THE BOOK--Playing The Percentages In Baseball

It’s funny that it took his 2006 season for him to gain national recognition.  From 2003-2005, by WARP3, he was about a 6- or 7-win player.  This past year, however--in part due to injury, in part due to his no power/low obp line--he was only worth about 3.

The Polanco free agent situation was just plain strange.  I read that there may have been some “clubhouse presence” concerns, but I also wonder if it reflects some sort of bias against second basemen in the market?  I know that came up in an earlier thread…

Yes, it was very strange.  He’s seen plenty of action at 3B, and some at SS, so he was definitely not your prototypical 2B.  He had a good UZR regardless of where he played.  His top sims are (good) infielders at every position:
http://www.tangotiger.net/scouting/sim2006_6049.html

A team needing any good fielder at any IF position would have been wise to go after him.

I don’t think it was clubhouse concerns or anything, just nobody realized how good a player he was.

UZR isn’t exactly mainstream.

All the same, the Fans see him as an above-average fielder in each of the 4 years I’ve run the Scouting Report.  I’d have to think that management views him the same way.  His batting stats are league average, meaning good for a middle infielder.  There’s nothing not to like about him.  (His UZR just reaffirms the scouting.)

I agree there’s nothing not to like about him, I said so on primer when he signed.

For the stat guys, you have above average offense for his position and great UZR.

For the traditionalist, you have high batting averages, low strikeout totals, and very few errors.

For the music fans, he’s a tenor who should sing the national anthem before games. (or at least with a name like his, he should be.)

Something for everyone.

It gets even better.  His full name is Placido Enrique Polanco.  Somebody protect the Queen from Reggie Jackson!

Baseball Prospectus (BP) has Jeter’s career as being +272 batting runs above average, on 7711 PA. Per 700 PA, that’s +25.  Polanco’s career is +6, on 4072 PA.  Per 700 PA, that’s +1.  The difference, hitting-wise, is a 24 run gap, according to BP.

BP also has Jeter as being -132 compared to a SS, on 1634 “games”, or -13 per full-season.  Polanco, on 887 “games”, is +82, or +15 per full-season.  Polanco is compared mostly to the average 2B and 3B and somewhat to the average SS.  The positional adjustment between the two would mean that Jeter is being compared to players that are, say, 4 or 5 runs better than who Polanco is being compared to.  So, Jeter’s -13 is more like -8, if compared against the typical 2B/3B.  Jeter’s -8 to Polanco’s +15, or a 23-run gap, according to BP.

That’s a 1-run gap, according to BP.  When they are both in the lineup, they are equivalent.

There’s a rub of course.  Jeter has been remarkably durable, and Polanco has not been.  Of course, Jeter doesn’t collide with his OF the way Polanco does. 

If you think that Polanco lack of games relative to Jeter means that it’s something you should expect in the future, that’ll knock 20% of his value, compared to Jeter.  If Jeter is a 10-mill/yr player, Polanco is an 8-mill/yr player. 

If you don’t think Polanco is more susceptible to injury, they’re equals.

One of the many reasons I’ve never been able to root for the Phillies as much as I’d like to is the kind of horrible personnel decisions they make.  One of the more notable blunders would be last year’s insane trade of Polanco to Detroit, receiving Ugueth Urbina and Ramon Martinez in return.

Enough to make you cry, really.

I’m reading the comments made on the BTF thread here:
http://www.baseballthinkfactory.org/files/newsstand/discussion/the_book_blog_tangotiger/

One group of comments is along the lines of “hey, your conclusion doesn’t make any sense”.  Where I’m from, conclusions follow from the evidence and the arguments.  Rather than take issue with the conclusion, why not take issue with the manipulation of the evidence? 

The biggest weapon in one’s arsenal is the durability factor.  And in fact, even Marcel The Monkey would agree with these people.  Marcel has Jeter with 22% more PA than Polanco entering 2006, and likely the same for 2007.  If one wants to claim that Polanco will play as many games as Jeter over the next 3 years, he better be ready to defend himself.

