Brian
I think there is a real good argument that WAA might be a superior standard for Hall of Fame than WAR. I like the "greatness vs value"
principle.

The problem is players who extend their careers by playing when they are in that area between average and replacement level. They still have value, and are helping their teams by playing, but they are damaging their career WAA.

I know you are not pushing this for Hall of Fame purposes, but I bet it would work pretty well as a HOF measure if you kept the total of the positve years but zeroed out the negative ones.
11:56 AM Jan 14th

KaiserD2
Dave Fleming is certainly correct that great players don't automatically win pennants, but I think when I do the numbers on that, as I plan to do, I will find that they help a lot. On the other hand, one of the striking questions I emerge from the study with is, how on earth could the Giants have won pennants in only 2 of the 12 seasons in which Willie Mays had over 5 WAA? (They won in 1951 and 1971 when he had less.) And incidentally, Victorino performed at the 5 WAA level for the 120 games in which he was in the line-up. He had a remarkable year. But yes, a lot of teams win pennants with a bunch of 3-4 WAA players. Yes, a whole lot of goodness will make up for a lack of greatness.

The question of how valuable a guy who remains at 2-3 WAA for 10-12 years is strikes me as one of the most difficult questions of all. Yes, he's a big help. But you won't win the pennant if he's the best guy on your team. And when I come up with the pitching data, you will be amazed at how many HOF pitchers fall into exactly that category. They were decent pitchers with durability who played for good teams--like Jack Morris.

DK


8:36 AM Jan 14th

sokho
Well, let's not go too far. I would be very surprised if, over the history of baseball, pennant winners did not have a greater-than-random share of excellent player seasons, probably materially greater (although I have not done [and will not do] the research).

Also, for planning purposes, I'd suspect that having a couple of five WAA players (well, players who can be reliably EXPECTED to add 5 wins above average) locked in for the following season generally makes the job of a GM much easier on a team thinking about contending...

My broader point below was just that value can reasonably be defined in different ways. The one that I'd usually employ if I were assessing career value or putting together a roster would be tied to some fungibility level of performance--but KaiserD2 uses a different definition of value to identify "greatness" as he defines it. So long as he's not saying it's like the speed of light or absolute zero (and I don't see any sign that he is), it's seems reasonable to me...
10:12 PM Jan 13th

steve161
"I'm trying to focus on greatness, defined by individual seasons, because it's what wins pennants."

I'm not so sure. If I understand sokho correctly, he's suggesting that a whole lot of goodness will make up for a lack of greatness, or at least a lack of excess of greatness. Looking at the World Series participants (not just the winners) of the last 20 years or so, I think he's got an argument.

I was thinking the same thing when Bob Gregory tallied up the decade results of the various votes in the Reader Posts. There doesn't seem to be much of a correlation between individual greatness and team success.

3:55 PM Jan 13th

sokho
Great points, rgregory1956 and KaiserD2, so long as we acknowledge that pennants are won by teams, all of which have average or below average performers who, nevertheless, nudge teams towards championships. In fact, from one perspective, the more performers there are on a team who are above average, the LOWER the need for any of those "peak" seasons from anyone (recognizing that, from another perspective, that's a sophist's arguement).

My thinking is pulled back to the Hall of Fame discussion by two unrealistically extreme illustrations that I find useful to frame my thinking:

Peak Value Player A: A guy comes up to the Majors as a 23 years old, plays 10 year, contributes 10 WAA (or, say, 12 WAR) for each of five years, plays another five as an undistinguished but full time player with 0 WAA (or 2 WAR) per year, then retires. 50 WAA, 70 WAR (a superpowered Albert Belle, perhaps)

Career Value Player B: A guy comes up at 19, never makes an all star team, never, in fact, is one iota above an average big league player, but does this full time, year after year, until he's 53--35 years of undistinguished, average major league performance. 0 WAA, 70 WAR (a more durable and consistent version of Julio Franco, maybe)

Is Player A a Hall of famer? From the discussion, seems like all the participants on this board would say yes. Is Player B? You bet your butt I'd vote for him. I'd be curious to the hear opinions of others.

Sokho.
11:28 AM Jan 13th

evanecurb
Oliverp:

A poem I learned to help with the Celsius scale:

30's hot
20's nice
10 is cool
and zero is ice

I prefer the Kelvin scale, which was named after Kelvin Bryant, who was a star running back with UNC in the 1980s and played in both the USFL and the NFL. Many thermostats still have the name "Bryant" posted on them in commemoration of the Kelvin scale.

I like the idea of absolute zero.

Steve161: Maybe we should measure players on the Kelvin scale: Wins above absolute zero.


