If I have a model that consistently beats closing lines on most games but happens to be losing - for example I beat closing lines on 75% of games but lose 65% of the time is it better to just hedge the bets out and take the consistent EG?

We've run RMSE numbers and opening lines have been better than closing lines this year in the NBA but closing lines are generally better in past years.

So my questions are: At one point do you decide to hedge all bets you're beating closing lines on? That means we might theoretically be letting the worse bets ride while taking the EV on the good bets when the good bets could have a higher potential EV? What are possible hedging strategies or hybrid strategies? Is what I'm describing even reasonable?

Let it ride.

If you're genuinely beating closing lines in a major sport, the odds on you being behind
after a significant sample size are very small.

Hedging when the line moves in your favour, unless you overbet in the first place or find a +EV hedge is a bad move. You're running bad, suck it up and wait for the tide to turn.

Thanks, I agree with your reasoning except this NBA season is much different than the past few seasons. I actually think the whole market may have the wrong read on what's happening - RMSE of closers is worse than RMSE of openers.

The sheer number of teams not playing all their players on B2Bs, the number of B2Bs, etc make me think that some of our fundamental assumptions are wrong this season. If you look at the past 3 or 4 seasons they look similar statistically. If you look at this year, it's an obvious outlier - HC advantage is way higher this year and from what I've seen, it's not evenly distributed.

Normally we stop a week short of the end of the season to avoid weird situations like teams tanking - given how poorly we're doing this year I'm tempted to shut it down and just skip this year.

If you know your system wins long term and you're still getting closing line value, is it reasonable to think the whole market doesn't completely understand the season and skip it? Or should I have more faith in the market and chalk it up to variance? Sample size is about 170 games and we're running at like 45%...

I am a big believer that the market is mostly efficient.

If there are more B2Bs then it could be a model that mostly uses data from one night stands isn't right?

The other possibility is LOL variance. I don't think you're running that badly over 170 games (5 games below expectation?)

We're 10-2-1 on line movement on today's NBA games. I placed all the bets when Pinny opened them - we'll probably end up going 4-9. In 2 of the games we've accurately predicted > 1 point movements and in 1 game it moved 1 point against us.

The model uses data from the whole season and, without giving too much away, also uses data from previous seasons. The problem is there's so much noise in this season's data. For example, normally we give it 200 games to converge and get the right sample sizes - this year those 200 games included back to back to backs, more 4 out of 5's, more b2bs, and scheduling artifacts, etc. On top of that, there are other scheduling artifacts that the baseline data does a poor job of predicting. Normally it's not as big of a factor because games play fewer B2Bs and 4 in 5s (we know we perform worse in these situations) so we have a better baseline of what teams actually are + there are fewer high variance situations.

I keep coming back to the same theory that maybe everyone's models are doing a poor job of predicting this season. The only other season in the last 6 seasons we would have run this bad is at the beginning of the 2006-2007 season where they changed the ball - in this season if you looked at RMSE's I think closing lines were also worse than opening lines until they changed the ball back.

shutting it down makes no sense to me. If you really are beating closers, at worst you could arb out or middle games.

I think you're probably on to something about the market being out of whack this year, though. Like you said, a lot of assumptions might not work this year,

Murdered line movement on Saturday and ended up going 9-2 then went 2-0 on Sunday going 1-1 on line movement. Went 8-2 on line movement today and ended up going 5-5.

Tomorrow we're 4-1 (4-2 if you count Minnesota but we halved that due to Kevin Love being out) on early line movement. We expect Milwaukee to come back due to Phoenix being on a 4 in 5.