Two days after Barack Obama’s reelection, ESPN compared me to Nate Silver due to Value Add Basketball. And after the next presidential election consumed me, Breitbart Sports unveils v4.0 and the top 10 players in the early going.
Click ValueAddBasketball.com to find your school’s hottest players among the 3,898 players who have taken the court early in the season. Value Add 1.0 first measured how many points each player adds to his team’s score and takes off his opponent’s score, and Value Add 2.0 incorporated suggestions from NBA Scouts and Basketball Prospectus. Version 3.0 honed the calculations to ensure the sum of player values equal the precise team values at www.kenpom.com. Version 4.0 measures the players’ actual value (3.0) and adjusts for how much the player would be worth if given the ball at least 16% of the time his team had it (thus adjusting for players for Kentucky and Duke whose minutes are limited on the bench behind other superstars).
The following constitute the top 10 players from the eight conferences calculated as five points or more better than the other 24 conferences (the Power 5 football conferences plus the Big East, American Athletic and Atlantic 10), followed by an explanation of the system:
1. Brandon Parrish #11, TCU, 12.17 Rating, was the only Horned Frog to start all 33 games last year and starts his senior campaign the same way he finished 2016, shooting 55.6% his last four games.
2. Markelle Fultz #20, Washington, 12.03 Rating, looked like James Hardin in scoring 30 and 35 points in his first two college games this month, putting the Pac-12 on notice.
3. Charles Cooke #4, Dayton, 11.02 Rating, is a returning all-Atlantic 10 player and defender.
4. Marial Shayok #4, Virginia, 10.93 Rating, is dominant the 10% of the time he has the ball, and with UVa dismissing Austin Nichols he could step up and be a superstar.
5. Devin Davis #15, Houston, 10.73 Rating looked good even as a freshman for Indiana three years ago, and his transfer gives Houston a great boost in the American Athletic.
6. Josh Hart #3, Villanova, 10.51 Rating, just crushed ACC foe Wake Forest for 30 points and the Wildcats shot at repeating as NCAA champs is possible when you consider his teammate Mikal Bridges is the 12th best player on this list.
7. DJ Johnson #4, Kansas St., 10.42 Rating looks like he might finally be at the superstar level everyone expected before he broke his foot in March Madness against Kentucky back in 2014.
8. Devin Robinson #1, Florida, 10.41 Rating shows this the sky is the limit for this player who was an all-SEC Freshman selection two years ago and then called the most improved Gator last year.
9. Malcolm Hill #21, Illinois, 10.36, B10, 10.36, 25%, SF*1, 1.00 – top 3% last year
10. ShawnDre’ Jones #3, Richmond, 9.97, A10, 8.31, 29%, PG*1.2, 1.00 – top 3% last year
Explanation of Value Add Basketball 4.0
Value Add 3.0 is still calculated and gives the actual offensive and defensive value of each player. Big Apple Buckets ran an explanation of that system here.
However, that system penalizes a player who does not get the ball as much because he is with a major program like Kentucky where he competes with other superstars for playing time and touches.
An excellent player can have the most impact if he plays about two-thirds of the minutes (resting one-third) and while on the court has the ball about 24 percent of the time (2/3 x 24% = 16% actual possessions). If a players does have the ball 16% or more then Value Add 4.0 calculates the same rating as Value Add 3.0. However, if a player on a stacked team only has the ball 8% of the time then Value Add calculates what his Value Add would likely be if he could have it 16%.
However, this is not as simple as doubling the value in his 8%, because Dean Oliver proved long ago that a player becomes less and less valuable the less rest he gets in a game and the more he has to handle the ball (e.g. he gets double teamed if he is the star, while a lesser player may only shoot when he is left open for a layup).
To measure how much more valuable the player would be if he could get the ball 16 percent of the time, Value Add 4.0 adopts the tables on page 234 and subsequent chapters of Basketball On Paper by Dean Oliver. Basically the same player who could average 131 points for 100 trips if he just handled the ball 1% of the time (meaning the defense ignored him, he only shot if open at the rim and he was well rested) would fall off to 122 if he had to extend and take 5% of his team’s possessions. The following table shows how much we would expect him to drop off if other players were injured and he suddenly had to take more and more of the load:
|Player Ortg Drop with higher Pos%||% Poss||% Min||Per Game||Off Rat|
This table is based on two players with great endurance who were All-Stars at the time. Obviously a lesser player like an eighth man who only had an Offensive Rating of 82 while handling 10% of possessions, would be very low if he suddenly had to start.
The lack of understanding about the how much a player falls off in efficiency the more he has to do is often the biggest stumbling block to understanding Value Add as well as why a player with a slightly lower Offensive Rating who can do it while playing many minutes and possessing the ball a lot is more valuable than a player with a higher rating who does not have to do that much.
In the top 10 above, Shayok only gets the ball 10 percent of time time in the University of Virginia’s games. He is worth 6.72 points per 100 trips down the court for UVa. The Value Add 4.0 sliding efficiency scale indicates multiplying his Value Add 3.0 by 1.48 indicates how much more he could do with the ball 16% of the time, while playing the shooting guard position (No. 2 spot on the court) makes him worth more (multiply by 1.1) for a Value Add 4.0 rating of 10.93.
Meanwhile Josh Hart of Villanova does handle the ball a lot more than most players — 20% of all of Villanova’s possessions — so his 9.55 Value Add 3.0 rating is the same as his 4.0 except that he also plays the Shooting Guard position to boost his value over a replacement player to 10.51.
So far this season point guards are worth 20% more (multiply Value Add by 1.2) than other players, shooting guards are worth 10% more (1.1), and small forwards, power forwards and centers all have equal value over their replacement players (1.0).
While many teams now play three or four guard line-ups, each team is considered to have one of each of these five positions on the court at all times — with a “swing man” who plays closer to the basket than the other two guards still being considered a “small forward” for purposes of these calculations.
As outlined as early as the Value Add 1.0 release in 2011, the ratings of players who need to play further on the perimeter are held down because they do not have as many chances for rebounds and risk more turnovers on the offensive end.
Our attempts to manually adjust for this factor was roundly rejected by other sports analysts at MIT and elsewhere, so it is now a mathematical confirmation of how many points each player is worth beyond the replacement player who would take over at his position.
A players position is calculated early in the year with the following equation: “<(Defensive Rebound Percentage / Assist Rate) + (0.2 x height in inches)> = Position Calculation.
The player on the team with the lowest “Position Calculation” is designated a point guard. If he plays 90% of the minutes he is likely the only point guard, but he plays 33% of the minutes and the next two lowest figures also both play 33% of the minutes then they would also be slated as point guards. the next 100% of minutes are the shooting guards, the small forwards, power forwards, and finally the highest figures are the centers.
In early going the average point guard only calculates a little over 80% (and the average shooting guard a little over 90%) of the average value of the other three positions, thus the 1.2 and 1.1 adjustment to their rankings. However, these figures change throughout the season and other seasons.
Historically the front line players had huge advantage due to hand checks and rough physical play, however, recent changes to allow “freedom of movement” gave all positions a more equal change to score.
The ratings will be updated regularly at www.valueaddbasketball.com.