By the Numbers, Vol. 14: Season Wrap Up
FINAL RATINGS First and foremost, congratulations to Minneota, Minnehaha Academy, DeLaSalle, and Hopkins for being the four teams to finish their season with three consecutive wins at the state tournament and come away as champions. They finished their seasons playing…
Access all of Prep Girls Hoops
Continue reading this article and more.
Continue ReadingFirst and foremost, congratulations to Minneota, Minnehaha Academy, DeLaSalle, and Hopkins for being the four teams to finish their season with three consecutive wins at the state tournament and come away as champions. They finished their seasons playing 31, 31, 29, and 32 games respectively. I had many more games than that, 5133 games to be exact. That is the total number of games, that I am aware of, that were played in the state of Minnesota from November 21 through March 16. Each game was predicted and the result was recorded. The table below shows the breakdown of how that went per month as well as during the state tournament.
Date | Predicted Record | Actual Record | Win Percentage |
All Season | 4309-824 | 4199-934 | 81.8% |
November | 289-70 | 256-103 | 71.3% |
December | 1172-240 | 1128-284 | 79.9% |
January | 1322-227 | 1310-239 | 84.6% |
February | 1377-248 | 1360-265 | 83.7% |
March | 149-39 | 145-43 | 77.1% |
State Tournament | 32-12 | 36-8 | 81.8% |
As you can see, the computer got better as the season went on, which is to be expected as more data is entered. However, it never quite lived up to the expectations as the predicted record was always better than the actual record. There are two possible explanations for this. The first is that the percentages are too aggressive and the favored teams are given too high of a chance to win. The other option is that the rating system isn’t perfect. To check this, I looked at the percent chance the favored team was given to win, and grouped them together.
Predicted Percent Chance | Record | Win Percentage |
50-60% | 305-249 | 55.1% |
60-70% | 355-221 | 61.6% |
70-80% | 482-200 | 70.7% |
80-90% | 723-156 | 82.3% |
90-100% | 2338-108 | 95.6% |
Ideally, the win percentages would fall in the middle of each range. This table rules out the first explanation of the percentages being off since the lowest percent chance was accurate and the highest percent chance was accurate. What I read from this table is that the rating system did fairly well as a predictor, but it isn’t perfect and it never will be. My goal is to get it as close to perfect as I can. I do have some ideas going forward to improve the ratings for next year to make them as accurate as possible. I plan to play around with the blowout multiplier and try a few different options as well as potentially add a fourth level of adjustments. There are a few things that the computer will never be able to fix completely due to the restrictions of the fairly closed scheduling system that most schools are bound to and the extreme differences in team quality, but there is always tinkering that can be done to attempt to make them better.
Also for next year, look for a Strength of Record rating for each team, which attempts to quantify the difficulty of obtaining a particular team’s record. This value would take wins and losses into account by comparing a team’s record to what would be expected if an average team played their schedule. Because of its emphasis on wins, the Strength of Record rating would be the more appropriate value to use in the seeding of teams. However, the efficiency rating would still be used for predictive purposes.
I want to thank everyone who took the time to look at the ratings this season as this was the first time they were able to be put in front of this large of an audience. I appreciate your patience in adjusting to the different school of thought that I come from and I understand if you disagree with me. I appreciate all your questions and feedback throughout the season and I plan to use that to make the rating system better.
Biggest Upsets
Below are the five biggest upsets this season according to the computer and the percent chance they would occur.
- St. Croix Prep 38, Southwest Christian 35 (0.62% chance of upset) – 1/12/19
- Park of Cottage Grove 56, Lake City 47 (1.12%) – 12/27/18
- Minneapolis Henry 50, PACT 49 (1.72%) – 1/3/19
- Greenway 38, Littlefork-Big Falls 35 (2.15%) – 12/13/18
- Redwood Valley 62, Eden Valley-Watkins (2.35%) – 3/8/19
Highest Scoring Games
Below are the five highest scoring games of the season. The average point total this season was 105.6 per game. These games far surpassed that.
- Park Center 99, Prior Lake 95 – 12/1/18
- Stillwater 101, Cretin-Derham Hall 93 – 2/19/19
- Osseo 98, Red Lake 94 (2OT) – 12/29/18
- Becker 101, St. Paul Como Park 86 – 2/18/19
- Chanhassen 94, Robbinsdale Cooper 92 (OT) – 2/22/19