Saturday, June 21, 2014


The two strains of neo-sabermetrics--"swagger" and "hysteria"--merged and metastasized in issues coalescing around starting pitchers.

The "swagger" involves all range and manner of analytic over-reach, from "new" and "better" ways to measure the true performance of a pitcher to the giddy quantification of any and all facets of the act of pitching.

The "hysteria" stems from the horrible awareness of risk that is first experienced, then exaggerated, and finally institutionalized within insider practices. It is further accelerated by the hyper-economic overlay that has overtaken the sport in the past two decades.

So much counting, so much modeling, so much measuring, so much over-investment in particular kinds of biased counter-intuitivity: we've met the enemy and it's ourselves.

And we are surrounded by holes. Black holes; wormholes; rabbit-holes.

Let's focus on "hysteria" today, as it has taken a back seat to "swagger" over the past half-decade or so, but has reared its head of late as a lamentable litany of young starting pitchers (who embody what we've termed, up top, as the "young phenom") has been providing the type of "night sky" show that enthrall astronomers but fill the neo-sabermetrician with dread.

Too much "swagger" about a subject that turns out to be only remotely understood--such as the continuing profusion of injuries to young pitchers (and in particular "young phenoms")--turns to a puddle of hysterical goo.

What's needed to at least allay these fears? Why, historical perspective, of course. Perhaps if these overcompensating careerists knew that the problem they so brazenly thought they could solve with a series of simple proscriptions (throw less innings, throw less pitches) was actually something that is more or less constant over time, they could find a different path through the forest.

And that's just what we hope to accomplish here, right now--by examining the "young phenom" from a larger historical perspective and with a complete absence of the over-ingenious measurements that actually get in the way of understanding the problem for what it really is.

The chart at right shows all of the "young phenom" seasons since 1901. We are defining the "young phenom" season as one in which a pitcher aged 23 or younger produces at least seven (7) Wins Above Replacement (WAR). There are 43 such seasons since 1901, produced by 35 pitchers.

The color-coding on the chart provides what we might call a "basic survival" rate for these pitchers, who were all worked hard at a young age--some more so than others. The pitchers coded in orange made less than 100 additional starts after their youthful peak seasons; the pitchers coded in yellow made less than 200 additional starts.

So we have a total of seven "young phenoms" whom we can say were, in some way, ruined by their early success. However, one of these phenoms (Babe Ruth) proved to have other more "marketable" skills, and was shifted off the mound. Another (Herb Score) was hit in the face by a line drive and could not recover from the trauma of that injury.

So that leaves five. Meaning that one out of seven, or about 14%, of the top young starters fell victim to arm injury and were unable to sustain their careers for more than a very few seasons after their breakout success.

We would expect the usual counter-argument: that this is a ridiculously small sample size; that we should use a much larger population of pitchers; that we should be trying to examine some more specific factors that point to cause-and-effect.

To which we reply, of course we should. But so many have done just that for so long and have so little to show for it, don't they?

It's better to look at the pitchers at the very top end of the spectrum, because the anxiety over "lost superstars" is the true locus of this hysteria.

If we know that the risk of virtually complete flameout (ironically, it's Bill James' namesake from the "Miracle Braves" who leads the pack here, having started only nine more games after his 1914 heroics at the tender age of 22...) is 14%, then we should be able to perform the simple act of subtraction to realize that six out of seven have at least had reasonably long careers.

In fact, the won-loss records for these 35 pitchers for the seasons immediately following their "young phenom" year adds up to the following: 4115 wins, 3125 losses, .568 WPCT, a 3.32 ERA.

That works out to an aggregate "rest of career" involving 250 additional starts (exact number: 249) beyond the peak (or, in this case, the latest) "young phenom" year (7+ WAR), producing an average subsequent won-loss record of 121-92.

The rest of the data for the "young phenom" group shows that more than a third of the them throw another 2000 innings over the rest of their career, just under a third of them win at least another 150 games, and less than a fourth of them win less than 50 games.

Jim Creighton: he died for your sins even before you committed them...
The problem--or, rather, the risk--involved with young pitchers who are extremely successful has been with us since the beginning of the game. We can only do so much to eliminate it. No matter what we do, it will fail to eliminate the problem. The best may be able to simply reduce the risk by 40% or so.

But it is beyond folly to have assisted in creating an environment where "hot-house flower" young flamethrowers are being created through mechanical techniques that make them virtual certainties for one or more surgical procedures, and then look the other way when assigning blame, or locating the root cause. The higher the K-rate, the greater chance that these altered mechanics are being employed; the higher the K-rate, the greater the chance for injury no matter the age of the pitcher.

The overuse of over-simplistic but so-called "advanced" measurement techniques has resulted in this "hysteria" within the "swagger." When the "swagger" identifies a problem and pig-headedly thinks it has solved it with a couple of math proofs that bear more resemblance to instant coffee than to Chateau Latour, we are greeted with hysteria and bewilderment. Quel surprise...

What's clear from the data above is that before, during and after all of these rather pathetic examples of overthink, the risk rate for "young phenoms" has been pretty constant over time. But, paradoxically, injury rates are arguably higher for lesser pitchers of all ages--and the effects of those injuries for lesser pitchers seems to be greater, more catastrophic (which makes perfect sense, since they have less ground to give before they are too ineffective to stay on a major league roster). However, we don't have a way to definitively quantify that, and until we do, "swagger" and "hysteria" will remain the reigning components in the ongoing banshee wail that passes for "sabermetric discourse."