3PT% Problem

[Sun Scorched]

We've all commented on problems relating to offensive efficiency before. I think the drive cap has helped a bunch. I think the pace cap unfairly limited backcourt players, but that's been resolved and it seems the "Sum 20" system is working.

But the 3PT% seems to be the biggest issue currently. I was interested to see how everything would shake out after a full season, but we're still locked in for the GREATEST shooting season from distance in the history of the NBA/CSL. So I wanted to propose the following:

A blanket revision downward of each players FG3 rating. Prolific 3pt shooters would still lead the league, they just wouldn't shoot 50% from distance. I think only a handful of players should be expected to shoot 40+% from range. I mean, IRL, Ray Allen shot 40% for his career. Stephen Curry is 44% so far, but he is best-of-the-best, one-of-a-kind talent.

Thoughts?

[Myles]

It would be like we are moving the 3 point line back. I'm in favor of it.

[KW]

3P% may be high, but 3PA are way, way down from where they are in the NBA. I think this is the issue. The average NBA team let it fly from deep 22.4 times per game. The CSL league leader in 3PA (Orlando) shot just 18.8 a game, which would rank 26th in the NBA below the Lakers and the fucking Hornets. The average CSL team shoots 13.9 3s per game, which would be dead last by more than 1 3PA in the NBA.

50 players in the NBA shot at least 300 3PA last season. 20 in the CSL.
57 players in the NBA shot at least 100 3PM last season. 37 in the CSL.

I think that with more 3-point attempts, the league becomes more balanced and closer to real life. I'd support a small shift in FG3 (-3 for everyone above a rating of, say, 42, and a -2 or -1 for everyone else), but everyone above a certain tendency threshold (so that DeAndre Jordan doesn't start jacking threes) should get a tendency boost for FG3s.

[Sun Scorched]

Yeah Biz - the answer to your questions is a function of preferences though, which individual GMs can impact directly. I mean, it would take an investment of time and player instructions, but you could theoretically shift Klay Thompson to a 100 3pt shot preference - he won't ONLY shoot 3s, obviously, but he will take many, many more than he would at whatever his current preference is.

And we would definitely need to discuss - I'm not sold on a % reduction or an increased reduction for better 3pt shooters. My goal here wasn't to impact 3pt shooter proportionally more than those without a 3pt shot, and I think the "(-3 for everyone above a rating of, say, 42, and a -2 or -1 for everyone else)" methodology does that.

It would be more of, say, a blanket 2 point downgrade for everyone. At least, I think that's what I'd lean towards, until I hear others' thoughts.

[Silogical]

wouldnt -2 on everyone destroy borderline 3 pt shooters. If it comes down, coming down by a % seems to be the most logical.

I think if the NBA shot as few 3's as CSL or CSL shot as many as the NBA the %s would be very similar. How do you raise 3 pt attempts in CSL? (not counting individual instructions)

A % tendency boost seems to be what youre looking for. As more shots are taken from three, the % of makes comes down.

[Sun Scorched]

wouldnt -2 on everyone destroy borderline 3 pt shooters. If it comes down, coming down by a % seems to be the most logical.

I think if the NBA shot as few 3's as CSL or CSL shot as many as the NBA the %s would be very similar. How do you raise 3 pt attempts in CSL? (not counting individual instructions)

Well, I disagree that more threes = reduced %s.

I mean, we have a full season of data now and various players have shot 100+ threes over the course of the season.

And I think it's important to keep every 3pt shooters' value constant, relative to players who don't shoot as well. I'm not a fan of the disproportionate revision.

[KW]

wouldnt -2 on everyone destroy borderline 3 pt shooters. If it comes down, coming down by a % seems to be the most logical.

I think if the NBA shot as few 3's as CSL or CSL shot as many as the NBA the %s would be very similar. How do you raise 3 pt attempts in CSL? (not counting individual instructions)

A % tendency boost seems to be what youre looking for. As more shots are taken from three, the % of makes comes down.

