 
|
||||||||||
|
||||||||||
|
||||||||||
|
MinV and MVP College Football Rankings |
||||||||||
|
Performance
|
||||||||||
|
Below is an assessment of the ranking systems posted on Kenneth Massey's ranking compilation page for this season. Using a simple form of Data Envelopment Analysis and software from Holger Scheel, each system is given an efficiency score representing how close its combined retrodictive and predictive performance is to the efficient frontier. Retrodictive performance is measured by the number of past head-to-head results that were correctly matched by the weekly ranking (the first number in the "Retrodictive" column below), while predictive performance is measured by the number of games in the following week that were correctly predicted (the first number in the "Predictive" column below). If one were to plot each system's performance, with retrodictive performance on one axis and predictive performance on the other, the efficient frontier would be formed by the most extreme systems in all possible directions. The efficient frontier approach is an arguably fair way to compare ranking systems regardless of whether they focus on retrodictive or predictive performance (or some combination thereof). The approach determines the implicit weight that a system places on each objective, and uses these weights (shown in parenthese below) to determine the appropriate comparison point on the efficient frontier. Each system is thus evaluated only against peers with a similar combination of objectives. Any system with 100% efficiency is on the efficient frontier, meaning that if the weights noted for that system are applied, no other system surpasses it. If a system has an efficiency less than 100%, then even if the weights noted for the system are applied, there is at least one other system (or a linear combination of other systems) whose performance is better. Columns can be sorted from high to low by clicking on the column header. Results shown are for games on or before December 2, 2017 using rankings posted through the games of the respective preceding week. Note that the system abbreviations are the same as those used by Massey, who denotes the MinV ranking as "CMV" and the MVP ranking as "MVP." |
||||||||||
Full Season Last Week
Rnk System Efficiency Retrodictive Predictive Efficiency Retrodictive Predictive
1 CMV 100.0% 4577 (.950) 536 (.050) 100.0% 662 (1.00) 11 (.000)
2 DOK 100.0% 3932 (.000) 544 (1.00) 90.39% 577 (.720) 11 (.280)
3 SMS 99.67% 4018 (.080) 541 (.920) 95.08% 578 (.690) 13 (.310)
