Stribor wrote:TTP wrote:bobbeaver wrote:im not so sure about that. His hot streak has been going on for 2 months now with a big 4-5 days break in the middle when he should have cooled off. Im not saying he will be shooting .435 he is doing in the last 2 months, but i dont think the cool off will be so sever. Mostly cause he gets better as the season comes closer to a close. But we will see. I think the number .38-.39% is sustainable.
But it also depends if he will continue to have open shots for 3, which i believe wount change. The fact he has shot some of the screen shots aswell tells me he locked the stroke in. But as i said we will see.
So in the last post you referenced a 19 minute sample size of good defense. In this post you're referencing 2 months of 3 point shooting. You place far too much emphasis on small sample sizes of data.
The last 2 months (since Nov 25), he's taken 133 three point attempts. A 133 attempt sample is nowhere near a conclusive sample - even a 750 attempt sample is ~50% noise. He had over 2.5x those attempts last year and shot 31.1%. In that sample, he's shooting 39.1%. So if you think 38-39% is sustainable, you are basically saying that you think his peak is sustainable. There's just no way that's smart forecasting.
Do you care to explain what do you mean by 50% noise? I assume you assume a stationary binomial probability distribution, but do not get what is the definition of noise? Are you relating to variance or to some significance bounds in Bayesian estimator of true probability parameter???
https://fansided.com/2014/08/29/long-take-three-point-shooting-stabilize/
