By R.J. Anderson //
The data used here is about a week old. My apologies for that, however not too much has changed.
You may or may not have caught some of the tweets I sent out about Tom Foley’s decision-making process and the Rays’ baserunning aptitude. If not, let me give you the rundown:
- Using Baseball-Reference, amongst other materials, I gathered each of the situations in which the Rays had a runner on second base. That includes four different base states (loaded, only on second, second and third, and first and second) as well as the varying degree of outs.
- From there, I looked at each of those situations where a single was hit. This does not separate infield singles from outfield singles, or even attempt to classify the batted ball type (i.e. a bloop single versus a frozen rope). I’m hopeful the distribution is normal, if not, oh well.
- To further the, “This is by no means a perfect analysis” points; I also did not adjust for the throwing arm or location of the fielder. Ichiro playing shallow on a ball that quickly reaches him or Johnny Damon on a ball that takes a good minute to reach his body count the same.
Here are the results:
Again, this doesn’t separate the variety of single, so a few of the categories won’t add up perfectly when you try to add the sent and stayed columns. Why? Because there are plays where the runner never even advanced to third base – think an infield single to the third baseman.
Now on to the analysis part where things get fun; after all, how do we know what a good success rate is? Well, obviously 100% is pretty nice, but without any idea of a potential breakeven point there’s no way of knowing whether Foley should be sending more runners or less. That’s why I used a Markov Chain to generate a run expectancy chart tailored for the Rays’ offense. From there, we can plug the numbers into a simple piece of algebra and find out the breakeven points.
Run expectancy of the runner staying at third = (Run expectancy of the runner going home and being safe * x) + (Run expectancy of the runner going home and being out*(1 – x))
Simply solve for x after inputting the RE and you’ll get the breakeven point. One more caveat about this process: I assumed that only our runner would advance an extra base. That means, on a single with a runner on first and second, only the runner on second is advancing to or past third. Again, this isn’t meant to be a perfect analysis.
Below you’ll see the situations with the breakeven points supplied. We still need one piece of information to really confirm/deny whether Foley does a good job. That piece of data is how often he sends his runners relative to other teams. Having a high success rate is great – it really is – but it’s not always the point. It could mean that Foley simply sends his runners only when they have a 95%+ shot at scoring and that would be the wrong play.
Take the situation _2_ with 2 outs. Foley has successfully sent 19 runners who each scored. That’s 100% success rate that far, far exceeds the necessary breakeven point. Maybe Foley should’ve sent those other five, right? Well, as it turns out, Foley is actually one of the most aggressive third base coaches in baseball. As a team, Tampa Bay sees their runners go from second to home more often than all but Colorado, St. Louis, and the Dodgers.
In an optimal world, third base coaches would send their runners until the breakeven point and success rate are nearly inseparable. However, we do not live in an optimal world. In the real world, media and fan perception – not to mention player perception – shift coaches and managers into a shell-like fixation on tradition and the accepted norms. The Rays are generally all about breaking that shell and Foley does a nice job. There might be a reason the Pittsburgh Pirates were interested in him as a manager a few years ago after all.