Fool's Gold - Who Knew?
wrwlumpy last edited by
OMG - Bill Self knows what he talking about.
Well, before you get out over your skiis too far, it’s never been about threes “made.” It’s about threes attempted.
Here are some stats for your consumption:
In 2014, Wisconsin made the Final Four shooting 21 three pointers per game during the tourney. Florida shot 16 per game during the tourney, also getting to the Final Four. And, of course, UConn shot 19 threes per game during the tourney, on its way to winning the national title. During the season, Wisconsin shot 20.7 threes per game and UConn shot 18.52 threes per game. Florida shot 17.74.
In 2013, Michigan shot 21 per game in the tourney on its way to beating KU and getting to the title game. Michigan shot 19.71 per game for the season. Wichita St. shot 20 per game on the way to the Final Four during the tourney, and 19.61 per game that season. Syracuse, who also made the Final Four (whose threes were down a bit in the tourney), shot 17.42 per game for the season. Louisville shot 15 per game during the tourney, significantly higher than our rate in our bad ball stretch. They also shot 17.27 per game during the season.
Now, why did you have to bring up Fool’s Gold on game day? We just made 9/13. Positivity, my man.
wrwlumpy last edited by
I hope Brannen doesn’t read this.
@HighEliteMajor I agree it’s the attempts, but it’s really the expected value per attempt, and the variance in that expected value, that matter.
The expected value of a shot attempt is the probability a shot goes in times the number of points it produces. In general, a 2-point shot has a higher probability of going in, but nets fewer points. In the regular season we made 190/496 3-point attempts (38.3%) for an expected value of 1.15 points per shot. We made 568/1224 2-point attempts (46.4%) for an expected value of 0.93 points per shot. Clearly, on average, a 3-point attempt is more likely to produce more points.
However, those expected values also vary game to game. We played 31 games and in 20 of those games each 3 point attempt netted more points than each 2-point attempt (points per shot). But in the other 11 games, when the long shots weren’t falling, each 2-point shot earned more than each 3 point attempt.
It’s an issue of variance, which is often overlooked when percentages are bandied about in sports contexts. Game by game, there’s more variability in our 3-point percentage than our 2-point percentage. That greater variability suggests the 3 is less consistent/predictable/reliable than the 2. The variance is higher for the 3. Specifically, the coefficient of variation (st dev/mean) over 31 games this season is 39% for our 3-point percentage and 22% for our 2-point percentage. The variance in shot percentage was nearly twice as high for the 3 than the 2 (0.020 vs 0.011), and the variance in expected points per shot was over 4 times as high (0.183 vs 0.042).
I think that’s what Coach Self means by fool’s gold. You can’t rely on the 3 as predictably. It’s more risky. On average it can lead to higher success. But it can kill you in a single game if you count on shooting your average and it just happens to be a night where you’re shooting the low end of the range of percentages. In a single-elimination tournament a good team may be better off with a less risky strategy, one with lower variance.
@HighEliteMajor I’m not saying your strategy is wrong, but that it’s just more risky . Your advocacy of the 3 seems consistent with your belief that a national championship is worth more than a string of conference championships (pardon me if I’m not remembering that discussion correctly). Greater risk can bring greater reward. But the risk also means you can fall flat on your face, and Self appears to favor the more conservative strategy.
But more attempts per game, as you’ve advocated, also may lower the variance. Part of the reason our 2 pt percentage is more consistent that our 3 point percentage is we take more 2-point shots than 3-point shots, so the per-game shot percentage will stay closer to the long-term average. So perhaps if we let them fly, it would be a more reliable strategy?
@tundrahok Of course, I don’t disagree with you. That analysis has been part of the discussion this season. The conversation from my end this entire year is that an increased amount of three point attempts is reasonable because 1) we don’t score reliably at the rim, and 2) we are excellent outside shooting team.
I have posted numerous times that if we’re scoring 65% at the rim, this conversation is different. We’re shooting 56.5% at the rim, and 35.2% on two point jumpers. And 37.5 on three point attempts. Self has actually gone the other way this season, shooting fewer three pointers than we normally do.
One thing that the “three pointer is too risky” crowd forgets is that we lose games missing close shots, too – most notably the Stanford game and the recent KSU game. The “bunnies” thing.
My view is that simply making the three ball a large part of your strategy, and the predominant one this season, would have made us a better offensive team. 20-25 is what I suggested back in late November. That’s the kind of team we have. But sure, we can win other ways.
Now, on to beating WSU … just win.
@HighEliteMajor Yes, I’m a few weeks late in joining this discussion and the timing is lousy to think too much about data analysis. I’m amped!
The key today is to not think about shots… just work hard and take what the defense gives. If they come out high guarding, then punch it in to Perry or Kelly drive the ball.
The Shockers are going to hustle… using their feet to make up for their size issue. Know that and be patient, work hard and we’ll find our shots.
HighEliteMajor last edited by
@drgnslayr And play defense. Dictate the game on the defensive end.
