I have tried searching the web a lot for this however I cannot find any results and was wondering if anyone here knew of any articles or statistical evidence they had seen.
I am currently working on a computer science course work project and I have chosen to design a machine learning model to predict NBA players performances in their next game.
Obviously the model should make predictions on how good the player is in general and how good the next team their facing is, but is their significant statistical and/or psychological proof that NBA players can peform relative to how they peformed in their previous games.
I chose to do this project with the NBA because of how much data there is, and compared to other sports, games are very frequent so there a greater chance of this being true.
I have watched a video regarding the “hot hand” and how it was proven to be a myth but then later proven to be true(very good video btw, heres the link if your interested https://www.youtube.com/watch?v=CR5vT44ZMK8). Could this be the same with game to game or over the course of several games.
Please let me know if you have seen any articles or statistics proving this. Also, forgive me if this is an obvious answer as I can’t watch the sport as much as I want to because here in the UK most games are started from 12am-3am. Thank you.
You must log in or register to comment.
Definitely. I mean are we forgetting Bubble TJ Warren? Harden had a crazy run in 2018 was it? Kobe, Dame. This will also depend on how many games they have to be on fire for. 2 just enough? or at least 4, 5. That’s up for debate.
As for watching the sports you always have highlights and stat sheets to analyze. Idk why people get so allergic to highlights.
In my opinion, I don’t think that is something that can be predicted. If you are doing the project just to learn ML that’s fine otherwise I would suggest you to try something different considering how much data is available.
Hot hand/Form more of a feel thing. There are times when you shoot you kind of get in a rhythm and that can happen on the first shot you take or may not happen in a game at all, it’s not something even the players can predict for sure.
How could you prove this?
Yes, players can play above their average level, for several games in a row:
• Kobe had a streak of scoring 40+ for 9 games in a row (Feb 2003).
• Wilt Chamberlain scored 40+ for 9 games, and he did that three times in his career.
•Jerry West scored 40+ in six consecutive games during the 1965 post season.
With that being said, what’s different (and more importantly, how could you prove), this was “being on fire”, versus statistical variance?
In order to help you, I think you need to have clearer definitions of what you’re looking for.
Probably some high level regression analytics and mathematical models that the NBA uses to predict stuff. But im too much of a dumbdumb to fathom how theyd do this