12/21/12 Takedown Wrestling Media Dual-Results ranking

[SetonHallPirate]

Note: Based on results of dual meets only. The basic idea is an iterative one. Begin with all wrestlers set to an even rating -- 100 in this case. Then, for each match wrestled, give each wrestler the value of their opponent's rating plus or minus a factor for winning or losing the match -- 30, plus one for each point the team won by. Total all of a wrestler's results, divide by the number of matches wrestled, and that's the end of a cycle. Then use those numbers as the start of the next cycle until you get the same results for each wrestler for two consecutive cycles.

[SetonHallPirate]

1	Oklahoma State	168.233
2	Ohio State	164.037
3	Missouri	162.353
4	Penn State	159.459
5	Illinois	153.923
6	Iowa	148.478
7	Virginia Tech	144.751
8	Cornell	140.387
9	Minnesota	138.592
10	Edinboro	134.959
11	Pittsburgh	133.767
12	Rutgers	131.744
13	Purdue	128.505
14	Penn	127.328
15	Nebraska	127.144
16	Kent State	127.007
17	Michigan	126.009
18	Maryland	123.028
19	Oklahoma	121.073
20	Wyoming	119.501
21	Central Michigan	119.261
22	Arizona State	117.491
23	Lehigh	115.726
24	Old Dominion	115.278
25	Bloomsburg	115.035
26	Wisconsin	113.871
27	Northwestern	113.122
28	Virginia	112.123
29	Michigan State	105.335
30	Oregon State	104.056
31	Rider	103.274
32	North Carolina	99.890
33	Army	98.662
34	Chattanooga	95.461
35	Iowa State	90.609
36	Clarion	88.626
37	Columbia	88.047
38	Stanford	87.880
39	Northern Illinois	86.977
40	Utah Valley	86.363
41	Binghamton	86.031
42	Bucknell	85.286
43	Brown	84.017
44	Indiana	84.001
45	Navy	83.855
46	West Virginia	83.743
47	CSU Bakersfield	79.617
48	The Citadel	79.569
49	Lock Haven	79.367
50	Ohio	78.479
51	Northern Colorado	78.144
52	Buffalo	77.591
53	American	77.054
54	South Dakota State	73.036
55	Northern Iowa	72.871
56	Eastern Michigan	72.535
57	North Dakota State	72.070
58	Hofstra	72.025
59	Boise State	65.555
60	Drexel	63.368
61	Princeton	62.711
62	Boston U.	62.024
63	Davidson	57.646
64	Duke	54.556
65	Cleveland State	54.151
66	Southern Illinois Edwardsville	54.003
67	Appalachian State	52.016
68	Campbell	49.690
69	Franklin & Marshall	42.016
70	Gardner-Webb	38.714
71	Cal Poly	25.993
72	George Mason	24.855
73	VMI	15.693
74	Sacred Heart	-16.668

[jammen]

Begin with all wrestlers set to an even rating -- 100 in this case. Then, for each match wrestled, give each wrestler the value of their opponent's rating plus or minus a factor for winning or losing the match -- 30, plus one for each point the team won by. Total all of a wrestler's results, divide by the number of matches wrestled, and that's the end of a cycle. Then use those numbers as the start of the next cycle until you get the same results for each wrestler for two consecutive cycles.

SHP, would you restate this using an example to illustrate.

[SetonHallPirate]

Begin with all wrestlers set to an even rating -- 100 in this case. Then, for each match wrestled, give each wrestler the value of their opponent's rating plus or minus a factor for winning or losing the match -- 30, plus one for each point the team won by. Total all of a wrestler's results, divide by the number of matches wrestled, and that's the end of a cycle. Then use those numbers as the start of the next cycle until you get the same results for each wrestler for two consecutive cycles.

SHP, would you restate this using an example to illustrate.

Sure. Navy has had one dual meet, and beat George Mason 33-4. George Mason's rating is 24.855. So add 59 to that (29 for the margin of victory, plus 30 for actually winning the dual), and you get Navy's rating of 83.855.

[AFHawk86]

Begin with all wrestlers set to an even rating -- 100 in this case. Then, for each match wrestled, give each wrestler the value of their opponent's rating plus or minus a factor for winning or losing the match -- 30, plus one for each point the team won by. Total all of a wrestler's results, divide by the number of matches wrestled, and that's the end of a cycle. Then use those numbers as the start of the next cycle until you get the same results for each wrestler for two consecutive cycles.

SHP, would you restate this using an example to illustrate.

Sure. Navy has had one dual meet, and beat George Mason 33-4. George Mason's rating is 24.855. So add 59 to that (29 for the margin of victory, plus 30 for actually winning the dual), and you get Navy's rating of 83.855.

