BIG BLUE FANS FOR
2011-12 Season Analytical Writings
Upsets are a curious phenomenon in sports. They give eternal fear to those that “have”, and they give eternal hope those “without.” Upsets are the product of being human, e.g. variability in performance levels, game to game. All of us have it; all teams have it.
Kentucky is now preparing for its final 4 games. Two at dangerous SEC venues at which the Cats will be vulnerable to being upset, and 1 at Rupp against Vanderbilt, and frankly, upset fever could rule the day in Rupp that day too. I think we all agree that this will be a very tough finish to a very successful season. The Cats will enter each of these events as the favored team, but such status does not guarantee victories, as we all know very well.
I define a team based on what it usually does, not just its best efforts, or its weakest efforts. I look at a team's average effort. All teams have a best, a worst, and an average. The spread between the extremes, on either side of the average is a measure of that team's consistency, or if you prefer, inconsistency. We can measure that spread; we can apply a statistical value to that spread, e.g. the standard deviation. Furthermore, if we divide the standard deviation by the average, the result is the variance. Therefore, we can define each team's profile by its mean and variance. This is true of shooting percentages, turnover rates, rebounding rates, pace, and this is true of Net Game Efficiency [NGE]. Teams that have higher average NGE, have played better basketball, start to finish over the course of a full season. But for the occasional upset outcomes, a team with the higher NGE will defeat teams with lower NGE.
However, no team can play their average every game every game. In fact, teams play their average very few times each season. Teams play about 1/2 of all games at levels above their average, and they play about 1/2 of all games at levels below their average. As the gap between the average NGE for any two teams widen, the probability that the better team will win increases from 50%, when the values are nearly the same, and approaching 100% as the gap becomes large. The predicted margin for any two teams is a function of the gap between their respective average NGE, and the database is now clearly sufficient to establish the relationship between predicted margin and winning percentage [thus probability of winning].
Within the limits discussed below, no team is immune to the upset; no team is unable to secure one. The vast majority of teams play a representative array of opponents, with several that are stronger and several that are weaker than they are. These teams will get an occasional upset, and these teams will be upset on occasion. They will win some they should have lost, and lose some they should have won. For the vast majority of teams, those bonus wins, and stinker losses tend to even out over the course of time.
However, there is a special exception to this general rule. Teams that find themselves at either end of the spectrum of talent and ability may only have one practical alternative in the upset saga. For teams at the very top, they seldom play teams that are clearly better than they are [there are not that many out there], therefore, those teams are vulnerable to be upset, but rarely get to enjoy the taste of getting a big upset of their own. For teams at the very bottom, the opposite is true.
Upsets happen! It is part of sport. For those on the outside looking up at those who have great talent, upsets are their source of eternal hope. The upsets are also the source of their eternal lament as well because these same teams that hang onto the possibility that they could snatch a big one from a team like Kentucky, also realize that their team has just dropped one to Georgia. Therefore, the lament is a common one, "If only my team could be consistent.” Nevertheless, all teams are consistent in their inconsistency.
So, upsets are part of the game. For Kentucky, the upset bug only bites one way. When the upsets occur, if they occur, they will still be the same team that has started 26-1, 12-0.
Submitted by Richard Cheeks
Submitted by Richard Cheeks