Some methodology recaps/explanations:
- Factors include: record, point differential (pythagorean expected wins), Home vs. Road games, and your strength of schedule to-date. *Note - unlike previous versions, they do not include strength of your future schedule, but I've included it in the picture for your information.
- How you weight each of these factors is difficult to determine. I tweaked until the rankings looked close enough to be defensible to my eye.
- Record and expected record are weighed the heaviest, then opp win%, then home/road splits
- The 'PowerScore' is what's used for the ultimate ranking. It is on a scale of Win%, so comparing the PowerScale to the actual Win% will give you an idea if a team is punching above/below their weight and if so, how much.
- I used a data tool to do the compilation, I think I got it set up so it's a quick download/execute moving forward. That means hopefully I can update after every sim. It also meant a ton of work going into it, so for points purposes I'd like some consideration
Also, if no one has ever used Alteryx, it's awesome, check it out
- Sure, the Raptors played the easiest schedule in the league, but they absolutely dominated it, winning 7 of their 8 games by more than 12 (and the 8th was by 12 even, just missing counting as a 'big win') - any model that doesn't have the undefeated team #1 is broken.
- The Bulls and Nets might be due for regression, as the model doesn't like their home-heavy winning sim against lesser competition.
- The 3-5 Magic and 2-6 Spurs might be better than their records suggest, and have probably gotten a little unlucky thus far. Look for these teams to rebound moving forward!