I am under the impression that a global change of -1 or -2 will not destroy a shooter previously shooting 35%, but it won't make a 48% shooter blink. Hence the need for some kind of scale to bring down the stupidly high 3P%s among the top 10-15 shooters, but doesn't affect a below average shooter by more than 1 or 2 points. It would be really counterproductive to go the other way, because a blanket -2 wouldn't have any affect on a guy shooting near 50%, but a -4 would bring an above-average shooter shooting 37% down to 33%, at which point, you're looking at less than 1.00 PPS

[Silogical]

wouldnt -2 on everyone destroy borderline 3 pt shooters. If it comes down, coming down by a % seems to be the most logical.

I think if the NBA shot as few 3's as CSL or CSL shot as many as the NBA the %s would be very similar. How do you raise 3 pt attempts in CSL? (not counting individual instructions)

Well, I disagree that more threes = reduced %s.

I mean, we have a full season of data now and various players have shot 100+ threes over the course of the season.

And I think it's important to keep every 3pt shooters' value constant, relative to players who don't shoot as well. I'm not a fan of the disproportionate revision.

The more 3's you shoot the lower your % becomes. Korver shooting 1 3/game would hit around 60% korver shooting 10/game would only hit around 30%

If you bring down a 34% 3 pt shooter using a fixed number equal to the 50% player you make that 34% 3 pt shooter worthless. He'll be so inefficient he'll have to stop taking 3's.

50% should come down to around 45%
40% should come down to around 38%
35% should come down to around 34%
33% should not come down at all

If -2 for everyone accomplishes the above than perfect. If not youd have to scale the downgrade like biz suggests.

[KW]

wouldnt -2 on everyone destroy borderline 3 pt shooters. If it comes down, coming down by a % seems to be the most logical.

I think if the NBA shot as few 3's as CSL or CSL shot as many as the NBA the %s would be very similar. How do you raise 3 pt attempts in CSL? (not counting individual instructions)

Well, I disagree that more threes = reduced %s.

I mean, we have a full season of data now and various players have shot 100+ threes over the course of the season.

And I think it's important to keep every 3pt shooters' value constant, relative to players who don't shoot as well. I'm not a fan of the disproportionate revision.

Interested to hear why you think raising the tendency wouldn't lower the percentage? I'll give you a real life example of my hometown boy Anthony Morrow- he was sixth in the NBA this season shooting 43.4% from deep on 325 attempts. A huge majority of those makes were assisted, as Morrow mainly shot the shots he was comfortable with, otherwise he looked for a better shot or passed it on.

Now ask Morrow to shoot 425 three-point attempts instead of 325. Now Morrow is taking 1.2-1.3 extra 3s per game that he wouldn't otherwise take - hurried shots with opponents' hands in his face, or 3s off the dribble, and instead of shooting 43% on the 325 attempts he's comfortable with, he only converts the other 100 attempts at a rate of 30%. His percentage drops from from 43.4 to 40.2.

Obviously, I don't expect we'd increase a player's total 3PAs by 30% like I did in the example, but you get the idea.

[Silogical]

Obviously, I don't expect we'd increase a player's total 3PAs by 30% like I did in the example, but you get the idea.

Why not?

using your numbers
CSL average 13.9 3's/game
NBA average 22.4 3's/game

seems like we should raise 3's by at least 30%. Which would cause all %'s to drop. Assuming DDS3 is somewhat realistic.

[KW]

Obviously, I don't expect we'd increase a player's total 3PAs by 30% like I did in the example, but you get the idea.

Why not?

using your numbers
CSL average 13.9 3's/game
NBA average 22.4 3's/game

seems like we should raise 3's by at least 30%. Which would cause all %'s to drop. Assuming DDS3 is somewhat realistic.

I forgot to edit that out- I edited my first example which had him shooting 500 attempts, increasing his attempts by 50% lol. 30% is a but drastic but it could work.

[Sun Scorched]

Well, as a former admin, I guess I'm coming at this from the perspective of how the back end FG3 ratings are structured. Regardless of how many attempts are taken, 50%+ shouldn't be possible, at least not in the context of the top shooters. Anyone in the 3p% leaders column has shot the qualifying number of 3 point shots to be there, so it's not like they've shot 5/10.

