Results of an "NCAA pool scoring" of each ranking listed in Kenneth Massey's Last Pre-Tournament College Basketball Ranking Comparison using the final 63 games of the 2018 NCAA Men's Basketball Tournament. |
1 PIG 78.000 (41/63 i.e. 22 10 4 2 2 1) 2 SFX 75.000 (42/63 i.e. 24 10 4 2 1 1) 3 DCI 72.000 (41/63 i.e. 23 10 4 3 1 0) 4 COX 70.000 (39/63 i.e. 23 8 4 2 1 1) 5 BNT 67.000 (40/63 i.e. 25 9 3 1 1 1) 5 EBP 67.000 (40/63 i.e. 25 9 3 1 1 1) 5 LOG 67.000 (40/63 i.e. 25 9 3 1 1 1) 5 SAG 67.000 (39/63 i.e. 23 10 3 1 1 1) 5 YAG 67.000 (40/63 i.e. 25 9 3 1 1 1) 10 Average 66.000 (39/63 i.e. 24 9 3 1 1 1) 10 BLS 66.000 (39/63 i.e. 24 9 3 1 1 1) 10 BUR 66.000 (39/63 i.e. 24 9 3 1 1 1) 10 DAV 66.000 (39/63 i.e. 24 9 3 1 1 1) 10 DUN 66.000 (41/63 i.e. 26 9 3 2 1 0) 10 LAB 66.000 (39/63 i.e. 25 8 2 2 1 1) 10 LMC 66.000 (40/63 i.e. 26 8 3 1 1 1) 10 TPR 66.000 (39/63 i.e. 24 9 3 1 1 1) 10 WLK 66.000 (39/63 i.e. 24 9 3 1 1 1) 19 CNG 65.000 (38/63 i.e. 23 9 3 1 1 1) 19 MAS 65.000 (40/63 i.e. 25 9 3 2 1 0) 19 SMS 65.000 (38/63 i.e. 23 9 3 1 1 1) 19 SP 65.000 (38/63 i.e. 23 9 3 1 1 1) 19 TRP 65.000 (38/63 i.e. 23 9 3 1 1 1) 24 7OT 64.000 (39/63 i.e. 26 7 3 1 1 1) 24 DOK 64.000 (38/63 i.e. 24 8 3 1 1 1) 24 FSH 64.000 (38/63 i.e. 24 8 3 1 1 1) 24 SPR 64.000 (37/63 i.e. 21 9 4 2 1 0) 28 DOL 63.000 (40/63 i.e. 25 10 3 1 1 0) 28 RT 63.000 (38/63 i.e. 25 7 3 1 1 1) 28 SMN 63.000 (38/63 i.e. 23 9 3 2 1 0) 28 ZAM 63.000 (38/63 i.e. 24 9 2 1 1 1) 32 DC2 62.000 (38/63 i.e. 26 6 3 1 1 1) 32 HAS 62.000 (36/63 i.e. 22 8 3 1 1 1) 32 MMG 62.000 (37/63 i.e. 23 9 2 1 1 1) 32 MOR 62.000 (39/63 i.e. 24 10 3 1 1 0) 32 RPI 62.000 (38/63 i.e. 24 8 3 2 1 0) 37 DII 61.000 (38/63 i.e. 24 8 4 1 1 0) 37 PRR 61.000 (39/63 i.e. 25 9 3 1 1 0) 37 USA 61.000 (36/63 i.e. 21 9 3 2 1 0) 40 BWE 60.000 (38/63 i.e. 24 9 3 1 1 0) 40 KPI 60.000 (37/63 i.e. 24 7 3 2 1 0) 40 PGH 60.000 (38/63 i.e. 25 8 2 2 1 0) 40 POM 60.000 (38/63 i.e. 24 9 3 1 1 0) 40 SGR 60.000 (39/63 i.e. 26 8 3 1 1 0) 45 KRA 59.000 (36/63 i.e. 22 9 2 2 1 0) 45 RTH 59.000 (37/63 i.e. 23 9 3 1 1 0) 45 SEL 59.000 (37/63 i.e. 24 8 2 2 1 0) 45 WIL 59.000 (36/63 i.e. 22 9 2 2 1 0) 45 WOB 59.000 (37/63 i.e. 24 8 2 2 1 0) 50 AP 58.000 (34/63 i.e. 20 8 3 2 1 0) 50 COL 58.000 (37/63 i.e. 25 7 2 2 1 0) 50 ESR 58.000 (37/63 i.e. 23 10 2 1 1 0) 50 KPK 58.000 (37/63 i.e. 24 8 3 1 1 0) 54 FAS 57.000 (37/63 i.e. 24 9 2 1 1 0) 54 Median 57.000 (36/63 i.e. 23 8 3 1 1 0) 54 STM 57.000 (36/63 i.e. 23 8 3 1 1 0) 54 WMV 57.000 (36/63 i.e. 23 8 3 1 1 0) 58 BIH 56.000 (37/63 i.e. 25 8 2 1 1 0) 58 DDB 56.000 (34/63 i.e. 21 7 4 1 1 0) 58 JNG 56.000 (36/63 i.e. 23 8 3 2 0 0) 58 STH 56.000 (36/63 i.e. 24 7 3 1 1 0) 62 ARG 55.000 (36/63 i.e. 24 8 2 1 1 0) 63 BBT 54.000 (36/63 i.e. 25 7 2 1 1 0) 63 REW 54.000 (34/63 i.e. 22 7 3 1 1 0) 63 RTP 54.000 (34/63 i.e. 23 6 2 2 1 0) 63 TRK 54.000 (36/63 i.e. 23 9 3 1 0 0) 63 WOL 54.000 (36/63 i.e. 25 7 2 1 1 0) 68 JJK 53.000 (35/63 i.e. 24 7 2 1 1 0) 69 DES 52.000 (32/63 i.e. 20 7 3 1 1 0) 70 SPW 50.000 (32/63 i.e. 21 7 2 1 1 0) Scoring: Round 1: 1 point. Round 2: 2 points. Round 3: 3 points. Round 4: 4 points. Round 5: 5 points. Round 6: 6 points. 'Average' is Massey comparison Mean ranking. 'Median' is Massey comparison Median ranking. |
