Is rj right about the hot hand and the nerds are wrong?

DistrictDeacon said:

Why is it stupid? Would you rather have a player (in this case Klay) who’s hit his last 4 shots, or one who has hit his last 5?

Click to expand...

I'd take either. There's no difference.

What's totally moronic is you are trying compare the same guy who has shown he's unconscious that night.

I can't edit my post above to say, you need to stop showing that you know nothing about the game and only want to play gotcha.

RJKarl said:

I'd take either. There's no difference.

What's totally moronic is you are trying compare the same guy who has shown he's unconscious that night.

Click to expand...

Klay is actually less likely to hit his shot after going 5-5 than after going 4-4. (See chart in the first post) Does that change your decision?

DistrictDeacon said:

It does say what’s more likely though. And because no one can accurately predict when a hot hand is going to start or end, long term 3P% is the best predictor of who’s more likely to make their next shot.

Click to expand...

Not necessarily. For truly random events with a known probability (like a coin flip or a die roll), the percentage of X over a large enough sample size should converge to the probability of X. So past percentage = current probability.

That isn’t true for a three point shot. The probability that a player will hit his next three point shot is not static from shot to shot. There are too many factors that go into a three point shot for it to be a truly replicable random event. Steph Curry in an empty gym is as close as you are going to get.

Shooting percentage is at best an approximation of the probability a player will make an individual shot, and for some players a very rough one at that.

We can come up with a much more accurate picture when we factor in location on the floor, who is defending, time and score, and yes, how hot or cold the player has been recently

District, you need to stop playing childish games.

Even if that chart had a great number of such events, the difference is 4% or one make every 25 times Klay gets to 4-4. That's definition of inconsequential and undeterminable.

My bad Juice, is 4% not 1 make in 25 attempts?

rj isn't correct, but the most rigorous stats do support a small improvement in shooting if the shooter has made previous shots, because they controlled for actual like, game events

if someone starts making a bunch of shots in a row, the researchers observed they'd feel more confident, take harder shots they might not always take or take them out of rhythm, ad likely the defense would tighten. that was leading to the false negative they'd been getting leading to the theory of the hot hand fallacy. if you control for that, the hot hand effect is pretty well demonstrated with large sample size data.

rj is still way off, but he focuses too much on what it feels like to heat up in a game, or on the observation we've all seen of a player going off. it's not nearly as predictive as it seems. there's still likely a lot of value in trying to run a decoy play for the hot hand and giving the ball to a higher % shooter. the latest data only really suggests if a player has made their previous shot 4 times in a row, on that fifth shot they are observed to hit it at 1% higher than their average. if that hot hand % only gets you to a below average shooter, you'll still wanna try to find a good shooter regardless of how hot or cold they are.

and notably, klay who has a lot of ridiculous clusters in his career also tends to get a ton of good looks even when he's hot because you can't really double him, his release is ridiculous, and he'll pass up a good shot for a better shot in the warriors system, which is light years ahead of dumb basic shit like "give the ball to the dude who's hot"

RJKarl said:

You can't have it both ways. Either you believe in the hot hand existing or don't. All else isn't equal THAT NIGHT.

We're not talking about who you'd want to shoot another night or if each has the same percentage that night.

Click to expand...

I believe it exists. I still think what you said is wrong.

I think a 25% shooter who is 4-4 from 3 on the night probably has a >25% chance of hitting his next three point shot in the same game.

I think a 40% shooter who is 0-4 from 3 on the night probably has a <40% chance of hitting his next three point shot in the same game.

I just don’t think either effect is typically large enough to make up that 15% point difference.

RJKarl said:

District, you need to stop playing childish games.

Even if that chart had a great number of such events, the difference is 4% or one make every 25 times Klay gets to 4-4. That's definition of inconsequential and undeterminable.

My bad Juice, is 4% not 1 make in 25 attempts?

Click to expand...

Even if it’s a small difference, wouldn’t you want to chose the one that has the greatest likelihood?

RChildress107 said:

Not necessarily. For truly random events with a known probability (like a coin flip or a die roll), the percentage of X over a large enough sample size should converge to the probability of X. So past percentage = current probability.

That isn’t true for a three point shot. The probability that a player will hit his next three point shot is not static from shot to shot. There are too many factors that go into a three point shot for it to be a truly replicable random event. Steph Curry in an empty gym is as close as you are going to get.

Shooting percentage is at best an approximation of the probability a player will make an individual shot, and for some players a very rough one at that.

We can come up with a much more accurate picture when we factor in location on the floor, who is defending, time and score, and yes, how hot or cold the player has been recently

Click to expand...

But none of that matters when picking between two players to take a shot. They have both faced variability in the past. You don’t need to know the exact probability of them making the shot, you just need to know who is more likely to make the shot.

Jimmy's gonna be the decoy!

DistrictDeacon said:

Even if it’s a small difference, wouldn’t you want to chose the one that has the greatest likelihood?

Click to expand...

I just looked at Klay's game log for this year. The most your concept could have happened is probably about six times in 64 games. Thus, it could very well take 4-6 YEARS for this to have any impact. Thus, it's totally inconsequential and would be to any rational person. No one would think about this.

Just admit the truth, it's about "getting" me no matter how ridiculous you have to make yourself look.

Haha RJ thinks statistics are about getting him, the ultimate everything is about me statement.

RChildress107 said:

I believe it exists. I still think what you said is wrong.

I think a 25% shooter who is 4-4 from 3 on the night probably has a >25% chance of hitting his next three point shot in the same game.

I think a 40% shooter who is 0-4 from 3 on the night probably has a <40% chance of hitting his next three point shot in the same game.

I just don’t think either effect is typically large enough to make up that 15% point difference.

Click to expand...

I think what you are saying is wrong. Shooting has a large confidence factor. When you get hot, the hoop looks bigger. You are more likely to make the shot.

If you haven't made a shot the entire game, you aren't going to be as loose. You are far more likely to continue missing.

RJKarl said:

I just looked at Klay's game log for this year. The most your concept could have happened is probably about six times in 64 games. Thus, it could very well take 4-6 YEARS for this to have any impact. Thus, it's totally inconsequential and would be to any rational person. No one would think about this.

Just admit the truth, it's about "getting" me no matter how ridiculous you have to make yourself look.

Click to expand...

No ones out to get you, I’m just a trying to figure out why you believe what you believe.

If every 4 years, I gave you the chance to pick an envelope, one that had 63% chance of containing $1000, and one that had a 67% chance of containing $1000, which one would you choose? Would you make the same choice every 4 years?

I would like nothing more than that for people to stop taking pot shots at RJ. Gossett did it here. He know others do it. It's a spectator sport. Sad.

It would be INCREDIBLE if this was last thread of this type that ever shows its face.

Maybe if you held others to the same standards you expect of RJ, it could happen. The problem there are many here who don't want this.

And you know that last statement is TRUE.

DistrictDeacon said:

No ones out to get you, I’m just a trying to figure out why you believe what you believe.

If every 4 years, I gave you the chance to pick an envelope, one that had 63% chance of containing $1000, and one that had a 67% chance of containing $1000, which one would you choose? Would you make the same choice every 4 years?

Click to expand...

I will never make a decision in a game on something on something that may happen once every 4-6 years. There is no logical difference for tonight. It's totally inconsequential.

I usually say that you should neither get sucked in nor play poker with palma. Buckets is getting close to that level. He's played you guys like a cheap fiddle.

DistrictDeacon said:

But none of that matters when picking between two players to take a shot. They have both faced variability in the past. You don’t need to know the exact probability of them making the shot, you just need to know who is more likely to make the shot.

Click to expand...

Que?

Player A is a 40% 3 point shooter. He shoots 46% from the right corner, 36% from the left, and 42% from the top of the arc. He’s been guarded by Robert Covington for most of the night. He’s 1-10 from three, has played 44 minutes on the second night of a back to back, and is dealing with a nagging ankle injury.

Player B is a 32% 3 point shooter. He shoots 42% from the right corner and <30% from everywhere else. He’s 6-7 from three on the night and shooting 44% over the last three games. He’s a 36% 3 point shooter at home, where you are playing tonight. He was a DNP-rest last night.


You have a play that you can run for either player that pretty much guarantees a clean look for that player from the right corner. Player A can also create his own shot from the top of the key. There are 5 seconds left and you need a three to tie.

Which of those three shots is most likely to go in.

Let’s start with the basic question. If you know the actual three point shooting percentage of two guys is 30 and 40 percent, who do you want taking a three randomly?

Also if the hot hand exists does it provide a ten percent boost to address the above if you choose the 30 percent guy above?

RChildress107 said:

Que?

Player A is a 40% 3 point shooter. He shoots 46% from the right corner, 36% from the left, and 42% from the top of the arc. He’s been guarded by Robert Covington for most of the night. He’s 1-10 from three, has played 44 minutes on the second night of a back to back, and is dealing with a nagging ankle injury.

Player B is a 32% 3 point shooter. He shoots 42% from the right corner and <30% from everywhere else. He’s 6-7 from three on the night and shooting 44% over the last three games. He’s a 36% 3 point shooter at home, where you are playing tonight. He was a DNP-rest last night.


You have a play that you can run for either player that pretty much guarantees a clean look for that player from the right corner. Player A can also create his own shot from the top of the key. There are 5 seconds left and you need a three to tie.

Which of those three shots is most likely to go in.

Click to expand...

Well you’re not talking about the hot hand anymore. Not shockingly, this is the analysis that coaches are paid to do, instead of just picking the player who has most recently made three in a row.