You know, Korver's FG3 rating isn't 53 or whatever ridiculous % he's shooting. It's several points lower than that. I'm just curious as to what is causing this inflated discrepancy between the actual rating and his monster 3p%. I think we can all generally agree that Gary kind whiffed on some of these things, hence why we have these conversations - I'm just interested in what the solution could be.

[Sun Scorched]

wouldnt -2 on everyone destroy borderline 3 pt shooters. If it comes down, coming down by a % seems to be the most logical.

I think if the NBA shot as few 3's as CSL or CSL shot as many as the NBA the %s would be very similar. How do you raise 3 pt attempts in CSL? (not counting individual instructions)

Well, I disagree that more threes = reduced %s.

I mean, we have a full season of data now and various players have shot 100+ threes over the course of the season.

And I think it's important to keep every 3pt shooters' value constant, relative to players who don't shoot as well. I'm not a fan of the disproportionate revision.

Interested to hear why you think raising the tendency wouldn't lower the percentage? I'll give you a real life example of my hometown boy Anthony Morrow- he was sixth in the NBA this season shooting 43.4% from deep on 325 attempts. A huge majority of those makes were assisted, as Morrow mainly shot the shots he was comfortable with, otherwise he looked for a better shot or passed it on.

Now ask Morrow to shoot 425 three-point attempts instead of 325. Now Morrow is taking 1.2-1.3 extra 3s per game that he wouldn't otherwise take - hurried shots with opponents' hands in his face, or 3s off the dribble, and instead of shooting 43% on the 325 attempts he's comfortable with, he only converts the other 100 attempts at a rate of 30%. His percentage drops from from 43.4 to 40.2.

Obviously, I don't expect we'd increase a player's total 3PAs by 30% like I did in the example, but you get the idea.

You know, I'm sure I don't know how the sim would handle increased %. Honestly. We all operate from the assumption that if a player shot 1,000 shots we would get closer to their intrinsic shooting ratings, adjusted for offense and caliber of players surrounding him.

But I don't think an extra hundred shots or so will suddenly turn Korver into a 45% 3pt shooter, at least not in the context of the direction this sim league is headed.

[KW]

Well, I disagree that more threes = reduced %s.

I mean, we have a full season of data now and various players have shot 100+ threes over the course of the season.

And I think it's important to keep every 3pt shooters' value constant, relative to players who don't shoot as well. I'm not a fan of the disproportionate revision.

Interested to hear why you think raising the tendency wouldn't lower the percentage? I'll give you a real life example of my hometown boy Anthony Morrow- he was sixth in the NBA this season shooting 43.4% from deep on 325 attempts. A huge majority of those makes were assisted, as Morrow mainly shot the shots he was comfortable with, otherwise he looked for a better shot or passed it on.

Now ask Morrow to shoot 425 three-point attempts instead of 325. Now Morrow is taking 1.2-1.3 extra 3s per game that he wouldn't otherwise take - hurried shots with opponents' hands in his face, or 3s off the dribble, and instead of shooting 43% on the 325 attempts he's comfortable with, he only converts the other 100 attempts at a rate of 30%. His percentage drops from from 43.4 to 40.2.

Obviously, I don't expect we'd increase a player's total 3PAs by 30% like I did in the example, but you get the idea.

You know, I'm sure I don't know how the sim would handle increased %. Honestly. We all operate from the assumption that if a player shot 1,000 shots we would get closer to their intrinsic shooting ratings, adjusted for offense and caliber of players surrounding him.

But I don't think an extra hundred shots or so will suddenly turn Korver into a 45% 3pt shooter, at least not in the context of the direction this sim league is headed.

I think I'll go ahead and suggest what I've been thinking, which is a mixture of both idea:

Everyone with a FG3 rating of 30 and below does not get affected.
Players with a FG3 rating of 31-36 are decreased by 1
Players with a FG3 rating of 37-41 are decreased by 2
Players with a FG3 rating of 42-45 are decreased by 3
Players with a FG3 rating of 46-50 are decreased by 4.

Players with 3P tendency below 15 are not touched. (these are guys that are barely taking threes anyway, like bigs)
Players with 3P tendency of 15-30 are increased by 15
Players with 3P tendency of 31-50 are increased by 25
Players with 3P tendency of 51-65 are increased by 15
Players with 3P tendency of 66+ are increased by 10 (there are very few players in this range anyway)

I'm not an expert on how much the tendency ratings actually affect the shooting perferences, but this is generally what I was thinking.

