First, our efficiency metrics may fall if this inconsistent play continues.
This is always true. We were roughly #10 on Barttovik at one point -- we are down to #14 now. If we keep repeating performances like the ones we had against TCU and Baylor, then obviously our computer ranking is going to head south quickly.
Second, it is a fine line. We’re basically talking a handful of outcomes. It doesn’t take much degradation in SD to have a swing that have a team on paper top-15, but change a great year record wise (plus regular season championship/great-seed-facilitating-run into a mid-pack), mid-seed 1st or second round exit.
The models are fun for seeing the teams on more of a continuum past simple binary outcomes, but everything that matters is still completely weighted on the ability to string isolated positive outcomes together.
Models are
fun but also
useful as predictive devices. No, they are not perfect, but Las Vegas makes a lot of money off having the best ones, so there are things to learn from them about the way the team is likely to play the 1.5 months of the season from here.
I know a season is judged and goes into the record books based on W-L binaries. Period. Having the best adjusted net efficiency in the Big 12 does not put a banner in the rafters. The point about the computers is not that they somehow undo a team that underachieves.
If anything here, I think they have two uses...
(1.) Predictions about how the team is likely to perform in games remaining. Towards this, the computer say this is a good team that should win many of those games.
(2.) Giving you some objective standard as to if a team
did under-perform or not. To judge if a team under-performed, you have to have some objective metric for how good it was in the first place. The model helps calibrate expectations for how a good a team was before you compare that standard to their actual W-L records. Was this a mediocre team who squeaked by a lot with a low ceiling and we should be happy to get what we get? Or is this a team who had a high ceiling but played bad games on occasion and therefore underachieved by losing a second round game in Columbus, Ohio to some #10 seed from a mid-major conference?
Particularly when we saw that team play like it could take on the world at times?
We use the models to have a more objective discussion of #2.
Every year there are teams that do well by metrics from not having a consistent weakness, but that’s often not enough to win the games commensurate of the metrical ranking.
This is exactly my point -- a much better summary of mine about the downsides of consistency at a lower rating versus the roller-coaster we are on right now.
They both have their good and their bad. I would rather roll with the upside, though.