THE BOOK--Playing The Percentages In Baseball

Here’s another thought.  Consider David Duval.  He was a big-time money leader from 1998 to 2001 (if not 1st in the world, at least challenging Tiger).  Since then, nothin’.  In golf, the payoff for winning is enormous.  The difference between 1st and 10th is more than between 10th and last place (or some such).

In baseball, because it’s a team sport, a run provided by Molina is worth exactly the same as a run provided by Pujols.  There’s an extremely linear linking between performance and value.

In short, golf, tennis, and single-player sports rewards in geometric proportion the player’s accomplishment.  In order to do that in team sports, we have to create a model that does that.  And the “great game”, or “incredible season” is what does that.

While there are plenty of other golfers who have a better record than David Duval, virtually none of them has his “greatness"… none of them will be mentioned in the same breath as Tiger Woods.

In 50 years, we’ll still be talking about Sandy Koufax, and it’s that “greatness” that needs to be modeled.  Even if we think that that “greatness” is really irrelevant.  Our job is to model it first, and pass judgement on it second.

This is all well and good, and I’d be a fool to deny the existence of it; I wasn’t around for Koufax, but I was for Pedro, or Bonds in 01-04 (assuming we can ignore his previous exploits).  It’s just the second step of judgement where I diverge from the mainstream, which is sort of what the tail end of my post was about.  But your efforts to codify what folks mean by “great” is certainly welcome.

I don’t really follow the NFL, but I’ve heard of Jim Brown (nine seasons in his career) being called the greatest football player ever:
http://www.pro-football-reference.com/players/BrowJi00.htm

Roger Federer could stop playing today, and others would bestow that title on him.  He’s 25.  At what age could Tiger have retired to be considered in the same inner circle as Jack?

These guys have 20 or 25 tournaments each year, in which they have a chance to reach the pinnacle, by themselves.  And, they’ve only got about 100 opponents in their competition class.

In team sports, you’ve got many more competitors, and only one or two “MVP”.

Jim Brown had a short career, but he is still 8th on the all-time rushing list and at the time of his retirement, was the leader.  The modern equivalent is Barry Sanders, who retired at thirty and yet is still third all-time.

I think that the individual sports performers like Tiger or Federer should be treated differently, because of the nature of those sports.  In golf and tennis, what is valued highly, by players and fans, is not winning money or even winning tournaments, it’s winning the majors.  And “career value” can be boiled down to “majors won”.  So Tiger, on the basis of winning 12, is already second all-time.

But in a team sport, value comes in helping your team win, and it takes a lot of players to do that, not just one superstar.  And so I think it’s completely appropriate to look at team-oriented measures.  Even superhuman performance by a Koufax doesn’t single-handedly win his team a pennant, whereas Woods IS the team.

I think this boils down to whether you care about, and think it’s appropriate to honor, individual excellence in a team sport.  I’m with Tango on this:  part of being great is being among the very best for some period of time. 

Carl Yastrzemski ranks 12th in runs created, but 151st in OPS+.  I’m not saying he’s the 151st greatest hitter ever—longevity counts—but I also could never rank him #12.  Could you, Patriot (assuming he was also 12th in RAA or RAR)?  Is Gibson’s 1.12 season really equal to three seasons of 2.70 on the historical “greatness” scales? 

If you take Patriot’s position to its logical conclusion (and I don’t know that he does), you shouldn’t care about black ink or any other measures of season-level excellence.  So NONE of the following should matter as HOF criteria: MVPs, Cy Youngs, batting titles, HR titles, ERA titles, etc.  I can’t agree with that approach.

I’d probably go even further, and argue that specific kinds of excellence are worth honoring, even apart from their contribution to creating or preventing runs.  For example, I think Nolan Ryan’s 7 no-hitters and 5000 Ks make him a little greater than a pitcher with the same RAA, but zero no-nos and 2400 Ks.  Watching someone dominate ML hitters and make them look like minor leaguers is part of the enjoyment of being a baseball fan. (But I can certainly understand someone not feeling that way.)

