Game Predictions | March Madness 2021

The table below provides our predictions for all games in the 2021 NCAA Tournament. The table will be updated after each round of the tournament, with new games/predictions added as they become available. A brief description of how to interpret the predictions is provided above the table, and a summary of the historical accuracy of each model is provided at the end of the post. See our related posts on our Full 2021 Tournament Bracket and our Tournament Simulation results.

Win Probability & Spread Predictions

Win Probabilities – For each game, “Team1 Win Prob” represents our model’s probability that Team1 (the better-seeded team) wins the game. Probabilities below 50 indicate that Team2 is more likely to win the game.

Spreads – The “284 Spread” is our model’s prediction of how many points Team1 will win by. For both Vegas and 284 Spreads, negative values indicate Team1 is favored, while positives indicate Team2 is favored.

Totals – The Vegas Total can be compared to our “284 Tot,” which is our model’s prediction of how many points will be scored in the game.

Team1Team2RoundTeam1
Win Prob
Vegas
Spread
284
Spread
Vegas
Tot
284
Tot
1 Gonzaga1 Baylor60.71-4.5-7.8159.5155.1
1 Baylor2 Houston50.64-5-3.4134.5134.5
1 Gonzaga11 UCLA50.79-14-12.6145153.7
2 Houston12 Oregon St40.80-8-12.6130128.9
1 Baylor3 Arkansas40.72-7-5.8149149.3
1 Gonzaga6 USC40.78-9-10.8154.5151.5
1 Michigan11 UCLA40.62-7.5-2.1136141.5
8 Loyola Chi12 Oregon St30.70-7-6.7125.5127.1
1 Baylor5 Villanova30.75-6.5-6.6140143.4
3 Arkansas15 Oral Roberts30.73-11-7.6159.5153.5
2 Houston11 Syracuse30.73-6-8.3139.5139.3
1 Gonzaga5 Creighton30.79-13.5-11.9158160.4
1 Michigan4 Florida St30.61-3-2.8144142.3
2 Alabama11 UCLA30.74-6-5.3146153.9
6 USC7 Oregon30.53-2.5-0.3139135.2
1 Illinois8 Loyola Chi20.63-7-5133129.4
1 Baylor9 Wisconsin20.76-6.5-7.7137137
3 West Va11 Syracuse20.54-4-1.6147146.9
3 Arkansas6 Texas Tech20.591.5-0.3140.5137.1
2 Houston10 Rutgers20.79-8-12132131.5
7 Florida15 Oral Roberts20.65-8.5-4.3149152.2
5 Villanova13 North Texas20.63-5.5-4.7126.5130.6
4 Okla St12 Oregon St20.60-6-2.5142.5139.3
2 Iowa7 Oregon20.65-4-3.6147145
1 Gonzaga8 Oklahoma20.82-14.5-14.2154156
11 UCLA14 Abl Christ20.49-5-0.3132.5134.7
5 Creighton13 Ohio20.68-5.5-5.3149.5147.5
1 Michigan8 LSU20.55-5-1149154.5
4 Florida St5 Colorado20.58-1-0.9139138.3
2 Alabama10 Maryland20.72-6-4.9137.5145.5
3 Kansas6 USC20.641.5-3134.5138.8
5 Colorado12 Georgetown10.68-4.5-5.1138143.2
4 Florida St13 UNC Greens10.69-11-6.4145.5147.6
3 Kansas14 East Wash10.68-10.5-6.1143147.1
8 LSU9 St. Bonaventure10.57-2.5-3144.5144.6
1 Michigan16 Texas So10.77-25.5-11.2142.5145.3
5 Creighton12 UCSB10.61-6.5-3.6137.5144.4
2 Alabama15 Iona10.82-16.5-11.3146149.4
6 USC11 Drake10.53-6.5-0.6135138.2
2 Iowa15 Grand Canyon10.73-15-8.9145144.9
7 UConn10 Maryland10.63-3-3.8129.5129.1
4 Virginia13 Ohio10.64-8-5.1131.5135.5
8 Oklahoma9 Missouri10.52-2.5+0.5141140.2
1 Gonzaga16 Norfolk St10.89-33-153.5-
6 BYU11 UCLA10.45-4+1.9138.5143.4
3 Texas14 Abl Christ10.59-9-3.6139.5142.8
7 Oregon10 VCU10.60-6-2.9137139.5
7 Florida10 VA Tech10.63-1-2.4137.5140.1
3 Arkansas14 Colgate10.65-8.5-4.4162.5153.8
1 Illinois16 Drexel10.80-22.5-11.6144.5138.3
6 Texas Tech11 Utah St10.68-5-5.2131.5127.1
2 Ohio St15 Oral Roberts10.72-16.5-7.3155.5150.2
1 Baylor16 Hartford10.84-26-13.7141139.3
8 Loyola Chi9 Georgia Tech10.67-1.5-4.1124.5126.9
5 Tennessee12 Oregon St10.70-7.5-6.1131.5134.1
4 Okla St13 Liberty10.54-9-0.9142136.6
8 UNC9 Wisconsin10.66-1-5.6139.5139.6
2 Houston15 Cleveland St10.85-20-15.5135133.4
4 Purdue13 North Texas10.6-7.5-2.4127.5127.4
7 Clemson10 Rutgers10.572-1.2126.5131.6
6 San Diego St11 Syracuse10.69-2.5-5.3139.5146.5
3 West Va14 Morehead St10.65-12.5-6.2136.5137.5
5 Villanova12 Winthrop10.59-6.5-4.9140151.3
16 Mt St Mary's16 Texas So00.541-0.8132.5135.4
11 Wichita St11 Drake00.45-2+0.8140.5139.6
16 Norfolk St16 App St0-3-134.5-
11 Michigan St11 UCLA00.59-2-2.4136141.8

Model Accuracy

Below is a summary of how each individual model performed on all 2017-2019 tournament games, with 2021 results being added to the tables as they become available. Note that spread/cover and total/over models are graded against the numbers as they are shown in the tables above. Additionally, links to detailed recaps of all predictions from previous tournaments can be found by following links in the header menu above.

Year Win Prob 284 Spread Cover Prob
Rec % Rec % Rec %
2017 47-15 0.76 34-26 0.57 42-21 0.67
2018 44-19 0.70 32-28 0.53 35-27 0.56
2019 45-18 0.71 24-36 0.40 27-33 0.45
2020
2021* 43-20 0.68 33-27 0.55
Total 179-72 0.71 123-117 0.51 104-81 0.56

Year 284 Total Over Prob
Rec % Rec %
2017 39-22 0.64 42-21 0.67
2018 38-25 0.60 36-27 0.58
2019 32-29 0.52 32-29 0.52
2020
2021* 36-22 0.62
Total 145-98 0.60 110-77 0.59

* Note that we have modeled the 2021 tournament using a different data/methodology than in previous years due to the various irregularities caused by COVID this season. Applying the historical data/methodology would have been a bit of a “square peg round hole” situation due to these irregularities. So, we decided to model the 2021 tournament with a clean slate, using data only from the 2021 season to generate our models/predictions. Additionally, note that we will not be producing the Cover Probability and Over Probability models this year, so the “284 Spread” and “284 Total” will be our main focus for spreads and totals in 2021.