(02-08-2020 09:28 PM)MaddDawgz02 Wrote:  

(02-08-2020 08:48 PM)pantone1935 Wrote:  

(02-08-2020 08:11 PM)MaddDawgz02 Wrote:  Some updated projections after this weekend's games. Your NIU Huskies now projected as a 3 SEED!! Despite the 3 seed projected, still have them down at 7th most likely to win MAC tourney.


NIU
Projected Conference Wins: 11.3
Projected Seed: #3
Projected First MAC Tourney Game: Quarterfinals #3 NIU vs #6 CMU/#11 EMU in Cleveland, OH
Chance for NIU to win MAC tourney: 6.1%

Rest of year mock lines:

@ Ball State, Ball State by 7.5
@ Miami-OH, NIU by 2
Central Michigan, NIU by 3.5
@ Eastern Michigan, EMU by 2.5
@ Western Michigan, EVEN
Toledo, NIU by 1
Ball State, EVEN

Projected wins in conference

1. BGSU 12.9
2. Akron 12.0
3. NIU 11.3
4. Ball State 11.2
5. Kent State 10.8
6. CMU 10.4
7. Buffalo 9.4
8. Toledo 7.4
9. WMU 6.7
10. Ohio 6.4
11. EMU 5.5
12. Miami-OH 3.9

Chances of NIU finishing with the following number of conference wins:

8: 1.1%
9: 6.8%
10: 18.7%
11. 28.4%
12. 25.9%
13. 14.2%
14: 4.3%
15: 0.6%

Percent Chances to Win MAC Tourney

1. Akron 20.8%
2. Kent State 17.7%
3. Bowling Green 17.1%
4. Ball State 13.4%
5. Buffalo 8.9%
6. Toledo 7.6%
7. Northern Illinois 6.1%
8. Central Michigan 4.3%
9. Eastern Michigan 1.5%
10. Ohio 1.3%
11. Western Michigan 0.2%
12. Miami-OH 0.1%

Just out of curiosity what parameters, input, etc. do you use to determine Akron has a 1 in 5 chance to go to the big dance whereas CMU only has a 1 in 23 chance? I know stats can be bent as most of the NET/Kenpom ratings seem to give more credence to early season
games results than current results.

GO HUSKIES!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! (just win baby)

I have a measure of offensive efficiency, defensive efficiency, and then a third measure which takes into accounts wins/losses. All 3 are adjusted by SOS. The sum of these 3 comes up to a power ranking. The power ranking is used in simulating head to head games. Those percentages to win tourney have a big weakness in that assumes the seeding is predicted perfectly. For example that 6.1% assumes NIU has a bye as a 3rd seed and there is no error in the seeding. The fuzzy math comes in since we wouldnt know NIU's opponent, I build a fake opponent that has power ranking characteristic made up of 70% CMU and 30% EMU ( assuming CMU has a 70 percent chance to beat EMU), and calculate the probability of NIU winning that game. The probabilities are calculated for each round until the finals to get the final probability. Obviously there are massive problems with it but I like to use it as a guide to determine realistic probabilities. I have equal weighting on every game of the year, so it wont see that Frye has left BG, etc. I think what is happening with Kenpom, is that even though NIU has won 6 games in a row, NIU's metrics that he is using are not showing much improvement from earlier in the year despite the wins. In other words, I think NIU is pulling games out of their rear that they probably have no business doing. If NIU was a stock, I hate to say, it might be smart to sell now! This kind of comes out in the above too as NIU, even with the bye, has lower percentages to win the tourney than a few teams who play a campus site game.

Thanks. Seems like a lot of work. I guess it still comes down to how the ball bounces (or goes through the hoop).

GO HUSKIES!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! (beat the birds)