9:31 AM Jan 12th

oliverp
But I was serious. The Fahrenheit scale is the preferred scale of myself and most Americans. All those other countries have it wrong and America is right about this issue. It's just like how a lot of countries foolishly are mad about soccer and don't even play baseball. Or countries that play cricket instead of baseball. All those countries that don't play baseball are just stupid.
7:48 PM Jan 10th

oliverp
I don't like the celsius temperature scale, it's stupid. Everybody should use Farenheit. Farenheit just makes sense. 100 is really hot, 32 is cold, and 0 is really cold. 50 is just about the tolerable level for being outside a long time. It just makes sense. I'm sick of all these foreigners trying to quote celsius temperatures when Farenheit is the real scale of temperature.
5:22 PM Jan 10th

KaiserD2
To Chill: your point about Barry Bonds and pennants is taking advantage of a semantic distinction. During Barry Bonds's career, to win the pennant, that is, to reach the World Series, you had to not only win about 90 games during the season at a minimum but win first one, and then two, rounds of playoffs. The playoffs are crap shoots and you can't blame one player for not winning them.

The Pirates, with Barry Bonds, reached the playoffs three times in a row. The Giants, with Barry Bonds, reached the playoffs 4 times, three times as division winners.

For Dave Fleming, I understand your point. I am not denying that an average player has value. I am saying that what you need to win 90-100 games is a number of well above average players, and thus, that's what I'm looking for. And, I really think accurately measuring replacement value is, to put it mildly, a gigantic task. My mind is more open on the question of whether shortstops or catchers should be measured against a different "average" on outfielders, but I haven't reached a firm conclusion yet.

This has been a lot of fun.

DK
9:56 AM Jan 10th

DaveFleming
On sprox's suggestion about booting the Maddux non-voters:

One of 'em is a long-time member, and fantastic participant in Reader's Section. He made his rationale for his vote clear. He should stay.

The other guy still owes someone a Don Mattingly autograph, to replace the one he ripped up. And I don't think he even pays the $3 bucks...I think there's some nepotism deal going on there. Let's revoke his voting privileges!
9:01 PM Jan 9th

chill
Reacting to KD2's repeated statement that to win pennants, you need players well above average...well, yes, but you need average players, too. Barry Bonds's teams won one pennant. One. Most people agree Ernie Banks was a pretty good ball player. How many pennants did his team win? Seattle with Junior, Edgar, the Unit, Alex-Rodriguez-before-he-was-ARod?

Average players are scarce and valuable, too.

Having enough players who are "average" for their position is a huge part of winning pennants.
8:47 PM Jan 9th

steve161
A few comments on David Kaiser's last post:

1) I don't really care where the midpoint of a metric is. The mean has this advantage over replacement level: it is well-defined. I reserve judgement on WAA until I understand it better, but sokho's point is well-taken: if it doesn't show the tightening about the mean he has observed in other metrics, that calls it into question.

2) I agree that Keith Hernandez is as deserving a HOFer as Eddie Murray (whose credentials I do not question for a minute). In general I'd like to see the two or three best defenders at each position in the Hall, just like Bill Mazeroski and Ozzie Smith, assuming one could figure out who they are.

3) Cal Ripken is indeed a very strange ballplayer. His highest career similarity score at BBRef (Dave Winfield) is only 789, and less than half of his top 10 comparables are middle infielders. His career numbers nosedive after age 30: only 3 of 10 seasons with OPS+ over 100 and one of them a half-season at age 38. I never held The Streak against him--his argument was that the Orioles didn't have anybody who even fresh would play as well, and I bought that--and we'll never know what effect it had on his production. But one has to wonder.
8:34 PM Jan 9th

KaiserD2
Yikes. I meant Murray, not Murphy--he was not that good--and I just missed that. Comes from doing so many things at once.

The reason I use wins above average is that I think it's a more reliable measurement, and also much more easy for me to visualize. It has been pointed out on the SABR list that there are several ways to calculate replacement, and therefore, different statements of WAR don't agree with one another. I know people claim that the trouble with WAA is that is makes people who are below average sound worthless. I know they are not worthless and that someone -1 WAA is a lot better than someone -3 WAA, but the fact is, what you need to win the pennant is a number of people significantly above 0 WAA. I even used a much less sophisticated version of WAA in my book Epic Season which was written in the late 1980s and published ten years later. As for win shares, I couldn't use Humphreys's fielding measurements with them, and I think they suffer from being based on actual wins instead of Pythagorean percentage. If your team loses a lot of one-run games, your win shares get affected that year.

I'm glad Murray is in the Hall. Again, Keith Hernandez's peak years are about as good as Murray's and he was on several pennant winners too. Not an accident.