@tundrahok Never too late to join the discussion … your analysis on the value of the three point is right on point. I agree with you on Self’s take on it. But I asked this question through the TT game … would you rather bet your life on Greene making a three or any of our bigs scoring on the post up? Self decided he didn’t like either idea and went with a primary strategy of attacking the basket, and lowering the number of threes.
Time to rock today,
JayHawkFanToo last edited by
Coach Self has always indicated that the best way to start on offense is with great defense, A good defensive stand will result in either a bad shot or turnover by the opposing team or a steal and a break away basket; the '08 team was particularly good at that.
As far as the 3s KU has proven that while good it is really not great at shooting them. It had some outstanding games but also some in which it did not score at all from the 3, in statistics it is called reverting/regresing to the mean. The idea is to find the sweet spot where you shoot at or above your average without getting to the point of diminishing results; it is also nice to know that if your 3s are not falling (we know this all too well) you have an inside game or penetration that can get you points and keep you in the game. Like all things in life, BALANCE tends to be the best approach. I believe this is what Coach Self has been trying to do all along and another Conference Title and a Coach of the Year award later, he seem to have been on the right track all along for the available personnel, which goes to show that he really knows more than we do…:)
And it wouldn’t be a discussion without @JayHawkFanToo citing Self’s resume. Diminishing returns? How is it then that our three point percentage went down the minute we started shooting less three pointers? Seems like we were getting pretty good returns, and we were 20-4. You speak of balance, but our threes decreased in a much larger proportion compared to prior seasons, when our drive it to the hoop game began. And you spoke many times defending Self on the constant post feed game saying we needed to score inside, yet Self abandoned that in large part, just as many of us suggested months ago. He chose a more singular focused drive it to the hoop game. Interestingly, our field goal percentage at the rim has not improved during that stretch, thus the inside game did not improve. Yet our record got worse during that stretch. That’s reality.
I’m done, and I’m going to enjoy the game today. Just win.
JayHawkFanToo last edited by
Well…you and I are going to disagree on this issue and we can both cite all kinds of statistics to justify our views but as Will Rogers famously said…there are lies, damn lies and statistics, so that is a no go.
On the other hand it is hard to argue with Coach Self’s success…and in spite of you saying all season long that he is doing it all wrong, most unbiased analyst called it his best coaching job ever, and in the words of the younger generation…HE"S GOT SCOREBOARD!!!
Lulufulu last edited by
@HighEliteMajor I am positively scared, like Michael
The thing with the Shockers is… they hustle and they never quit. If we bring the hustle on defense and shut down their offense… we will still need to score points at a basic efficiency rate.
The Shockers like to play bad ball, too.
Now you have two teams that prefer a low scoring game.
It really will be about who figures out enough offense to pull away (if the game is not strictly a nail biter).
If it is tight all game, then which team knows how to close the game out.
Possessions will be king. Which team hustles more for 50/50 balls and rebounds? Which team turns the ball over more?
I like our chances at bad ball. I like our chances if we play it right… low TOs, rebound well, and play tough defense. I like our chances that we will figure out more offense than the Shockers. All we have to do is limit VanVliet.
Crimsonorblue22 last edited by Crimsonorblue22
Crimsonorblue22 last edited by
@drgnslayr thought you would want to know, don’t be sorry.
@HighEliteMajor Hey, you’re right!
3 point accuracy increases with increased number of attempts (each point is a different game this year). I didn’t expect that!
And variance seems to decrease with a greater number of attempts too…points are closer to the line when attempts>18. Which violates an assumption of regression, but what the hell.
Sorry, I know you just want to enjoy the game today. I’m actually at work, in my office, failing miserably at focussing on all the work I’d hoped to get done before the game. But you proposed a hypothesis, and I had the data at hand and couldn’t resist taking a look.
Maybe I’ll just give up and go home and have a beer.
JayhawkRock78 last edited by
@tundrahok I can’t remember the formula, but this looks like an out of control statistical process chart that Demming used.
@JayhawkRock78 It’s just a straight linear regression of % vs attempts, ignoring chronology. Simple, probably simplistic, but interesting to me.
jaybate 1.0 last edited by
I have had to learn a good deal of QA over the years. I’m glad you ran the numbers and showed the chart.
This is exactly why I have argued that to really be successful with Trey balling a coach and team need to be schemed to take at least 75% of their FGAs as treys, preferably more.
All core strategies have to be primarily used to become reliable. Self Defense would never become so effective if he did not recruit guys that can play it, and then develop and rely on it.
Every strategy requires tireless development to execute it well.
The only way to get the most out of the Trey ball is to develop it the most…
All balanced attacks still have a cornerstone.
@tundrahok “Maybe I’ll just give up and go home and have a beer.”
Great idea. Cheers and RCJH!!
HighEliteMajor last edited by
@tundrahok Thanks for the analysis … you know I’m going to cite that!
Everyone enjoy the game. Time to put this Shocker nonsense to rest, and on to Cleveland!
VailHawk last edited by
Are you going to the game?