Do Not Chew Your Stolen Bases

One of the problems faced by Luddites and other haters of Sabermetrics is that they are forced to fight gun battles with knives. The numbers crunchers speak in the language of research, data and measurable results. Cement heads talk about tradition, unwritten rules and a hatred of math.

Keith Hernandez is not a Luddite or a hater of Sabermetrics. In fact, he's one-third of the best local TV broadcast team in the Majors. Along with Gary Cohen and Ron Darling, Hernandez provides more insight in one inning than in an entire week's worth of White Sox telecasts.

But even Hernandez is susceptible to misrepresenting the basics of new analysis. On a recent broadcast, Hernandez claimed that Sabermetrics eschews the stolen base.

Sabermetricians doesn't eschew stolen bases. They don't chew them and they don't es them. They don't cashew stolen bases. Or almond or peanut them. Seamheads don't apply any nut or legume to stolen bases. If you can count, you love stolen bases as much the innumerate fan.

Stolen bases are good. They take runners from first and put them on second. Yay.

On average, teams score .941 runs when they start an inning with a runner on first and none out. With that runner instead on second, they score 1.17 runs. That's an extra run every fifth game. If your opponent offers you an extra run every fifth game, take it.

What Sabermetrics doesn't like is getting thrown out trying to steal. Outs are bad. Far worse than advancing a base is good. In fact, research (no! not that!) shows that getting thrown out attempting to steal offsets slightly more than two successful thefts. 

If that same runner gets nailed instead, his team's run expectancy that inning drops from .941 to .291. He costs his team two-thirds of a run. Not only is he no longer on base, but his team forgoes a precious out. They only get 27 of them.

Which is why stat guys look at steals with a jaundiced eye. The year Ty Cobb set the steals record with 96, he ran into 38 outs. His baserunning, for all the added bases he took, was worth all of two runs. Rickey Henderson is the current record holder with 130 steals in 1982. He was cut down 42 times. He netted his team about nine runs.

Those are commendable, but hardly worthy of a record. For most of my youth, Maury Wills was the steal king. In his 104-steal season of 1962, he was caught just 13 times, a net benefit to his team of 22 runs. So although Henderson broke his record, it came at a stiff price. For his extra 26 steals, Henderson got thrown out 29 times, a miserable tradeoff.

Henderson is the single season base thief god, but Tim Raines was more valuable stealing bases for six consecutive seasons -- in which he swiped 454 of 513 -- than Henderson was in his record-breaking year. (To be fair, Rickey had six better seasons on the basepaths too, including the following year, when he stole 108 of 127. Absent his 1982 season, that would stand as the record without the caveat.)

So Sabermetricians are being rational when they look askance at SB numbers without knowing the CS numbers too. Of course, being rational is another thing they do that annoys Luddites.