No, I couldn’t rank Yaz as the “12th greatest”, because I try to avoid using that term, specifically to avoid the kind of objections you and others would raise.  I would try to use “most valuable” or “best”, or, as I did in the series on pitchers ranked by career WAR I have written up, “My Top 60 Pitchers of All-Time”.  I purposefully included “my” so that I could set down my definition and avoid wrangling about what others think of it (from the semantics level--anyone is perfectly free to think that my definition tells you squat).

Throughout your post, Guy, you are talking about “greatest”.  That is precisely why my post was intended to wash my hands of greatness.  I’m not claiming that ranking players by career RAR would give you a list of “greatness”. 

If you take Patriot’s position to its logical conclusion (and I don’t know that he does), you shouldn’t care about black ink or any other measures of season-level excellence.

For the record, I do, with the caveat that clustered performance is more valuable to team success then scattered performance.  If you gave me two identical pitchers, each with 300-250 records, and one won 5 Cy Youngs and the other 0, I’ll take the guy with 5.  I might take him if he was fairly close in career value to the other guy.  But I would take Sutton over Koufax, assuming we are valuing from the replacement baseline (I used .380 for starters, Tango uses something close IIRC). 

As far as Ryan and his exploits go, the strikeouts can be a point in his favor, but that’s more of a question of how do you evaluate pitchers to me, getting into DIPS, trying to seperate pitching from defense, etc.  (And I would point out that Ryan single-handedly put more people on base without his defense missing an opportunity to make an out then anyone else in baseball history too).

But the no-hitters mean nothing to me, other then saying that maybe he had 7 of the best 200 or so best pitched individual games ever.  That doesn’t do anything to make me think he had a more valuable career then Jim Palmer. 

Also, you said “specific kinds of excellence are worth honoring, even apart from their contribution to creating or preventing runs”.  I’m all for honoring no-hitters, or 20 strikeout games, or 60 homer seasons etc.  I just draw the line at honoring them by making them a criteria in ranking players or choosing Hall of Famers.  But those things are certainly worthy of a display in the Hall.

But in a team sport, value comes in helping your team win, and it takes a lot of players to do that, not just one superstar.  And so I think it’s completely appropriate to look at team-oriented measures.  Even superhuman performance by a Koufax doesn’t single-handedly win his team a pennant, whereas Woods IS the team.

If golf and tennis was nothing more than the Davis Cups and Ryder Cups, then Tiger would not be TIGER.  And he would not have achieved “greatness” until well into his 30s.  Imagine a golf team with Tiger, Craig Perks and a couple of other guys.  It’s be like Mario Lemieux in his first 6 years.

When I speak of greatness, it transcends “value over baseline”, whatever that baseline happens to be.  I’m talking about the talent level that has manifested, compared to the best among his peers.  Sandy Koufax did it, and Sutton barely did.

In hockey, we have Mike Gartner:
http://www.hockeydb.com/ihdb/stats/pdisplay.php3?pid=1855

His 708 goals would put him in elite company (sixth all-time), but among the “greatest” hockey players, he wouldn’t even rank in the top 100.

Greatness: how high a peak did he reach, and how real was it.

If one stipulates that:
1) The correct criteria for admission to the HOF is greatest value (or being the “best), and
2) Value/best = most wins,
then of course you will not care about peak, and there’s really no room for argument.  I don’t have a problem with that point of view.  But it IS a point of view.  Perhaps I’m misreading you, but you seem to be saying that my (or Tango’s, or anyone else’s) definition of “greatness” is subjective, while your definition of value (or “best") is objective.  I don’t think that’s right—both are subjective.  To me, being one of the all-time best (let’s use that word) means that you were one of the best at some particular time, that is, over the course of some seasons.  To you, that’s irrelevant.  Fine.  But those are both, ultimately, just opinions.  Neither view is correct (or incorrect), or objective. 

