depress those Win Share totals some more. Then at the end of his career cut him loose after 1945 since the Red Sox are about to power to the 1946 pennant. Johnson, unwanted but still productive, heads back to the high minors for part-time duty which easily could have been done in the majors to pad his career totals.

Bob Johnson, wrong place, wrong time in baseball history.

The "the players must get credit for those wins" argument is specious because the run and win values of a player's individual contributions is context-sensitive because the relationships between hits & runs and runs & wins is not linear. Berra's individual accomplishments get more traction because he is playing with Mantle and Ford. Berra's performance creates more additional wins for the Yankees than it would for the Browns. These effects are quite small in any single season, but they get more noticeable on extremely good and extremely bad teams or if they add up over the course of a career. For most players, these effects even out over the course of their careers, but for a few players (like Berra or Charlie Keller or Bob Johnson or Ralph Kiner), they don't even out, and over the course of a career they may be worth 10-20 win shares above or below what a comparable player would earn on an average team.

There is no real evidence to support this view. If a player X creates 100 runs over the course of a season, this has roughly the same value on a good team and a bad team. The only way this might be true is if good teams regularly exceeded their pythag expectations by at least 5 to 10 games every season - which they don't. If there is an advantage, it might ad up to 1-2 win shares over the course of a career.

Besides, isn't it a good thing that a team a player plays on constantly wins? Isn't that the goal of the game? Do we really want to start penalizing players because their teams won too many games?

If a player X creates 100 runs over the course of a season, this has roughly the same value on a good team and a bad team.

If I understand WS correctly, this isn't accurate. I think it would be more accurate to say that it would have the same value if the teams scored the same number of runs. Otherwise, the team scoring more runs would disperse fewer BWS to the 100 RC player than the team scoring fewer runs.

A lot of athletic plays get made behidn the plate.

Not that many any more. Independent putouts (CPO-K) have been trending downward across time, as have non-CS assists for catchers.

-- MWE

There is no real evidence to support this view. If a player X creates 100 runs over the course of a season, this has roughly the same value on a good team and a bad team.

The runs created formula contains a final step that adjusts its result to give the individual player a percentage of his team's actual runs created that is equal to his percentage of the teams' theoretical runs created, so the number the RC system gives us as the player's runs created is already influenced by his team context. By the formula, the individual player actions that produce 100 runs created on one team might have led to more or less runs created on another team.

If anyone wants to see some evidence that strongly suggests that win shares does not reward equal performance with equal win shares regardless of the quality of the player's team, please go to the 1958 Ballot Discussion thread. In posts 72, 87, and 100, I present evidence, based on comparisons of offensive performances that are highly similar when sorted by the players' offensive stats, that shows that win shares rewards players for these seasons disproportionately in ways that correlate strongly with the performance of the player's team.

By making this argument, I'm not saying, "don't use win shares." I use win shares: it's my primary metric. I'm saying, "don't assume win shares is a perfect system or that its claims to be fair to all players (i.e. context neutral) are entirely borne out by its results.

Do we really want to start penalizing players because their teams won too many games?

This is silly, both rhetorically and conceptually. We are "penalizing" players if we reduce win-shares' estimates of their quality only if the win shares' assessment of that quality is accurate. A win shares total is a product of a system, not an inherent value produced by the player. Why should we necessarily consider win shares to be the best measure of a players' merit, just because the win shares earned by players on a team are pro-rated to add up to team wins? That element in the system provides a check on the magnitude of possible errors in the system, but it doesn't guarantee that the win share is the most accurate, meaningful measure of merit possible.

In the HoM project, we want to determine a player's merit. Few players, in most of baseball history, have picked either their team, or their teammates, and so the quality of their teammates should not be considered to be part of their merit. If there's evidence that the win shares system rewards players with the identical individual performance differently based purely on how the rest of the team performed (and I believe that there is clear evidence of this), we should take that evidence into account when making use of that system's results as a measure of merit.