Some posts look only at 1-yr of data, which is ridiculous.  What we care about, only, is their true talent level going forward, and the playing time to display that talent.  We look at the historical record to inform us to that effect.  This is why people get hired to do jobs: can he do it for a set period of time?

Most of the other posts completely disregard the gap between Jeter and Polanco’s fielding, and it’s the biggest part of the comparison!  I’m claiming there’s a 20-25 run gap in their fielding, which offsets their 20-25 run gap in their hitting.  UZR is showing that Polanco has the best infielder stats of 2000-05, after Rolen and Everett.  It’s very comforting that UZR sees Rolen and Everett as the top two, since this agrees completely with the Fans.  Polanco being #3 is a surprise to Fans, but not a big surprise.  Fans see him as an above-average fielder, a step below the great ones, and not in the company of them.  Jeter is on the flip-side, where Fans see him as an average-fielding SS, and not one of the worst in the league.  That’s the gap, is it a 5 run gap in fielding between the two, or a 30 run gap?  But, quoting a VORP, which makes no mention whatsoever on a player’s fielding completely misses the point.

When they play, the evidence shows that they are in the same ballpark in terms of the impact of their offense and defense.

The difference between the two is durability.

Argue the merits, not the conclusion.

I also forgot to mention that I am a leading proponent for Jeter as MVP!  I certainly don’t have it in for him.  Jeter deserves the extra credit he’s done this year for:
a - producing above his true talent level
b - producing even above that high level, when his team needed him

a) is worth an extra +1.5 wins
b) is worth an extra +2.5 wins on top of that

Jeter has added +4 wins this year, over and above what his talent level dictated, and the timing of his performance.  It’s a 10 million$ performance that justifies his salary (for 2006).

A-Fraud on the other hand performed at 1 to 1.5 wins below what his talent level dictated as a hitter, and another 2 to 2.5 wins below for his poor timing.  His fielding was also atrocious for him, and that was probably 1 win below what his talent level dictated.  That’s a 4.5 win performance this year below what his true talent dictated.  A-Fraud owes NY 10 million$.

Why call him “A-Fraud” if all available knowledge suggests that he was simply the victim of bad luck in his timing. I realize that some people think his poor clutch performance was due to being booed by the NY fans, etc., and was therefore a genuine choke. But I’m skeptical, and don’t buy into that. And even if there is some truth there, I don’t think that makes A-Rod a poor clutch performer--I think it simply shows that anyone can be beaten down. That is, if you are going to evaluate clutch performance as something other than pure luck, you must consider the full context. Put Jeter into ARod’s context of being skewered, and who knows how his performance might suffer.

At the same time, why should we be lining up to kiss Jeter’s ass when he most likely just had a great run of good luck.

I don’t have a problem with giving WPA a strong weight in an MVP discussion. But I’m certainly not going to let it color my opinion of the intestinal fortitudes of these players.

WPA or some other “clutch” measure has EVERYTHING to so with MVP discussions.  The only role that context-neutral or overall stats have is to give the voters a “feel” for how good a player is in general.  Voters cannot remember every time a player has or has not been clutch or when a player has been good in low leverage situations.  Voters look at overall stats before they vote for the MVP, so whether justified or not (probably not), overall and context-neutral stats play a significant role in the voting process.

I see absolutely no reason why we should assume by default that A-Rod suffered adversely from all the negative publicity.  I think we have to assume that his non-clutch play was almost completely luck-driven, since we don’t see much clutch/non-clutch skill in general.  Until we do some “experiments” (retrospective empirical ones of course, as always) to show that some players do in fact “fold” under pressure, we have to assume that that is not the case and was not the case with A-Rod.  Since we have already “proven” that there is little clutch skill in the population of baseball players, I think the burden of proof in any particular case (such as this) is clearly on the argument that a player has shown some clutch or non-clutch “skill”, such as A-Rod folding under pressure.  And of course, ALL deviations from a player’s true talent is ALWAYS a combination of luck and skill…

Jeter and A-Fraud deviated in TWO aspects:
1 - Their overall performance was different from their true talent levels as we’ve understood it, coming into 2006
2 - Their timing in close situations was much different than their overall

I don’t see why we should credit Jeter the extra win or two of his overall performance in 2006 being better than expected, while not also crediting his clutch performance.  Both were “luck”.