Thanks SHP. Now, show us how GM's rating of 24.855 comes into being.

[lightweight]

What does a negative score indicate? Does it mean that Sacred Heart would be better off if they just didn't get on the bus?

[SetonHallPirate]

What does a negative score indicate? Does it mean that Sacred Heart would be better off if they just didn't get on the bus?

LOL...no, it just shows that Sacred Heart is 100 points lower than average.

[SetonHallPirate]

SHP, would you restate this using an example to illustrate.

Sure. Navy has had one dual meet, and beat George Mason 33-4. George Mason's rating is 24.855. So add 59 to that (29 for the margin of victory, plus 30 for actually winning the dual), and you get Navy's rating of 83.855.

Thanks SHP. Now, show us how GM's rating of 24.855 comes into being.

Virginia: 112.123 rating minus 68 (loss 41-3) = 44.123
Gardner-Webb: 38.714 rating minus 42 (loss 24-12) = -3.286
Appalachian State: 52.016 rating minus 30 (loss 22-22, criteria point is not counted) = 22.016
Navy: 83.855 rating minus 59 (loss 33-4) = 24.855
Drexel: 63.368 rating minus 35 (loss 24-19) = 28.368
American: 77.054 rating minus 44 (loss 23-9) = 33.054
Those six dual ratings add up to 149.130, and divided by six (number of duals) equals 24.855

[VakAttack]

I think this would play better if you normalized the numbers over a 1-100 scale. I still appreciate the system, but I think it would be more attractive and understandable to people who tend to think in Base-10.

[SetonHallPirate]

I think this would play better if you normalized the numbers over a 1-100 scale. I still appreciate the system, but I think it would be more attractive and understandable to people who tend to think in Base-10.

The first column of numbers in the next post will be based on the margin of victory rating (ie. if a team wins by five, the factor is five points), and the second column is based on the win/loss-specific rating (where the factor is 30 points, regardless of whether the team wins the dual by criteria or in a 60-(-1) stomping). That's why I keep it in the form it is.

[Pinnum]

Correct me if I am wrong, but didn't you not use to factor in the 30 points for a win?

Also, I see not all the teams are ranked. I assume this is due to them not having duals yet and thus all of the numbers are skewed and need more duals to be blended better.

[SetonHallPirate]

1	Oklahoma State	78.275	89.957
2	Ohio State	76.050	87.987
3	Missouri	73.578	88.775
4	Penn State	84.529	74.931
5	Illinois	77.335	76.588
6	Iowa	75.906	72.572
7	Virginia Tech	69.842	74.910
8	Cornell	68.684	71.702
9	Minnesota	70.708	67.884
10	Edinboro	63.261	71.697
11	Pittsburgh	63.665	70.102
12	Rutgers	61.543	70.202
13	Purdue	61.784	66.721
14	Penn	56.379	70.949
15	Nebraska	60.849	66.295
16	Kent State	58.845	68.162
17	Michigan	60.179	65.830
18	Maryland	54.077	68.951
19	Oklahoma	61.866	59.207
20	Wyoming	61.060	58.441
21	Central Michigan	58.706	60.554
22	Arizona State	59.866	57.625
23	Lehigh	56.585	59.141
24	Old Dominion	49.659	65.619
25	Bloomsburg	53.228	61.807
26	Wisconsin	50.330	63.541
27	Northwestern	57.066	56.057
28	Virginia	60.901	51.222
29	Michigan State	51.487	53.848
30	Oregon State	60.274	43.783
31	Rider	48.216	55.058
32	North Carolina	55.481	44.409
33	Army	46.501	52.161
34	Chattanooga	46.830	48.632
35	Iowa State	46.841	43.767
36	Clarion	46.651	41.975
37	Columbia	41.368	46.679
38	Stanford	40.388	47.492
39	Northern Illinois	38.640	48.337
40	Utah Valley	42.048	44.315
41	Binghamton	45.195	40.837
42	Bucknell	42.400	42.886
43	Brown	39.427	44.590
44	Indiana	40.354	43.647
45	Navy	52.550	31.305
46	West Virginia	41.803	41.941
47	CSU Bakersfield	38.037	41.580
48	The Citadel	34.732	44.836
49	Lock Haven	44.372	34.994
50	Ohio	40.620	37.860
51	Northern Colorado	41.849	36.295
52	Buffalo	42.613	34.978
53	American	41.762	35.292
54	South Dakota State	33.728	39.308
55	Northern Iowa	39.330	33.541
56	Eastern Michigan	40.719	31.817
57	North Dakota State	48.958	23.111
58	Hofstra	38.275	33.750
59	Boise State	47.250	18.305
60	Drexel	32.485	30.884
61	Princeton	32.733	29.978
62	Boston U.	34.311	27.713
63	Davidson	25.599	32.047
64	Duke	33.321	21.234
65	Cleveland State	33.192	20.959
66	Southern Illinois Edwardsville	27.805	26.198
67	Appalachian State	24.943	27.073
68	Campbell	22.868	26.822
69	Franklin & Marshall	27.427	14.590
70	Gardner-Webb	26.659	12.055
71	Cal Poly	19.761	6.232
72	George Mason	23.550	1.305
73	VMI	22.956	-7.263
74	Sacred Heart	-9.987	-6.681

