Playing with Volleyball Stats

I am a volleyball nut. I have been caught by my wife watching a college volleyball match on TV and having a browser window open to an international match on youtube. I am also a bit of a geek. I'm dipping my toes in the programming pool, and of course I am trying to figure out how I can use this new knowledge to make my use of stats even better.

Perhaps I should mention that I spent the last few months coaching a high school JV girls volleyball team, and I am looking at possibly being the boys varsity head coach in a couple months. Combine all that with my past experience with college teams and their use of stats, and we have an odd geek niche that I am looking to explore.

So tonight I was having a bit of trouble sleeping. I decided to jump on the computer and make a little volleyball simulation that takes a team's side out percentage, their opponent's side out percentage, and the length of the season, and it gives you a win-loss record. I have been told that if both teams have an identical side out percentage (let's just say 50%), they will end up with about 15 wins and 15 losses. I have also been told that if you can improve to 51% and hold your opponent to 49% (but it may have been 52% and 49%), your record improves to about 20 wins and 10 losses. I wanted to set up a simulation that demonstrates that.

So I made this thing that does a coin toss to determine who serves, and it then plays a game with the receiving team having their side out percentage as a probability to win the point. Side out and the serve alternates. Serve, score, rinse, repeat. When the score hits 25, and the team in the lead has at least a 2 point lead, that team wins the set. The fifth set only goes to 15, and first to 3 sets wins.

I dialed up the season length from 30 to 30,000 matches to even out any wild swings in the random numbers. At a 50% side out percentage for both teams the win-loss record was not surprisingly 15-15 (after dividing by a 1000 to bring the superlong season back to reality). For 51%-49% it changed to 18-12. 52%-48% went 21-9. 53%-47% was 24-6. That is a pretty big spread in wins and losses for such a small difference in side out percentage. The numbers are pretty close to what I was told before.

After spending way too much time playing with that, I thought I could alter it in the future to include the different side out percentages for each serve receive rotation to see how changing the initial serve receive rotation alters the win-loss record. Maybe on another sleepless night...