We’re certainly living in strange times when a story predicated on the Runs Expected Matrix attributing “runs” to isolated events (stolen bases) appears in a publication that isn’t published by SABR or Baseball Prospectus. Just a few years ago such a story as this one would have been so mind blowing that it might have been possible not to quibble. But, quibble we must.
I think he should have put in context the number of runs attributed to Patterson this year that didn’t come from steals, and referenced the league and team leaders so that we had some idea what it all meant. Maybe Patterson shouldn’t be running, but it’s really hard to argue against stealing when you’re safe nearly every attempt.
Peter, do you have any idea where they get this “run expectancy data” from. I find their results as regards to Patterson rediculous.
BP publishes a Run Expected chart on their stat page. It’s for members only. The value of Patterson’s steals doesn’t strike me as off base, at least not off the top of my head. But I generally know the context of all this stuff and I’m a little baffled by his story anyway. Someone who is coming to it fresh will be mystified, except for the part where he tells them that Patterson’s stolen bases aren’t worth much because of all the home runs being hit.
So, does runs expected go something like this? The chance of a runner scoring from 1st with one out is x%.
The chance of a runner scoring from 2nd with one out is x+%…
Or do they actually take into consideration line-up, defensive strategy, who is pitching, likely bullpen moves, 1st base open—>IBB the next guy, double play in order/not in order..etc, etc?
My guess is the first and that doesn’t seem too useful to me, but I’ll gladly stand corrected. I just can’t imagine that if Patterson put himself into scoring position an extra 25 times, it only led to 4 extra runs…but…could be.
It works this way. If Patterson gets on first base with one out, the expected runs for that inning would be .22 (I’m making these numbers up for expediency sake.). If he stole second, so there was a runner on second with one out, the run expectancy would go up to .52. Patterson would be responsible there .3 runs.
It doesn’t take into account who bats behind him or who is pitching, though somebody could certainly push the fallacy of small sample sizes and make adjustments for all that.
Well, if that’s the way it works, then I put zero credence in it.
Because if they are not allowing for the difference between stealing 2nd with Juan Pierre batting or Derek Lee batting, then it makes no baseball sense. And that’s so obvious as to question why someone would come up with “run expectancy” in the first place.
I first came across Expected Runs in Craig Wright’s The Diamond Appraised. It is a tool for evaluating all the base/out situations and helping to determine the validity of strategies like stealing the base, the sacrifice bunt, etc. There are a number of people who are using it to determine performance issues. BP uses it to give values to relief performances (the pitcher gets credit for improving the base/out run expectancy, a debit for hurting it). I suspect that situational issues (Pierre v. Lee on deck) matter, but to some extent even out when comparing like players. But Patterson’s stolen bases when batting down in the lineup would have a different value than if he has Melvin Mora and Miguel Tejada hitting behind him.
So, their statement that Patterson has netted the Orioles an extra 4 runs is equivalent to “Any runner stealing 26 out of 27 bases on any team no matter where he hits in the order will net their team an extra 4 runs”.
Does that not seem like a horribly misleading blanket statement?
Do they feel the same about singles vs. doubles, bunting etc. I can see it now, Rowand hits a leadoff double and Bell gives himself up with a grounder to 2nd. High fives in the dugout, but the fans Boo! waving their data sheets crying, “Wasted out, on average that will only create and extra .2 runs you &!%$*”