Odds of going 16 for 25 on threes

[Suebird10fan]

For all you stat geeks ... if you assume that every shot is an independent event (i.e. you don't believe in the idea of a "hot hand"), and you assume Louisville normally shoots 33% on threes (slightly above their percentage for the year, but that included one player going 3 for 21), then the odds of going 16 for 25 are almost exactly 1 in 1,000.

 

[Phil]

There's a lot of analysis suggesting that the so-called hot hand is a statistical myth, but I don't buy it. It seems unlikely that the probability of hitting a three would be stationary. There are lots of reasons to think it could vary. If it didn't vary, we would expect that the distribution is Poisson (which is how I assumed you did the calculation) but if Poisson, then the variance of the ratio over all games would be 33%. I'm too lazy to look it up and do the calculation, but I bet it is higher, in which case a negative binomial would be a better choice. Would reduce the odds, but it was still a phenomenal achievement.

 

[60'sfan]

The amazing thing for me was the most of the threes were contested. If they were getting open threes I might have been less impressed.

 

[Tenspro2002]

Don't believe in a hot hand? Just ask the kid on Michigan yesterday.

That said, the chances for Louisville to shoot 16/25 was astronomically small.

 

[wire chief]

It shows that Jeff is one up on Geno, for the latter has yet to show he can teach his team to make 2/3 of their 3's.

 

[stamfordhusky]

There's a lot of analysis suggesting that the so-called hot hand is a statistical myth, but I don't buy it.

I don't buy it either. Every one of the analyses I have seen were deeply flawed in one way or another. For one, players who have a hot hand often start taking tougher shots. Also they start facing much tougher defensive pressure. Both tend to push their shooting pct back towards the average - even if they do have a hot hand.

 

[Icebear]

There are lots of reasons to think it could vary. If it didn't vary, we would expect that the distribution is Poisson (which is how I assumed you did the calculation) but if Poisson, then the variance of the ratio over all games would be 33%.

Poisson? Does that mean its fishy?

 

[JoePgh]

Not to mention that one player's shooting is hardly independent of her teammates' shooting -- from casual observation, it seems that a team can be often collectively hot or cold. If a team relies heavily on 3-point shooting and their best shooter misses a few, the others get panicky (especially if they are behind on the scoreboard) and start feeling anxiety to make the shots that their best shooter is missing. 3-point shots made after the game is decided are much easier (and I bet that an empirical analysis would demonstrate a higher percentage of makes in that circumstance).

That implies that Louisville's shooting last night was especially noteworthy because those shots were made under pressure, with a huge game far from decided.

 

[Jim]

It shows that Jeff is one up on Geno, for the latter has yet to show he can teach his team to make 2/3 of their 3's.

Not so. Holly Rowe tweeted from yesterday's practice -- "Huskies track rapid fire 3 point shooting drill. Made 111-175 64% WOW. Kaleena Mosqueda-Lewis 24-31."

 

[Icebear]

Not so. Holly Rowe tweeted from yesterday's practice -- "Huskies track rapid fire 3 point shooting drill. Made 111-175 64% WOW. Kaleena Mosqueda-Lewis 24-31."

Yikes! Kaleena at 77%.

 

[JoePgh]

Needless to say, that wasn't done under game conditions against the #1 team in the country, in a lose-and-you're-out situation. I think wire_chief is right.

 

[intlzncster]

There's a lot of analysis suggesting that the so-called hot hand is a statistical myth, but I don't buy it.

From an ex statistician, this is where statistics come up short, trying to quantify the un-quantifiable. There's plenty of demonstrated research about 'the zone' athletes get into, particularly in basketball, where players often describe the basket looking as big as the ocean. It becomes very interesting when a team becomes hot, players feeding on each other's success.

 

[doggydaddy]

Needless to say, that wasn't done under game conditions against the #1 team in the country, in a lose-and-you're-out situation. I think wire_chief is right.

I think thinking that tweet about the 3 point shooting was real is pretty darn funny.

 

[DobbsRover2]

Even statisticians fall victim to the fallacy of the mean (not to be confused with the fallacy of being a mean person, as they just slurp sour lemonade). Just because a team shoots 33% on 3s for a season does not mean that they are that likely to shoot that 33% in any given game. Over the course of a season a team will have games where they shoot abysmally (like UConn's 0% in an ND game) and just marvelously (like when UConn hit 15 of 23 for 65.2% against Marquette). UConn did shoot 37% on 3s so far, but it still is in the ballpark with Louisville and it is not true that the Cardinals have a 1 in a 1000 probability and UConn maybe 1 in 900 of shooting around 67% from behind the arc. The odds are far far greater because statistically sports action is a series of peaks and valleys rather the steady plane that the overall stats seem to indicate. And just like in UConn's Marquette game and Louisville's Baylor game, the hot shooting happens far more often than we think. Note that UConn also had a 12-20 game for 60% against Hartford this year.

And for those who say it is harder to do it under the pressure of a big game, maybe that's also when it's easiest for players to will themselves to a great performance.

 

[intlzncster]

And for those who say it is harder to do it under the pressure of a big game, maybe that's also when it's easiest for players to will themselves to a great performance.

The better the player, the more likely this is. Big games sharpen concentration/focus. Some players get the yips, but the best generally shine (that's what makes them the best).

 

[Suebird10fan]

I agree that there must be some sort of hot hand effect. How could there not be? But I suspect it's smaller than most of us would think. Anecdotally I find that any time I look at the frequency of rare events in baseball, the actual frequency corresponds much closer to theoretical (i.e. assume no hot hand effect) than I would think.

The odds (assuming no hot hand effect) of 12 out of 20 for UConn based on a 37% season average is 1 out of 50. Much different than 1 out of 1,000.

My guess is that Walz got his players into a zone of feeling that the pressure was all on Baylor. But I'm also guessing that pure luck was a much bigger factor.

 

[alexrgct]

What were the odds of hitting that spinning back to the basket layup around Griner that Shoni threw up? That might have been 1:1000.

 

[UConnChapette]

What were the odds of hitting that spinning back to the basket layup around Griner that Shoni threw up? That might have been 1:1000.

I still can't figure out how Shoni got that shot to fall. Defies logic.

 

[msf22b]

She deserved to yap after making that.

 

[Samanthabrown3]

She deserved to yap after making that.

And this is why I read the Boneyard. I love you guys!

 

[huskyharry]

For all you stat geeks ... if you assume that every shot is an independent event (i.e. you don't believe in the idea of a "hot hand"), and you assume Louisville normally shoots 33% on threes (slightly above their percentage for the year, but that included one player going 3 for 21), then the odds of going 16 for 25 are almost exactly 1 in 1,000.

Baylor's defense helped a lot...no adjustments were made when they saw that Louisville was red hot. Baylor's perimeter defenders really needed to contain the UL players, not attempt to trap the ball and steal it from them. BG could have still be a roamer inside the 3 pt line but her player needed to be covered by the others and rather than attempt traps and steals they needed to stay in front of the UL players...I, for one, am very glad that they did not!