One more comment: Cal Ripken's WAA numbers are good but I can't look at his record without thinking they should have been much better. They are strange. He had three terrific seasons: 7.4, 6.9, 5.1 WAA. Not only that, he had two of those when he was 21 and 23. But beyond that, he never made it to 2.5 WAA, He had ten full seasons of average or below average. I have to think he could have been much more valuable if he hadn't focused on Gehrig's record.
DK
2:04 PM Jan 9th

rgregory1956

Not that it's overly noteworthy, but I noticed that the BBWAA averaged 8.4 names per ballot. The BJOL voters averaged that same 8.4 names per ballot. I'll let someone else decide if this has any significance or meaning.



12:53 PM Jan 9th

DaveFleming
KD2: You've mentioned Eddie Murray twice now: he was elected by the BBWAA in 2003, his first year of eligibility.

I wonder if you mean Murphy...Dale Murphy.
12:45 PM Jan 9th

sokho
Interesting discussion, Gentlemen.

(1) Sounds like everyone's in rough agreement with the interpretation of Gould's point (that the decreasing frequency & extremity of real outliers performances above the mean indicates that the mean is rising towards the notional "right wall" of distributed human performance [or vice versa]). I would not be as confident as you are, Mr. Kaiser, based on the data he's shared, in stating that Gould is wrong--I've seen the distribution of various performance metrics plotted year by year over time and they almost all tighten about the mean (i.e. all have shrinking standard deviations) although I, admittedly, haven't seen it for Wins Above Average. It is, nevertheless, an interesting point.

(2) I assume that when you present or publish, you'll share all the ins & outs of your methodology. I'm curious if you give us a previous on the following question: why do you use contribution above average performance as your standard for value as opposed to contribution above some point of fungibility (wins above replacement, win shares, etc.)? Both are common metrics, but there are many reasons (widely discussed, and I won't repeat here) that value above a replacement threhold is used more often both in baseball operations and in retrospective analysis. Obviously, defining "greatness" as performance above average would skew it much more heavily towards the peak performances you identify than would defining it as performance above replacement, which would tend to identify players with greater lifetime totals...

Thanks, again, all, for this discussion.

Sokho.
12:30 PM Jan 9th

ventboys
Clearly Whitaker was voted in based on his WAR, and that's probably what will get his case to the top of the pile in the 2020s, when the VC gets serious about the 1980s.
8:47 AM Jan 9th

DaveFleming
(Edit: Steve161 did a better job of paraphrasing Gould's argument than I did, for those who haven't read the article.)
9:54 PM Jan 8th

steve161
"And Palmeiro’s comments to Congress aside, there is strong evidence suggesting that Will Clark’s alter-ego not only used performance enhancing drugs, but participated in commercial endorsement for them."

Loved this. Of course, the drug Palmeiro endorsed may have helped him score, but it didn't necessarily help him hit.

David K: I think you've misunderstood Gould's argument--at least, you have not refuted it by noting constancy in the number of superstars. What Gould asserted--which I think is obviously correct--is that today's level of competition is higher not because of the best players but the worst. That is, while the top 50% of a roster may be comparable across eras, the bottom 50% gets stronger as you get closer to the present day.
9:42 PM Jan 8th

ventboys
Jeter's WAR value probably overrates him, because it is based on batting with a shortstop's glove on more than how well he played shortstop. The more defensive minded metrics tend to underrate him because they do the opposite: they make him bat without any adjustment for defensive position, while they compare his defensive value to other shortstops, most of whom are far less productive hitters.

The answer? Well, find it and you'll get rich.
7:45 PM Jan 8th

KaiserD2
Dave,

Thank you.

You were half right.

I checked my spreadsheet, and I have to apologize. The problem is not with my system, the problem was with my data on Jeter. I have developed this data by starting with spreadsheets from baseball-reference.com, and then moving some columns around, entering the Humphreys fielding data, and using a slightly different formula to calculate WAA. One aspect of that formula is to subtract the fielding runs data provided by baseball-reference, and then to add in Humphreys's data. For some reason, the row in Jeter's entry that was supposed to be filled with the baseball reference data was blank. I have checked nearly all my spreadsheets including about 80% of my data and I don't seem have done that with anyone but him. How it happened, I don't know. I'm glad you forced me to take a look.

And yes, having corrected it, the results are different--although not that different.

The recalculated results show Jeter's four best seasons at 3.9, 3.2, 3.1 and 3.0 WAA, in 1999, 1998, 2006, and 2005. Here are the rest of his WAA figures, from top to bottom: 1.9, 1.8, 1.2, 0.8, 0.5, 0.4, 0.4, -0.3, -0.7, -0.8, -1.2, -1.6, -1.9, and -2.2. So the corrected figures show him rounded off to average or below average in exactly half of his 18 seasons in the major leagues.