Same with A-Fraud.  His year with the bat, overall, was below his norm.  His poor clutching was even worse.  And, don’t forget, he had probably his worst year with the glove in his whole life.

It’s the whole kit and kabootle that was +4 for Jeter and -4 for ARod, relative to the OVERALL CONTEXT-NEUTRAL expected performance, coming into 2006.

One of the interesting things and which complicates the whole matter and introduces a metaphysical aspect to this whole thing is that we are by no means sure that Jeter was X amount better this year, by luck, than his true talent.  Only that he was X amount better this year than our estimate of his true talent going into this year.  For all we know, this year WAS his true talent level and in previous years he was lucky.  Or something in between (he was a little lucky this year and a little unlucky in previous years).  In fact, after this year, we have to update our estimate of his true talent (it is now higher than we thought - we were wrong!), so we now KNOW (actually, again, suspect) that he WAS a little unlucky in previous years.

So to say that someone has exceeded his true talent (such as Jeter this year) or that someone has underperformed his treu talent (as in A-Rod ) this year, and that the difference is “luck’ is disengenuous and in fact should NOT be couched that way.  The corrrect way to couch it, which is what I usually do even thouugh it tends to be awkward and wordy, is that Jeter exceeded our ESTIMATE of what we thought his most likely true talent level was going into the season, and in doing so raised our current estimate of his true talent (and what we think he is most likely to do next year).

Obviously we are trying to communicate the same thing and we are on exactly the same page, but one way is not couching it quite the way we mean it and the other way is.

One can also add that there is a finite possibility that this year’s performance IS Jeter’s true talent level, or last year’s, or his career thus far, but that according to the models which best fit reality so far (in major league history or at least in recent major league history), the BEST estimate of what a player will do in the near future, and it is only an estimate and by “best estimate” we mean the estimate among the infinite number of estimates that has the greatest probability of “coming true”, is a weighted average of his history, adjusted for context, age adjusted, and regressed appropriately (to account for sample error) toward the mean of the population from whence we think the player comes.

The bottom line is, and I’ve said this many times and in many different ways, we can “credit” (or discredit) a player for anything we want in retrospect and it doesn’t mean diddlysquat.  In the “real world” we “credit” people for things accomplished without nary a thought as to how much of it was “luck” and how much of it was skill.  We, as forecasters, sabermetrcians, statisticans, or what have you, tend to forget this.  When I coached and played baseball at the youth and adult level and someone had a great day at the plate or on the mound or in the field, I said, “Nice job, beautiful game, congratulations,” or whatever.  I did not say, “I know you are a bad player and that today’s stellar performance was just luck,” or something like that.  It just doesn’t work that way in the real world.  If a player was stellar but did not contribute to a win or even the propsect of win (say he went 5-5 in a blowout), he still got props and deservedly so.  If he did well in a clutch situation or situations, he may have gotten more props.  Either way he got props.  That is the way it works in life.  Heck, when someone tells us we have a cute pet, we say “thanks” as if we did something admirable.

