Sunday, May 6, 2007

Possibly stupid softball question

For the last few years I've been on softball teams made up partially of novice softball players---usually foreign student who have never played softball before. At least for their first year, they typically only reach base on errors. Along with these newer players there are usually a few really good players who almost never get out as well as a number of ok players who hit safely somewhere between 1/3 and 1/2 of the time.
The question I have is what is the optimum lineup for such a situation? The rules in the league are that everyone at the game gets to bat. The two approaches I see most often are stacking the front of the order and interspersing the good players among the bad. I don't really have a feel for which works better, so I'm throwing the question out to anyone who may know "the answer" or at least have an idea. I'm wondering if the answer changes depending on the relative numbers of each set of player or if the final answer is that batting order doesn't matter at all.


Alan Rosenwinkel said...

In general (for baseball) each out in an inning costs less (as measured by run expectancy) than the previous one. Also, an out with a man on base costs more than with the bases empty. I suspect putting all the bad players together would minimize the chances that they would come up with men on base and maximize the chances for them to get out in the same inning, thus minimizing the total cost of their outs. I'll have to run my baseball simulation on this later though, and see what it comes up with :-)

dave hiller said...

Ok Matt, I just told a story yesterday to people on my softball team about you running a simulation to answer this question. As I recall (if this ever happened) you said dividing the lineup into two halves, with good players at the front of each half, was the best. This strategy was especially useful in my league, since we have to bat three men, two women, then repeat. So I hit possibly our 2nd best male hitter 8th, so he would hit after two men instead of two women.

The more general case is more difficult I think. In baseball it doesn't make much difference - and presumably the biggest contribution is getting some extra at bats. I found this study you can look at about the 2000 Blue Jays. However, in softball, we have really extreme performances, so I wouldn't be surprised if you could make a big difference. When you figure it out let me know.

Matt said...

I don't think I've run the simulation before, but now that you've said I have I guess I have to.

Matt said...

of course, I may have in the past, but my memory is shot these days. I was looking for an excuse to write a softball statistics simulator anyways.

Alan Rosenwinkel said...

I already have a baseball simulator. I'll run the simulations, but it will have to wait until after the wedding.

php hit counter