Hi I am looking for some help regarding probabilities and probability distributions. I'll try and just break down my situation the best I can:
I am interested in determining the probability of an comeback in a basketball game based off of the time in the game and the teams playing. Let me break that down a little further.
So I have a code running through every game of the last five NBA seasons. It is finding every time a team went up by 10 and recording the following four things in a large matrix- 1) the time in the game when a team went up by 10 2) did the other team ever comeback (bring the game back to a tie) 3) the strength of the leading team (based off of tiers which is based off of vorp) 4) the strength of the losing team.
I need help with the next step. I was hoping to use this historical data to predict probabilities of future games. Say I was watching a game and a team went up by 10, then in a new program I could input the time in the game, and the strengths of the teams and it would output my probability of a comeback. What I'm confused about is how I can make a probability distribution of the event from the 2/3 variables* and then how to fit a model to that.
*Time is one variable, strength of each team could be 2 more variables or I've played around with he idea of using the difference between the two strengths to combine those into one variable.
Also fyi, I can only code in matlab so knowledge of any built in statistic functions would be great but a push in the right direction is really all I am hoping for.
Any help is greatly appreciated,