The Flat Earth Society

The Moneyball Backlash is in full effect these days, with no shortage of announcers and writers jumping on the anti-sabermetric bandwagon. Locally, even esteemed Yankee announcer Jim Kaat spent a good time bashing Michael Lewis’ and the use of statistical analysis during the Yanks-Red Sox series, and he’s been at it a few times since then, making YES telecasts just that much less enjoyable (the quality of the Yanks’ play notwithstanding). When I tune in I expect Michael Kay to blow hard, but hearing Kaat head for the dark side is quite dismaying.

The other night ESPN Baseball Tonight’s Harold Reynolds unleashed an anti-on-base percentage rant that was so bitter, I had to TiVo it so that I could drag his ignorant screed out into the harsh light of day. No, I don’t expect any better from him or ESPN these days, but still, somebody ought to whack Reynolds upside the head with a fungo bat. To set up the context, Reynolds’ rant occurred towards the end of the show. Chris Berman had just displayed a chart of the AL On Base Percentage leaders:

Frank Thomas  CHI  .494

Ron Belliard CLE .481
Lew Ford MIN .473
Jason Giambi NYY .453
Melvin Mora BAL .442

Berman then asked his fellow Baseball Tonight panelists, “Should we care about that stat?” This was apparently a setup for the show’s “3 Up, 3 Down segment,” so it was more or less written beforehand. Here is Reynolds’ response, transcribed to the best of my ability. You’re really missing his squeaky inflection and condescending tone — his fractured style is more or less captured, however:

I think it’s overrated. I don’t think it indicates how the game is played at all. There are certain roles — guys that get their man over in certain situations — you’re not going to get an on-base percentage for that. I think it takes away from the game. And the other thing is a lot of the guys with high on base percentages, they just clog the bases. Talk about Frank Thomas.

Corey Paterson — these are guy who have bad on-base percentages right now — .309 [OBP]. This guy is going to score runs for you. On this list I’ve got Corey Patterson, I’ve got Jimmy Rollins, and I have Derek Jeter.

Jimmy Rollins last year had a .320 on-base percentage scored 85 runs. He hit 8 home runs. Take away those home runs, he was on base 211 times and he scored 77 times. That ain’t no .500 on-base percentatge, but he’s scoring a heck of a lot of runs.

Derek Jeter ,well we know, .259 [OBP], we know his at batting average struggle and all that. He scores 33 percent of the time he’s on base and that will change as the season goes on because he’s going to be on base. Guys that don’t clog the bases are going to go base to base.

Now I have a problem with everybody saying, oh this is such a great stat. Jason Giambi, if he hits the ball out of the ballpark, that’s great. But if he’s on first base, he ain’t scoring on the gapper, its taking two hits to score him. To me that’s the difference in the game today. Everybody’s saying on-base percentage is the greatest thing ever. Jason Varitek, this guy last year, he scored 63 runs. All right? That’s great, he had almost a .400 on-base percentage. He scored 63 runs! I mean, I don’t get it. I just don’t get it.

Then Frank Thomas, we talked about the Big Hurt. .494 OBP, .489 last year, he doesn’t extend the plate that much. 216 times he’s on base last year, he scored 45 runs. He’s scoring one out of five times he’s on base. I think we’re getting carried away with this on-base percentage thing, because it doesn’t tell the true story of a full game. I’m not going to pitch to Frank Thomas in a situation when I know I got a base open and he’s not going to score on a gapper. I don’t want him hitting the ball out of the ballpark. It changes the way the game is played. I think we’re taking numbers and we’re forgetting all the things that go in to making baseball what it really is.

Whew! Holding up Corey Patterson and Jimmy Rollins as “guys who are going to score some runs” while disdaining the likes of Jason Giambi, Frank Thomas, and Jason Varitek is really something coming from a man with a career .327 OBP and 21 homers. Reynolds seems to think scoring runs has nothing to do with where in the batting order you are and who’s hitting behind you in the lineup, and he doesn’t have much use for those lousy runs scored by big slow guys who can hit the ball over the fences.

Here’s a chart of the players in question. TOB is Times on Base, which is H + BB + HBP – HR. R – HR is the number of times a guy scored besides his own homers, and %R is the frequency with which he did so. HR and R are added for some sorely needed perspective.

           PA    OBP  Out  TOB  R-HR   %R    HR    R

Rollins 689 .320 491 211 77 36.4 8 85
Patterson 347 .329 243 101 36 35.6 10 49
Jeter 542 .393 345 202 77 38.1 13 87

Thomas 662 .390 415 216 45 20.8 42 87
Giambi 690 .412 416 243 56 23.0 41 97
Varitek 521 .351 352 181 38 21.0 25 63

The “score some runs guys” scored a bit over one in three times they got on base, while the “clog the basepaths guys” only did so about one in five — a decent point worth making, to cut Reynolds some slack. The BUT in that statement is bigger than Jennifer Lopez’s derriere, however. First off, Reynolds completely dismisses the value of a home run — that’s a run in the bank, while a runner on base is merely a potential run. The cloggers outhomered the scorers by a margin of 108 to 31, about 26 homers per player, to say nothing of the benefits of driving in other runs. Second, even with Jeter missing a month of the season and Patterson missing about half, the “score some runs” guys used up many more outs than the cloggers, 1259 to 1183 — about one game’s worth of outs per player — and they scored 26 fewer runs overall. The scorers are helped immensely by Jeter, who really is an OBP machine compared to the other two. Compare Jeter and Rollins, who scored the exact same number of runs without homers. Rollins used up 146 more outs and only got on base nine more times than Jeter. Harold, HOW IN THE HELL DOES THAT HELP AN OFFENSE?

