It's time for...
PLAYING. WITH. NUMBERRRRRRSSSSSS!!!!!
This is the game in which Brad uses random statistical tools to try to make sense of stats and figures that most likely won't produce any reasonable outcome! This week's game will be centered around the most exciting part of All-Star weekend...
Not the Slam Dunk Contest.
Not the All-Star Game.
THE 3-POINT SHOOTING CONTEST!
Before I continue on this charade that I'm calling an article, these stats were derived from the players in-season numbers (for the entire season in full) during the season in which they competed in the 3-point contest. I've pulled together shooting stats from the past 10 years of the 3-point contest, focusing on the 3-point attempts per game and the 3-point shooting percentage.
Without further ado...
PLAYING. WITH. NUMBERRRRRRSSSSSS!!!!!
This is the game in which Brad uses random statistical tools to try to make sense of stats and figures that most likely won't produce any reasonable outcome! This week's game will be centered around the most exciting part of All-Star weekend...
Not the Slam Dunk Contest.
Not the All-Star Game.
THE 3-POINT SHOOTING CONTEST!
Before I continue on this charade that I'm calling an article, these stats were derived from the players in-season numbers (for the entire season in full) during the season in which they competed in the 3-point contest. I've pulled together shooting stats from the past 10 years of the 3-point contest, focusing on the 3-point attempts per game and the 3-point shooting percentage.
Without further ado...
The basic rundown of the 3-point contest field is as follows:
From the names alone, you might be able to tell why this years contest is far more exciting than any of the other ASW festivities. Not only is the field star-packed, but the players competing are legitimately the best three-point shooters of their generation. Kyle Korver in particular is on pace for a 50-50 year (50% FG% and 50% 3PT%), a feat that hasn't been accomplished since Steve Kerr in '95-'96 (>50 3PA).
Now, since the field is filled with such amazing shooters, I thought it might be fun to look at who has the highest chance of winning the contest, based solely on a regression from the past shooting contests. I spent much more time pulling individual season stats for each of the players in the contest for the past 10 years and compiled my own regression to try to predict who has the highest odds.
Disclaimer: The actual correlations on these stats was actually pretty god-awful, but it's still fun to see if any of it holds true.
The base equation that I ended up with was as follows:
- Marco Belinelli, SAS, 3.5 3PA/G, 38.1% (Winner from last year)
- Steph Curry, GSW, 7.9 3PA/G, 39.9%
- James Harden, HOU, 6.9 3PA/G, 38.3%
- Kyrie Irving, CLE, 5.3 3PA/G, 41.4%
- Kyle Korver, ATL, 5.8 3PA/G, 52.3%
- Wesley Matthews, POR, 7.6 3PA/G, 39.8%
- J.J. Redick, LAC, 5.5 3PA/G, 43.6%
- Klay Thompson, GSW, 7.0 3PA/G, 44.0%
From the names alone, you might be able to tell why this years contest is far more exciting than any of the other ASW festivities. Not only is the field star-packed, but the players competing are legitimately the best three-point shooters of their generation. Kyle Korver in particular is on pace for a 50-50 year (50% FG% and 50% 3PT%), a feat that hasn't been accomplished since Steve Kerr in '95-'96 (>50 3PA).
Now, since the field is filled with such amazing shooters, I thought it might be fun to look at who has the highest chance of winning the contest, based solely on a regression from the past shooting contests. I spent much more time pulling individual season stats for each of the players in the contest for the past 10 years and compiled my own regression to try to predict who has the highest odds.
Disclaimer: The actual correlations on these stats was actually pretty god-awful, but it's still fun to see if any of it holds true.
The base equation that I ended up with was as follows:
Exp. Win = -.044 * [3PA/G] + .95 [3PT%]
Basically, there was a very high influence by the overall 3PT%, but the more shots that you put up per game had a very small negative influence. It makes a bit of sense, given that the strict shooting percentage should have the most influence, but players that more or less only shoot threes might not fare as well in a contest that has much more pressure than a normal game (at least it makes sense to me).
When we give each player their expected wins, then take the overall expected wins for the player, given the rest of the field, we come up with the following win-percentages:
A bit shocking that Curry ends up in last, but he takes the most threes of any player in the field. Korver holding the best odds shouldn't shock anyone .
In order to control for individual numbers and skewed fields, I also ran a regression for the rank of the individual player with respect to the rest of that year's field. The following is the regression equation for the ranks:
When we give each player their expected wins, then take the overall expected wins for the player, given the rest of the field, we come up with the following win-percentages:
- Korver - 23.36%
- Belinelli - 20.06%
- Redick - 16.68%
- Irving - 15.52%
- Thompson - 10.76%
- Harden - 5.98%
- Matthews - 4.41%
- Curry - 3.24%
A bit shocking that Curry ends up in last, but he takes the most threes of any player in the field. Korver holding the best odds shouldn't shock anyone .
In order to control for individual numbers and skewed fields, I also ran a regression for the rank of the individual player with respect to the rest of that year's field. The following is the regression equation for the ranks:
Exp. Win = .058 * [Rank 3PA/G] + -.01 * [Rank 3PT%]
The equation is consistent with the coefficients from the original (remember that the lower the number of the ranking, the better). There is a bit more weight in this equation for the number of 3s shot per game, but since the actual numbers can be so drastic, it makes sense. This equation should also be a better fit for the linear regression we performed before.
For the rank regression, this is what we end up with for expected win-percentages, using the same process as before:
Geez. Curry is hurting bad (Chris won't be happy about this).
And, to even out for both models, we average the two percentages:
And there yah have it folks! Belinelli is your expected winner, with Korver, Irving, and Redick having the next best shots at winning (respectively).
So who do YOU all think will win?
For the rank regression, this is what we end up with for expected win-percentages, using the same process as before:
- Belinelli - 22.22%
- Irving - 21.44%
- Redick - 18.66%
- Korver - 16.55%
- Harden - 9.12%
- Thompson - 9.00%
- Matthews - 2.90%
- Curry - 0.12%
Geez. Curry is hurting bad (Chris won't be happy about this).
And, to even out for both models, we average the two percentages:
- Belinelli - 21.14%
- Korver - 19.95%
- Irving - 18.48%
- Redick - 17.67%
- Thompson - 9.88%
- Harden - 7.55%
- Matthews - 3.65%
- Curry - 1.68%
And there yah have it folks! Belinelli is your expected winner, with Korver, Irving, and Redick having the next best shots at winning (respectively).
So who do YOU all think will win?
Mosby:
If numbers aren't your thing, here's my analysis: Steph Curry will win the 3-point contest because he is the central name drop for the Drake song "0 to 100". Also, he's the most likely MVP-candidate. AND EVERYONE ELSE ISN'T THAT GOOD. THEY'RE NOT EVEN THAT GOOD. AND NUMBERS ARE DUMB AND EYEBALLS DO NOT LIE, RIGHT CHARLES? I BET BRAD WROTE THIS BECAUSE HE CAN'T GET GIRLS AND HAS NO TALENT AT BASKETBALL.
If numbers aren't your thing, here's my analysis: Steph Curry will win the 3-point contest because he is the central name drop for the Drake song "0 to 100". Also, he's the most likely MVP-candidate. AND EVERYONE ELSE ISN'T THAT GOOD. THEY'RE NOT EVEN THAT GOOD. AND NUMBERS ARE DUMB AND EYEBALLS DO NOT LIE, RIGHT CHARLES? I BET BRAD WROTE THIS BECAUSE HE CAN'T GET GIRLS AND HAS NO TALENT AT BASKETBALL.