As the season moves at a brisk pace, I attempt to do the same. On Sunday the Portland Trail Blazers take on the Sacramento Kings. Here’s my prediction for Damian Lillard’s first quarter shot selection:
(click to enlarge)
First, let’s address something. You may be thinking: “Ooh, two shots, what a bold prediction. Not impressed.” Fair, maybe. However, the process I’ve adapted here is becoming increasingly detailed and increasingly better – linking the how many with the where is not exactly a blindfolded dart throw (although if this fails that may be the next strategy). Further, less is more in cases like this. It’s relatively easier to over-predict, or just shoot for buffered averages that encompass nearly everything. What I’m attempting to do is find patterns of behavior interlaced among all of Lillard’s activity; I would describe it as a signal within the noise, but I feel like that’s already been taken.
This leads me to my second point.. but why two shots? Admittedly, I’ve been a little sparse in some of the details lately (which may or may not directly correlate to my annual bitterness/motivation from lack of attendance at the MIT Sloan conference); I tended to elaborate more on the Melo predictions. So, consider this:
- There is an increasingly negative correlation between a previous game’s shot count and the next game’s 1st quarter shot count; as the previous game shot count goes up, Dame’s next game goes down (he took 22 shots on Friday; on the high end for Lillard)…expect a low quarter on Sunday
- This season, following a 5-shot 1st quarter (like Friday vs. OKC Thunder), Dame never shoots more than that in the next game’s 1st quarter…expect less than 5 shots in the 1st quarter on Sunday
- Further, this season that following game’s 1st quarter has followed a rotating pattern of 4 shots/1-2 shots (the latest occurrence was 4 shots)…expect 1-2 shots during the 1st quarter on Sunday
- There is another growing negative correlation: previous game’s minutes played and the next game’s 1st quarter shot count…he played almost 37 minutes Friday, on the high-ish end for Dame, expect low shots in the 1st quarter Sunday
- And finally, some icing on the cake: during Dame’s career vs. Sacramento, in the 1st quarter he either takes 4-5 shots or 1-2 shots… ever
The shot locations don’t change much. At the rim and a wing 3-pointer remain constant; and realistically they likely always will, when the shot count is expected to be so low. If, IF, Dame were to take a third shot that quarter, I would expect it near the foul-line; that’s his next most common spot on the floor vs. the Kings. While the locations seem boring, they are the product of intense examination of where Lillard is most active (and successful) in these situations.
Next week full 1st half predictions begin, pending disastrous results – so far each predictions is getting better and better (nothing like linking your own tweets). These improvements are expected; the goal is to shorten the learning curve every time. Finally, on the topic of famous schools in Boston and predictive analytics, if you’re wondering if a computer could do this better than me, I’d argue yes, probably – I’m getting there, eventually.