The NCAA Selection Committee is in love with the big win. The problem, though, is this– they way overvalue it. Literally nothing Arizona State did after beating Xavier and Kansas by mid-December mattered. They lost a whopping nine games to teams not good enough to make the field.
It’s a philosophy that serves to benefit the power conference teams as a whole, even though one or two members here and there may be aggrieved in a given season. But I can show that it is a philosophy that is not grounding in reality of picking the best candidates. How? As I ranted yesterday on Twitter, by looking at actual results.
If simply taking the bubble teams with the best wins was the best way to do things, you would think that those teams would then be good bets to advance, at least compared to their peers that could not win the big one in the regular season. In fact, the opposite is true.
None of the Cinderella stories came from teams that look like Arizona State (wins over Xavier and Kansas), Providence (two wins over Xavier, one over Villanova), Virginia Tech (wins over North Carolina, Virginia, and Duke), or NC State (North Carolina, Duke, Arizona). None of them looked like the Big 12 bubble teams with all their Quad 1 wins. Of the eight teams to advance to an Elite 8 as an 8 seed or lower, 5 of them came from outside the six power conferences, and the 3 that were from power conferences went a combined 0-11 against the RPI Top 12. They often got seeded down because they couldn’t “win the big one.”
If we expand it to 7 seeds, we add four more Elite 8 runs. None of those teams had multiple big wins against the Top 12 in the regular season either. Altogether, the 12 lowest seeded teams to advance to an Elite Eight were a combined 3-26 against the Top 12 teams in RPI.
Here is a summary:
I think it’s fair to wonder if a lot of those teams would have even made the field with how this current committee is making decisions. Even though most of our Cinderella runs have come from teams that look more like Middle Tennessee and St. Mary’s, the Committee struggles to put their records in perspective.
That 2008 Davidson team? Sure, they were 26-6 and tested themselves by playing Duke, UNC, and UCLA. But who did they beat? They went 0-4 in Quadrant 1 games. Sorry, Stephen Curry, you are out of this field.
2011 Butler? They got an 8 seed, but I’m not sure they even get in now. I mean, compare their resume to Middle Tennessee. Their only Quad 1 win was at Cleveland State. Not many people have been able to stop Brad Stephens, but Bruce Rasmussen’s gang could have, by not putting them in. VCU was a controversial selection in 2011, but would have no chance today.
2013 Wichita State and 2014 Dayton? Sorry, didn’t beat anyone that was all that good. Heck, Xavier from just last year bears a striking resemblance to Louisville, in RPI, Quad 1 record, and no top wins. Syracuse two years ago was 0-4 against the top teams.
Let’s also take a quick look at the opposite. Here are the 11 teams that had an RPI of 40 below but had two or more wins against the Top 12 teams and got in the field in the last decade:
Those teams that got their seed boosted or got in the field because of big wins, despite a mediocre RPI? They went a combined 3-8 in the first round, and none advanced to the Sweet 16. So much for being teams that rose to the occasion.
I’m going to spit some fiery truth here. Those teams got in because of what they did in 10% of their games. But if you are on the bubble, you are flawed in some way. Those teams were not tournament teams the other 90% of the time, and lo and behold, when they got in, they played more to their usual form. The Selection Committee way overvalues greatness by association and fails to properly put wins in context based on volume of opportunity. You always hear “body of work,” but putting Arizona State in comes from looking at part of the face but completely ignoring the body.
I know there’s this big “death to the RPI” movement, and believe me, I’ve led the charge. But the biggest problem isn’t the RPI, it’s the underlying philosophy. It’s over crediting one or two wins and not accounting for bad losses by major conference teams. It’s thinking Kansas is the #1 seed because they have 12 Quad 1 wins but not considering that they created their own cottage industry of putting teams in contention on the bubble by losing 4 games to teams that aren’t in the field (plus Arizona State). It’s thinking Michigan State is not because they don’t have enough Quad 1 wins, when the worst loss is to Michigan, a #3 seed, twice.
If you replaced the RPI with Pomeroy’s rankings tomorrow as a sorting tool, but didn’t change the philosophy, nothing would change. Michigan State would still have a 4-4 Quad 1 record, while Kansas would be 12-4 against Quad 1. Saint Mary’s would still have a 2-1 Quad 1 record but only the win at Gonzaga if we ignore everything else. Arizona State would still have those wins over Xavier and Kansas, and have 9 losses in Quads 2 and 3.
I’m going to continue to stump for “Wins Above Bubble” as a philosophy, and hope someone with way more clout picks up the cause. Wins Above Bubble is the amount of wins you have, minus the wins a typical bubble team (with average luck) would have playing your schedule, and you can see Seth Burn’s calculations here.
It is legitimately crazy to me that if Arizona State had gone 20-11, but they had won at Oregon State and beaten Oregon at home, but lost to Xavier and Kansas, that they are viewed so differently. You are, as Bill Parcells is famous for saying, what your record says you are. But there’s no doubt they would have had zero chance of being in the field with the exact same record, but a different ordering of specific game results.
And what Arizona State’s overall body of work says is that they aren’t a tournament team. They were 1.6 wins below a bubble team. A tournament team would have been expected to win 22 games against Arizona State’s schedule. Yes, Arizona State’s non-conference was great and it shouldn’t be ignored. Wins Above Bubble would have them about 3 wins more than expected for a bubble team in the non-conference. But with their conference schedule, where they played Washington State and California–who are two of the three lowest-rated power conference schools along with Pitt–and did not have to play UCLA and USC on the road, a bubble team would have gone 12-7 in the 19 games they played. The cushion they built up would have allowed for less than that. But not 4 games less. It’s as if those losses didn’t count.
The Committee keeps making the same mistakes over and over. Someone argued with me on Twitter that I couldn’t look at what happened after seeding to say they made mistakes, because who deserves to be there and who can win are different things. I disagree. You shouldn’t overreact to one or two events, or one or two tournaments (or, you know, one or two big wins). But you should look at outcomes over time to see if a rationale was supported or not. Our definition of accomplishment should be tied to doing the things that most indicate quality of team, which should then lead to showing that going forward.
If you claim that big wins were the best indication of resumé quality, to decide who the best teams were, you might want to test to see if that held up. Since the teams that have been successful near the bubble don’t support that, maybe it’s time to admit we actually need to look at the body of work, and not just one part of it.