Handicap Index for Past Month: Who's Doing Well?

[DobbsRover2]

In another thread concerning ND and UTenn, the question was brought up about a certain team that seems to kind of play to the level of its opposition a lot and what this says about game preparation for top teams. That led me to wonder whether it is true that certain teams either underachieve or overachieve on a regular basis.

Unlike the NFL's Against the Spread stats, I'm not aware of any stats that rate how WCBB teams perform versus expectations. One rating system that does make daily predictions for WCBB scores however is Massey, and it still has prediction records going back to Dec. 15. So I ran a quick scan for a bunch of top WCBB teams to see how they've rated against the Massey spread for about the last month (36 days in all but with the holiday break). Which ones are pretty stable, which ones are volatile, and when they are well off predictions which side do they veer toward?

A few notes. Not everyone maybe agrees with the accuracy of Massey predictions, but they seem to be reasonably good compared to other systems, and they helpfully do have at least some past predictions still available to view. The predictions for first month of the season when teams were sorting themselves out are no longer available on Massey, so this only covers more recent times when teams have developed more of a rhythm. Of course handicapped prediction systems penalize you for doing well by upping the ante for you in following games, so any Husky fan who expects UConn would look good in the results will be somewhat disappointed. If you beat ND on the road by 18 and smacked Duke by 31 at home, you are expected to beat Tulsa at home by 54 even if the regulars aren't in for much of the game's last quarter. Lastly, the W is the biggest thing for a team in any game, and in only two games for the 8 teams reviewed did underachieving lead to a loss, for ND at Miami and Kentucky at LSU.

The teams selected were the top 7 in the rankings plus a Kentucky team that to me has always epitomized the crazy boom and bust type team that can knock off top-3 teams and also lose to #50 and #68 teams. Six of the teams played 8 or 9 games in the period, while UConn had 10 and Maryland 7. The stats below divide the results into three types: games where a team exceeded predictions by 6 or more points, the games where they pretty much met expectations by exceeding or missing the prediction by 5 points or less, and the slacker type games where they were 6 or more points off projections. An average margin versus the spread is also given:

1. South Carolina: 5-3-1 (8.6)
2. Maryland: 4-2-1 (7.4)
3. Baylor: 4-4-1 (3.2)
4. UConn: 4-3-3 (0.5)
5. Tennessee: 3-4-2 (-2.0)
6. Louisville: 3-1-5 (-3.3)
7. Notre Dame: 1-4-3 (-5.6). One unexpected loss
8. Kentucky: 1-2-5 (-7.0). One unexpected loss

The Gamecocks have been blasting a number of teams in a somewhat fluffy stretch of schedule by bigger than expected margins, and the Terps have done well against a slightly tougher lineup. Baylor has been very stable and missed expectations by a lot only in a 2-point win over Syracuse, and UConn needed to paste St. John's, Temple, and Tulsa by much bigger margins. Tennesse has well exceeded expectations against Stanford and Texas A&M while underperforming against Wichita and Missouri, and Louisvile can be way up against Old Dominion and Georgia Tech but also slack off against a lot of weaker teams. ND has been unusually jumpy probably largely because of the Reimer situation, and KY as usual is just crazy, and on the crazy bad side.

 

[itsasport]

One thing that makes this a little murky is in the instance when a team is behind by only a couple of points at the end of thae game and starts committing intentional fouls. For an example, in the recent Baylor/Kansas game, Baylor was ahead by 2 points with a bout a minute to go. Kansas used the fouling tactic and the mead free throws put Baylor up by 8 in the final score.

Even that score should considered under-achieving; but, the actual performance of Baylor in that game should be based on a win of about two points. Perhaps the ESPN play-by-play could used to discover a more accurate spread.

 

[DobbsRover2]

One thing that makes this a little murky is in the instance when a team is behind by only a couple of points at the end of thae game and starts committing intentional fouls. For an example, in the recent Baylor/Kansas game, Baylor was ahead by 2 points with a bout a minute to go. Kansas used the fouling tactic and the made free throws put Baylor up by 8 in the final score.

Even that score should considered under-achieving; but, the actual performance of Baylor in that game should be based on a win of about two points. Perhaps the ESPN play-by-play could used to discover a more accurate spread.

Yeah, and obviously another instance was ND and UTenn where the Vols were down by 6 and then started fouling like crazy in the last minute and their margin of defeat grew to 11. That changed their margin of difference from being +3 to -2 since ND was projected to win by 9. That didn't put either team in a new category here since it was all in +5 to -5 area which is around meeting expectations. For Baylor it did move them into the middle "made expectations" category at -4 from the -10 they were at around the minute mark.

It does change the average MOD and sometimes switches categories here, but end-game fouling is always a gamble and sometimes it works and sometimes it doesn't, and the Irish's or Bear's margins could have been shrunk if they missed FTs or the Vols and Wildcats made baskets. There was the famous case with ND and DePaul this year when the fouling was highly effective and even turned the game around, so things can even out.

 

[RockyMTblue2]

One thing that makes this a little murky is in the instance when a team is behind by only a couple of points at the end of thae game and starts committing intentional fouls. For an example, in the recent Baylor/Kansas game, Baylor was ahead by 2 points with a bout a minute to go. Kansas used the fouling tactic and the mead free throws put Baylor up by 8 in the final score.

Even that score should considered under-achieving; but, the actual performance of Baylor in that game should be based on a win of about two points. Perhaps the ESPN play-by-play could used to discover a more accurate spread.

As Dobbs points out, the end of the game fouling gamble that takes place in close games...well, it is part of the game. So is the team that is down four and jacks up wild 3 pointers for the last two minutes and loses by 14. Now, if we start weighting statistical results for "free throw parades", should be other late game phenomena that should be measured to add to "validity" of the metrics, such as "aberrant star foul outs with 10 minutes to go", or So & So ref'd that game, diminish foul shots made by 6.... Is this a bad spoof, yes. Do I hate statistics, yes. Do they have their place, absolutely. The remaining fans of baseball and all sorts talk radio would disappear without them.

 

[DobbsRover2]

Actually, the most relevant factor for UConn at end of game would be that the reserves are often playing for 7 to 10 minutes, and that can turn a margin that's heading toward 50 to settle in around 35. These type of performance-versus-the-spread analyses are not really useful for games with margins projected over 25, and that unfortunately encompasses a good many UConn games.