Jeremy here for a rare PB appearance. I wanted to share some numbers that a friend of mine - we'll call him Doug - put together using a derivative of PB Mark's Real Standings. We all - or at least I do - love the computer rankings, and a cursory review of Sagarin or Pomeroy or Greenfield shows KU to be very well represented. However, computers, like humans, can be fooled. The computers love Duke, but is that because they're a legitimately great team, or is it because they leave their starters in longer than they need to and run up the score against their opponents at home? Would a great team lose multiple games by double digits to ok teams? That's what we're trying to figure out here.
REALly great teams do more than beat up on patsies. REALly great teams win on the road. So let's take a look at how teams did in tough games versus how they might be expected to do. The results are revealing.
So how we judged. An elite team should...
1) win all home games 2) win all non-top 25 neutral court games 3) win all non-top 50 road games 4) split top 25 neutral court games 5) split #26-50 road games 6) lose all top 25 road games
By applying those rules to each team in the top 25, we came up with an expected number of wins and losses. If a team would have been expected to have won 6 games and lost 4, but went 4-6, it's an indication that they are not as good as a team which did the opposite. As Doug indicates:
Under that system, KU would have been expected to go 3-6 against the top 25 and 24.5-6.5 overall but was actually 8-1 for a +5 'score' against the top 25 and +4.5 for the season.
Syracuse would have been expected to have a 3-2 top 25 record, but was actually 4-1 for +1. For the season, they would have been expected to go 27-4, but were 28-3, so also +1.
Duke would have been expected to go 3-4 against the top 25, but was actually 4-3 for +1. For the season, they would have been expected to go 26.5-4.5, but were 26-5, so -0.5.
Kentucky would have been expected to go 1-1 against the top 25, which was their actual record for +0. For the season, they would have been expected to go 29-2, so were again +0.
The table below shows how each team in the top 25 compares in Actual versus Expected wins in "Losable" games.
Tennessee, Mizzou, Pittsburgh and Wisconsin all stand out as maybe not so good as the computers think.