The answer is that it doesn’t. Yet still some persist in similar lines of reasoning.

Last week ESPN’s Buster Olney reintroduced a “stat” called the Productive Out Percentage which he introduced last fall. According to the article, a productive out is defined as either:

1) a baserunner advancing with the first out of an inning

2) a pitcher sacrificing with one out;

3) a baserunner driven home with the second out of an inning

Productive Out Percentage is the percentage of productive outs divided by the total number of outs. Writes Olney:

… Boston plays the “Moneyball” style — never bunt, don’t take chances on the bases, sit back and let your hitters hack away and do the work regardless of the game situation, regardless of the identity of the opposing pitcher. Other teams — the Anaheim Angels and the Florida Marlins, most notably — prefer to use their outs productively, by bunting, employing the hit-and-run; they put runners in motion and emphasize aggressive base-running as part of a larger strategy to put pressure on the opposing pitcher and the defense behind him.

…it will be interesting to see if, eventually, this passive-aggressive approach hurts Boston, especially with the shift in the team’s makeup. The Red Sox nearly bashed their way to the World Series last year, but they improved their pitching for 2004, shed Todd Walker, added light-hitting glove whiz Pokey Reese, and have been playing without injured Nomar Garciaparra and Trot Nixon, whose rehabilitations are being closely monitored.

… The Marlins and Angels have fully diverse offenses: some excellent power hitters, an essential element; some patient hitters who draw walks, also crucial; they have hitters who make contact, advance runners efficiently; and they run the bases.

The offenses of the Red Sox and Athletics, on the other hand, are effectively two-dimensional, eschewing the productive out within their philosophy. Boston has one sacrifice bunt, Oakland zero, and through games of April 26, the Red Sox rank next-to-last in productive out percentage — a statistic developed by the Elias Sports Bureau and ESPN — at .200; Oakland is last, at .137.

Productive out percentage is the ratio of productive outs — generally, advancing runners with the first out in an inning, or driving home a run with the second out. Last season, Anaheim ranked fourth overall in this statistic, at .347, the Marlins fifth, at .334. Juan Pierre ranked third among individual players, with a POP of .545.

Accompanying this is an Elias-generated list of the top six and bottom six teams in POP through April 26:

1. Detroit Tigers .430

2. Arizona Diamondbacks .417

3. Pittsburgh Pirates .417

4. San Diego Padres .400

5. Texas Rangers .365

6. Houston Astros .349

25. Seattle Mariners .229

26. San Francisco Giants .226

27. Cincinnati Reds .225

28. New York Yankees .210

29. Boston Red Sox .200

30. Oakland Athletics .137

Are the alarm bells sounding yet? They should be. Trumpeting a stat in which the Tigers lead the majors after four weeks is just silly, potential 25-game improvement or not. In all likelihood the three lowest-ranked teams in this stat are going to the playoffs, while the three highest-ranked will be making tee times by August. Furthermore, what of the other 18 teams? It would be helpful to know, for example, how well the division-contending teams are doing, even at this early juncture. It would be even more helpfpul to have a full season’s data, or several full seasons of data to look at so that we can better evalulate the veracity of the stat. Why we don’t have that, here or anywhere else, is a topic to which I’ll return later.

One big problem with the productive out concept is that trading a base for an out is not, on the whole, a good payoff. Looking at a run expectancy matrix such as this one, which was compiled by TangoTiger based on 1999-2002 data, we have (reading across is the number of outs, down is the baserunner situation):

          0     1      2

0 0.555 0.297 0.117
1 0.953 0.573 0.251
2 1.189 0.725 0.344
3 1.482 0.983 0.387
1+2 1.573 0.971 0.466
1+3 1.904 1.243 0.538
2+3 2.052 1.467 0.634
1+2+3 2.417 1.650 0.815

The expected yield of a runner on first with no outs is 0.953 runs. Use up an out to move him to second and the expectancy drops to 0.725 runs, a loss of 0.228 runs. The expected yield of a runner on second with no outs is 1.189, use up an out to move him to third and that drops to 0.983, another 0.206 runs lost. And so on.

