- Joined
- Aug 26, 2011
- Messages
- 9,346
- Reaction Score
- 23,550
Pardon the deviation from the Ollie saga, but I'm already angry about the living contradiction that is the NCAA Tournament selection committee. They glorify the "human component" and value the "basketball people" on the committee, because apparently, statisticians and odds-makers are not qualified to rank the field despite doing that sort of thing for a living. Let's look at some of the outright crime that is being forecast for tomorrow:
Xavier is apparently LOCKED FOR A ONE SEED
This is where one can argue that relying strictly on the good old RPI will lead you astray. The Musketeers are no strangers to the formula and scheduled only two games against opponents ranking below 200 in the RPI. They piled up a staggering 10 victories over teams ranked 51-100 and won the conference title despite being swept by the only other ranked team. In total, they have ONE win against top 25 competition (Cincinnati at home way back in December) and place 14th on KenPom. The analytics suggest that their profile is closer to that of a four seed and they may very conceivably open as underdogs against whoever they draw in the sweet sixteen. Madness.
Rhode Island is an eight and Michigan State is a four despite having virtually the same resume
See if you can spot the difference:
Rhody - #14 RPI, 1-3 vs. top 25, 3-4 vs. top 50
MSU - #15 RPI, 2-4 vs. top 25, 2-4 vs. top 50
When you add in other relevant data, it becomes clear that Michigan State is far better. But if we're going to reward Xavier for gaming the RPI, we should be consistent, right?
What to do with Arizona?
So, uh, here is one where they can either be complete hypocrites or they can stay true to their word and just get slammed by everyone for 48 straight hours. This is a team that some, myself included, had preseason #1. Their roster is silly loaded...and that was before they signed the best big man prospect since maybe Shaq. The problem is that they stumbled around OOC (they finished 8th out of 8 in their preseason tournament) and then dropped four more games in what a lot of metrics ranked the 7th best conference. They're currently 18th in RPI (update before tonight's win over #34 USC) and 21st on KenPom. They have zero victories over top 25 teams. There is objectively no reason to seed them as better than a four (I know some of their players have missed a couple games, but nobody else is being extended that benefit of the doubt).
There are more examples but you get the point. It's one conflicting data point after another, and while you can observe some effort on the part of the committee to meld them together, the absence of a universal formula - even if you're only using it to cluster teams to preserve the tension of the selection show - leaves the process vulnerable to bias and human error.
Xavier is apparently LOCKED FOR A ONE SEED
This is where one can argue that relying strictly on the good old RPI will lead you astray. The Musketeers are no strangers to the formula and scheduled only two games against opponents ranking below 200 in the RPI. They piled up a staggering 10 victories over teams ranked 51-100 and won the conference title despite being swept by the only other ranked team. In total, they have ONE win against top 25 competition (Cincinnati at home way back in December) and place 14th on KenPom. The analytics suggest that their profile is closer to that of a four seed and they may very conceivably open as underdogs against whoever they draw in the sweet sixteen. Madness.
Rhode Island is an eight and Michigan State is a four despite having virtually the same resume
See if you can spot the difference:
Rhody - #14 RPI, 1-3 vs. top 25, 3-4 vs. top 50
MSU - #15 RPI, 2-4 vs. top 25, 2-4 vs. top 50
When you add in other relevant data, it becomes clear that Michigan State is far better. But if we're going to reward Xavier for gaming the RPI, we should be consistent, right?
What to do with Arizona?
So, uh, here is one where they can either be complete hypocrites or they can stay true to their word and just get slammed by everyone for 48 straight hours. This is a team that some, myself included, had preseason #1. Their roster is silly loaded...and that was before they signed the best big man prospect since maybe Shaq. The problem is that they stumbled around OOC (they finished 8th out of 8 in their preseason tournament) and then dropped four more games in what a lot of metrics ranked the 7th best conference. They're currently 18th in RPI (update before tonight's win over #34 USC) and 21st on KenPom. They have zero victories over top 25 teams. There is objectively no reason to seed them as better than a four (I know some of their players have missed a couple games, but nobody else is being extended that benefit of the doubt).
There are more examples but you get the point. It's one conflicting data point after another, and while you can observe some effort on the part of the committee to meld them together, the absence of a universal formula - even if you're only using it to cluster teams to preserve the tension of the selection show - leaves the process vulnerable to bias and human error.