Analytics and 3-point shooting

In regards to "why not just stop shooting 3's once you miss 4 in a row", law of averages.

 

Why worry about it. We had the third best season in Franchise history. Only the 94 championship team and the 97 WCF team with the big 3 did better. The only thing we need is someone like Korver who is money from 3 every game.

 

Stop shooting 3's... and do what, exactly?

 

So far in the playoffs Harden, Smith, Brewer, Terry, and Ariza are collectively averaging 40.18% from 3 and 45.08% from 2. They also average a total of 15.2 free throws made per game.

Let's assume they shoot 100 from each.

(100*.4018)*3= 120.54 points via 3 pointers
(100*.4508)*2 + 15.2= 105.36 points via 2 pointers and free throws

At their current averages shooting 3s is far more efficient, even when points from free throws are added to the 2 pointers.

This is a common misconception. The law of averages refers to large sample sizes for example an entire season. It doesn't work for smaller sample sizes such as a single game. This belief is considered the Gambler's Fallacy.

 

You never know when you are gonna come out of your slump. So yes you just have to keep firing. The gamblers fallacy is in play to a degree.

Compound that with the fact that our offense is designed to shoot those 3's.

 

Analytically, it makes sense to keep shooting because a shot:

1) Guarantees your possession doesn't end in a turnover
2) Even if missed, has a chance to result in an offensive rebound

The only issue is if long misses lead to fast breaks for the other team (partly due to players trying to go for offensive rebounds), which seemed to have occurred numerous times in Games 3 and 4 vs Dallas.

 

Sorry to bum you out, but the law of averages doesn't apply and the Rockets don't go by it.

The statistics that apply here are:

If you flip a coin 9 times and it's heads 9 times, then the chances of it being heads or tails the 10th time doesn't change at all. In basketball, every shot has the same chance of going in, regardless of what has happened in the game to that point - the only thing that significantly changes the likelihood of this going in is how open they are and how good they generally are as 3pt shooters.

The way that the Rockets decide to take 3's is: if you're a decent 3pt shooter on average and that average is acceptable in terms of points per shot then you shoot when you're open. That's it.

If you are money from mid range and your points per shot from there is very good, go for it. But we don't have that type of player and they are rare so it would be idiotic to take more of those or run plays for those. The only time we should be taking those shots is if the player is insanely wide open, but that seems unlikely because...

The best shot in the game is a shot at the basket. So you try to make a shot at the basket and if that's not available you go for the 3pt shooters. Those are the best two options and because the distance between the rim and the 3pt line is such that the other team will inevitably find it difficult to guard both really well then one of these two options is almost always available.

So, yes, statistics does apply here but I have no idea where you're getting law of averages from. Maybe you just heard it around the board or you're friends, but I'm not sure why you would think there's any truth to it.

So if we're not making 3's and we're open on the next 3... we can either go to midrange where we suck or we can go inside where the team has decided to pack the paint. What do we do?

 

Most of the time, slumping happens for no apparent reason. No injuries or anything. The shot feels okay, you aren't doing anything different, but you just can't hit shots you usually hit.

There is no explanation for it. You just have to keep shooting and your shot will come back eventually.

 

You could write a function to determine the estimated shooting average considering all those independent variables.

Collect the data on the shooter in each of the variables you mentioned (injury, defender's skills, fatigue etc.). Run a regression and use those coefficients in your function.

shooting average = intercept coefficient + x1(fatigue coefficient)+x2(defender's skill)+x3(injury) etc.

Just brainstorming though. Would be interesting to see if this was viable. I don't have time today to test it out

 

so when we hit 4 in a row, we should stop? :grin:

morey also said in a perfect offence their wouldn't be a 3 point shot

 

I think a while back there was an actual study done on "hot and cold" streaks in basketball and they found that previous makes/misses have no bearing on the chance of the next shot going in. Obviously there is going to be effects from defense, fatigue, etc. however what we are seeing is really just statistical variance in terms of successes or failures of the shot. Same thing as how you can play poker and run good or bad for months at a time. Even years.

 

The thing is each shot is independent of every shot you've taken, so just because you missed 100 shots in a row doesn't mean you will miss the next one. Feeling good or feeling bad is all mental that's why everyone tells you to keep shooting, because the idea is you are supposed to be a good shooter so the shot should go in, its not like we have De Andre Jordan out there shooting 3s.

Also the law of averages is part of analytics.

 

Not quite sure what you're talking about from my understanding the average here is the players 3 pt %. So if on average the guy has been hitting 40% his entire career then he misses 8 in a ros then chances are he's gonna hit the next 1. Also the Rox use the law of averages...they use it to determine who is good from the 3 and should be shooting them.

 

again you are falling for the gambler's fallacy. law of averages is for larger sample sizes than a single game.