4 FPI 99.53% 3947 (.080) 541 (.920) 89.71% 571 (.720) 11 (.280)
5 BRN 99.37% 3952 (.080) 540 (.920) 89.71% 571 (.720) 11 (.280)
6 PIG 99.17% 4019 (.080) 538 (.920) 87.88% 575 (.740) 10 (.260)
7 Con 98.88% 4122 (.090) 535 (.910) 95.54% 582 (.690) 13 (.310)
8 DII 98.79% 4079 (.090) 535 (.910) 98.50% 588 (.670) 14 (.330)
9 RUD 98.62% 4160 (.090) 533 (.910) 96.21% 588 (.690) 13 (.310)
10 BWE 98.59% 3983 (.080) 535 (.920) 89.94% 573 (.720) 11 (.280)
11 BIL 98.38% 4125 (.090) 532 (.910) 93.81% 587 (.710) 12 (.290)
12 ARG 98.38% 4046 (.090) 533 (.910) 92.34% 574 (.700) 12 (.300)
13 PIR 98.35% 4109 (.090) 532 (.910) 95.87% 585 (.690) 13 (.310)
14 KAM 98.14% 3930 (.080) 533 (.920) 92.34% 574 (.700) 12 (.300)
15 SAG 97.95% 4001 (.090) 531 (.910) 88.44% 580 (.740) 10 (.260)
16 PAY 97.83% 4102 (.090) 529 (.910) 94.74% 575 (.690) 13 (.310)
17 MAS 97.80% 4089 (.090) 529 (.910) 94.15% 590 (.710) 12 (.290)
18 LAZ 97.63% 4088 (.090) 528 (.910) 94.52% 573 (.690) 13 (.310)
19 DES 97.62% 4006 (.090) 529 (.910) 87.88% 575 (.740) 10 (.260)
20 DCI 97.62% 4005 (.090) 529 (.910) 86.71% 574 (1.00) 9 (.000)
21 DWI 97.38% 4050 (.090) 527 (.910) 93.25% 582 (.710) 12 (.290)
22 KPK 97.30% 4011 (.090) 527 (.910) 95.31% 580 (.690) 13 (.310)
23 BDF 97.16% 3867 (.080) 528 (.920) 84.89% 562 (1.00) 9 (.000)
24 MOR 96.97% 3933 (.080) 526 (.920) 85.35% 565 (1.00) 8 (.000)
25 CGV 96.85% 3958 (.090) 525 (.910) 93.50% 564 (.680) 13 (.320)
26 LOG 96.33% 3954 (.090) 522 (.910) 86.86% 575 (1.00) 9 (.000)
27 DEZ 96.25% 3996 (.090) 521 (.910) 90.17% 575 (.720) 11 (.280)
28 HOW 96.18% 4041 (.090) 520 (.910) 94.97% 577 (.690) 13 (.310)
29 CTW 96.03% 3971 (.090) 520 (.910) 88.33% 579 (.740) 10 (.260)
30 PGH 95.83% 4198 (.090) 516 (.910) 96.67% 592 (.690) 13 (.310)
31 KEL 95.71% 4057 (.090) 517 (.910) 95.08% 578 (.690) 13 (.310)
32 BBT 95.54% 3976 (.090) 517 (.910) 92.00% 571 (.700) 12 (.300)
33 MAR 95.41% 3997 (.090) 516 (.910) 94.18% 570 (.680) 13 (.320)
34 RWP 95.35% 4050 (.090) 515 (.910) 95.65% 583 (.690) 13 (.310)
35 RTP 95.34% 4125 (.090) 514 (.910) 92.31% 594 (.730) 11 (.270)
36 ENG 95.30% 4023 (.090) 515 (.910) 93.25% 582 (.710) 12 (.290)
37 HAT 94.84% 4128 (.090) 511 (.910) 95.54% 582 (.690) 13 (.310)
38 MGS 94.51% 4132 (.090) 509 (.910) 93.13% 581 (.710) 12 (.290)
39 RTB 93.31% 4271 (1.00) 359 (.000) 99.43% 657 (.750) 11 (.250)
40 NUT 93.28% 3866 (.090) 505 (.910) 92.46% 575 (.700) 12 (.300)
41 PPP 92.72% 4082 (.090) 499 (.910) 95.99% 586 (.690) 13 (.310)
42 TPR 92.57% 3849 (.090) 501 (.910) 89.15% 566 (.720) 11 (.280)
43 BAS 89.73% 3863 (.090) 484 (.910) 87.20% 569 (.740) 10 (.260)