But that’s only one part of the matter at hand. Proponents of the productive out tend to decree the walk-wait-wallop model of offense in part because strikeouts don’t advance baserunners, which is certainly true. But strikeouts also prevent even more detrimental events such as double-play grounders,. It’s been shown — most recently by Ryan Wilkins at Baseball Prospectus — that far from the conventional wisdom that batter strikeouts are worse than other outs, they have a slight positive correlation with measures of offensive performance such as OPS and Marginal Lineup Value Rate which correlate well with scoring runs.

Olney’s introduction of the POP stat came during his postmortem of the World Series, when he pointed to Aaron Boone’s at bat in the 11th inning of Game Four, with the bases loaded and one out, when Boone struck out. Olney pointed out the so-called significance of productive outs:

There have been 142 post-season series since 1969. In 130, one team or another has had an advantage in Productive Outs — and in 62.3 percent of those 130 series, the team with the advantage in Productive Outs has prevailed. Factor in the 12 series in which opposing teams have tied in Productive Outs, and it can be said that teams with a deficit in POs have won 34.5 percent of post-season series.

The problem is that even in considering the results of a short series, Productive Outs is an unproductive indicator relative to other statistics. A man named Mitchell Below writing a now-defunct blog called Tribescribe did a handy little study of those 142 series, the results of which are here:

Adv.  Winner  Loser   Neither

R 78.9% 17.6% 3.5%
HR 61.3% 26.1% 12.7%
PA* 57.7% 33.1% 9.2%
PO 57.0% 34.5% 8.5%
BB 56.3% 35.9% 7.7%

Productive Out advantage predicted the winner in 57% of postseason series (62.3% if you exclude the no-decisions as Olney did above), a rate exceeded by advantages in plate appearances (for these purposes simply at bats + walks), homers, and runs. In other words, thanks for nothing. Yes, there are certainly times where a productive out comes in handy, and the Yanks might have had themselves World Championship number 27 if Boone had been able to provide one. But such anecdotal evidence isn’t what holds water in this battle. The real question is do productive outs correlate with scoring runs, or don’t they?

I don’t have an answer for that, as I lack the facility to process play-by-play data — it will take a Keith Woolner, a Tangotiger, an MGL to answer that question. But all of this brings me back to what I was talking about above with regards to the incomplete reporting of the team POPs in Olney’s article. If this stat is so damned important, then why isn’t it being calculated on a daily basis? The answer may lie with the players involved.

The Elias Bureau has a very colorful history in its proprietary dealing with statistics; if you’ve ever read any Bill James, you know that Elias’ obstinance is what led to the founding of STATS Inc. and James’ alliance with it. Back when the first Olney article was published, Baseball Musings’ David Pinto, who used to work for STATS, had this to say about POP:

This is Elias playing politics. The Elias Sports Bureau cannot survive without the support of the leagues. What they see is themselves being made irrelevant by the likes of Billy Beane and Theo Epstein, who look to non-Elias people for information. If I’m an owner, I have to start asking why MLB is paying the Hirdts big money to keep stats, when others can do it as well and cheaper. So Elias has decided to appeal to all those GMs who think Beane is wrong.

Pinto, who now works for a company called Baseball Info Solutions along with other STATS vets, has been critical in the past of Elias for being more interested in peddling trivia than in educating fans or media clients about the game. And I’m afraid that short of a usable, testable statistic to add to our arsenal, that’s just what we have here: trivia. I’m not saying that productive outs here and there aren’t important or that they won’t win you a game, but creating a stat which one columnist occasionally pulls out of his ass to selectively support his arguments is irresponsible at best and intellectually dishonest at worst.

That’s the battle sabermetrics faces these days. Grinding persuasion won’t work on some people, while others see the use of statisics as a license to selectively pull numbers out of the air without regard to their foundation. For every convert that Moneyball made, there’s a guy with a career .397 OBP who’s joined the Flat Earth Society and declared that chemistry is what matters, and another who declares that really, it’s what guys hit on alternate Tuesdays when they’ve got the platoon advantage in a road game.

Back to Jim Kaat, Larry Mahnken had some good lines the other day, which I’ll close with because they really sum this whole mess up quite well:

Listening once again to Jim Kaat spout more half-assed comments about how the game is supposed to be played, I’m starting to come to that realization I suppose all people come to as they get older, that you can’t change the minds of the previous generation, you have to win over the minds of the next one. I should let Kaat and Kay’s foolishness roll off my back a little, and focus more on presenting information to the casual observer of sabermetrics in a way that might be more appealing. Let the media and fanboys fawn over Derek Jeter’s defense; we can’t change their minds, it’s the unbiased who we have to educate.

Amen to that.

Postscript: BP’s Derek Zumsteg has been beating POP like a rented bat-boy over at his USS Mariner blog. Rather than making this article even longer by attempting to incorporate his take, I highly recommend you read what he has to say as well.

Post-Postcript: One more thing. After the commercial break following Reynolds’ rant, BBTN went to a quick segment on Rickey Henderson, who signed with the Newark Bears of the independent Atlantic League again. The irony is killing me; somehow, I think it went right over Reynolds’ head.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Spam Protection by WP-SpamFree