ESPN   CBS Sports   SI   Ken Pomeroy   Prediction Tracker  Massey  

NCAA Basketball Analytics

Main Page   Todays Games   Yesterdays Games   Conference Rankings   Win Prob Calculator   Team Stats   Raw Data   Blog   About

New Feature Update - 10/24/2013

A major part of the site redesign are the new "game pages". In the past, these pages had a pre-generated win probability chart and some other game info (including WinProbAdded by player). These pages are now dynamic and interactive. Currently, you can see the score and play for each data point - I want to add the Game Time of each play as well. I need to format the X-axis better, but other than that, the charts are helpful. As we get into the season I will add more auxiliary data to each of those pages. This page here:

Server Migration Update - 10/24/2013

Team pages now feature Player and Schedule history back to 2008. Some player stats still need to be updated, and I only have player metadata (Height, Weight, Hometown, etc.) back to 2012. You can see past games for each season, but I still need to load some historical data from before 2011 to get the game charts to work.

Server Migration Update - 10/20/2013

The site is updated for the 2014 season. I am working on loading history (back to 2008). The main concern is maintaining a fast site with all this data. I have compliled conference metadata (Commissioner, Year Founded, etc.) and will be adding more data to the conference pages as the season goes on. I have also created a list of every coaching change, so those should be online very soon.

New Server Updates - 10/13/2013

This new server migration will drastically change both my ability to make changes and the back-end design of the website. In the past, all pages were generated overnight and FTP'd to a server - there was no backend online database generating the pages. This is no longer the case - I now have MySQL backend with PHP script generating most of the pages. Team pages have been redesigned, as have each of the games pages. When complete, the Win Probability graphics will be interactive and informative. Stay tuned as the other updates come online.

Website Updates

As part of a new server setup for the 2013-2014 season, I have updated all of the new teams and updated conference alignments. The WAC is now a shell of its former self, the Conference USA is a mess, and several other smaller schools have moved around. We have four new teams in Division 1: Abilene Christian and Incarnate Word into the Southland, UMass Lowell into the America East (replacing Boston U who moved to the Patriot), and the (for-profit) Grand Canyon Antelopes into the WAC. In addition to being a for-profit university, the Antelopes are notable for their famous coach - Dan Marjele.

Geographic Recruiting Diversity by Conference

Updated 2.26.12

I had always been under the assumption that public schools were more likely to recruit in state and that private schools were more likely to go outside their local area to find talent. The table below shows the average # of different states/countries represented on each team by conference. The Ivy and Patriot leagues (both located predominately on the east coast - where there are smaller states - and have excellent academics) lead the pack with averages of 11 and 10, respectively. The average was approximately 7.75, with the Atlantic Sun, Big West, and Southland at the bottom - featuring an average of less than 6 teams each.

Conference NameAvg # of States Per Team
Missouri Valley8.8
America East8.4
Atlantic Ten8.4
Big East8.4
Conference USA8.3
Big South8.09
Ohio Valley8.0
Mountain West7.7
Horizon League7.7
Big Sky7.6
Big Ten7.5
Big XII7.4
Great West7.3
West Coast6.6
Atlantic Sun5.7
Big West5.3

Most Unlikely Upsets

Updated 2.23.12

A previous post discussed the most unlikely comebacks. This post is about highlight three of the most unlikely upsets of the year. I limited my dataset to conference play only - and wanted to find the games with the highest predicted margin where the home team lost. There might've been games where a heavy road favorite lost - but losing on someone's home court is not as interesting.

Home TeamVisitor TeamGame DateFinal ScorePredicted MarginVariance
CharlestonUNC Greensboro2012-01-1266-7724-31
Texas A&M-CCUTSA2012-01-1149-5020-21
UMBCNew Hampshire2012-01-0276-8220-26

Home Court Advantage By Conference

Updated 2.17.12

A common belief on the difference between good teams and just average teams is their ability to win on the road - in conference play. This made me decide to examine Home Court Advantage - by Conference - to see which conferences have the strongest effect. This difference could be due to lots of average teams (like Arkansas who can't win on the road) or lots of "tough" teams (apparently tough means they can win on the road to most people). The rankings below are for all games thru Wednesday January 16th, 2012.

RankConference Name# of Wins# of GamesWin %
4Mountain West243764.8%
8B1G Ten497763.6%
10Big XII416563.0%
11BIG EAST6610562.6%
14Great West111861.1%
17Ohio Valley477860.2%
18 ACC386459.3%
23 Big South417157.7%
25A-Sun 407255.5%
27Big West285154.9%
28America East34 6353.9%
29Big Sky305851.7%
30NEC 387451.3%
31Sun Belt387451.3%

Quantifying "Garbage Time"

Updated 2.15.2012

By looking at cumulative win probabilities I am able to quantify "garbage time" and see what percent of each team's game are spent in garbage time. My definition of garbage time is any play where either team has a 98% chance of winning. The top 5 Teams in Postive Garbage Time (when they are predicted to win) and Top 5 Teams in Negative Garbage Time can be seen below.

