Wednesday, November 27, 2019

Pistol Pete's Offense vs. UVa 2019 or Jordan's Defense; 96 Great Teams Calculated for Game

After using modern tempo-free stats to create player cards for 48 great 21st century teams, the focus turned to estimated the tempo played by 20th century teams to create accurate cards for Value Add Basketball Game. Click here for playing cards for the 48 great 20th century teams.

The defenses of Michael Jordan's 1982 UNC title team (average score 67-55) and last year's UVa champions (71-56) were obviously much better than the defense of Pistol Pete Maravich's 1970 LSU team (93-89 average score). While Pistol Pete is the only player in the game to get the ball half the time he is in the game and the offense was elite, based on the number of shots taken we know LSU went up and down the court many more than the "average" 66 trips for each team in VABG and the team's LSU faced that year were not nearly as good as the opponents faced by UVa and UNC.

For fellow wonks, the complicated math appears below the list of teams in the game. For everyone else, if all teams played 66 possessions in the game (rather than a clock) against the same competition in the average year, we would expect 1970 LSU's elite offense with Pistol Pete to average a 80-74 win. That would give them a better offense than 1982 UNC or 2019 UVa, but playing the same competition in the same era we would expect those two to average winning by 73-55 and 71-56 respectively.

In the game, after each team's players ability to steal, block shots, avoid fouling and turnovers, rebound and hit various shots after getting open for three-pointers, two-pointers, draw fouls, and hit each or drawing fouls, a "dunk/layup" and "adj/opp. dunk layup" range is established for the whole team to adjust for tempo and strength of schedule. In early years, we estimate the number of shots were three-pointers, and the number of steals and blocks - but the dunk adjustments make the cards for each team as good as they should perform in real life.

We believe the game is as accurate as you can get playing teams from various eras. However, this does NOT mean we actually believe a team from 1950 would get in a time machine and stand a chance against a modern team, any more than Babe Ruth could step into a time machine and hit modern sliders. However, if we treat each eras as equal, we believe the teams will perform very close to how the would over the course of a season - with the normal ups and downs from game to game that occasionally mean Evansville can beat Kentucky and Stephen F. Austin can beat Duke.

For fellow math geeks, the math is below the tables - but here is the score we would expect each team to average if all played the same competition. The first score is how many points they would average scoring, then allowing, and then the difference. The teams are grouped in six modern conferences for the game, but the final column shows the conference in which they actually played.

YearACCAdj PtsAllowedNet/66 posActual Conf
1974North Carolina State81.062.918.1ACC
1982North Carolina73.355.417.8ACC
2005North Carolina79.362.816.5ACC
2009Pittsburgh73.160.013.0Big East
2003Syracuse75.262.912.4Big East (West)
2004Georgia Tech72.260.012.2ACC
1987Syracuse74.262.711.5Big East
2005Wake Forest77.166.011.1ACC
1996Wake Forest65.454.710.7ACC
1970Notre Dame79.369.110.2Ind
YearBig 12Adj PtsAllowedNet/66 posActual Conf
1957Kansas74.657.517.1Big 7
2008Kansas74.658.116.6Big 12
1946Oklahoma St76.162.214.0MVC
1982Missouri72.459.213.2Big 8
1985Oklahoma81.468.712.7Big 8
2004Oklahoma State71.559.212.3Big 12
1959West Virginia80.368.411.9Southern
2019Texas Tech66.854.911.9Big 12
2004Saint Joseph's71.860.011.8A-10 (East)
1966Texas-El Paso68.957.911.0Ind
2003Texas73.963.011.0Big 12
2016Oklahoma72.461.511.0Big 12
2008Kansas State73.163.99.2Big 12
2011West Virginia68.259.19.2Big East
YearBig EastAdj PtsAllowedNet/66 posActual Conf
2018Villanova75.960.215.7Big East
2013Louisville73.058.114.9Big East
1999Connecticut72.157.314.7Big East
2002Cincinnati72.057.614.4CUSA (American)
1984Georgetown71.457.713.7Big East
2004Connecticut73.960.413.5Big East
1985St. John's (NY)73.861.012.8Big East
2007Georgetown67.756.411.3Big East
1989Seton Hall70.859.611.2Big East
2014Creighton72.762.110.6Big East
2003Marquette72.663.49.2CUSA (American)
1985Villanova68.960.18.8Big East
YearBig TenAdj PtsAllowedNet/66 posActual Conf
1976Indiana76.654.522.1Big Ten
1960Ohio State83.663.120.4Big Ten
1979Michigan State73.956.517.4Big Ten
1963Loyola (IL)81.864.916.9Ind
1969Purdue82.665.716.8Big Ten
1989Michigan76.560.915.7Big Ten
2015Wisconsin70.956.714.2Big Ten
2005Illinois71.958.213.7Big Ten
2018Purdue72.759.113.6Big Ten
2007Ohio State71.659.012.6Big Ten
2013Michigan73.460.812.6Big Ten
2002Indiana68.057.011.0Big Ten
1954La Salle72.261.311.0Ind
2009Michigan State71.260.510.7Big Ten
1979Indiana State75.164.510.7MVC
2002Iowa69.162.66.5Big Ten
YearPac-12Adj PtsAllowedNet/66 posActual Conf
1991Nevada-Las Vegas79.761.418.3Big West
1956San Francisco68.454.613.9CBA
1959University of California68.055.113.0PCC
2011San Diego State68.057.610.4MWC
1998Utah66.256.59.7WAC (Mountain)
2013Wichita State68.159.98.2MVC
2018Loyola (IL)64.659.65.0MVC
YearSECAdj PtsAllowedNet/66 posActual Conf
1996Kentucky80.961.019.9SEC (East)
1994Arkansas77.663.614.1SEC (West)
2006Florida73.560.712.7SEC (East)
2006Louisiana State71.260.211.0SEC (West)
2017South Carolina66.457.98.5SEC
2008Davidson70.462.48.0Southern (South)
2006George Mason66.659.37.4CAA
1970Louisiana State79.774.15.6SEC
2011Virginia Commonwealth68.664.44.2CAA