Now, once you make that subjective judgment, I think we both agree that we should use metrics that rationally and accurately measure those things.  And 5-year peak may in fact not be a very good way of measuring excellence at the seasonal level.  We should definitely argue about those kinds of issues.  But the prior question—does seasonal performance even matter?—will always be a subjective judgment.  But I think it is objectively true that yours is a minority view, and that the large majority of fans do consider a player’s peak (however measured) to be part of what they mean by “the best.” Most of our baseball memories are at the game or seasonal level—Gibson in ‘68, Gooden in ‘85, Pedro in ‘99—and that shapes our definition of greatness (sorry, “bestness"). 

* *

You make a good point on no-hitters, Ks, etc.  They should be recognized at the HOF, but not a significant criteria (if at all) for admission.

Chris Dial started a thread on BTF trying to define greatness.  A consensus was never reached.

http://www.baseballthinkfactory.org/files/dialed_in/discussion/greatness_how_do_we_calculate_it/

When I speak of greatness, it transcends “value over baseline”, whatever that baseline happens to be.  I’m talking about the talent level that has manifested, compared to the best among his peers.  Sandy Koufax did it, and Sutton barely did.

Understood, and accepted.  The distinction I was trying to draw is that golf and baseball are different sports, with different objectives.  The objective in golf is to win, and the objective in baseball is to win.  But in golf, you compete as an individual.  In baseball, you compete as a team.  Therefore, in golf, greatness as you define IS value, as I define it.  Or certainly it is a lot closer to being equivalent then it is for baseball.

you seem to be saying that my (or Tango’s, or anyone else’s) definition of “greatness” is subjective, while your definition of value (or “best") is objective.

I think my position can better be summarized thusly: wins, losses, and pennants are objective.  (Aside: we may (are) not be able to perfectly, objectively, break up team wins and pennants into player wins and pennants.  And of course we are making subjective judgments about how to weight wins against losses.  These are the complications of taking the theory or the definition and applying it practically).

Any other way of defining value, other then by wins, losses, and pennants, is going to introduce even more subjectivity, or crystallize the subjective judgments of others.  For instance, we could evaluate value in economic terms--marginal revenue added.  But if we did that, we would be incorporating whatever subjective decisions fans, advertisers, etc. make when deciding how much they are willing to pay for baseball.  This is a fine way for a team owner to determine how much to pay a player, but I don’t see any reason to think about who is the best baseball player in those terms.  To define the best player as anything other then the one who does the most to help his team win games and championships is to ignore the real on-field objective of the game of baseball. 

I should also point out, as I did in my post, that I have no problem with making a list of the best seasons, or the best five consecutive seasons, or what have you.  It is when you ask me to use one of these standards to rank someone that I lose you.  I don’t see any reason why I shouldn’t evaluate Hall of Fame candidates on the body of their work rather then some subset of it.

"I think my position can better be summarized thusly: wins, losses, and pennants are objective.... Any other way of defining value, other then by wins, losses, and pennants, is going to introduce even more subjectivity, or crystallize the subjective judgments of others.”

I think fans generally define value the same way, as do I:  the best players are those who do the most to help their team win.  Where most fans differ from you is in caring whether a player is the best at SOME MOMENT IN TIME.  To use a fictional example, when Roy Hobbes’ manager tells him “you’re the best I ever saw,” we can be pretty sure that’s a peak, not career, evaluation (since Roy only had one season!).  Being the best is important to Hobbes, and also to most fans. 

Now, would we admit a real-world Hobbes to the HOF?  No.  But suppose Ted Williams had been injured in Korea and never played again.  Would we still say, based on his first 10 seasons, that he was one of the very best hitters who ever lived?  I think most fans would, while I assume you wouldn’t.  I’d say he’s still an easy choice for the HOF, but using career wins/pennants he’d probably be a borderline case. 

