Share this:

Efficiency has direct correlation to national championship teams

By RICHARD CHEEKS, Contributing Writer

The Men’s Basketball championship is in the books for the 2018-19 season, and Virginia has taken home the trophy. Their path to the end prize was anything but easy, and as the old adage goes, nothing worth gaining comes easy.

However, I believe Virginia’s work toward this end result did not begin with the first round of this tournament, or with any subsequent round of the six-game tournament journey. Virginia earned this championship with a season long effort that resulted in their team having the best adjusted net efficiency in the country for their entire body of work.

This is not an isolate occurrence. Since 2002 (18 champions), ten of the 18 champions have ended the season as the #1 efficient team based on each of those team’s full body of work for these respective seasons. 10 of 18. Furthermore, 3 of the last 4 champions were #1 efficiency team, and 6 of the last 10 champions have been #1.

When the #1 team does not survive the gauntlet of this tournament, which has happened 8 of the last 18 years, the #2 team has won the crown 2 times, and either the #3 or #4 team has won the championship 3 times.

That is a total of 15 of 18 champions from the top 4 most efficient teams. I am not talking about seeds. I am talking about rank. One of the top 4 teams has won this championship 15 of 18 years. If the seeding in this tournament had been based solely on efficiency ranking, a #1 seed would have won this tournament 15 of 18 times.

Yes, there are 3 exceptions to this rule. In 2003, Syracuse won the tournament as the #8 most efficient team in the nation (equivalent 2 seed); in 2011 #10 UConn won it all (equivalent 3 seed), and in 2014 #15 UConn (equivalent 4 seed) won the tournament.

None of this means that teams that are not in the top 4 efficiency are chopped liver. That is absolutely not the case. Upsets are a part of competitive endeavors. The 2018-19 season is now in the books, and the upset rate for this season was exactly as it has been each of the prior seasons.

This year, there were 5,603 D1 college basketball games, and 1,476 of those games ended with the team favored to win based on efficiency lost to the team that was predicted to lose on the same basis. That is an upset rate of 26.3% for the season.

The upset rate is a function of the theoretical margin for a particular game based on this game by game efficiency analysis. For example, if the analysis projects a theoretical margin near 0 points, a true “pick em” game, the upset rate should be about 50 percent, and as the theoretical margin for a game rises from 0 points, the probability that the favored team will win rises from 50 percent until it approaches and becomes asymptotic to 100 percent with very large theoretical margins.

The upset rate in the NCAA tournament this year was lower than the theoretical upset rate indicated in the table, and the average upset rate for the regular season. This year, for example, the experience of the regular season indicated that there should be 22 upset outcomes among the 67 NCAA tournament games. However, this tournament had only 17 upset outcomes, start to finish. 10 of those 17 upsets occurred in the first weekend. There were 3 upsets in the sweet 16, and 3 upsets in the elite 8. There was 1 upset in the final four, and the championship game did not end with an upset.

The upset rates also varied by probability as discussed above for the entire season. For games with theoretical margins of 0-2 points, there were 13 games in the tournament, and 6 ended with upset outcomes. For games with theoretical margins of 2+ to 5 points, there were 15 games with 6 ending in upsets. For games with theoretical margins of 5+ to 10 points, there were 19 games, 7 ending in upsets. For games with theoretical margins > 10 points, there were 20 games with no upsets.

Some said prior to this tournament that they expected more “chalk” this tournament, and the 17 upsets this tournament is substantially below the number of upsets in the 2018 tournament that ended with 25 upset outcomes out of 67 games, against a theoretical 23 upsets (as compared to this year’s 22 theoretical upsets). 2019 had more “chalk” than last year. Yet, last year’s champion was also the #1 efficient team in the nation, which indicates that the best of the best can and often do survive this gauntlet despite the raging upsets that swirl about this tournament.

The bottom line: There are no guarantees, but when a team is the best team (efficiency based), it will win this tournament over 50 percent of the time, and the top four teams will produce a champion over 80 percent of the time.

I want my team to be in this very exclusive club, and to be in this very exclusive club, it must have an ANE > 0.3 ppp, and have offensive and defensive efficiencies in the top 20. These parameters should be guiding lights for any program during every season.

5 comments

Skip to comment form

  1. This fails to take into account that Virginia did not have to play Duke again; a team that had beaten them twice.
    It also fails to take into account the fact the Refs completely failed to make the walking call on Virginia. If they make the call, game over; Auburn moves on.
    I agree that Efficiency plays a huge part in what a team does in the tournament; but so do many other factors such as injuries, good calls and no calls have a huge impact, experience cannot be measured, and some nights a team may be off or on fire; no way to factor any of these things into the results.
    A few years ago, UConn did not have a good season, but got on a roll and won it all.

    1. Only thing I can add is Auburn got robbed.

  2. Virginia had a nice team…not great. They were LUCKY to win it all, and with similar positive breaks from the refs, we would have been in the finals.

  3. Good article Professor. Knowing the factual content will be very helpful in evaluating next year’s field. The eye test seems great but it’s hard to argue with 50% prediction rate on the winner based on #1 net efficiency

  4. #1 1 spot 10/18 55.6% 1.8:1 odds
    #2 1 spot 2/18 11.1% 9:1 odds
    #3 and #4 2 spots 3/18 or 1.5/18 each spot 12:1 odds
    #4 Through #16 12 spots 3/18 or 0.25/18 each spot or 72:1 odds
    #17 and higher, 0/18, odds of such a team winning this championship is too large to state.

    I want my team in that top 4 group, preferably #1

    Pomeroy provides the easiest way to track this but the Pomeroy values are not based exclusively on the current season until a team has played 20 games, so temper the values you see there until at least mid-January each season.

    Kentucky finished the year #8, and the E8 finish is right on their full body of work. Teams can over achieve or under achieve this in the tournament by about +/- 1 round, thus a Sweet 16 or FF exit would not be out of the question for this team, but teams fall out 2 or more rounds early on rare occasions, and advance 2 or more rounds deeper than this on rare occasions.

Leave a Reply