Prejducie Against Batting Average

[John Marsh]

The revised post is below.

[John Marsh]

If one thing unites various baseball stat heads like Bill James, Pete Palmer and John Thorn, or Jim Albert and Jay Bennet, it is their conclusion that managers, sportswriters, broadcasters, and fans habitually overestimate the value of batting average, either as a statistic that reflects the offensive contribution of a player or that can predict the number of runs a team will score. Consequently, stat heads’ favored general managers, like the Oakland A’s Billy Beane for example, have taken advantage of this routine overvaluing of batting average--and consequent undervaluing of other abilities, like drawing walks--to construct unorthodox but quite productive offenses. Moreover, if most OOTP general managers are at all like me, they too have followed Beane and stat heads in neglecting a players’ batting average or ability to get hits and instead focused on other abilities like hitting home runs and taking base on balls. In short, I believe that approach to be misguided, and since it is a tendency I have had to correct in my own decisions about players, I thought I would offer my conclusions to this forum. Assuming my own general manager strategies to be representative, then, we have a collective prejudice against a batters ability to get hits and a pitchers ability to prevent them.

Like most prejudices, the one against batting average has little validity. Take, for example, an utterly average batter, who will have the following ratings:

Hits: 4.5
Doubles: 4.5
Triples: 4.5
Homer Runs: 4.5
Walks: 4.5

Such a player, over the course of 550 plate appearances, will on average produce the following statistics:

AB H 2B 3B HR BB OBP SLG OPS

AVG 496 130 23 4 16 54 .335 .421 .756


We know a player with “4.5” ratings across the board will produce these statistics since anyone who has ever experimented with the player editing function (you know who you are) learns that ratings are determined by various, knowable performances. For example, a “4” rating in the ability to get hits implies a batting average between .237 and .261. A “5” rating implies an average between .262 and .286. Averaging these numbers gives us a batting average of .2615. Following this out for each category, we can determine the average number of doubles, triples, home runs, and base-on-balls such a player will likely produce. Moreover, the “percentage” numbers--on-base-percentage, slugging percentage, and OPS--derived from this hypothetical average player match almost exactly the actual league averages in those categories. In other words, the numbers seem right.

But if a general manager is interested in obtaining the most productive offensive players, which of the five batting categories--hits, doubles, triples, home runs, or walks--most influences a player’s productivity? One way to answer that question would be to improve the player to a “6” (Good) and an “8” (Brilliant) rating in a single category while keeping all other categories “average” and see what happens to his OPS. Since Brilliant ratings in hitting are so rare in OOTP, they are not included; so too any evaluation of abilities in hitting triples, which is an infrequent and largely insignificant ability. Regardless, here is what happens:

AVG 496 130 23 4 16 54 .335 .421 .756
HITS (6) 496 155 23 4 16 54 .380 .472 .852
2B (6) 496 130 33 4 16 54 .335 .442 .777
HR (6) 496 130 23 4 23 54 .335 .464 .789
BB (6) 472 124 22 4 15 78 .367 .422 .789
2B (8) 496 130 43 4 16 54 .335 .462 .797
HR (8) 496 130 23 4 30 54 .335 .506 .841
BB (8) 448 117 21 4 14 102 .398 .420 .818

If we rank these numbers by OPS, here is what we get:

HITS (“6”) .852
HR (“8”) .841
BB (“8”) .818
HR (“6”) .799
2B (“8”) .797
BB (“6”) .789
2B (“6”) .777
AVG .756

What these results suggest is that if you could improve one category of a batter’s ability from “Average” to “Good,” increasing his ability to get hits to “Good” would make a player more productive than increasing any of his other categories. Perhaps most surprisingly, increasing his ability to get hits to “Good” makes him more productive than increasing either his Home Runs or Walks to “Brilliant.” (To be sure, these numbers assume a low “Brilliant” rating of “8”; some players have considerably higher “Brilliant” ratings.)

Even if we are allowed to increase multiple categories, however, we achieve the following still surprising results:

HR/BB (“6”) 472 124 22 4 20 78 .367 .453 .820
HR (“6”) BB (“8”) 448 117 21 4 19 102 .398 .453 .851
HR (“8”) BB (“6”) 472 124 22 4 27 78 .367 .498 .865
HR/BB (“8”) 448 117 21 4 26 102 .398 .500 .898

Increasing both HR and BB figures to “Good” or HR to Good and BB to “Brilliant” still produces ratings (.820 and .851) below that of a “Good” rating in hitting (.852). Indeed, the only way to improve over a player who has a “Good” ability to get hits is to increase HR to “Brilliant” and BB to “Good” or both his home runs and base on balls to “Brilliant.”