Again, I think both points of view are reasonable.  But I don’t think either is more “objective” than the other.

Regarding peak and longevity, I think intuitively we know the former to be the more valuable, but both factors matter.  Much like we know OBP to be more important than SLG, that doesn’t mean if taken to its logical conclusion that SLG isn’t important.  In a perfect world we’d want a player with great OBP and great SLG.  In an imperfect world we say OBP is, what, 1.4 times more valuable than SLG and adjust accordingly.  Our greatest baseball players have peak and longevity.  Peak alone makes for an interesting footnote, and longevity alone just doesn’t cut it.  (I think Jay Jaffe’s JAWS system, imperfect as it is relying on WARP, is the best-fit model I’ve seen yet to pin potential HoFers.)

I have Williams’ first ten seasons as being worth about 105 wins better then a replacement level hitter.  That is more then any HOF left fielder has for their career save Musial and Yaz.

I realize you picked that as an example, and I’m sure you could find one that would better illustrate your point and that you and others would say was in and I would say was out.  But I think it is somewhat illustrative, in that when you come across a true great, in the sense that you or Tango might use the term, it doesn’t take that long for them to compile a career WAR that will stand up with anyone.

A better case might be if Pujols was to pull a Judge Crater.  I would have him definitely out.  Pujols, though, has certainly demonstrated his greatness, so I assume somebody out there would have him in.  But if he plays at the same level as he has, he would be qualified for mine in four seasons too.  Heck, three seasons and he’d be in the same neighborhood with Palmeiro, McGriff, Mize, and Keith Hernandez among first baseman.  My point is that career value should not be confused with longevity; there’s a positive relationship, but brilliant performance for a relatively brief period of time can shoot you up those lists. 

I can agree that neither way of looking at is objective; however, to me it is abundantly obvious that the starting point should be the player’s entire body of work.  To choose a peak criteria is to willfully ignore the majority of Don Sutton’s career.  Most people, you included I would surmise, do attempt to balance the two, and so you are not ignoring anything, you are just putting a higher weight on some of them.

Actually Williams is the perfect example if you say WWII instead of Korea.

If all you had to evaluate were Ted’s first 4 seasons, including the .406, could we still say he was one of the greatest hitters ever?

I think we could.

I am with you, Patriot, on evaluating a player’s entire career.  I usually find myself in the minority on these issues as most people seem to go for the peak.

I have room for Sutton and Koufax in my Hall, and I’m generally a small HOF guy - borderline guys like Murphy or Rice do not make my cut.

I don’t disagree with the idea that career value is our logical starting point.  But I can’t completely agree with this: “I don’t see any reason why I shouldn’t evaluate Hall of Fame candidates on the body of their work rather then some subset of it.” Baseball seasons have meaning and integrity—they are not some arbitrary “subset” akin to ranking composers or painters based on their best 12-month output.  To me, it takes great seasons to make a great player; or, if you prefer, a player must be one of the best players for several seasons to be considered one of the best of all-time.

The Pujols case is the good one.  If he quits now, is he one of the greatest players of all time?  His wOBA is around .450 in a league of .340, and he has 4300 career PA.  He is 14 standard deviations from the mean, making him undoutedly not an average player.  If he was truly a .425 hitter, his performance would be 3 standard deviations from that point.

A .425 wOBA would translate into an OPS+ of 155, as opposed to his career of 170.

Frank Thomas, from 1990-1997 had a better career OPS+ to that point, with about 500 more PA.  (He’s obviously not as good a fielder.)

I’d have had Thomas as a sure-fire HOF, and his performance since then is completely irrelevant to his “greatness”.

I think this is a bullsh*t topic for analysts to “anal-yze” over.

Give me WAR, WAA, pennants-added, postseason performance, etc. If you want to give me wins over .600 (or whatever Tango was talking about), I’ll look at that, too.

