Hooligans Sportsbook

Matty Rain's 2024-25 NHL Thread

  • Start date
  • Replies
    1,272 Replies •
  • Views 20,609 Views
Can't you use serious maths to figure out the likelihood the model is trash?

If the model has a 5% edge, the minimum to be profitable or whatever, the likelihood of going 5-5744764 is X?

Something like that?
Absolutely. P-value. The P-value of all my bets so far (sample size of about 1500 between baseball and hockey) is high, meaning that it is likely that my results are due to random chance, i.e. I likely don't have a significant edge. Mind you, like VD said, I would need more data to confirm. The lower your edge, the bigger the sample size you need.
 
Shots on goal are prob poisson distribution. I'd find league average shots allowed, divide the nights opponent shots allowed by average and multiply player average shots per game by the inverse of that number. That's the players average shots on goal for that opponent (are they better or worse than league average?). Then I'd poisson the line and convert to odds. If the line you're getting is off by some percentage (i arbitrarilly choose 20%) expected I bet it.
 
I took all my monies out. I could still place paper bets but I dunno. I don't think my shit is working tbh.
At least you ended up on the + side!
292-327 +3.88u :clap::clapping::clapsir:
Nah, forget the paper bets, GL standards don't adhere to Monopoly betting. Plus you probably end up like me when doing so and WIN and be pissed at yourself.