Now, forecasting someone’s performance in the future or estimating someone’s “value” with an eye toward the future is a completely different animal.  The two have to be separated.  The problem of course, is that the average person, or even the above-average one, cannot separate the two.  This is why announcers, fans, and the teams themselves are so misguided in their thinking and their subsequent actions.  Of course it is only a “problem” if you are a team owener and are trying to maximize your chnaces of winning or making money.  Otherwise it ain’t a problem which is why when someone implies or suggests that my mission is supposed to be to “educate” the masses about sabermetrics and make it more palatable and understandable to them, I say, “Hooyey!” I doubt I would even be doing them a favor if I could do that.  There is merit to the statement, “Ignorance is bliss,” and I mean that literally and sincerely.  The world would NOT be a better place if everyone understood the “truth” about sports. In fact, a world where everyone thought the same way, right or wrong, would not be very interesting at all.  And in the case of sportrs, we are NOT trying to cure a disease.  I suppose you can always invoke the “The truth will set you free” argument, but frankly in this case, I won’t buy it.  Just a 4:00 in the AM rant…

MGL, your 1st post is correct, and I stated as much when I said:

Their overall performance was different from their true talent levels as we’ve understood it, coming into 2006

Obviously, the more words I can use, the more correct I can make my statement.

***

Going back to A-Rod, if you watched his games this year, you could try to separate his “clutch situations” not based on the game state as we normally would, but how he himself looked like he approached the PA.  When he himself says he’s “pressing”, and it’s an LI of 1.0, well, to him, it’s an LI of 3.0, or 0.3.  His performance there, in that situation, is atypical for him.  And therefore, when we weight his various PA, we don’t necessarily mark each one as a “1” (i.e., equally-weighted), nor do we necessarily mark each one based on the LI of the situation (i.e., including clutch). 

If we determine that half of ARod’s PA were very un-ARod like in approach, then we need to discount alot of those PA, so that when we come up with ARod’s true talent level, we come up with something consistent of a guy who has a great all-round hitting game and will end up with 600 HR.

The other dimension to the true talent level, which we typically mix-in as part of his true talent, but we should really separate, is his “drive” or “ability to realize potential”.  Up until this year for ARod, and for most ballplayers in general, we give them “95-100%”, since a player will need to perform at his best to stick around.  ARod maybe gets an “80%”.

Just as clutch ability is not necessarily a persistent skill, neither is it likely that a great player’s “drive"-level will be as low as it was for ARod this year.

What sabes naturally do is simply assume everyone is at “95-100%” drive, and just reduce their true talent levels so that it matches.

Again, it is a complicated issue, but one thing we have to understand is that a player’s varying drive, transient physical and pyschological states, etc., are already factored in to his mean performanc level to some or even perhaps to a large extent.  A player may actually be capable of hitting .450 but the fact that he sometimes feels worse than his peak, momentarily loses his concentration, gets something in his eye, etc., makes him a .300 hitter.  These things tend to occur randomly and we simply lump them (and correctly so) into the “luck” category and “create” the random variance of the OBP or BA binomial.  Remember that pitching and hitting a baseball (the result, that is) is NOT a random event centered on a certain mean.  The things that make it “random” are actually things that could theoretically be measured and articulated.  You have to be careful about saying things like, “There is the natural randomness around a mean of an OBP of .350 AND there are things about a player that fluctuate his true talent,” as if those are TWO SEPARATE things.  They are not!  They are inextricable!  It is like the roll of a die.  That is not in any way shape or form a random event centered on a known distribution.  It is just that the way we throw that die and the way it bounces, etc. is so non-repeatable and so unbiased that it simulates nicely a random event centered on a certain distribution.

The only thing we KNOW is that the actual mean of a batter or pitcher’s performance does NOT jump around like crazy, otherwise we would see a much higher variance than the binomial model would suggest.  But that does NOT mean that all the things that traditional baseball fans talk about does not in essence “create” the mean that we see.  I’m not sure I am explaining myself well and I’m not sure I even understand it well enough to be able to explain it well.

I think we agree when you say:

...a player’s varying drive, transient physical and pyschological states, etc., are already factored in to his mean performanc level ... These things tend to occur randomly and we simply lump them (and correctly so) into the “luck” category and “create” the random variance of the OBP or BA binomial.