Don’t give me your subjective definitions of greatness. K Wood’s 20 K game in 1998 was about as great as you can get, but it doesn’t mean squat to me in evaluating Wood for the HOF. I am not interested in where someone wants to draw that line. I’m not interested in a statistical technique which tells me how likely it is that a pitcher was really a .350 player, or a .650 player, based on their performance and PA totals. I am just as impressed by a player who can provide value at age 40, as I am by a player who was “great” at 25, but flamed out at 30 (Koufax). Durability, and the ability to maintain skills upon ageing, is a skill, IMO, and is undervalued in these discussions of greatness.

For a period of 2 years in my mid-20s, I was the most devastating foosball player you ever saw. If my front line got the ball, it was an automatic goal because my moves were unstoppable. The same was true in my mid-teens with board hockey. But, I certainly don’t consider myself the best ever. I had no longevity.

Longevity for a baseball player, is IMO an apriori indicator of greatness, instead of some obstacle to be overcome by focusing on a smaller subset of performance.

Good, we have two opposite points of view. 

I couldn’t care less about tieing longevity as a skill to greatness.  The hockey world certainly doesn’t do it with respect to Bobby Orr and Gordie Howe.  Baseball has this attitude of “accumulating” stats that the other sports simply don’t have, rewarding longevity because they can compile counting stats.

I do prefer David’s argument as he has put it forth, in that he explicitly establishes that longevity is not just about adding stats, but something real beyond that.  But, most people don’t recognize that, as Koufax will always be remembered as a great pitcher and Sutton won’t be.

That’s my point.  I’m modeling greatness the way a large group (a majority I’d say) sees it.  How high a peak did he reach, and how real was it.

As for your foosball point, I’m pretty sure ten million people can claim such a thing, and that’s just in Brazil.

More comments on Patriot’s site:
http://walksaber.blogspot.com/2007/04/what-is-greatness-or-what-is-peak.html

Rich Lederer on greatest living hitter:  http://sportsillustrated.cnn.com/si_blogs/baseball/fungoes_blog/2007/04/baseball-beat-greatest-living-hitter.html.  Rich considers both peak (OPS+) and career (RCAA).  However—consistent with Patriot’s point—it doesn’t give you very different results than just looking at career.  Only noticeably change from including peak is bumping McGwire up a bit higher on the list.

One thing I noticed was Rich saying that an OPS+ of 150 meant an OPS 50% higher than average.  In fact, it’s more like 25% above average.  I suspect this is a pretty common misunderstanding (Rich’s statement would be true if applied to ERA+).

Yes, one of the problems with OPS+, but in its redeeming nature it scales to RC+.

ERA+ of 150 is not 50% better than average, but 33% better than average.  ERA of 3.00 in a league of 4.50 gives you ERA+ of 150.  While all other measures are metric/league, this one is league/metric, all in an effort to get “bigger is better”.

Right.  My bad.

Actually, I’ve never liked the league/metric definition, since an 80 ERA+ is really much worse than a 120 is good.  I wish it were defined as 100 * (1 + (league - player)/league). Then each point would represent an equal amount of RA.

Your equation boils down to:
200 - 100 *(player/league)

So, when player = league, you get 100.  When player is half the league, he’s 150, and when he’s 1.5 times the league, he’s 50.

I’m not really opposed to it, but why not just have 100 * (player/league), so the corresponding numbers would be 50 and 150 (but the other way)?  Is it really necessary to subtract those numbers from 200?

Doesn’t matter to me, but I can see the argument that people are used to higher=better.  In any case, too late to change it now. 

If we’re going to introduce a new pitching metric, I’d want something that uses pythag to generate a win% for each pitcher (assuming average offensive support).  So, in a 4.00 league, you get these ERA/ERA+:
2.50/160
3.00/133
3.50/114
5.00/80

If we calculate win% (treating ER as R for simplicity), we get:
2.50/.689
3.00/.621
3.50/.558
5.00/.398

It’s not linear, so great pitchers should be valued a bit more than under a straight RAA or RAR metric.  But ERA+ goes too far the other way, exaggerating the value of the great pitcher by shrinking the denominator.

