Livefyre Profile

Activity Stream

I think that your conclusion that “There is no room for argument. ……..It's not actually an argument worth having; it's dumb and if you disagree, you're wrong. This is not a point for discussion.” may be unwarranted. And, I'm not arguing baseball, I'm arguing metrics.

 

Unless I’m missing something, the data that you present as a basis for concluding that bunting a runner from first to second results in fewer runs being scored – i.e., teams with runners on first and no outs have a 0.941 run expectancy for that inning, and teams with runners on second and one out have a .721 run expectancy for the inning – does not allow us to draw this conclusion, and in fact, is biased (perhaps significantly) in favor of drawing that conclusion.

The bias results from the fact that the 0.941 expectancy includes, by definition, all runs that are scored (from second base with one out) as a result of a successful sacrifice and subsequent RBI hit. So, some portion of the 0.941 percentage is a direct result of the success of the sacrifice strategy, and it is therefore, in essence, an “inflated” metric for the purpose of making this specific comparison.

 

If we want to examine the use of a sacrificing strategy with a runner on first base with no outs, it seems to me that we should be comparing (a) the expectancy of scoring a runner from first base with no outs, absent both successful and failed sacrifice attempts, versus (b) the expectancy of scoring a runner from first base given any (whether successful or failed) sacrifice attempt.

 

With respect to (b), the metrics you cite indicate that if the sacrifice is successful there is a 0.721 expectancy of scoring a run (i.e., from second base with one out), and if it fails there is still a 0.562 expectancy of scoring a run (i.e., from first base with one out). Now, if we know the percentage of sacrifices that are successful vs. failed, we can compute a simple expected value for the sacrificing strategy. I’m guessing that this data may be readily available. Getting the relevant data for (a) may be more problematic, as we need to be able to remove the effects of both successful and failed sacrifice attempts from the 0.941 metric. 

 

2 years, 6 months ago on I have an idea to help the Pirates' moribund offense: STOP GIVING AWAY OUTS | May 2012

Reply