My point is simply that if we can pick out certain players where things didn’t occur randomly, like we suppose with ARod, then we should keep that to the side as a separate variable.  This would probably only affect 10% of players, due to injury or mental anguish or what have you.

(I guess Rick Ankiel is another case here.  Chuck Knoblauch, too.)

I’m by no means any expert on the mental part, but if someone wanted to say that ARod had the anguish, I’d include it.  So, if ARod is a true talent “90” on a scale of 0 to 100, with a league mean of 50, the “anguish” factor could be “80%”, and therefore, drops him from a 90 to a 72.

Now, if ARod always suffers from this anguish, such that it’s now a part of him (like Ankiel, Knoblauch), then, ARod’s “true talent” is now a 72, and we don’t need the “anguish” variable for him.

Similarly, Jeter could be a true talent 70, and he has an “overachieving” parameter that bumps him up to an 80.  (An overachieving that doesn’t manifest itself all the time.)

I wouldn’t get too carried away with it, otherwise the 10% of ballplayers that deserve these “one-time” variables can end up being 50% of all ballplayers.

But, that is the reality of it.  It’s a matter of determining is it now ingrained in him, persistent, ala Ankiel, or, is it transient.

If we’re not smart enough to figure it out, then, we simply ignore the idea that ARod may have been anguished, and simply readjust his true talent level from a 90 to an 82, because we’re not sure if he’s really still a 90, or if he might really be a 72.  That is, we have more uncertainty around ARod than we would have around, say, Soriano.

I think my point is that we have to ignore it until and unless we find it in the population (like clutch hitting) and then either regress everyone, most likely heavily, such that it becomes inconsequential, or we discover a way to identify the very few players for whom it is significant, which IMO, is not likely.

You (Tango) seem to cavalierly assume that A-Rod has suffered true-talent-wise from fan and media abuse rather than just had a bad spate of luck.  As always, don’t we have to establish the spread of “fan-abuse talent’ in the population BEFORE we can assume that anyone has suffered it?  If not, don’t we HAVE to assume that ALL deviation from a player’s estimated norm is unexplainable ("random") and treat it appropriately as such?  Have we found a “mental” skill that I am not aware of?  Of course, that would be difficult as we would have to study historical players and their stats and somehow create a subset of performance such that a player is possibly in a bad or good mental state without looking at the results of that performance first.  And of course the bad hitting, or at least part if it, always comes before the abuse, and by randomness alone, half the players after the abuse will continue at some level below their norm and half would “redover” by doing even better than their norm after the abuse.  So, as usual, even if there were no such thing as a player being affected by fan and media abuse and “pressing” there would be a boatload of players (almost half who suffered the abse in the first place) who would be falsely accused of “pressing” when in fact their continuted suckiness were simply bad luck (a Type II error I guess).

Yes, very cavalier.  I’ll say that Ichiro does have a true reverse platoon split, and that ARod was affected by something-whatever-that-he-should-be-able-to-overcome.  I’ll say that Roger Clemens is a Nolan-Ryan freak when it comes to aging.  I’ll say that UZR/Dewan completely miss the boat on Betancourt.

It’s completely unscientific, and I have no strong basis for my statements, other than “gut feel”.

I also wouldn’t bet on anything I say.  Which of course means that everything I say has to be construed as b.s.

Seriously (I am not sure how much of your last post is sarcastic), there is definitely merit to observation and common sense.  To add to your examples, Eric Chavez has always had a failry large L/R platoon split I think and when you watch him hit, you can see him bailing out on almost every pitch from a lefty.  For other lefty batters, after only a few PA, you can see that they “hang in there” versus a LHP.  So basically observation (or other “data") can help us overcome small or even medium and sometimes even large sample problems.  Same thing with “common sense.” I am reasonably sure that A-Rod must be affected by all the abuse.  How much he is affected, I have no friggin’ idea.  I would be comfortable nicking his true talent by 5% but not much more than that.  And at what point, if we are doing a projection, does he get used to or learn to overcome the abuse?  Does he start fresh next year or will be there some lingering effects?  So do we just use a basic Marcel for his projection next year or do we still add a “pressing factor?” And of course, trying to use common sense to alter sabermetric projections and other sabermetric tenets and conclusions is a dangerous thing.  Isn’t that basically what causes the public to get almost everything wrong - using so-called common sense (surely batter/pitcher matchups and hot and cold streaks must have predictive value, surely you should often IBB a team’s best batter, surely you should use your closer with a 3-run lead, etc.)?

