I was watching “Baseball Tonight” recap the Phillies and Cardinals game yesterday, and I heard John Kruk explaining a graphic that showed that with Shane Victorino on 2nd with 0 outs in the top of the 7th inning, there was a 66% chance of the Phillies scoring. That graphic was followed by highlights of Phillies players flying out to right and grounding out to third, as further graphics flashed across the screen showing the percentage chance of scoring falling. Then came the next highlight, as with 2 outs, Ben Francisco hit the game deciding 3-run homer.
All of that was fine and good until Karl Ravech said out of nowhere that sabermetricians would have “bunted the runner over to 3rd” to “increase the odds of scoring“, and that the homer “threw all the metrics and ‘Moneyball‘ and all that out the window“. Kruk followed that up with an enthusiastic “OUT THE WINDOW” and Mark Mulder mumbled something worthless like he always does.
I did think it was odd that “Baseball Tonight” kept showing the percentages of scoring, but now I get why they did it, as they were trying to show how probability can’t predict baseball and the talking heads then segued it into a piece about how stat nerds suck.
Because statistics and stat nerds are full of shit, you see.
The problem, of course, is that it’s complete horseshit, and anyone who remotely follows stat nerds would why.
For starters, they never compared the probability of scoring between a runner on 2nd and 0 outs with a runner on 3rd and 1 out, all they did was show the probability of scoring for the runner on 2nd with 0 outs, 1 out, and 2 outs. To put it lightly, no stat nerd in his right mind would ever suggest that bunting in that situation was the statistically sound decision.
Runner On 2nd/0 Outs – 1.06 Expected Runs Scored
Runner On 3rd/1 Out – 0.89 Expected Runs Scored
Runner On 2nd/0 Outs/Top 7/Tied Game – 60.4% Chance Of Winning
Runner On 3rd/1 Out/Top 7/Tied Game – 58.7% Chance Of Winning
As always, those odds are assuming the bunt is successful, which is hardly a guarantee.
I know it’s almost boring to bag on ESPN at this point, and I’ve seriously learned to let a lot of stupidity slide, but if you’re going to insult statistics and stat nerds for being wrong in a sample size of one type of scenario, then the least you could do is figure out what stat nerds would actually endorse.