Knicks411
Living Legend
- Joined
- Aug 29, 2011
- Messages
- 13,873
- Like
- 12,573
There are more sophisticated ways of doing it. Unfortunately, I haven't really done anything that meaty with statistical models in like 8+ years, so I don't exactly remember what. I want to say though that you could look at historical spread of all records and "fit" your projections to match up with that.
Maybe that one did, but just eyeballing I doubt it because the highs aren't high enough and the lows aren't low enough.
Yeah I'm presuming that's just a model where you sim a season a million times or something and just take the average of all of those trials. Which is obviously going to regress things a ton.