[Silogical]

Using Korver as an example. Someone would have to go through the other top 3 pt shooters to see if there is any trend.

2015 3.8 Attempts .523
2014 4.5 Attempts .499
2013 4.9 Attempts .458

more attempts lower %

[Silogical]

Everyone with a FG3 rating of 30 and below does not get affected.
Players with a FG3 rating of 31-36 are decreased by 1
Players with a FG3 rating of 37-41 are decreased by 2
Players with a FG3 rating of 42-45 are decreased by 3
Players with a FG3 rating of 46-50 are decreased by 4.

Players with 3P tendency below 15 are not touched. (these are guys that are barely taking threes anyway, like bigs)
Players with 3P tendency of 15-30 are increased by 15
Players with 3P tendency of 31-50 are increased by 25
Players with 3P tendency of 51-65 are increased by 15
Players with 3P tendency of 66+ are increased by 10 (there are very few players in this range anyway)

I'm not an expert on how much the tendency ratings actually affect the shooting perferences, but this is generally what I was thinking.

[/quote][/quote]

I dont know much about the sim so dont know if the numbers are correct, but that concept looks right. It maybe could go even deeper maybe 8 layers deep. if you can go into the decimals. .5 decrease 1.5 decrease ect

[KW]

I dont know much about the sim so dont know if the numbers are correct, but that concept looks right. It maybe could go even deeper maybe 8 layers deep. if you can go into the decimals. .5 decrease 1.5 decrease ect

Definitely can't do decimals, but at that point, we're nitpicking anyway. On the same amount of attempts, it's pretty much a toss-up on who's a better shooter between someone with a FG3 rating of 40 and someone with 41. They aren't guaranteed to shot that percentage.

[Dennis]

I believe nothing said here is really spot on. Yes it is possible that players have an insane good shooting year, but it is also possible they don't hit a single shot at all for a season. Csl is just not the best example as Gary added fluke seasons. So looking at one season is like a sample size of 1. Nll is a better sample size. Compare the shooting %s of players over there over the time of 5-6 years. It shows better how dds works. The shoot more 3 theory is no valid argument because of the fact I explained before. Dds doesn't work that way. If a players takes infinite shots he will shoot exactly his FG3 rating. I believe 3 point shooting doesn't need to be fixed, at least I don't see any problem there. If we would fix that we also would have to fix DRFL. That is the real problem. We have like 20 players shooting more than 10 FTs per game.. That's even way more ridiculous. Is dds realistic? No. It has its flaws. But I think one season isn't a good sample size. Just my thoughts. From the top of my had I believe we have 4 or 5 players with a FG3 rating of 44 or better and like 20-25 with 40+. But I'm not 100% sure. Also most of them are players with very low score ratings.

[Silogical]

Even with 1 season you can see 3PA are low, that isnt going to change much year to year.

If DDS doesnt understand the concept of diminishing returns on 3 pt shots. Then adding more 3's would be a terrible idea. At least in its current form the %'s are realistic. If the NBA shot as many 3's as CSL they would have what seems to be high % on made shots too.

Two options
Raise 3PA and Decrease 3P rating. Tough to do because im sure no one knows the best ratios.
Leave it alone - Just pretend in this alternate reality players/coaches only like really good 3 pt looks. So they shoot less but make more/shot. In this alternate reality teams are run by Karl Malones and Charles Barkleys who dont understand analytics.

OT how do we see these stats FG3? FG3 tendency? ect someone wihtout access to these stats are at a huge disadvantage.

[GreenBear]

Admittedly I didn't read all of the posts, so I don't know if this has been said already or not, but I think part of the reason that 3P% is so high is because a lot of teams run zone defense. You stop running zone defense, the 3P% will drop.

I agree with Dennis in that I don't think it's a big problem. Lowering the 3P rating of players is only going to encourage more zone. And there's a reason NBA teams rarely use zone defense. It would get destroyed by any team with a good outside shooter. But teams in the CSL use zone to cut down on fouls and to counteract teams that are good at driving to the basket.