----"I’m modeling greatness the way a large group of people sees it.”

And I’m sure you’ll do a good job at it. But the first task IMO is to objectively evaluate whether this popular concept of greatness holds logical water, and is therefore worthwhile to model. Unless I missed it, I don’t recall seeing that from you....

Another popular concept is a pitcher’s “command”. I’ve been playing around with ways to estimate that from the official stats, but first I looked at whether there is really any “meat” to the concept of command, and decided that there was. (Yes, I’m aware it’s a “junk” stat, but the top 3 on my active list are M Rivera, Maddux, and Glavine.) But as to *greatness*, I’m still trying to decide “where’s the beef?”.

Like I said, after Wayne Gretzky’s 5th year, he’d already established himself one of the best ever.  Bobby Orr had 8 MVP-like years, and he’s #2 all-time.  To me, that’s greatness.  And to the people who voted Bobby Orr #2, ahead of Gordie Howe, they believe it as well.  (By the way, Orr pleaded with The Hockey News that if he didn’t win, put him at #3, behind Gordie Howe.)

-----"Like I said, after W Gretzky’s 5th year, he’d already established himself as one of the best ever.”
___________________

But where do you draw the line? If 5 years is sufficient, then why not 4 years, and so on.

Given all that we knew about him, it can be argued that after his dominating age 22 season in 2003, M Prior had legitimately established his greatness. Seriously. Not that he had great potential, but that he was already a great pitcher. Look at his stats from that season. The only thing that was gonna stop him from being a 1st ballot HOFer was injuries. Prior’s career (to date) is simply a speeded-up version of Koufax’ career--he was quicker to establish his greatness, and quicker to flame out due to injury.

So, where does Prior fit in to the “popular” conception of greatness that people want to model?

Like I said, I don’t base it on number of years as a threshhold.  Simply “what is the chance that his performance is that of one of the best ever”.  And Gretzky’s performance after 5 years is a performance that could only have been achieved by one of the best.

It’s simply: number of standard deviations from a certain baseline.  That is based on how much he’s played, what his performance is, and what the baseline is that establishes greatness. 

So, the “line” would be whatever number of players you want to be in the inner circle.

-----"and what the baseline is that establishes greatness”

And there’s the rub. You have yours, I have mine, the public has its own, not to mention those of B James, MGL, Tom House, and Joe HOF voter.

And that’s why “greatness” doesn’t work.

This would apply to any measure.  If my inner circle only has 10 great players, and someone else is allowed 200 great players, then it is what it is.  By that definition, nothing ever works.

All I’m saying is that the way I construct my list is based on the likelihood that such and such a performance was put up by the inner circle.  If I predefine my inner circle as a number of 10, then that’s what I use.

Other people will say “Wins over replacement”, which is fine.  But then, how many make the inner circle?  10?  200?  You have to draw the line.  I draw it before, and others draw it after.

No biggie…

Tango, your definition suggests an interesting exercise:  who was the greatest pitcher ever, defined as reaching a specified confidence level (say, 95%) that he was X good or better?

Right, that’s what I’m saying.

Let’s take a quick example.  Koufax from 1962-1966 had an OBP of around .240 (5400 PA), when the league average was around .325. 

If we make the “greatness” level 60 OBP points better than average (.265), then how many SD is .240 from that?  That performance is 4.2 SD.

Now, let’s look at Dwight Gooden, 1985.  His OBP was .254, on 1065 PA, when the league was .320.  That makes the greatness level .260.  That’s only 1 SD from the greatness level. 