The point is that (in my play at least) I have consistently undervalued a player’s ability to get hits and consistently overvalued a player’s ability to hit home runs and draw walks. I hereby exorcise Billy Beane and Earl Weaver from my OOTP soul!

But we should not also cast Billy Beane into the base on ball circle of hell. Beane general manages in a league where few care about a players’ ability to draw base on balls or a players’ OPS, but in which *many* care about whether a player hits .300. Beane therefore takes advantage of an irrational market, one which undervalues base on balls. But OOTP is another league all together, one which *does* care about the value of base on balls and a players’ OPS. It is impossible, therefore, to take advantage of any market irrationality.

The conclusions to be drawn from this study, then, are as profound as they are dull: Seek players who have “Good” abilities to get hits, or players with a “Brilliant” rating in Home Runs or Base On Balls and at least a “Good” ability in the non-“Brilliant” category. Or, best of all, players with “Good” talent in hitting and “Good” or better ability in HR and BB.

We reach similar conclusions in evaluating a pitcher’s value. Instead of starting from a pitcher with “Average” ratings, however, let’s start with a pitcher with “Good” ratings across the board. He would produce the following statistics:

AB H 2B 3B HR BB RUNS IP R/9

6-6-6-6 490 123 20 3 15 60 67.9 122.1 5.00

Pitchers do not have ratings in “Triples,” but that number can be derived from the number of other extra base hits they surrender. In any given season, the number of triples will be roughly .075 the number of doubles and home runs combined. The same should apply to a given pitcher. The RUNS figure is based on Bill James’s simplest Runs Created formula, and is the number of runs a pitcher is likely to surrender given the number of other events he surrenders. IP is at bats minus hits (the number of outs) divided by 3, and R/9 is the number of runs a pitcher can be expected to give up over the course of nine innings. It is similar to ERA, but more valuable, I believe, since ERA masks the number of runs a pitcher actually gives up. It doesn’t matter how many runs a pitcher “should” be credited with giving up, but with how many run-producing events he actually does give up. A pitchers’ performance after errors is at least as important as his performance before them and should not be ignored. ERA assumes a perfect world where errors never happen. But errors happen all the time--they are as predictable as civilian casualties in a bombing campaign. We should therefore not pretend that when they happen they are a “mistake.”

Nevertheless, what happens when we increase each pitching category--hits, doubles, home runs, and walks--to “Brilliant?”

AB H 2B 3B HR BB RUNS IP R/9
8-6-6-6 490 98 20 3 15 60 48.5 130.2 3.34
6-8-6-6 490 123 10 3 15 60 61.2 122.1 4.50
6-6-8-6 490 123 20 3 8 60 57.6 122.1 4.23
6-6-6-8 513 129 21 3 15 37 60.7 128 4.27

For pitchers, an increased ability to prevent hits is even more valuable than a batter’s ability to make hits, preventing, on average, nearly a full run more than a pitcher with “Brilliant” abilities in HR and BB and more than a full run than a pitcher with a “Brilliant” ability to prevent doubles. For pitchers, the ability to prevent walks is about as valuable as the ability to prevent home runs, which is less the case for batters, where the ability to hit home runs is considerably more valuable than an ability to draw base on balls. But the conclusions are the same: Seek pitchers with a “Brilliant” ability to prevent hits (or a “7” rating in that category) and, secondarily, pitchers with a “Brilliant” (or “7” rating) to prevent home runs and base on balls.

These conclusions may seem obvious to many OOTP players, but for someone who has only recently begun playing, it took longer than I would like to admit to shed my own prejudices and biases. So I hope this research helps. I welcome corrections. Best,

John Marsh

[Henry]

John,

Excelent analysis

I noticed the same thing when I did my ballpark factor analysis. "Hits" (or AVG as it is represented for ballparks) is an "acculmulated" factor.... meaning if you raise average, you are ALSO raising doubles, triples, and homeruns per the players ratings ratio. You get "more for your buck" so to speak.

When you increase "only" doubles, triples, or homeruns - and "hits" (or AVG) stay the same, your simply replacing one type of hit with another - thus a smaller effect.

[JAttractive]

Great work John!

(note you should just edit the first post and erase it to tell people to view the second one... Most people like me will read it through only to notice the proper one below).

[Sudy Nym]

While I certainly haven't read them all, I would have to say that this is the best study I have ever seen here at the .400 Studios forums. No doubt that it could be done in even more depth, but its conclusions are concise and absolute.

Brilliant job. You're right that most of it might actually be common sense, but this only verifies our suspicions.