About Irod and 3000 hits, I was thinking last night of a spectacular defensive player who had 2400 hits after his age 33 season and a history of durability. He was a LOCK for 3000 hits, and 3300 hits would not have been an outlandish prediction. Three years later he retired with 2750 hits and a ranking of about 5 to 10 all-time in his position.

Roberto Alomar was a small to medium level better than Irod up to age 33, but they are interesting comps, particularly if Irod collapses by age 37 as some here might predict.

If I understand WS correctly, this isn't accurate. I think it would be more accurate to say that it would have the same value if the teams scored the same number of runs. Otherwise, the team scoring more runs would disperse fewer BWS to the 100 RC player than the team scoring fewer runs.

This is incorrect. The team scoring more runs has more BWS to disperse (assuming the same context of year, park, etc).

I am not sure we can compare Johnson (generally a LFer) to Miller (the A's CFer) here. Wasn't Miller a hell of a CFer as well? Even if Johnson could have handled the position adequately I am not sure the A's would have gained too much. Haas on the other hand...

If anyone wants to see some evidence that strongly suggests that win shares does not reward equal performance with equal win shares regardless of the quality of the player's team, please go to the 1958 Ballot Discussion thread. In posts 72, 87, and 100, I present evidence, based on comparisons of offensive performances that are highly similar when sorted by the players' offensive stats, that shows that win shares rewards players for these seasons disproportionately in ways that correlate strongly with the performance of the player's team.

Chris,

I am willing to keep an open mind on this, so I reviewed the 1958 ballot discussion thread (post 100 is really the relevant one). It requires some checking because there may be some other stuff going on that causing the correlation without it necessarily being a flaw in the system. The Foxx comparison of 1935 and 1938 is questionable, because Foxx was not the same player in 1935 and 1938 - in 1938 he had more plate appearances, far more total bases, 60 more RBI, and 26 more runs created on his own than he did in 1935 (I understand the context was different). OPS+ may not be the best metric to use in general, since its a rate statistic and win shares is not.

Still, its an interesting study, an one that's worth delving into further when I have the time. I'm still not 100% convinced, but I'm willing to say you might be right.

Still, its an interesting study, an one that's worth delving into further when I have the time.

If you have time to delve further, I'd be very interested to see what you find. I readily concede that the study as I have taken is so far is only suggestive, not conclusive, and I haven't had time to do anything more with it myself.

This is incorrect. The team scoring more runs has more BWS to disperse (assuming the same context of year, park, etc).

Maybe I'm wrong, but I still think that's not quite right. WS are at the top level based on how many wins the team compiles. You can score 1000 runs, but if you're a .500 team, you still only have about 120 batting win shares to spread. If you score 500 runs and you're a .600 team, there's ~150 BWS to go around. So in one case you've created 1/10th of your offense with 100 RC and in another 1/5th. That won't result in the same value for those 100 runs.

In addition, it's even more complicated. WS is based on 500 runs, so in the 1000 run scenario if the marginal line is 500 (say) and in the other scenario it's 300, that will put further distance between the two 100-rc-sluggers.

Maybe I'm wrong, but I still think that's not quite right. WS are at the top level based on how many wins the team compiles. You can score 1000 runs, but if you're a .500 team, you still only have about 120 batting win shares to spread. If you score 500 runs and you're a .600 team, there's ~150 BWS to go around. So in one case you've created 1/10th of your offense with 100 RC and in another 1/5th. That won't result in the same value for those 100 runs.

In addition, it's even more complicated. WS is based on 500 runs, so in the 1000 run scenario if the marginal line is 500 (say) and in the other scenario it's 300, that will put further distance between the two 100-rc-sluggers.

This isn't how its done. If a team scores 1000 runs, but is only a .500 team, than most of its win shares will be offensive; there will be something like 150 to 200 batting win shares. If a team scores 500 runs but is a .600 team, then almost all of its win shares will be defensive, and you won't have many batting win shares at all.

I am almost positive that Dizzy is right about this Doc.

Also,

If we are going to adjust the WS for players on extreme teams should we only take pitching and defense into account? I don't think that how good an opposing team's lineup is really affects that players personal performance, whereas playing against a great pitching staff and a great defense does. Now most extreme teams are either good or bad at both, but we may find some non-extreme teams that also distort WS somewhat.

In the words of Anthony Daniels, I have "been known to be wrong . . . from time to time."

; )

"If a team exceeds their pythags, the players on that team will be more valuable than would be suggested by their cumulative statistics, by definition. The fact that win shares takes this into account is a strength of the system, not a weakness IMO."

I like this guy :-)

"And good teams, more often than not, to exceed their Pythags, so in reality, despite what Bill James claims, and ignoring if it's correct for them to do so, players DO benefit in win shares by playing on good teams."

Good players are what make teams good. Also, I thought great teams generally undershoot their pythag record, not overshoot it. That's what Neyer and Epstein say in the Dynasties book, right?

"We came across this with Bob Johnson vs. Joey Medwick, where because Ducky was always on winning teams and Indian Bob always on losing teams their competition was not equal."

I could definitely see this as being an issue. OPS+ corrects for this (the B-R version anyway) because the park factors there take quality of a team's own batting into the pitching park factor and vice-versa.

"I'm saying, "don't assume win shares is a perfect system or that its claims to be fair to all players (i.e. context neutral) are entirely borne out by its results."

I wholeheartedly agree with this.

"OPS+ may not be the best metric to use in general, since its a rate statistic and win shares is not."

OPS+ is not something to use for systemic analysis on individual players. It ignores way too much. It's a good eyeball metric, but it doesn't account for things like baserunning (SB, GIDP), and more importantly, it underweights OBP vs. SLG (especially in high offense eras).

It could be used for comparing groups of players if precision isn't important. But I think for individuals too many of the things it doesn't measure do not even out.

"Maybe I'm wrong, but I still think that's not quite right. WS are at the top level based on how many wins the team compiles. You can score 1000 runs, but if you're a .500 team, you still only have about 120 batting win shares to spread."

Dizzy is right on that. It depends on how good the pitching staff is. If you score 1000 runs, you'll have many more than 120 WS to spread, unless you are playing in a league that averages about 12.3 R/G.

Players don't benefit in win shares by playing on good teams.

I haven't had a chance to read the whole thread, so if Chris Cobb or others have covered this point for me, I apologize for the redundancy. (I pointed this out in a previous election, though I don't remember the year.)

Good teams get more wins because they don't play themselves. Bad teams get more losses for the same reason.

Look at the 1932 AL.
NYY 107 47 .695
Bos 43 111 .279

For common opponents, there are the other 6 teams, 132 games. For the other 22, the Yanks got to fatten their stats against a sub-300 team. Meanwhile the Sox got to play a .700 juggernaut. This is a problem because the Yanks don't have to play the juggernaut and the Sox don't get to play the replacement level team.

We can compensate for this with the following hypothetical construct. Add 22 additional games to each schedule; the Yanks play themselves and go 11-11, similarly for the Sox. If you do that to each schedule, and then reduce back to 154 games, the Yankees lose 11 Win Shares (its benefit from not having to play against the World Champs); the Red Sox gain 13 Win Shares. This works out to about 1 Win Share for each regular (8 everyday players and 4 starters). Not a big deal for any one season; a problem when comparing say Charlie Keller and Indian Bob Johnson; one systematically benefited during his career and the other systematically got shorted.

Note that a .500 team is completely unaffected by this effect and it's correction.

Good teams get more wins because they don't play themselves. Bad teams get more losses for the same reason.

I agree with your theory, but I think you are opening up a much larger can of worms than you realize. Every statistic, not just win shares, is affected by the fact that a team doesn't play itself. A hitter that plays for a good pitching team benefits from the fact that he doesn't have to face his own pitching staff. The 1990's Atlanta hitters statistics were inflated by the fact that they never had to face their own pitching staff.

I have no idea how you would correct for this sort of thing, but I don't think its that simple.

>Look at the 1932 AL.
NYY 107 47 .695
Bos 43 111 .279

The problem with this logic is it presumes that teams are .695 teams and .279 teams intrinsically and that ball games are just sims.

But the .695 team is a .695 team because it beats .279 teams on the field. It IS not a .695 team, it becomes a .695 team, and it earns WS by becoming such. They make the WS the old-fashioned way, they earn them.

But the .695 team is a .695 team because it beats .279 teams on the field. It IS not a .695 team, it becomes a .695 team, and it earns WS by becoming such. They make the WS the old-fashioned way, they earn them.

That's true for the _team_, but any individual good player short of a Barry Bonds/Ted Williams type could be traded from the .279 team to the .695 team (say, for example, Red Ruffing) for an average player, and the .695 team would still be great, and the .279 team would still suck, and win shares would tell you that the player became better, because now he didn't have to play against a bunch of other good players in the league, because they were all his teammates. And the player who went in the other direction would lose a win share or two every year because now he had to play against those old teammates.

Baseball is a team game, but we are studying the records of individual players. It only makes sense to normalize the level of competion within individual seasons, when we can, to take account of the fact that a good player on a bad team faces tougher competition than a good player on a good team.

BBref's OPS+ and ERA+ normalize for this, for goodness sake! All jimd and I and others are saying is that win shares should be normalized the same way.

It's not going to turn good players into scrubs, but it's going to make 20-30 win-share swings when comparing the Bob Johnsons of the world to the Charlie Kellers, players who spent their careers on outstanding teams or on poor teams.

That's true for the _team_, but any individual good player short of a Barry Bonds/Ted Williams type could be traded from the .279 team to the .695 team (say, for example, Red Ruffing) for an average player, and the .695 team would still be great, and the .279 team would still suck, and win shares would tell you that the player became better, because now he didn't have to play against a bunch of other good players in the league, because they were all his teammates. And the player who went in the other direction would lose a win share or two every year because now he had to play against those old teammates.

As I mentioned above, this isn't just true of win shares - its true of every statistic you can name. Its true of WARP3, its true of linear weights, its true of OPS+, its true of hits, its true of home runs, and its true of walks. People are implying that this is a weakness particular to win shares, and its not.

There may be other reasons why win shares overrates players on good teams relative to other statistics, but this isn't it.

As I mentioned above, this isn't just true of win shares - its true of every statistic you can name. Its true of WARP3, its true of linear weights, its true of OPS+, its true of hits, its true of home runs, and its true of walks.

No, it isn't. WARP3 and OPS+ on bbref, at least, adjust for team competition context. That's why the batting and pitching park factors at bbref are not always the same. I don't know linear weights well enough to speak for its adjustments here, but I wouldn't be suprised if it does.

No, it isn't. WARP3 and OPS+ on bbref, at least, adjust for team competition context. That's why the batting and pitching park factors at bbref are not always the same. I don't know linear weights well enough to speak for its adjustments here, but I wouldn't be suprised if it does.

I'm confused. Baseball reference doesn't include WARP3 (that I can see), and I've read the method for determining OPS+ on the site several times, and it does not mention team competition context, as far as I can see. Am I looking in the wrong place?

Never mind - I found it.

Two clarifications:

1) In the phrase "WARP3 and OPS+ on bbref," I meant "on bbref" to apply only to OPS+, since that stat may be available elsewhere, calculated in different ways.

2) The adjustment in OPS+ for team context is handled as part of the calculation of the park factor that is applied within the OPS+ formula. If you go follow the link to the explanation of park adjustments, you'll find that the calcuations include two parts, first the calculation of a raw park factor, and then the application to it of a second calculation to arrive at independent "batter adjustment factors" and a "pitcher adjustment factors."

