If you ask yourself one day, “Who’s the best defensive rebounder in the country?”, what do you do? Well, you click over to the Kenpom defensive rebounding percentage leaderboards. The guy who’s first isn’t necessarily the answer, but even with some serious teammate and opponent troubles, it’s tough for that top guy to escape the top ten or fifteen. If you ask yourself, “Who’s the best shotblocker?”, the same is true. There’s a statistic built to answer that specific question.
So what statistic answers the question, “Who’s the best three-point shooter in the country?”
Well, it’s not three-point percentage. Let’s set the bar at, say, a minimum of 100 three-point attempts, and look at last season’s three-point percentage leaders.
1. Jordan Burgason, Lipscomb (71-135, 53%)
2. Joe Ragland, Wichita State (59-117, 50%)
3. Jordan Hulls, Indiana (72-146, 49%)
4. Tyler Murray, Wagner (74-151, 49%)
5. Brendon Lavender, Arizona (55-113, 49%)
6. Doug McDermott, Creighton (54-111, 49%)
7. Nick Barbour, High Point (108-223, 48%)
8. Drew Hanlen, Belmont (94-195, 48%)
9. Gary Bell, Gonzaga (51-107, 48%)
10. Mardracus Wade, Arkansas (70-147, 48%)
I mean, obviously it’s not terrible, or anything. As long as you put in some kind of attempts floor, you’ll find something semi-reasonable. My point is more that three-point percentage isn’t built to find the best three-point shooters, in the same way that pure field goal percentage doesn’t give you the best scorers in the country.
So I made up a stat that does: Three-Point Score (3PS).
We divide up every possession with the player on the floor into three categories: Made threes, missed threes, and everything else. Then we give the player credit for three points when he makes a three, 0.3 points (the likelihood of getting an offensive rebound, 30%, times the expected points per possession, 1.00) for a missed three, and one point for any other possession with the player on the floor. That’s it. The formula looks like this:
3*3PM + 0.3*(3PA – 3PM) + (TmPoss*%Min – 3PA) 3PS = ------------------------------------------------- TmPoss*%Min
It basically gives you the offensive rating for a completely average team, considering that the player in question shot and made threes at the frequency he did.
So here’s the Top 10 from last season in 3PS among players who played at least 60% of their teams’ minutes:
1. Nick Barbour, High Point (1.084)
2. John Jenkins, Vanderbilt (1.078)
3. Reggie Chamberlain, UMKC (1.072)
4. Brian Sullivan, Miami OH (1.070)
5. Keaton Cole, Western Carolina (1.070)
6. Brady Heslip, Baylor (1.068)
7. Parker Smith, North Florida (1.066)
8. Drew Hanlen, Belmont (1.062)
9. Isaiah Canaan, Murray State (1.062)
10. Kris Davis, SIU Edwardsville (1.062)
And here’s that same list, adjusting for schedule (by dividing 3PS by the SOSD found on each team’s Kenpom page):
1. John Jenkins, Vanderbilt (1.108)
2. Kenny Boynton, Florida (1.100)
3. Brady Heslip, Baylor (1.097)
4. Keiton Page, Oklahoma State (1.088)
5. John Shurna, Northwestern (1.087)
6. Jordan Hulls, Indiana (1.084)
7. Ryne Smith, Purdue (1.083)
8. Doron Lamb, Kentucky (1.077)
9. Darius Johnson-Odom, Marquette (1.073)
10. Kam Cerroni, Green Bay (1.072)
Which is actually a pretty reasonable list of the best shooters in America last season, although Canaan gets beat up bad by the schedule adjustment.
It works identically for two-point percentage, with the only change being that you give the player two points for a make now. Here’s the schedule-adjusted top 10 for 2012:
1. Anthony Davis, Kentucky (1.095)
2. Doug McDermott, Creighton (1.088)
3. Kevin Jones, West Virginia (1.086)
4. Thomas Robinson, Kansas (1.084)
5. Deshaun Thomas, Ohio State (1.083)
6. Jae Crowder, Marquette (1.080)
7. Rodney Williams, Minnesota (1.076)
8. Jared Sullinger, Ohio State (1.075)
9. Otto Porter, Georgetown (1.073)
10. Andre Drummond, Connecticut (1.073)
Which also looks pretty good as a list of the most effective scorers in the country inside the arc – though people often fold drawing fouls and making free throws into that conversation, as Drummond’s inclusion here makes abundantly clear.
We took a step in the right direction in separating two-point shooting from three-point shooting. There’s no reason we can’t take another step, weighting volume and efficiency together.