But how do the top 25 of the rankings stack up? Here is the pool scoring based solely on the top 25 teams in each ranking, treating the other teams as a large 26th-place tie. (This is inferior scoring of the pool other than to compare impartially with a ranking that only ranks 25 teams, such as AP or USA: see scoring explanation below.) |
1 PIG 79.997 (42.997/63 i.e. 23.997 10 4 2 2 1) 2 SFX 79.925 (42.953/63 i.e. 23.960 10 4 2 1.993 1) 3 Allsame 77.627 (41.854/63 i.e. 23.735 9.225 3.888 2.200 1.862 0.945) 4 DCI 73.189 (42.189/63 i.e. 24.189 10 4 3 1 0) 5 COX 71.203 (40.203/63 i.e. 24.203 8 4 2 1 1) 6 SAG 67.615 (39.615/63 i.e. 23.615 10 3 1 1 1) 7 LOG 66.809 (39.809/63 i.e. 24.809 9 3 1 1 1) 8 BNT 66.622 (39.622/63 i.e. 24.622 9 3 1 1 1) 9 SPR 66.600 (39.600/63 i.e. 23.600 9 4 2 1 0) 10 BUR 66.585 (39.585/63 i.e. 24.585 9 3 1 1 1) 11 TPR 66.279 (39.279/63 i.e. 24.279 9 3 1 1 1) 12 BLS 66.203 (39.203/63 i.e. 24.203 9 3 1 1 1) 12 CNG 66.203 (39.203/63 i.e. 24.203 9 3 1 1 1) 12 EBP 66.203 (39.203/63 i.e. 24.203 9 3 1 1 1) 12 TRP 66.203 (39.203/63 i.e. 24.203 9 3 1 1 1) 16 YAG 66.131 (39.131/63 i.e. 24.131 9 3 1 1 1) 17 LAB 65.953 (38.953/63 i.e. 24.953 8 2 2 1 1) 18 Average 65.809 (38.809/63 i.e. 23.809 9 3 1 1 1) 18 WLK 65.809 (38.809/63 i.e. 23.809 9 3 1 1 1) 20 DAV 65.790 (38.790/63 i.e. 23.790 9 3 1 1 1) 21 SMS 65.753 (38.753/63 i.e. 23.753 9 3 1 1 1) 22 SP 65.587 (38.587/63 i.e. 23.587 9 3 1 1 1) 23 MAS 64.928 (39.928/63 i.e. 24.928 9 3 2 1 0) 24 HAS 64.569 (38.569/63 i.e. 24.569 8 3 1 1 1) 25 USA 64.543 (39.543/63 i.e. 24.543 9 3 2 1 0) 26 SMN 64.468 (39.468/63 i.e. 24.468 9 3 2 1 0) 27 LMC 64.228 (38.228/63 i.e. 24.228 8 3 1 1 1) 28 FSH 64.182 (38.182/63 i.e. 24.182 8 3 1 1 1) 29 DUN 63.875 (38.875/63 i.e. 23.875 9 3 2 1 0) 30 DOK 63.841 (37.841/63 i.e. 23.841 8 3 1 1 1) 31 7OT 63.500 (38.500/63 i.e. 25.500 7 3 1 1 1) 32 MMG 62.694 (37.694/63 i.e. 23.694 9 2 1 1 1) 33 ZAM 62.545 (37.545/63 i.e. 23.545 9 2 1 1 1) 34 RPI 62.303 (38.303/63 i.e. 24.303 8 3 2 1 0) 35 RT 62.223 (37.223/63 i.e. 24.223 7 3 1 1 1) 36 AP 62.046 (38.046/63 i.e. 24.046 8 3 2 1 0) 37 MOR 62.004 (39.004/63 i.e. 24.004 10 3 1 1 0) 38 DOL 61.906 (38.906/63 i.e. 23.906 10 3 1 1 0) 39 DII 61.203 (38.203/63 i.e. 24.203 8 4 1 1 0) 39 KRA 61.203 (38.203/63 i.e. 24.203 9 2 2 1 0) 39 WIL 61.203 (38.203/63 i.e. 24.203 9 2 2 1 0) 42 POM 60.809 (38.809/63 i.e. 24.809 9 3 1 1 0) 43 KPI 60.722 (37.722/63 i.e. 24.722 7 3 2 1 0) 44 RTH 60.659 (38.659/63 i.e. 24.659 9 3 1 1 0) 45 BWE 60.548 (38.548/63 i.e. 24.548 9 3 1 1 0) 46 DC2 60.428 (36.428/63 i.e. 24.428 6 3 1 1 1) 47 PRR 60.279 (38.279/63 i.e. 24.279 9 3 1 1 0) 48 Median 59.767 (37.788/63 i.e. 23.809 8.979 3 1 1 0) 49 WOB 59.370 (37.370/63 i.e. 24.370 8 2 2 1 0) 50 DDB 59.166 (36.996/63 i.e. 23.826 7.170 