One of the things that got me going on this was the Bill James question fro the Keltner test: "If this guy were the best player on your team, could you win the pennant?" Bill did not say, how many times? But this goes to your question of Jeter winning pennants. The Yankees did win the pennant in his four seasons of >3 WAA and he was a major contributor. He may have been the most valuable Yankee in 1999 and 1998; he certainly was not in 2006 or 2005.

I apologize for the mistake, and I will recheck my data very carefully before publishing it, but my evaluation of Jeter was not that far off--unless you want to argue that Humphreys's fielding data are for some reason particularly unkind to him. They are 84 runs worse, over the course of his whole career, than the baseball-reference estimates--that is, about 4 runs a season worse.

At some point I plan to run the correlations between individual records of 4, 5, etc. WAA on the one hand, and total team wins on the other.

David K

6:39 PM Jan 8th

KaiserD2
As I have mentioned, I'm in the midst of a long-term project to identify the greatest players, based on seasonal wins above average. The point of baseball is to win pennants, and the way players help their teams do so is to perform at a much-higher-than average level. I hope to present the whole method and some results at a SABR convention soon.

It is clear that the voters on this ballot, like the BWAA voters in some years, pay far more attention to LIFETIME totals than to best seasons. This leads to VERY different results, as I will show with one or two examples.

My initial, arbitrary standard of greatness was four seasons over 5 WAA. The record for seasons over 5 WAA, I am sorry to say, is held by Barry Bonds, with 15. Henry Aaron ranks second with 14. Willie Mays, Ted Williams, Stan Musial, Ty Cobb, Babe Ruth and Lou Gehrig were all over 10. (This list is not exhaustive.) Mike Schmidt and Ricky Henderson (the greatest players of the Boom generation) had 9 and so does Albert Pujols.

(I should mention that these figures do not match the ones at baseball-reference because I've made some adjustments but this isn't the time to go into that.)

Now to begin with, although my standard is MUCH higher than traditional HOF standards, there are two people who meet my standard who are NOT in the HOF. One of them is Bobby Grich, with 4. That ties him with Rod Carew. Lou Whitaker, on the other hand, is a perfect example of the kind of player who does much better by lifetime measurements than by mine. Never did he have a season with even 4 WAA, by my method. He did have six seasons of three or more. But according to my method, there is no way that Whitaker could be mentioned in the same breath with Eddie Collins, Nap Lajoie, Joe Morgan, Rod Carew, and Charlie Gehringer.

I won't take the time to go through everyone mentioned in the original post here, but Greg Maddux and Curt Schilling are both VERY great according to my method. Frank Thomas qualifies by my method as great. Glavine topped 5 WAA only once, 4 twice. Mussina topped 5 twice and 4 twice more, making him a stronger player, for me, than Glavine. (Again my figures on pitchers will differ, but only very slightly, from baseball-reference.)

Last, but hardly least, Craig Biggio (sorry to be a year late) is, by this method, one of the weakest potential members of the Hall of Fame. His four best WAA marks were 8.5, 3.2, 3.1, and 2.4. This is partly because, according to the Humphreys fielding runs stats, which I use, he was a very poor second baseman. I might add that that makes Biggio substantially superior to Derek Jeter, whose for best WAA seasons are 4.7, 3.8, 3.8, and 1.2, and who has been a below average player for most of his career. And who will undoubtedly enter the Hall of Fame the first time he is eligible.

That's enough for now, but I hope I got people thinking. One more thing--another player who just barely missed my standard of greatness, and who by my lights has a much better record than most of today's marginal candidates, was Eddie Murray. Keith Hernandez is in the same class.

David Kaiser




1:54 PM Jan 8th

ventboys
There are always a coupla mild surprises; this year mine are Morris dropping off completely (not that it matters, he's a VC guy anyway), McGriff's relatively poor showing (I think we kind of forgot him for a minute), and Mussina beating Schilling to the alter (barely). Overall, though, I think Dave's ballot works exponentially better than the real one. Well done again, Dave. We'll meet at the same joint next year, and I'll bring the hot wings.
8:09 AM Jan 8th

MarisFan61
Comiskey
Minoso

(Those aren't votes, they're spelling corrections.) :-)

I wouldn't have said anything about Comisky, but I couldn't bear seeing "Minosa."

Minnie's a Hall of Famer. (I mean in my book. I know he isn't in Cooperstown's.) I hope he stages a rally and wins one of these write-ins some year.
2:33 AM Jan 8th

evanecurb
Hi Dave,

Thanks again for doing this. It's always fun. How does the write in process work?
11:46 PM Jan 7th