Right, that’s the point I was getting at.  Roger Clemens is unlike any other 40+ yr old pitcher.  I wouldn’t want to age him the same way as I would other pitchers that age.  But, how sure am I of that?  Who knows.  And that’s the dangerous part, when we allow our human intuition to creep into these things, what’s stopping our egos from taking over? 

On the other hand, if I was alotted say “10 intuitions” that would be considered gold, and therefore, become a fabric of a forecasting system because I “know” more than the performance data is telling me, then I would probably use it on those things I said. 

But, say we do that with 400 observers like me, and we each get 10 intuitions.  What happens if 90% of the observers say that Clemens will age like any other 40+ yr old, and 10% say no? 

Or, if 90% do agree that LH Ichiro has real reverse-splits (.866 v LP, .793 v RP, career, and has had reverse splits every year, since his 2nd yr in 2002), but they say that because they look at the numbers, and not how they look at him?  Do we tag him as being a real reverse-split guy?  (In THE BOOK, Andy does have Ichiro as the guy with the smallest platoon gap for a LH, but even then, not a reverse-split, just a very small normal split.)

Or Eric Chavez’s 2004 split, and to a lesser extent 2005.  Do we then think he has turned the corner then?  2006 would then say no.

In the end, I’ll grant anyone 10 trump cards that overrides the data, just because I don’t want to argue against it, with the lack of data I have.  Ichiro has a real reverse-split, Betancourt is 20 runs better than UZR says he is, and Clemens will age like Gordie Howe.  But, we can’t get carried away and have a strong opinion on every player.  You get 10, and that’s it.

This year (as of Sept 24), Polanco has 622 PA in 137 G.  His WPA/LI (LWTS by Game State, de-leveraged) is +1.9 wins.  (.339/.387/.459)

He also has +1.2 clutch wins.  Tigers fans rave of his fielding, and is probably +1.0 to +1.5 wins.  As a 2B, he plays a “neutral” position.

Jeter has 694 PA in 152 G, with WPA/LI of +2.45 wins, and he’s been clutch-neutral.  (.319/.386/.440) Even though they have virtually the same hitting line, Jeter has performed more optimally in game situations (i.e., changes his approach based on the game state). 

Yankee fans are a bit down on his fielding, and is probably -0.5 wins as a SS according to them.  As a SS, he gets an extra +0.5 win boost.  So, his fielding+position is average (+0 wins), though UZR and others may disagree with that.  Let’s give him the benefit of the doubt there.

While normally we may want to ignore clutch, this is Jeter.  And, just because someone is having a better clutch year than Jeter doesn’t mean now we don’t get to count it. 

Jeter does have 15 more games.  And, since the replacement level is around -.016 wins per game, that gives Jeter a +0.2 win boost.

Let’s give Jeter a 0.5 win advantage in baserunning.

Let’s tally it up:
Polanco is -0.5 wins as a hitter
Polanco is +1.2 wins as a fielder
Polanco is +1.2 wins in the clutch
Polanco is -0.5 wins as a runner
Polanco is -0.2 wins in durability

Add it up, and Polanco is +1.2 wins compared to Jeter. 

You keep saying Jeter.  I’ll keep saying Polanco.

”...replacement level is around -.016 wins per game...”

If an average player is 2 WAR, then that would equate to 125 games (2/.016). Would an average player be closer to 2.5 WAR then?