The page explains the reason that it uses "batter adjustment factor" and not simply the raw park factor as follows: "The batter adjustment factor is composed of two parts, one the park factor and the other the fact that a batter does not have to face his own team's pitchers."

This is the crucial normalization to which I was referring. I don't have time right now to dig corresponding language out of the WARP glossary, but maybe tonight.

As I mentioned, I found it on baseball reference. I spent a while working through the formula, which was difficult, because the page is riddled with errors, where they say one thing, and then you realize they mean something else by the example.

One of the things thats interesting when you go through the exercise is how little impact the adjustment has in all but the most extreme cases. I'm assuming your statement about the 20 to 30 win shares is just a guess. Running through the formula, it appears the effect of this calculation would be much smaller in almost every case. More work is necessary, but we shouldn't assume that this adjustment would have a substantial effect for anyone in particular.

Sorry to have completely hijacked this thread, but...

I did a comparison of two teams with the same number of runs scored, but massively different numbers of runs allowed and wins - the 2001 A's and Rangers. The Rangers scored 890 runs and the A's scored 884 runs, in a slightly better pitchers park, according to baseball reference. The A's gave up 645 runs, while the Rangers gave up 968 runs, leading to a difference of 29 wins between the two teams. If players on good teams were getting too many win shares, it would follow that the A's would be getting more win shares for their offense than the Rangers, by a pretty substantial amount, especially if it is to make a difference in the individual player win shares.

It didn't happen. Oakland was assigned 147.4 offensive win shares, while Texas was assigned 146.2, a difference of 1.2 win shares. Those 1.2 win shares would have to be divided among many players, making a totally insubstantial difference.

Obviously, this is only one example, but if it holds in other cases, it suggests win shares are better than what they are being given credit for.

I'm assuming your statement about the 20 to 30 win shares is just a guess.

No it's not. For a .700 team (or .300 team) in an 8 team league, the effect is about 1 win share per regular. Nobody plays on .700 teams for their entire careers, but many Yankees amassed career records around .600, and Bob Johnson was just above .400. For these players, the effect is about .5 WS per season, or 7-11 for a career, relative to a .500 player, so the overall discrepancy for this effect can be around 20 when comparing Keller-Johnson. Chris has noted other problems with WS, so I'll let him defend the rest of the "20-30" estimate.

>Look at the 1932 AL.
NYY 107 47 .695
Bos 43 111 .279

In Win Shares, the primary inflation comes from the extra Wins. If we gave the Yankees and Sox identical schedules (they play themselves and they split), the Yanks would go 118-58 .670, and the Sox would go 54-122 .307. The Yanks should get 310 Win Shares (67% of 3x154), not 321. The Sox should get 142 WS, not 129.

I'll repeat. It's a small effect on any given season, but it can add up over careers. Babe Ruth plus or minus 10 Win Shares also doesn't matter. For Charlie Keller and Bob Johnson, it does.

it suggests win shares are better than what they are being given credit for

I use Win Shares in my evaluation system. I also use WARP, and I'll confess to weighting the latter more heavily. Win Shares has a lot of good things going for it, but it also has flaws, as does WARP. There's a thread devoted to WS-WARP debates if you are so inclined.

Welcome aboard, diz. This project has been a blast for me. I hope you enjoy it as much.

No it's not. For a .700 team (or .300 team) in an 8 team league, the effect is about 1 win share per regular. Nobody plays on .700 teams for their entire careers, but many Yankees amassed career records around .600, and Bob Johnson was just above .400. For these players, the effect is about .5 WS per season, or 7-11 for a career, relative to a .500 player, so the overall discrepancy for this effect can be around 20 when comparing Keller-Johnson. Chris has noted other problems with WS, so I'll let him defend the rest of the "20-30" estimate.

I don't believe this is true. My early indications of going through the data suggest that for a .600 team, the effect is nowhere near .5 win shares per player. I don't have the exact figures as of yet - I'll write in again when I have more.

I have demonstrated above that the win inflation due to inequal schedules is about 12 WS for a .700 team or about 6 WS for a .600 team. I don't know what else you are investigating.