Top Postive Garbage Time Teams

RankTeam NameGarbage Winning PlaysGarbage Losing PlaysTotal Garbage PlaysTotal PlaysPostive Garbage %Negative Garbage %Total Garbage Percent
1Ohio St2639442683696537.8%.6%38.5%
3North Carolina27602122972798834.5%2.6%37.2%

Top Negative Garbage Time Teams

RankTeam NameGarbage Winning PlaysGarbage Losing PlaysTotal Garbage PlaysTotal PlaysPostive Garbage %Negative Garbage %Total Garbage Percent
1Grambling St1138513862201.04%62.9%62.9%
2Texas Southern26152415503143.8%48.5%49.3%
3Miss. Valley St20159716173326.6%48.0%48.6%
4Alabama A&M01000100023470.0%42.6%42.6%
5Alcorn St072072016900.0%42.6%42.6%

Top 3 Craziest Comeback of the Year (so far)

Updated 2.14.2012

One of the great things about my system is that it makes it very easy to quantify the most unlikely comebacks. I'm not totally satisfied that the current design credits a comeback as "bad luck" by the losing team, but until I can develop a better system, its the one I'm using. These are the games in my database where the winning team had the lowest cumulative win probability over the course of the game.

#1 Manhattan 75, Iona 72, January 12, 2012

Iona, undefeated in MAAC play at the time (5-0), had a 17 point half time lead, and still had that with 7:58 to play in the game. Manhattan went on 21-4 run to eventually tie the game. A Lamont Jones jumper tied the game with 3 seconds left, and a three pointer by Emmy Andujar with 1 second left gave Manhattan the most unlikely win of the year. The cumulative win prob for Manhattan was a measly 9.1%.

ESPN Box Score

#2 UC Riverside 64, Washington St 63, November 27, 2011

This game was much earlier in the season, and unlike Iona's undefeated conference record, the only thing on the line here was 7th place in the 76 Classic. Wazzu had a 7 point lead at the half, and led by 17 with 7:05 to go. UC Riverside went on a 14-0 run over the next 3 minutes to get back into the game, and scored the winning bucket with 13 seconds left. UC Riverside's cumulative win probability was 11.4%

ESPN Box Score

#3 Purdue 63, Xavier 66, December 3, 2011

With 10:44 left in this game, Purdue took a 55-36 lead. Senior Guard Tu Holloway scored 21 points, including 3 straight 3s in the last two minutes to bring Xavier back into the game. Xavier took the lead at 61-60 with 1:06 left after losing the lead at 2-0 early in the game. Xavier's CWP for this game was 21.9%

ESPN Box Score

7OT Player of the Year

Updated 2.14.2012

In an effort to identify my own "Player of the Year" candidates, I wanted to find the ten guys who had the most Win Probability Added, but also played for "good" teams. My criteria for good became a team who was more than 3 games over .500 overall and more than 1 game over .500 in the conference. This dropped out players like Shane Gibson at Sacred Heart who are putting up gaudy numbers on subpar teams.The list:

1Doug McDermottCreighton15.25
2John JenkinsVanderbilt14.15
3Dominique MorrisonOral Roberts13.32
4Kyle WeemsMissouri St12.55
5Ramone MooreTemple12.51
6Toure' MurryWichita St12.20
7Isaiah CanaanMurray St11.97
8Dee BostMiss St11.60
9Kevin MurphyTennessee Tech11.57
10Lamont JonesIona11.49

Most Clutch Players

Instead of filtering for players on "good" teams, I wanted to see which players were the most successful in the clutch so far this season. I defined "clutch" as Average WPA in games where the margin was within 5 points with less than 10 minutes to go in a game. Players needed 75 plays to qualify, and are sorted by Average WPA, not Total WPA as they are above. The top ten:

RankPlayerTeamAverageWPA# of Plays
1Zack RosenPennsylvania.05084
2Khalid MutakabbirPresbyterian.04976
3Tashan NewsomeGarnder-Webb.04678
4Shane GibsonSacred Heart.04595
5Mike MooreHofstra.04585
6Kent BazemoreOld Dominion.043104
7Julius MaysWright St.04287
8Terrance HenryOle Miss.04186
9Torrey CraigUSC Upstate.04189
10Kevin DillardIona.04189

Copyright: Mark Bashuk, 2011
E-Mail: bashukma "at symbol" gmail "dot" com
Twitter: @tbbBaseball

Copyright: Mark Bashuk, 2013
E-Mail: bashukma "at symbol" gmail "dot" com
Twitter: @SevenOvertimes