The idea is that good is driven by metrics, not perception, I think your misinterpreting what he's saying. It's like saying a team that has gone 0-10 sucks, they won't be rated in any AP poll ever, but if they're playing a team that is 10-0 the 10-0 is obviously favored right? Well what if the 0-10 team has played the hardest SOS, and the 10-0 the easiest, his ratings are meant to be predictive of who will win that 0-10 vs 10-0 matchup, not tell you that the 0-10 team sucks.
i get not everybody loves KenPom, margin of victory is confrontational (even if it's been proven to matter) among other things. But it's better than everything else out there, there is a very small deviation between KenPom predictions and vegas lines, it's been shown to be the best predictive system out there in determining ncaa tournament games, so there's clearly some value. If you blatantly dismiss it, that's fine. But kenpom is working as a consultant for numerous ncaa teams, so there are a lot of people who do believe in what he is doing.
The best thing he's done is help move the narrative from simple raw number statistics and into rebounding %, TO%, FT/fGA and such.
I love KenPom.
But there are simply things that numbers can't capture and can never capture. What he calls "Luck" I call "Competitive Fire". There is a reason why teams are consistently "Lucky" and it's not because they're rubbing rabbit feet. Similarly, there's a reason why teams are consistently "Unlucky". It's my same issue that I have with metrics in baseball. I get that BABIP. There's a norm and a standard deviation. It works in the aggregate. But when I see a pitcher or batter who is consistently, season-to-season, outperforming the norm... it's something that guy is doing, not simple good fortune.
Some teams and players are winners and some teams and players are losers.
I was listening to an NBA guy on the radio today (may have been Reggie Miller) and his comment that of the 10 guys who are on the court in a close game, only 3 of them want the ball in their hands to take the last shot... and you just hope your team has 2 of them. This type of mental toughness and competitiveness and thriving on pressure is what separates some coaches from others and it permeates the culture of a team. It is why when KenPom's numbers tell me that Michigan State and Tennessee are equivalent teams this year I roll my eyes (or that Stanford is #35 and Colorado is #64). By necessity, he's got to write overperformance or underperformance as "Luck" because it's probably impossible to statistically account for. But when teams consistently find a way to win, even if it's winning ugly and winning close, that means something. That "Luck" becomes "Fact".
As much as the metrics are useful tools for uncovering tendencies, revealing team strengths and weaknesses, giving insight into styles of play, and pointing out which factors have greater relevance in leading to wins... it's still sports. It's still competition. The metrics may also be useful as predictors. But they don't know how to quantity the "Competitive Fire" as a modifier.
Last, as you brought up, MOV is a highly questionable statistic. In many cases, it might point to a team having more offensive firepower so when they have things going they really get out of hand... especially against bad teams.
The other side of that is when you have a coach like Tad who plays things very conservatively in endgame situations. If he's up by 7 with 4 minutes left, he's going to look at it from the point of view that if he can take 30 seconds per possession there are only 4 possessions per team. Even if his team is taking difficult shots that might lower CU's PPP to a 0.8, he's still at a +10. He'll take that and bet that his defense isn't going to give up a 2.5 PPP to the opponent. And if the other team decides to extend the game by fouling, that plays into his hands too. Because it virtually guarantees that the Buffs will be over a 1.0 PPP and he trusts his defense (even if the opponent is making miracle shots) not to give up so much above that to cause a loss. Will it yield more 1-5 point wins than 10-15 point wins? Sure. But that doesn't make the Buffs a worse team than if they'd gone for the blowout. At the end of the day, all that matters is the W or L. Tad gets that, but his Buffs get penalized in the metrics and rated the #64 team. Buffs aren't 10-2 in games decided by 7 points or less by accident or "Luck". To accomplish that with one of the youngest teams in the country shows a winning culture. Likewise with Pitt only being 7-7 in those types of games even with a veteran team. I couldn't give a **** less whether Pitt beat Savannah State by 30 or 15.
P.S. I don't gamble any more, so the predictive measures from this stuff are inherently less meaningful to me. I only care about the W or L, not the style points. But I would assume that over the course of most seasons that Tad's teams are a solid bet against the spread since they win games they "shouldn't" and rarely lose games they "should win". I'd think that would usually outweigh the fact that they're not always delivering the margin they should or, in a season like this one, where a young team let some scores get away from them.