Also note that the effects in a 16 team league are going to be about the magnitude of those in an 8 team league.

Also note that the effects in a 16 team league are going to be about HALF the magnitude of those in an 8 team league.

I disagree - I don't think you have proved what you think you have proved. There is a relatively easy way to investigate this issue. Take a group of matched pairs of teams, in which the offensive performance is very similar (runs scored, league context, and park effect). Second, ensure that one of each set of pairs has a substantially better pitching and defensive performance. If win shares overestimates the value of players on good teams, then the total offensive win shares of the teams with good pitching should be substantially higher than on those with worse pitching. If the offensive win shares are independent of the pitching and defense, then win shares is not overestimating the effect.

Dizzy,

That is kinda like what I have been suggesting. However I didn't have any concrete details of how to do it and I am also too lazy to do it myself so nothing really got done. I think there will be a small difference still but not that large.

Yeah, I don't disagree that there's theoretical difference. But in real life...

What is suggested above that Bob Johnson might have lost as much as 1 WS per year if he played on a .300 team. Chris says that adds up to 20-30 WS but that seems to me to be a theoretical boundary. In 13 ML seasons at .400, I'd guess it probably had more like 6.5 WS worth of impact. That's a guess. But 20-30 doesn't seem credible to me.

jimd's case is mathematically sound and models precisely the adjustment needed to equalize the level of competition faced by each team in a league. It proves what it sets out to prove.

If you can actually find a .650 team and a .350 with similar runs scored or runs allowed in similar parks in the same season, it would be interesting to see how their win share totals match up, but I don't think you're likely to find many cases. If you can find any, the results would be interesting.

On the 20-30:

I'm talking about a swing between 2 players, one gaining from great teams, the other losing from terrible teams, so a 20-point swing would mean the player on terrible teams loses 10 win shares, while the player on great teamas gains 10. Upon reflection, a 30 win-share swing is highly unlikely, but I think 20 win shares is more of a practical maximum than a theoretical boundary.

With a bit of work, it would be easy to examine specific cases.

Chris, Diz already did that in #82. Okay it wasn't 65-35 but it was 29 wins different, which I would say is a more real-world eg anyway. Yes it's just one eg, but right now the eg's are 1-0 in favor of no significant impact.

I'm not sayin' there's no impact. And maybe Bob Johnson is in fact the one case in 140 years of MLB where a 30 WS swing occurred.

But the burden of proof is on the FOBJ.

Baseball is a team game, but we are studying the records of individual players. It only makes sense to normalize the level of competion within individual seasons, when we can, to take account of the fact that a good player on a bad team faces tougher competition than a good player on a good team.

BBref's OPS+ and ERA+ normalize for this, for goodness sake! All jimd and I and others are saying is that win shares should be normalized the same way.

Things that should be involved in normalization:

1. League
2. Park
3. Platoon Advantage
4. Competition level
5. Home Field Advantage
6. Luck/Other

All 4 taken together will give you "context" or "environment" for a player.

Almost all 'advanced' calculations adjust for league. Most adjust for park.

For some reason, platoon advantage/disadvantage is typically overlooked in 'comprehensive' metrics, although vs. LH and vs. RH stats are certainly looked at by even casual fans.

Palmer's Linear Weights and bbref OPS+, among many, adjust for the competition level to the extent that they attempt to adjust for pitchers not facing their own team's hitters and vice versa. Win Shares does not.

WARP takes the competition level a step farther, and tries to adjust for not just the "not playing your own team" effect, but for the entire league's "skill" level relative to all other leagues. Win Shares does not.

Only SUPER-LWTS and maybe some of the BP metrics, that I know of, adjust for the mix of teams played against beyond the "not playing your own team effect". For MLB thru 1968, teams generally played the same slate of "other" teams, but 1969 thru today the mix can be important, especially with 3 unbalanced divisions and interleague play. It's also important for leagues such as the Negro Leagues, where teams sometimes played very different mixes of opponents.

