Hey McKealty.
Do you have a distribution of scores over the lifetime of all games for both players?
I'm guessing that this sub-forum is where such a question would be appropriate.
I play a simulation sports game on my tablet and I'm trying to figure out the better scorer between two players.
Granted, one player's sample size is smaller than the other but I hope it can be solved.
So here it goes:
"Player A" scores 7.5 points per game with a scoring percentage of 0.676 across 36 games.
"Player B" scores 10.3 points per game with a scoring percentage of 0.743 in 980 games.
Is it possible to project where Player A will be at the same point in his career, that being approximately 1,000 games, as Player B?
Separately from that, what is Player A's projected points per game and scoring percentage across 1,000 games?
Thank you!
Well the best you could probably do is to use the averages.
If past averages don't give away information for future averages, then the estimate of the next game is the average of the current one. These processes are known as Martingales of which this is just a normal martingale.
One way of prediction would be to solve ax = by = 1000 where a and b are the points scored per game (using the averages as an estimate) and x and y are the number of games for each respective player to reach that.
Since you don't have player data, there is basically no way to give meaningful projection models or to do any kind of inference whatsoever and at this point, the whole thing is rather moot.