Why I Think La Salle Should Shoot More Threes

[JoeFedorowicz]

Feb 20, 2024 19:32:17 GMT -5 GlitterBro #2 said:

Anyone else notice that we shot below our season average for number of 3 point attempts on Saturday and won the game, or was it just me?

Nearing ever so closer to your dream of finishing 14th in the Atlantic 10

[diehardexplorer]

Feb 21, 2024 8:00:44 GMT -5 JoeFedorowicz said:

Feb 20, 2024 19:32:17 GMT -5 GlitterBro #2 said:

Anyone else notice that we shot below our season average for number of 3 point attempts on Saturday and won the game, or was it just me?

Nearing ever so closer to your dream of finishing 14th in the Atlantic 10

triggered 

[GlitterBro #2]

First game against Rhode Island:
- 31 3's attempted. 8 made
- Result - Loss by 2

Second game against Rhode Island:
- 26 3's attempted. 12 made
- 6 of those were with less than 4 minutes left and up by 20 and the team was just burning time and chucking - went 1 for 6 during the final 4 minutes from 3. Take those last 6 away and the team shot 11 for 20 on high quality attempts
- Result - 23 pt win.

But by all means....shoot more 3s, right Joe?

[lasallejohn]

Think the real difference has been defense as opposed to 3s vs. 2s

[JoeFedorowicz]

Feb 25, 2024 17:01:54 GMT -5 GlitterBro #2 said:

First game against Rhode Island:
- 31 3's attempted. 8 made
- Result - Loss by 2

Second game against Rhode Island:
- 26 3's attempted. 12 made
- 6 of those were with less than 4 minutes left and up by 20 and the team was just burning time and chucking - went 1 for 6 during the final 4 minutes from 3. Take those last 6 away and the team shot 11 for 20 on high quality attempts
- Result - 23 pt win.

But by all means....shoot more 3s, right Joe?

We’re not going to agree. I’m glad they’re winning games. I don’t think it translates to tournament wins, but we’ll see. Rhode Island looked like a high school team today.

[MisterD]

There's been a positive correlation between 3s attempted per 40 and win percentage for atleast the last three seasons. I'll go back further if someone else feels like copying down all the data from sports-reference.

[GlitterBro #2]

Feb 25, 2024 20:09:39 GMT -5 MisterD said:

There's been a positive correlation between 3s attempted per 40 and win percentage for atleast the last three seasons. I'll go back further if someone else feels like copying down all the data from sports-reference.

Have that for conference games by any chance? That's the barometer for me.

[MisterD]

As in 3s are bad in conference but good out of conference?

[MisterD]

If you want to find the data for just in-conference records and stats, I can run the numbers (or have my intern do it).

[GlitterBro #2]

A link of you have it to see the 3s per 40 shots you mentioned. You mentioned a correlation but wasn't sure if you meant all games or in-conference only. I'm more curious about the in-conference.

[MisterD]

I just pulled the total stats from sports-reference (here: www.sports-reference.com/cbb/seasons/men/2024-advanced-school-stats.html) but those are totals. There are in-conference records but not stats splits.

(If it does anything for you, teams' in-conf winning percentages also positively correlate to overall 3s per game)

[GlitterBro #2]

Feb 25, 2024 20:09:39 GMT -5 MisterD said:

There's been a positive correlation between 3s attempted per 40 and win percentage for atleast the last three seasons. I'll go back further if someone else feels like copying down all the data from sports-reference.

Your analysis is completely incorrect here.  I extracted the numbers from the link you provided to Excel and ran a quick formula to calculate 3's per game for each team and then run the correlation coefficient (5 minutes of work in total).

Maybe you aren't remembering your basic stats course, but you might recall a correlation coefficient can go from -1 (indicating a perfectly negative linear relationship) to +1 (indicating a perfectly positive linear relationship. A value of 0 means no relationship exists. 

Well, my friend, running the numbers for this season gives a correlation coefficient of 0.16. As you may recall from stats...anything below 0.20 is considered a "very weak" correlation...and essentially negligible for this sample size of 362 teams.