I set the replacement level for a non-pitcher as .380 winning percentage, meaning that a team of non-pitchers, with average pitching, will win .380 games. 

There are 9 nonpitchers per game in the AL, and 8.35 in the NL (pinch hitters being the 0.35 part).  With 14 teams in the AL, and 16 in the NL, that gives us 8.65 non-pitchers per game.  If each of those non-pitchers was a .486 player, you’d get a team record of .380.

(That’s .486/(1-.486)= player odds of .946, which you raise to teh power of 8.65, to give you team odds of .616, and to convert to a percentage, that’s .616/(1.616)=.380.)

So, I guess I should have said -.014 wins is the replacement level per game (or -2.0 wins per 143 GP or per 620 PA).

***

If we want to make it -.016, that implies a nonpitching team replacement level of .365, which is certainly reasonable as well.

The linear expression is also close enough: -.014 x 8.65 = -.121, or .379.

Willie Bloomquist by the way is pretty much at this level.  Career WPA/LI of -3.52 in 1176 PA.  Per 4.3 PA (one game), that’s -.013.  His fielding/position is league average.

Since this thread was started, this is how Polanco and Jeter have done, through May 7, 2008:

Polanco,.330/.382/.447, 166 games started
Jeter, .319/.379/.440, 183 games started

The difference is 17 games for Jeter (that’s worth +0.25 wins), and the “ability” to play SS, against Polanco being a much better fielder.

It’s not even close at this point…

From out perspective, there was little doubt that you were right about their relative values.

Even in the face of the clear evidence you point out above, how many baseball insiders, commentators, and journalists would accuse you of knowing NOTHING ABOUT BASEBALL if you asserted publicly that Polanco has had WAY more value than Jeter over the last two years, and that it could have been easily predicted as such?  I would say greater than 99%.  There is something seriously sad about that.

Again, I have approximately 47 players above Jeter in my sorted list of player values (rate) at the start of the 2008 season.  47 players!  I doubt it was less than 20 players in the last 10 years.

Since I started this thread two years ago:

Polanco,.323/.369/.436, 284 games played (195 R, 108 runners driven in)
Jeter, .311/.375/.430, 307 games played (190 R, 120 runners driven in)

A match offensively.  Jeter may play the tougher position, but Polanco is a much better fielder.  Jeter’s got the extra 23 games.

I mean, that’s about as TO-MAY-TO, TO-MAH-TO as you can possibly get, and I think I’m being a bit generous toward Jeter.

Since I started this thread, here is the WAR for Polanco and Jeter, 2007-2009 (through games of yesterday)

9.1 wins Polanco (337 GP)
9.7 wins Jeter (361 GP)

Polanco is at 4.37 WAR per 162 GP, while Jeter is at 4.35 WAR per 162 GP.

It’s a horse race.  Polanco is not having a good year batting, while Jeter is having a good one. 

Jeter gets a slight bonus (0.6 wins in 2+ seasons) for his durability.  It’s still a very, very close race.

It’s also interesting to compare Polanco and Manny Ramirez over the same time period.

Fangraphs lets you see the leaders in WAR from 2007-2009:

http://www.fangraphs.com/leaders.aspx?pos=all&stats=bat&lg=all&qual=n&type=6&season=2009&month=12

Pujols, Utley, and David Wright are the leaders among non-pitchers.

Among pitchers, it’s a two-horse race with Halladay and CC:

http://www.fangraphs.com/leaders.aspx?pos=all&stats=pit&lg=all&qual=n&type=6&season=2009&month=12

Speaking of Polanco…

Sky’s got a graph up at BtB that I think you’d like.

http://www.beyondtheboxscore.com/2009/6/12/906547/graph-of-the-day-placido-polanco

Very nice graph!

Would anyone like to see the bold red Total RAR line be Total RAA (above average) instead, so that all lines would use 0 as average (instead of one using it as replacement)?  You’d have to mentally calculate RAR, though, although RAR would be the distance between the red and blue lines…

Or don’t you care?