Notes on the math:

Tempo Free Stats - When we calculated the 21st Century cards, we used the tempo free team ratings and stats at

However, we switched to the incredible Simple Ratings System (SRS) at basketball reference for development of the 20th century player cards and then updated the other cards to make the all one system.

We started with the offensive and defensive ratings from SRS, however we noticed rules changes such as freedom of movement allowed the better teams in modern eras and other times to distance themselves by greater margins that teams in more "grind it out" eras.

To address this, we adjusted the Offensive SRS so that the Net Average each year for the 10th best team would be +16.6.

We did not allow any team who did not win the title (e.g. 1991 UNLV) to have a net rating higher than +25.0, so a few were adjusted down a couple of points but still very strong.

On the flip side we improved a couple of elite championship teams by a couple of points - including the undefeated 1976 Indiana champions.

However the vast majority of teams have the exact calculation that also results in the 10th best team the season they played at +16.6.

These numbers are then converted back to equivalent ratings based on adding the adjusted SRS Offensive number to 100 to get the estimated offensive efficiency rating, and subtracting the adjusted SRS Defensive number from 100 to get the estimated defensive efficiency rating - even for the modern teams where the calculated number was used rather than the number to adjust for eras.

The numbers that appear above are that number multiplied by .66, to calculate the number of points the team would score in a 66 possession Value Add basketball game (the game actually assumes a 20-20 tie after 22 possessions, and then dice are rolled to play out the final 44 possessions of each game.

The best defensive teams do not necessarily have the best "Adj opp Dunk/layup" range. If a team were to have a great rating, but also have a very high number of dice rolls resulted in steals, blocked shots and defensive rebounds, then all of those things might take away so many points from the opponents that the actual "Adj opp Dunk/layup" range would need to ADD points.

However, normally the best defenses have the best dunk range - making the cards accurate against an average defense in an average era.

A "Adj opp Dunk/layup" of -2 for a team would mean that 2 numbers were taken off the numbers on the opponent's "dunk/lay-up" range.

Likewise, the best offensive teams do not necessarily get the highest dunk range. A team that avoided turning the ball over, drew a lot of fouls, dominated offensive rebounds, and hit a high percentage of shots could calculate as scoring so many points that their actual dunk range was lower, perhaps 51-51 instead of the average 51-53.

However, normally the best offenses have the best dunk range - making the cards accurate against an average defense in an average era.

If an offense's dunk range is 11-57,and the opposing defense's adj opp dunk range is -3, then the dunk range for the game would be 11-54.

However, if at least one of the teams calculated dunk range is higher than 51-56, or lower than 51-50, then it works best to adjust both team ranges the same number.

For example, if one team's calculated dunk range was 51-64 and one was 51-61, you could lower the ranges to 51-53 and 51-50.

No comments:

Post a Comment