Don Sutton’s career OBP is .286 on 21,631 PA, when the league OBP was I dunno, let’s say .320.  Greatness level would be .260.  So, we can’t look at his whole career.  That doesn’t help him.  I doubt you can pick out enough of his seasons that would let him reach a greatness level.

So, that’s the exercise.  I don’t know what the greatness level would be.  And I don’t know how many SD from the mean you’d need.  But, this is the model.  This is what makes Bobby Orr the #2 player of all-time.  This is what would make Gretzky a HOF after 5 years (if not less).  Guy Lafleur for example made the HOF (and the top 20 or 30 list of all-time) on the strength of six MVP-level seasons:
http://www.hockeydb.com/ihdb/stats/pdisplay.php3?pid=2902
His other 7 seasons were pretty good years, but really irrelevant to his candidacy.
Same should apply to Frank Thomas’s start to his career.

This is the model I’m talking about.

I don’t think the comparison in Tango’s #31 is exact.  Tango chooses 65 points from league average as is threshold, which is personal and arbitrary. 

The value above baseline standard, be it WAR or WAA or Pennants Added, on the other hand, is neither personal nor arbitrary.  Sure, we all might be using different baselines or different definitions of what constitutes value in baseball.  But once we agree on what we want to measure, we can come to a conclusion on what the specific definition or baseline should be.  The specifics can spring from the definition.  For instance, we may decide that value is constituted by being better then a AAA player.  We may then wrangle about how good a AAA player actually is, but we are all coming from the same place.

With greatness, we can all agree with Tango’s definition of being a likelihood of a given level of performance.  But what we choose as the given level of performance will not have the theoretical underpinnings that the value-based definition *can* have.  The greatness can never have those theoretical underpinnings, because there is no definition, no starting point, with which to define it.  Greatness, is as Tango says, “number of standard deviations from a certain baseline”.  There is nothing there to alow us to determine what that certain baseline is.

I don’t agree.  Tango’s method would allow you to sequentially rank every pitcher in baseball history, in terms of their highest demonstrated peak performance.  And that will have clear theoretical underpinnings.  Indeed, he’s measuring value exactly as you do—winning games—but he’s looking for the best performance at some moment in time, rather than cumulatively. 

Now, drawing the “greatness” line is still arbitrary, but a career value metric doesn’t avoid that problem—you too have to draw a line for HOF admission. 

I’m not completely endorsing Tango’s definition—I’d give some weight to career value for HOF admission.  But his definition is no more arbitrary than what you and David want to do.

We could also consider SD from mean.

Don’t forget that in SD, the playing time gets square root the weight.  So, in the Koufax / Sutton cases, Sutton has 4 times the playing time, so that’d double the effect.

If Koufax is 80 OBP points better than mean, and Sutton is 40 OBP points better than mean, then they come out as equally great.

In terms of “value”, Sutton likely provided more.  But, that’s not what I’m talking about.

Now, drawing the “greatness” line is still arbitrary, but a career value metric doesn’t avoid that problem—you too have to draw a line for HOF admission.

But where I draw the HOF line is completely independent of how I define the metric.  You and I could agree competely on the details of the metric down to the third decimal place on the coefficient for a home run, and yet you could choose to draw the line at the top 100 players and I could choose the top 500.  Either way, our lists would be identical.

But in the SD approach, Tango said it himself:

All I’m saying is that the way I construct my list is based on the likelihood that such and such a performance was put up by the inner circle.  f I predefine my inner circle as a number of 10, then that’s what I use.

You have to decide ahead of time what the greatness level is, and then rank the players based on that standard.  So in that case, the defining of the HOF standard comes before the determination of the player’s ranking.  In the WAR case, we’ve already defined the methodology, then after the fact we can make our subjective decision, without making any changes the approach that we agreed upon before hand.

That is the point I was trying to make in #34.  For clarity, I am not saying that the subjectivity in defining the rankings means that the approach has no merit; only that there is a fundamental difference between drawing the line in the greatness case and drawing the line in the WAR case.