4 1 1 0) 51 PGH 59.161 (37.161/63 i.e. 24.161 8 2 2 1 0) 52 ESR 58.906 (37.906/63 i.e. 23.906 10 2 1 1 0) 53 SEL 58.767 (36.767/63 i.e. 23.767 8 2 2 1 0) 54 STM 58.606 (37.606/63 i.e. 24.606 8 3 1 1 0) 55 SGR 58.428 (37.428/63 i.e. 24.428 8 3 1 1 0) 56 COL 57.827 (36.827/63 i.e. 24.827 7 2 2 1 0) 57 KPK 57.809 (36.809/63 i.e. 23.809 8 3 1 1 0) 57 WMV 57.809 (36.809/63 i.e. 23.809 8 3 1 1 0) 59 JNG 56.600 (36.600/63 i.e. 23.600 8 3 2 0 0) 60 FAS 56.585 (36.585/63 i.e. 23.585 9 2 1 1 0) 61 STH 56.479 (36.479/63 i.e. 24.479 7 3 1 1 0) 62 DES 56.246 (36.246/63 i.e. 24.246 7 3 1 1 0) 63 REW 55.582 (35.582/63 i.e. 23.582 7 3 1 1 0) 64 TRK 55.548 (37.548/63 i.e. 24.548 9 3 1 0 0) 65 RTP 55.296 (35.296/63 i.e. 24.296 6 2 2 1 0) 66 ARG 54.809 (35.809/63 i.e. 23.809 8 2 1 1 0) 67 BIH 54.512 (35.512/63 i.e. 23.512 8 2 1 1 0) 68 SPW 54.127 (35.183/63 i.e. 23.239 7.944 2 1 1 0) 69 BBT 53.931 (35.931/63 i.e. 24.931 7 2 1 1 0) 70 JJK 53.031 (35.031/63 i.e. 24.031 7 2 1 1 0) 71 WOL 52.534 (34.534/63 i.e. 23.534 7 2 1 1 0) Scoring: Round 1: 1 point. Round 2: 2 points. Round 3: 3 points. Round 4: 4 points. Round 5: 5 points. Round 6: 6 points. Top 25 of rankings only. Games with both teams beyond 25 or teams of same rank given a calculated partial credit. 'Average' is Massey comparison Mean ranking. 'Median' is Massey comparison Median ranking. 'Allsame' is ranking with all teams ranked equally. |
This pool scoring is inferior, other than to compare against rankings that rank only 25 teams.
Scoring method: when a ranking ranks at least one game within its top 25, that ranking's implied choice of winner is obvious. For a game in which a ranking ranks neither team as "top 25", that ranking is given partial credit: a fraction of what correctly calling the game would have earned it. To be impartial, for each game, there is one fraction used for any ranking that doesn't call that game with their top 25 rankings.
The fractional credit for a game is the number of full rankings that correctly called that game, e.g., if there were 50 full rankings and 40 of them called that game, the fraction would be .8, an estimate of the probability that some random ranking method would have called the game correctly. Thus full rankings do affect this scoring, but in the same way for all the rankings including themselves.
I adopted this type of scoring because the difference between ranking a team and not (e.g., ranking it 24 versus ranking it 27) otherwise affects such a "top-25 only" pool results a lot, even when those teams ranked 24 and 27 never actually play each other, the typical case. I adopted a fraction based on expectations of which way the game would go because it seems wrong to give a ranking much credit for an upset between two teams they didn't rank in their top 25 or to penalize them very much on a game likely to be called correctly.
This method is imperfect, but so is any method of scoring such a pool based on incomplete rankings.
Wobus Sports: www.vaporia.com/sports