44 OSC 89.60% 3557 (.080) 487 (.920) 95.31% 580 (.690) 13 (.310)
45 PFZ 88.85% 3600 (.080) 482 (.920) 88.25% 558 (.710) 11 (.290)
46 ACU 88.81% 4065 (1.00) 460 (.000) 95.65% 583 (.690) 13 (.310)
47 DOL 88.60% 4055 (1.00) 399 (.000) 96.44% 590 (.690) 13 (.310)
48 LSD 88.55% 4053 (1.00) 459 (.000) 95.54% 582 (.690) 13 (.310)
49 COL 88.44% 4048 (1.00) 431 (.000) 96.21% 588 (.690) 13 (.310)
50 CSL 87.00% 3982 (1.00) 424 (.000) 95.99% 586 (.690) 13 (.310)
51 BSS 86.34% 3952 (1.00) 402 (.000) 95.54% 582 (.690) 13 (.310)
52 WIL 86.32% 3951 (1.00) 363 (.000) 97.23% 597 (.690) 13 (.310)
53 JTR 86.24% 3947 (1.00) 411 (.000) 94.86% 576 (.690) 13 (.310)
54 BIH 86.02% 3937 (1.00) 367 (.000) 97.00% 595 (.690) 13 (.310)
55 FEI 85.46% 3754 (.090) 460 (.910) 90.87% 561 (.700) 12 (.300)
56 FMG 85.16% 3898 (1.00) 373 (.000) 96.33% 589 (.690) 13 (.310)
57 JNK 85.14% 3897 (1.00) 364 (.000) 96.10% 587 (.690) 13 (.310)
58 WEL 84.68% 3876 (1.00) 361 (.000) 95.65% 583 (.690) 13 (.310)
59 HKB 84.66% 3875 (1.00) 374 (.000) 95.76% 584 (.690) 13 (.310)
60 PCP 84.51% 3868 (1.00) 375 (.000) 91.98% 591 (.730) 11 (.270)
61 RME 84.33% 3860 (1.00) 363 (.000) 93.36% 583 (.710) 12 (.290)
62 FAS 84.12% 3850 (1.00) 375 (.000) 94.97% 577 (.690) 13 (.310)
63 RT 83.79% 3835 (1.00) 381 (.000) 92.91% 579 (.700) 12 (.300)
64 WLK 82.81% 3790 (1.00) 379 (.000) 95.54% 582 (.690) 13 (.310)
65 SOR 82.57% 3583 (.090) 445 (.910) 97.12% 596 (.690) 13 (.310)
66 YAG 82.35% 3059 (.000) 448 (1.00) 88.47% 560 (.720) 11 (.280)
67 SEL 81.47% 3729 (1.00) 337 (.000) 96.21% 588 (.690) 13 (.310)
68 EZ 81.21% 3717 (1.00) 327 (.000) 97.00% 595 (.690) 13 (.310)
69 ABC 81.12% 3713 (1.00) 331 (.000) 96.78% 593 (.690) 13 (.310)
70 S&P 80.99% 3707 (1.00) 434 (.000) 89.97% 553 (.690) 12 (.310)
71 ISR 80.84% 3700 (1.00) 328 (.000) 96.67% 592 (.690) 13 (.310)
72 ASH 80.53% 3686 (1.00) 342 (.000) 95.76% 584 (.690) 13 (.310)
73 MRK 80.45% 3682 (1.00) 329 (.000) 95.99% 586 (.690) 13 (.310)
74 MvG 80.40% 3680 (1.00) 327 (.000) 95.76% 584 (.690) 13 (.310)
75 COF 80.14% 3668 (1.00) 341 (.000) 95.31% 580 (.690) 13 (.310)
76 MCK 80.10% 3666 (1.00) 294 (.000) 94.71% 627 (1.00) 10 (.000)
77 RSL 80.05% 3664 (1.00) 374 (.000) 92.12% 572 (.700) 12 (.300)
78 HEN 79.97% 3660 (1.00) 336 (.000) 95.42% 581 (.690) 13 (.310)
79 MGN 79.13% 3622 (1.00) 333 (.000) 91.89% 570 (.700) 12 (.300)
80 KEN 79.05% 3618 (1.00) 329 (.000) 94.41% 572 (.680) 13 (.320)
81 SGR 78.50% 3593 (1.00) 335 (.000) 93.25% 582 (.710) 12 (.290)
82 HNL 78.41% 3589 (1.00) 339 (.000) 94.86% 576 (.690) 13 (.310)
83 SFX 77.58% 3551 (1.00) 374 (.000) 88.13% 557 (.710) 11 (.290)