[MisterD]

No, my analysis is fine and your reading of or into it is where things go wrong. A positive correlation is a number that starts with a plus sign rather than a negative correlation (which would start with a minus sign) and, for three straight years, the number starts with a plus sign. I intentionally went back three years rather than relying on one because it was low and since all three stayed around the same range, that's when I stopped and posted here. At no time did I say it was a strong correlation or imply anything more than the very literal "positive correlation".

(I also took the extra step, as stated above, of running it per 40 rather than per game to normalize against OT outliers. And even doing that extra step, it took me a lot less than five minutes per year.)

[diehardexplorer]

[GlitterBro #2]

Feb 28, 2024 11:15:48 GMT -5 MisterD said:

No, my analysis is fine and your reading of or into it is where things go wrong. A positive correlation is a number that starts with a plus sign rather than a negative correlation (which would start with a minus sign) and, for three straight years, the number starts with a plus sign. I intentionally went back three years rather than relying on one because it was low and since all three stayed around the same range, that's when I stopped and posted here. At no time did I say it was a strong correlation or imply anything more than the very literal "positive correlation".

(I also took the extra step, as stated above, of running it per 40 rather than per game to normalize against OT outliers. And even doing that extra step, it took me a lot less than five minutes per year.)

A statistically insignificant positive correlation, defined as "very weak" by stats textbooks. Why you'd use this in an argument as to why La Salle should shoot more 3's is puzzling. Maybe it's my bias working in pharmaceuticals, but the NIH defines a result below 0.30 as a "negligible correlation". (and we haven't even touched on causality, but that's another discussion for another time.

A better metric would be correlating 3pt% to winning%. That jumps the correlation coefficient to 0.50 - a much stronger metric to use. What that tells me, when looking at the two, is that more effective, higher percentage 3s correlates to more wins. 

And that is what this thread is about. Joe's statement of "just shoot more 3s" doesn't work if the team is not shooting effective 3s by driving and creating more open looks on the perimeter as the defense collapses. And, the data for La Salle shows we win more by shooting fewer 3s more effectively. (43rd in 3's attempted per game but 143rd 3pt %).

[MisterD]

Correlating it to 3 pt % might be a great thing to look into but has little to do with the entire point of this thread which was about volume.

Also, and I'm sure this is obvious, if the correlation were like 0.70 or something, we wouldn't be having this discussion because teams would entirely give up on all but the most open 2s. Anything discussed here will be in the margins and, at its most basic, in-conference La Salle is averaging 1.04 points per 3 attempt and 0.98 per 2 attempt assuming the A-10 website is up-to-date. To Joe's point, an increase in volume would lead to an increase in points all things equal. (And no, I'm not doing "more twos would equal greater three efficiency" any more than I'd claim "more threes will open the lane". That's all chicken vs egg.)

[JoeFedorowicz]

Feb 28, 2024 13:38:27 GMT -5 MisterD said:

Anything discussed here will be in the margins and, at its most basic, in-conference La Salle is averaging 1.04 points per 3 attempt and 0.98 per 2 attempt assuming the A-10 website is up-to-date. To Joe's point, an increase in volume would lead to an increase in points all things equal.

Thanks for the grunt work. 

[GlitterBro #2]

Feb 28, 2024 13:38:27 GMT -5 MisterD said:

Correlating it to 3 pt % might be a great thing to look into but has little to do with the entire point of this thread which was about volume.

Which is, in and of itself, an absurd way to look at this. We don't have unlimited possessions and if you become a one-trick pony you become easier to defend. Despite the "grunt work" the reality is La Salle has more losses in conference in games where they shoot a higher number of 3s...because the quality isn't there for a variety of reasons mentioned in this thread by multiple people.

Feb 28, 2024 13:38:27 GMT -5 MisterD said:

 To Joe's point, an increase in volume would lead to an increase in points all things equal.

