Page 1 of 1
Is it just me....
Posted: Mon Jan 04, 2021 3:38 am
by vzografos
...or are game results completely out of whack this season?
Last season around the same time (70-80 games played in total) the markets were predicting correctly close to 70% of the games. This time around I am only seeing around 50%, and some games are really weird. Many sub 30-20% win chance games are been won by the underdogs and some times by large point spreads.
Maybe it is a small sample effect or maybe this COVID season is turning out to be really weird. I wonder how the absence of the home crowd advantage (or other psychological reasons) will affect players and game outcomes. Let's see
Re: Is it just me....
Posted: Mon Jan 04, 2021 12:52 pm
by rainmantrail
It's definitely not just you. This is certainly a strange start to the season. Although I think it was somewhat predictable after watching games from the covid bubble last season. Players are on edge and team chemistries are getting messed with. Covid is having a massive effect on everything around us, and it appears to be taking its toll on the NBA as well. Playing basketball without fans is strange. I considered shorting the win totals market this season almost across the board, betting overs on the bad teams and unders on the good teams because I thought that these trends we saw from the bubble very well might continue through this season. Ultimately, I chose not to place those wagers because I couldn't justify my hunch with data, but watching things play out as they have thus far makes me wish I had.
Re: Is it just me....
Posted: Thu Jan 07, 2021 9:17 pm
by liminal_space
vzografos wrote: ↑Mon Jan 04, 2021 3:38 am
the markets were predicting correctly close to 70% of the games.
you talk a lot about your probability theory and math knowledge but think this is an appropriate way of measuring the market? lol
Re: Is it just me....
Posted: Tue Jan 12, 2021 9:08 am
by vzografos
rainmantrail wrote: ↑Mon Jan 04, 2021 12:52 pm
It's definitely not just you. This is certainly a strange start to the season. Although I think it was somewhat predictable after watching games from the covid bubble last season. Players are on edge and team chemistries are getting messed with. Covid is having a massive effect on everything around us, and it appears to be taking its toll on the NBA as well. Playing basketball without fans is strange. I considered shorting the win totals market this season almost across the board, betting overs on the bad teams and unders on the good teams because I thought that these trends we saw from the bubble very well might continue through this season. Ultimately, I chose not to place those wagers because I couldn't justify my hunch with data, but watching things play out as they have thus far makes me wish I had.
So here is a list of an up-to-date predictor tracking for the season so far
https://www.thepredictiontracker.com/nbaresults.php
As you can see ther results are abysmal.
I have (I may have ) managed to remove the home advantage bias by simply predicting a game as normal as well as if the teams switched and then averaging the probs (I know it is simple). And it does improve the accuracy and Brier but it does not go to upper 60%-70%s like last year. Again accounting for a small sample to-date, maybe the home advantage absence is not the only parameter here. As you said other psychological factors may be also at play.