84 KLK 77.10% 3529 (1.00) 342 (.000) 92.23% 573 (.700) 12 (.300)
85 VRN 75.66% 3463 (1.00) 291 (.000) 96.44% 590 (.690) 13 (.310)
86 KEE 75.11% 3438 (1.00) 303 (.000) 95.54% 582 (.690) 13 (.310)
87 LMC 75.09% 3437 (1.00) 341 (.000) 91.78% 569 (.700) 12 (.300)
88 BMC 75.01% 3117 (.090) 406 (.910) 93.13% 581 (.710) 12 (.290)
89 MJS 74.94% 3430 (1.00) 285 (.000) 95.42% 581 (.690) 13 (.310)
90 GLD 74.90% 3428 (1.00) 290 (.000) 95.08% 578 (.690) 13 (.310)
91 YCM 74.87% 3427 (1.00) 327 (.000) 94.18% 570 (.680) 13 (.320)
92 MDS 74.46% 3408 (1.00) 289 (.000) 95.20% 579 (.690) 13 (.310)
93 TRP 72.56% 3321 (1.00) 303 (.000) 90.28% 576 (.720) 11 (.280)
94 BOW 70.59% 3231 (1.00) 295 (.000) 96.10% 587 (.690) 13 (.310)
95 WOL 69.87% 3198 (1.00) 258 (.000) 96.67% 592 (.690) 13 (.310)
96 RTH 69.83% 3196 (1.00) 257 (.000) 96.55% 591 (.690) 13 (.310)
97 REW 69.46% 3179 (1.00) 255 (.000) 91.30% 585 (.720) 11 (.280)
98 KNT 69.22% 3168 (1.00) 260 (.000) 96.33% 589 (.690) 13 (.310)
99 GBE 68.71% 3145 (1.00) 254 (.000) 95.65% 583 (.690) 13 (.310)
100 PIT 68.39% 3130 (1.00) 365 (.000) . . (. ) . (. )
101 LAB 67.90% 3108 (1.00) 256 (.000) 92.23% 573 (.700) 12 (.300)
102 BCM 65.94% 3018 (1.00) 302 (.000) 92.68% 577 (.700) 12 (.300)
103 MTN 61.85% 2831 (1.00) 222 (.000) 95.08% 578 (.690) 13 (.310)
104 GRR 59.71% 2733 (1.00) 218 (.000) 100.0% 581 (.000) 15 (1.00)
105 AND 54.95% 2515 (1.00) 178 (.000) 97.00% 595 (.690) 13 (.310)
106 JWN 52.33% 2395 (1.00) 180 (.000) 95.90% 565 (.670) 14 (.330)
107 DUN 51.93% 2164 (.090) 281 (.910) . . (. ) . (. )
108 NGS 50.27% 2301 (1.00) 231 (.000) 91.10% 563 (.700) 12 (.300)
109 WWP 45.12% 2065 (1.00) 139 (.000) 95.31% 580 (.690) 13 (.310)
110 KH 44.19% 2010 (.100) 237 (.900) . . (. ) . (. )
111 ATC 39.17% 1793 (1.00) 166 (.000) 88.56% 581 (.740) 10 (.260)
112 KRA 36.12% 1653 (1.00) 97 (.000) 97.00% 595 (.690) 13 (.310)
113 WOB 35.83% 1640 (1.00) 95 (.000) 96.78% 593 (.690) 13 (.310)
114 HUF 34.91% 1598 (1.00) 98 (.000) 92.57% 576 (.700) 12 (.300)
115 TFG 34.08% 1560 (1.00) 103 (.000) 90.42% 557 (.700) 12 (.300)
116 RBA 33.58% 1537 (1.00) 97 (.000) 87.91% 555 (.710) 11 (.290)
117 CI 32.82% 1502 (1.00) 163 (.000) . . (. ) . (. )
118 SP 29.56% 1308 (.090) 159 (.910) . . (. ) . (. )
119 STH 12.54% 574 (1.00) 14 (.000) 96.92% 574 (.670) 14 (.330)
|
||||||||||
|
|
||||||||||
|
B. Jay Coleman, Ph.D. Richard deR. Kip Professor of Operations Management & Quantitative Methods Department of Management | Coggin College of Business | University of North Florida | Jacksonville, FL 32224 jcoleman@unf.edu Disclaimer |
||||||||||