Beyond even this, I think SUPER-LWTS and some BP measures even adjust for the actual quality of individual pitchers and batters faced.

Almost no methods adjust for home field advantage. Home field historically yeilds about a 10% statistical advantage. For MLB, since teams typically play the same # of home and away games, it's assumed to "even out" in most methods. However, for leagues such as the Negro Leagues, where a team might play ALL of their games on the road, or have 70% of their games at home, it's a needed adjustment.

LUCK/Other are adjustments such as "clutch" hitting (that IS in Win Shares), leverage index for pitchers, regression to the mean, DIPS, etc type adjustments that attempt to account for actual timing of events (ironically sometimes to include and sometimes to exclude, depending on the method!), sample size issues, season length, etc.

In 13 ML seasons at .400, I'd guess it probably had more like 6.5 WS worth of impact.

Agreed. And in 13 ML seasons at .600 it's the same the other way. So if your comparing Keller and Johnson, this effect alone accounts for 13 WS.

Please note that I do not have Bob Johnson on my ballot (at least not yet). I'm just enjoying a good argument here.

Great post KJOK.

Only SUPER-LWTS and maybe some of the BP metrics, that I know of, adjust for the mix of teams played against beyond the "not playing your own team effect".

Do you know the methods employed? I've always assumed this would require iteration techniques that would converge to a solution; IOW, there is no exact formula (unlike what can be can be done to adjust for a balanced schedule).

Similar things need to be done to adjust "park factors" for the imbalanced schedule. Some NL teams get to play in Coors 10 times, others only 3.

OTOH, WS does not even do the simple adjustment that will normalize raw park factors for a balanced schedule.

As for the RH/LH split, I do know that MGL (the man behind Super LWTS) denies that RH batters have different splits. He says that the numbers say that given enough at bats RH hitters will all hit about 9% better against LHP than RHP.

As for the RH/LH split, I do know that MGL (the man behind Super LWTS) denies that RH batters have different splits. He says that the numbers say that given enough at bats RH hitters will all hit about 9% better against LHP than RHP.

I agree with MGL, but if one right handed batter faces only LH pitchers, while another faces only RH pitchers, then that's a 9% bias that would need to be adjusted for, correct?

Only SUPER-LWTS and maybe some of the BP metrics, that I know of, adjust for the mix of teams played against beyond the "not playing your own team effect".

Do you know the methods employed? I've always assumed this would require iteration techniques that would converge to a solution; IOW, there is no exact formula (unlike what can be can be done to adjust for a balanced schedule).

I believe you are correct - iteration techniques such as some of the college football computer ratings systems use. From a practical standpoint, it becomes a question of what you adjust first, and how many iterations you do.

If I understood SUPER-LWTS corretly, MGL compiles every batter-pitcher matchup (Helton batting in 3 PA's against Pedro in Coors Field, for example) and somehow adjusts to the expected result of an "average" batter batting 3 times against Pedro in Coors field. But the devil is in the details of the calculation, of course. How do you figure out Pedro's expected results allowed? Is that number biased by what batters HE faced, and where he faced them? How do you figure out the expected "true" impact of Coors field? Is that expected Coors impact number effected by Helton's batting? Is it effected by Pedro having 2 starts in Coors while Kris Benson has zero? If so, how do you avoid "double/over adjusting" since Pedro's performance could be impacting both the competition adjustment and the park adjustment numbers?

I just want to mention again that I have interviewed Yogi Berra on a couple of occasions, and he seems every bit the modest gentleman you suspect he is.

A year ago, he gave a speech to a crowd of high school kids, and his voice cracked as he recalled not getting past the 8th grade, and how good baseball had been to him over the years.
We've gotten used to politicians, entertainers, and even sports stars being so conniving these days that it's quite striking to see someone so utterly genuine.
No bonus points for any of that from Hall of Merit voters, but he needs no help anyway to breeze into our entrance door...