Monday, March 24, 2014

Trying to fix the SuperBracket

I really like the idea of the SuperBracket...but it's not right yet. At the moment, here's where we're at:


On the top we have the values I originally created. By Friday I had identified some problems and wrote about them. I tried a band-aid fix by creating alternate values. I think it's in improvement, but the whole thing is still screwed up.

I had an idea that I thought might help...which is each round is worth points, plus you get a bonus for seed differential versus chalk. The first round is obvious. A good exmaple of what I mean is this year's Tennessee-Mercer game. Penciling in a 11seed Tennessee to get to the Sweet 16 is bold. You shouldn't get punished that Mercer beat Duke. So it would consider Tennessee round of 32 game to be a 11-3, for an 8-point differential.

The problem with this is still that the relative values are messed up. 12 over 5 is way over valued, especially compared to 13 over 4.

The more I thought about it, the current system really kills the value of a 1 seed. Is it really worth three times more to advance a 3 to the Sweet 16 than a 1?

Then I had another thought, borrowing from NFL Calcutta...

What if instead of using the actual numbers of the seeds, what if I grouped them into seed values.

Thinking out loud here...

1-4 is worth 1pt
5-8 is worth 3pts
9-12 is worth 5pts
13-14 is worth 7 pts
15-16 is worth 10pts

Something like that.

Even if the values aren't perfect, it's making more sense to me. But...I see a problem already. Picking a 8/9 to make the Sweet 16 should be worth 5 times as much as picking a 1. But picking the 8/9 first game shouldn't be worth that much.

To get really precise, you would need to arrange seed values that are based on round...

This is getting out of hand.

1 comment: