- Joined
- Jan 6, 2008
- Messages
- 6,081
- Reaction score
- 6,381
Here's a question I'd be interested in the answer to. Let's assume a prior win proability of .92, which I think results in about a 20 percent chance of 19-0. What is the likelihood that team will finish 18-1?
A few teams have finished 18-1. I'm trying to get a mathematical idea of how much more rare it should be for a team to finish 19-0.
Replace one of the P(WIN) for an individual game with 1-P(WIN), so 18 times P(WIN) times 1-P(WIN). But there are 19 places in the schedule they could have lost, so it becomes more complicated as you have to start to do combinatorics.
There are 19 ways to lose one game, so you have to add the probabilities of each of those seasons (e.g., loss in first game OR loss in second game, etc). So 19*P(loss in game 1, win in remaining games) . You can add them because the outcomes are independent (like the probability of rolling a six or a one: you can add the individual probabilities because there is no overlap in the sample space).
Anyway, doing the calculations, the probability of 1 loss, assuming .92 prob of win per game, is:
0.34.
Probably of winning all 19, assuming same prior is 0.205.
I think that's it anyway. Someone let me know if there is a mistake.
It would be cool to make a 3-d plot: probability of win on x axis, number of games lost on y axis, and z axis is plot of the probability of that combination. It would look sweet. It would take me an hour to put it together though...so not unless I really need to procrastinate... It will get even more complicated combinatorically when you have more than one loss (there are lots of ways to have two losses in a season: first and second game, first and third, second and third, etc etc). It would become a mess.