Links to Sports Illustrated, ESPN, NBC Sports and Fox Sports show wide praise for the of www.valueaddbasketball.com database to pinpoint a player’s value, but today announces major improvements to answer valid criticisms that have circulated at the annual MIT Sloan Sports Analytics Conference. You can say the same about Jerian Grant (pictured), who Value Add pegs as a possible Player of the Year candidate in the Top 25 table–just a year after being dismissed from Notre Dame.
Grant has been worth a 9.91 point swing for Notre Dame this year, a big enough impact that Notre Dame would be favored in just three ACC games instead of 10 if Grant did not play. Gonzaga’s Kevin Pangos has been even better, with a 10.94 Value Add on the table below. Four big improvements are being introduced today to make Value Add once again a precise measurement of each player’s value–those are reviewed below the table of the hottest 25 players to date. Until January we refer to the leaders as the “hottest” because this early one big game by a mid-major player against a great opponent can make the list–but by January the top players are mostly major conference players who have proven themselves over several tough games.
|Hottest 25 Players/Jersey No.||Rank||Team||Ht||Yr||Off||Def||PG/Per||Value Add|
|23 Zikiteran Woodley||1||Northwestern St.||6′ 2″||So||10.45||-0.64||SF+.55||11.64|
|24 Rayvonte Rice||2||Illinois||6′ 4″||Sr||7.41||-3.47||SF+.55||11.43|
|4 Kevin Pangos||3||Gonzaga||6′ 2″||Sr||9.82||-0.87||PG+.25||10.94|
|23 Wesley Saunders||4||Harvard||6′ 5″||Sr||7.28||-3.09||SF+.55||10.92|
|3 Corey Hawkins||5||UC Davis||6′ 3″||Sr||8.78||-1.32||SG+.63||10.73|
|40 Josh Scott||6||Colorado||6′ 10″||Jr||9.64||-0.84||C+.04||10.52|
|20 Levi Randolph||7||Alabama||6′ 5″||Sr||9.48||-0.43||SF+.55||10.46|
|22 Jerian Grant||8||Notre Dame||6′ 5″||Sr||10.81||1.15||PG+.25||9.91|
|5 A.J. English||9||Iona||6′ 4″||Jr||8.95||-0.58||PG+.25||9.78|
|31 Ron Baker||10||Wichita St.||6′ 4″||Jr||7.24||-1.57||SG+.63||9.44|
|5 Traylin Farris||11||Incarnate Word||6′ 8″||Fr||7.76||-1.58||PF-.04||9.31|
|4 Sterling Gibbs||12||Seton Hall||6′ 2″||Jr||9.02||0.36||SG+.63||9.29|
|10 Sheldon McClellan||13||Miami FL||6′ 5″||Jr||7.53||-1.04||SF+.55||9.12|
|22 Drew Brandon||14||Eastern Washington||6′ 4″||Sr||6.77||-2.02||PG+.25||9.04|
|15 Alan Williams||15||UC Santa Barbara||6′ 8″||Sr||5.06||-3.91||C+.04||9.00|
|55 Egidijus Mockevicius||16||Evansville||6′ 10″||Jr||5.96||-2.90||C+.04||8.90|
|55 Delon Wright||17||Utah||6′ 5″||Sr||7.00||-1.60||PG+.25||8.85|
|11 Marcus Marshall||18||Missouri St.||6′ 3″||So||8.60||0.40||SF+.55||8.76|
|15 Willie Cauley-Stein||19||Kentucky (*1.5)||7′ 0||Jr||3.61||-2.18||C+.04||8.75|
|1 Isaiah Williams||20||Iona||6′ 7″||Jr||7.50||-1.17||PF-.04||8.62|
|1 Devin Booker||21||Kentucky (*1.5)||6′ 6″||Fr||4.18||-0.93||SG+.63||8.62|
|44 Frank Kaminsky||22||Wisconsin||7′ 0||Sr||5.94||-2.63||C+.04||8.61|
|13 Phil Forte||23||Oklahoma St.||5′ 11″||Jr||6.32||-1.60||SG+.63||8.55|
|22 Tim Huskisson||24||Northern Colorado||6′ 5″||Sr||7.50||-1.05||PF-.04||8.51|
|21 Emani Gant||25||Texas St.||6′ 8″||Jr||3.25||-5.24||PF-.04||8.45|
The improvements to the system being introduced today are as follows:
1. VALUE ADD ADJUSTED FOR POSITION. Since the introduction of Value Add, the major criticism reported to me from MIT and other analytical circles has been the “PG/Per” rating that was added subjectively for points guards and perimeter defenders. Today that figure is being replaced by a precise adjustment based measuring the player against others who play the position. A rating of +3.0 Value Add is supposed to indicate a player is good enough to start for a major conference or very good mid-major team. In theory, this means a player would have to be one of the top 100 players at his position in college basketball.
This year we have taken the extra step of identifying the primary position of each player and determining the rating of the 100th best player at each person to use as the landmark. The adjustments are now:
Point Guard +0.25 — top 100 Point Guards have ratings of 2.75 or better before the adjustment
Shooting Guard +0.63 — top 100 have 2.37 or better
Swing/Small forward +0.55 — top 100 have 2.45 or better
Power Forward -0.04 — top 100 have 3.04 or better
Center +0.04 — top 100 have 2.96 or better
2. YEARLY ADJUSTMENT FOR POINTS PER 100 TRIPS. Another weakness in the system was revealed with the much higher level of scoring in 2014 due to rules changes. Defenders were suddenly not allowed to place a hand on point guards for positioning and players had to virtually jump out of the way of driving guards because charges were almost never called. In evaluating the level of defense a player had to overcome, Value Add assumed the average defense gave up about 100 points per 100 trips down the court. In 2014, this resulted in the system believing almost everyone was facing very weak defenses in almost every game, so ratings came out only 75% as high as in past years. We evened out the overall average by multiplying each Value Add by 1.33, but in fact found that led to distortions in exactly how many points each player was worth. This year the programming has allowed the average defense in the country to be measured on the front end, which has once again made the result precise. Interestingly, referees are still not allowing hand checks, but they have gone back to giving defenders credit for drawing charges even if their feet are not glued to the floor like they had to be in the slow moving game in the 1940s. Here are the average points allowed per 100 trips:
2009 to 2013: between 98.9 and 101.3 every year
3. KENTUCKY VALUE ADD. I readily admitted in early posts that the system did not work for Kentucky, or any other team that had more than five NBA prospects, because players would not get enough playing time no matter how well he plays. The Kentucky Kernel was one of the first to take note of the system. This year the team has taken it to a new level with two full teams of future NBA players splitting time as squads. In doing calculating based on Basketball on Paper, it appears each Kentucky players’ Value Add would actually be right at 50% higher if they were to play for any other team where they were not limited to half a game, so I have made that adjustment for this year’s rating.
4. JERSEY NUMBERS. One final adjustment is for the fans rather than for my fellow stat geeks and NBA scouts. Fans wanting to access the database at games have wanted the number of each player included so that they can immediately spot the star player I have identified by clicking on Value Add and then punching in the team name. The table above, and the database update being made shortly thereafter, will include the jersey number.
We are also working on the mobile version with fewer columns since most are now viewing it on a phone rather than a wide screen computer.
The updated results for all players will appear at www.valueaddbasketball.com shortly after this piece is published.