Well everybody, with the ECAC schedule starting this Tuesday, we're ready for a new season of wildly ridiculous prognostications based on everybody's favorite retrodictive (NOT predictive) rating system, KRACH. Using KRACH, we can try to predict how many points a team will take in each game. Every team has played a game, so it's time to get underway.
Standings
1. Quinnipiac (32)
2t. RPI (31)
2t. Brown (31)
4. St. Lawrence (28)
-----
5. Cornell (25)
6. Clarkson (23)
7. Harvard (21)
8. Yale (20)
-----
9. Union (17)
10. Colgate (15)
11. Princeton (14)
12. Dartmouth (6)
(Rounding Error = -1)
Note: KRACH doesn't do ties. KRACH doesn't do prediction, it does retrodiction. KRACH cares not about your momentum, injuries, home-ice, etc. Also, for now, most teams don't have finite, real KRACH values. I've taken to calculating my own, and I'm getting around the problem with the hypothetical tie against a 100 team.
Now, in addition to the projected final point total standings, I have a burgie12 style projection of final ranks by Monte Carlo! Based on 1,000,000 simulations (which is obviously a tiny, tiny, tiny fraction of the 955,004,950,796,825,236,893,190,701,774,414,011,91 9,935,138,974,343,129,836,853,841 different ways an ECAC season can play out):
Notes: 0.0 means the team did end up that rank, but fewer than 500 times. If there is an x, it never happened. I'm not fancy enough to break ties the right way, so this breaks ties randomly. I'll try and make that better as the season goes on, but by the time they really matter, I'm sure burgie will have his super awesome posts up.
Standings
1. Quinnipiac (32)
2t. RPI (31)
2t. Brown (31)
4. St. Lawrence (28)
-----
5. Cornell (25)
6. Clarkson (23)
7. Harvard (21)
8. Yale (20)
-----
9. Union (17)
10. Colgate (15)
11. Princeton (14)
12. Dartmouth (6)
(Rounding Error = -1)
Note: KRACH doesn't do ties. KRACH doesn't do prediction, it does retrodiction. KRACH cares not about your momentum, injuries, home-ice, etc. Also, for now, most teams don't have finite, real KRACH values. I've taken to calculating my own, and I'm getting around the problem with the hypothetical tie against a 100 team.
Now, in addition to the projected final point total standings, I have a burgie12 style projection of final ranks by Monte Carlo! Based on 1,000,000 simulations (which is obviously a tiny, tiny, tiny fraction of the 955,004,950,796,825,236,893,190,701,774,414,011,91 9,935,138,974,343,129,836,853,841 different ways an ECAC season can play out):
Code:
| KRACH | 1 2 3 4 5 6 7 8 9 10 11 12 |Avg Rk --------------------------------------------------------------------------------------------- Qu | 443.2 | 32.6 27.2 20.5 12.0 5.0 1.9 0.6 0.2 0.0 0.0 0.0 x | 2.39 RP | 429.8 | 32.7 26.4 20.0 12.3 5.4 2.1 0.7 0.2 0.1 0.0 0.0 x | 2.42 Br | 416.8 | 24.2 25.5 23.1 14.7 7.5 3.3 1.2 0.4 0.1 0.0 0.0 0.0 | 2.73 SL | 303.8 | 8.7 14.8 21.0 25.7 15.6 8.2 3.8 1.6 0.5 0.1 0.0 x | 3.77 Cr | 224.5 | 0.7 2.5 6.2 13.5 24.0 21.8 15.4 9.3 4.5 1.7 0.4 0.0 | 5.74 Ck | 185.7 | 0.8 2.5 5.7 11.6 19.0 21.8 17.3 11.6 6.3 2.5 0.7 0.0 | 5.99 Ya | 140.5 | 0.2 0.7 1.9 4.9 10.2 15.9 20.7 19.4 14.1 8.3 3.5 0.2 | 7.24 Ha | 157.6 | 0.1 0.4 1.2 3.6 8.6 14.1 18.8 20.4 16.8 10.7 5.1 0.1 | 7.58 Un | 106.0 | 0.0 0.1 0.4 1.4 3.6 7.3 12.6 18.7 23.6 19.9 11.5 0.9 | 8.55 Cg | 82.7 | 0.0 0.0 0.0 0.2 0.7 2.3 5.8 11.6 20.4 30.0 25.8 3.1 | 9.57 Pr | 71.2 | 0.0 0.0 0.0 0.1 0.4 1.2 3.0 6.6 13.3 25.2 43.9 6.3 | 10.13 Da | 24.0 | x x x x x 0.0 0.0 0.0 0.2 1.4 9.0 89.4 | 11.88
Comment