But all things aren't equal. Put in shot clocks, opponent's abilities, our own team's health, game dynamics, etc, etc, etc. The whole premise of "shoot more 3's" made no sense, and even the attempt to contextualize it with screenshots from one game that show nothing of the flow also made no sense. The data has proven that out for La Salle this season...and last season.

[JoeFedorowicz]

If I cared more I’d run through all the games and remove three point attempts in closing minutes while down X points. But basketball Jesus could come down and say that what I say makes sense and you’d still come up with another reason to continue this fight. But that’s what you do. Here we are.

[MisterD]

Feb 28, 2024 17:36:01 GMT -5 GlitterBro #2 said:

Despite the "grunt work" the reality is La Salle has more losses in conference in games where they shoot a higher number of 3s...because the quality isn't there for a variety of reasons mentioned in this thread by multiple people.

This is like saying "the Phillies have more losses in games where they use more relief pitchers" so therefore they should stop going to the pen so much. When you're trailing, you shoot more of the shots that are worth more points. If suddenly the NCAA instituted a 5 point shot from 45 feet out, guess what trend you'd see from teams trailing late. And even knowing all that, there's still a multi-year positive correlation between volume of 3s and winning percentage.

(I assume the reason you haven't posted anything about in-conference shooting percentages to support your case here is because we're 14th in 2 point FG% and 14th in 3 point FG%?)

[diehardexplorer]

Feb 28, 2024 18:07:56 GMT -5 JoeFedorowicz said:

But basketball Jesus could come down and say that what I say makes sense and you’d still come up with another reason to continue this fight. But that’s what you do. 

pot meet kettle

[jb71]

I'm getting dizzy reading this thread... I'm expecting we'll soon be analyzing 3pt Launch Angles and Exit Velocity..

[JoeFedorowicz]

Feb 28, 2024 18:48:58 GMT -5 diehardexplorer said:

Feb 28, 2024 18:07:56 GMT -5 JoeFedorowicz said:

But basketball Jesus could come down and say that what I say makes sense and you’d still come up with another reason to continue this fight. But that’s what you do. 

pot meet kettle

I’ve been much better the last couple years

[GlitterBro #2]

Feb 28, 2024 18:07:56 GMT -5 JoeFedorowicz said:

 But basketball Jesus could come down and say that what I say makes sense and you’d still come up with another reason to continue this fight. 

I think Jesus was more of a Cross Country kind of guy

[MisterD]

La Salle was up at the half and bled out like -16 over the next 9:30. Over that span they shot four 3s. Definitive proof that you lose when you don’t shoot enough from deep???

[GlitterBro #2]

Feb 28, 2024 21:46:27 GMT -5 MisterD said:

La Salle was up at the half and bled out like -16 over the next 9:30. Over that span they shot four 3s. Definitive proof that you lose when you don’t shoot enough from deep???

Sure. If you want to look at an N of 1 to build your case even though it isn't statistically relevant, have at it. Even a broken clock is right twice a day.

[JoeFedorowicz]

And in that time they took about 8 bad, contested twos and turned the ball over six times.

[MisterD]

Feb 28, 2024 22:23:55 GMT -5 GlitterBro #2 said:

Sure. If you want to look at an N of 1 to build your case even though it isn't statistically relevant, have at it. Even a broken clock is right twice a day.

I know the guy who jumped into this discussion with his “only in-conference matters” after nine (9) games isn’t talking about sample size.

[diehardexplorer]

Feb 28, 2024 22:35:39 GMT -5 MisterD said:

Feb 28, 2024 22:23:55 GMT -5 GlitterBro #2 said:

Sure. If you want to look at an N of 1 to build your case even though it isn't statistically relevant, have at it. Even a broken clock is right twice a day.

I know the guy who jumped into this discussion with his “only in-conference matters” after nine (9) games isn’t talking about sample size.

glitter is going to have to put on his coaching hat and figure out how to handle this double-team. 

[MisterD]

He loves 2s so as long as no one else jumps in he’ll be fine.