Kevin Durant started the season off by scoring 20+ points in the first 14 games. In fact, in the single game he scored less than 20 points so far this season, he only played 32 minutes. The question is: was his 14 game streak an anomaly or something we would expect him to do?
Our original stat prediction model (explained in detail here and here) predicts he’ll score between about 17 and 43 PTS per game, and there’s only a small chance he scores less than 20 in any given game.
We can extend it to predict streaks. Essentially, any question around streaks can be asked:
How likely is it James Harden makes at least one 3PT in every game this season?
Can Damian Lillard foul out 3 games in a row?
What’s the chance Dwight Howard goes 3 games without an assist?
So returning to our original question: our model gives Kevin Durant a 65% chance of starting off the season with 14 games of 20+ points. On one hand, that’s incredible. On the other hand, maybe not so surprising. One important caveat is the model (see here for technical details) assumes 36 minutes per game. A more dynamic model could incorporate fluctuations in playing time. However, as it stands, the model is saying if Kevin Durant stays healthy and plays a reasonable number of minutes per game, it’s very likely he can keep a 20+ point streak going for long stretches of the season.
What about other players? Here's the probability the top scorers in the league have a 14 game streak of 20+ points.
And just for fun, how likely is it Dwight Howard goes 3 games without an assist? 11%.
And what about a more outlandish one? The probability Steph Curry gets 40+ points 4 nights in a row? 2%. Not bad.
A quick note about the newsletter
There’s been a dramatic uptick in subscribers recently. Thank you so much for giving up your email and signing up. It means a lot to me. Every time I get another notification about new subscribers, it makes me happy that someone finds value in this content. I hope I can uphold my end of the bargain and keep supplying content you are interested in. My day job as a data scientist is much less rewarding than the hours I spend working on these models and sharing them with you all. Please feel free to reach out with anything! Hearing from subscribers is something I’ll never get tired of.
I have a lot more models in the works, and I’m hoping you’ll enjoy them.
If you’re new here, and want to keep up on my models and analysis, feel free to subscribe.