UNBAKING THE PIE/1: IS BATTING AVERAGE "MEANINGLESS"?

Caution: Tango Love Pies are more radioactive than ever...

YOU already know our position on the Tango Love Pie: don't swallow, and if you must chew, be ready to spit it out at a moments notice. The damage inflicted on baseball analysis by the Elon Musk of sabermetrics continues unabated by a talented but increasingly unhinged "data architect" who's been given far too much leeway to shape post-neo-sabe discourse and the date it uses to do so. We've been seeing it coming for more than a decade, and this two-part exposé will hopefully put more of that history into focus for you.

Beware the emboldened "barrelmeister" who remains unrepentant in making outrageous claims about traditional statistics in order to feather his own nested constructs--some of which are interesting and useful...but many more of these "advanced stats" are simply reconfigurations of existing metrics that don't advance our knowledge and understanding one iota. Beware such sweeping statements that fly from his fingers whenever there is some chemical reaction that occurs when someone challenges him...

...and one of these cropped up just recently, as he deemed batting average to be "meaningless" in conferring any form of useful information for analyst and fan alike. This is nothing more than a feverish throwback to the overheated "everything you know is wrong" stance that brought us the original neo-sabe movement (take a bow, Baseball Prospectus).

NOW please understand that we are not suggesting that batting average is the best possible statistic for determining the value of a hitter. We've been in the field (but only intermittently under house arrest) for thirty-odd years, which makes us slightly aware of the classical sabermetric work that created bedrock measures such as runs created and on-base plus slugging (and the park adjustments, such as they are, that have supplemented our ability to better discern the value of a hitter's performance).

What we're strenuously objecting to is the Love Pie's patently absurd claim that batting average is "meaningless." We've never seen him quantify this--which is odd, because that is the stock-in-trade of the "take no prisoners" approach that dominates post-neo "praxis" at this point in time. So we'll trot out a few charts that will put the lie to this claim, and we'll expose it as being part and parcel of an ideology that brought the game to a crisis point in the past five years--one that the powers within MLB have been forced to address with a series of rule changes that only band-aid the central ongoing dilemma within the structure of the game.

If we were to accept this hyperbolic claim, we'd be forced to presume that a hitter with a .300 batting average has only a random chance of being a "good" hitter. That there's something like a 50/50 chance that such a hitter is helping his team by, say, hitting at or above league average. 

It's a seductive thought that's floated around in the field for almost fifty years, ever since Bill James suggested it in the late 1970s. But James never suggested that batting average was "meaningless." (By the way, Bill plays some odd form of "rope-a-dope" with the Tango Love Pie over at Bill's site; it's an ongoing siege that is the sabermetric analogue to the Hundred Years War. It's not as interesting as it sounds, and you're not encouraged to plunk down your hard-earned money to witness it.)

SO now let's look at the first chart (at right, above) that we've generated to begin a pushback against the notion that batting average is "meaningless." What is this? It's an historical look at the yearly percentage of hitters with .300+ batting averages who've had an adjusted on base plus slugging (OPS+) of under 100 (in effect, under league average). The red line across the chart shows the overall historical percentage of .300+ BA hitters who've been less than average: that value is almost exactly five percent (5%).

Turning this around, that tells us that .300+ hitters are "good" or better hitters 95% of the time.

Hardly seems "meaningless," now, does it?

The yearly fluctuations on the chart as they move up and down over time are interesting, and could probably benefit from some interpretation. The high-water mark of below-average .300+ hitters occurs in the 1920-1940 time frame, when batting average was at its height (recall that the entire National League hit .303 in 1930). After WWII, there is a long period where .300+ hitters are all above league average (as represented by the cluster of scatter plot points shown at zero on the chart). With the offensive explosion that kicks in during the mid-1990s, however, we see a rise in the number of below-average .300 hitters. 

As pitching finally evolves in a way that brings the offensive explosion to an end, .300+ hitters decline sharply, and those who remain are primarily singles hitters: that combination causes the percentage of such hitters to rise. We can get a better look at how the phenomenon has played out by looking at a companion chart that shows the data in running five-year averages over time. 

This is a much more satisfying depiction of the prevailing trends regarding the frequency of "bad" .300+ hitters over the course of baseball history. What's important to note is that even at the zenith of .300+ hitters whose OPS+ was less than league-average (<100), that figure was only around 10-12%.

It should be clear to you now that hitters with .300+ BAs are, over the course of baseball history, 95% likely to be at least league-average. That is meaningful, even if it remains an incomplete picture of what comprises a hitter's value. We'll stop here now, and resume with a look at the other side of the coin--the ongoing ideology-based skirmish over low batting-average hitters with high isolated power--in our next installment.