Computer Program for (NHL Moneylines)

Josu · 03-13-10, 05:37 PM

So you got any stats on how this computer program is doing this season? Or are you just starting tracking tonight?

Canada · 03-13-10, 05:41 PM

i like the computer's thinking.

cecil127 · 03-13-10, 05:53 PM

canes showing their moxy lately against some top notch opponents. yotes on a 3 game ws fresh off a big win vs the nucks. should be a good one

not going near it

PRC · 03-13-10, 09:25 PM

i'd love to see more about your computer system.

Flight · 03-14-10, 09:09 AM

You said "computer"

We need more info. Seriously computer models are great but generally will need lots of tuning to become profitable. I'm sure we're all interested to hear what statistics you use to make up your prediction. There are several publicly available statistical based systems as well, one that I am fond of is the Massey Ratings. Takes historical data for the season and comes up with offensive and defensive ratings and then comes up with a most likely outcome.

This approach is not profitable as I have tracked. The problem is the historical data does not predict future results. A markov model or other state machine based approach to the game of hockey is the only computer model I have seen with true promise for predicting hockey outcomes. I do not believe it has been done yet because it applies better to other sports like baseball and football.