[SetonHallPirate]

Correct me if I am wrong, but didn't you not use to factor in the 30 points for a win?

You're right, that 30 point factor for the win was added in for the 2010-2011 season (actually, it was a rare midseason adjustment, after Cornell dropped in the now-margin-of-victory subranking after winning National Duals).

As you mentioned, the first number in the most recent post is the margin of victory subranking, the secocnd number is the win/loss specific subranking.

Also, I see not all the teams are ranked. I assume this is due to them not having duals yet and thus all of the numbers are skewed and need more duals to be blended better.

True in part. Air Force, Harvard, and North Carolina State have yet to wrestle any duals against Division I opponents, and therefore are not ranked. I wouldn't say all of the numbers are skewed, but some of the teams do have a somewhat small subset, and therefore their ranking is more an indication of the schedule they have wrestled than their actual performance to date.

[Pinnum]

Thanks for the info. I am always a big fan of these rankings.

[LtWtLooker]

What is the correction for teams that overwhelm a poor schedule early in the season and then win or lose by very tight margins when they hit the meatier part? Or a really strong team in a mediocre conference that loses or wins narrowly to the better teams and finishes by killing its weak division opponets at end of season

[SetonHallPirate]

What is the correction for teams that overwhelm a poor schedule early in the season and then win or lose by very tight margins when they hit the meatier part? Or a really strong team in a mediocre conference that loses or wins narrowly to the better teams and finishes by killing its weak division opponets at end of season

I don't understand your question, but there's a win/loss component to it, and a margin of victory component, both of which have underlying strength-of-schedule components figured in. If you're pummelling soft opposition early in the season, that'll be borne out by a strong ranking on the margin of victory side, and a weaker ranking on the win/loss side. (if you're wondering why Penn State is only ranked fourth, it's because of a weak strength of schedule)

[LtWtLooker]

You beat your first 6 teams, who are in the bottom 10 on the chart by 40+ pts and then you split against the next 6 who are top 10 by a score of16-15. I do realize that scenario would never happen but am using it as an illustration.

[jammen]

Maybe this will make sense later in the season, but right now I have a hard time believing that Michigan State is a better dual team than Oregon State.

Unless you're saying that Michigan State isn't a better team, just higher ranked.

[SetonHallPirate]

Maybe this will make sense later in the season, but right now I have a hard time believing that Michigan State is a better dual team than Oregon State.

Unless you're saying that Michigan State isn't a better team, just higher ranked.

Keep in mind that North Dakota State and Boise State are both ranked in the bottom 20, which hurts Oregon State badly, and Minnesota is a bit underranked by the formula as well.

The Margin of Victory ranking probably does a better job than the Win/Loss ranking this early in the season, and of course, Oregon State is nearly ten points better than Michigan State in the MoV ranking.

[SetonHallPirate]

You beat your first 6 teams, who are in the bottom 10 on the chart by 40+ pts and then you split against the next 6 who are top 10 by a score of16-15. I do realize that scenario would never happen but am using it as an illustration.

In the current rankings, that team would have a ranking of roughly 130 (primarily because the bottom is so skewed negatively from a normal bell curve, which I would expect to shake out over the course of the full season), which would put it between Rutgers and Purdue (12th and 13th).

[rossel3]

What is the correction for teams that overwhelm a poor schedule early in the season and then win or lose by very tight margins when they hit the meatier part? Or a really strong team in a mediocre conference that loses or wins narrowly to the better teams and finishes by killing its weak division opponets at end of season

I don't understand your question, but there's a win/loss component to it, and a margin of victory component, both of which have underlying strength-of-schedule components figured in. If you're pummelling soft opposition early in the season, that'll be borne out by a strong ranking on the margin of victory side, and a weaker ranking on the win/loss side. (if you're wondering why Penn State is only ranked fourth, it's because of a weak strength of schedule)

---------------

so.... you're telling me Iowa isn't the only d1 team with a "cupcake" schedule?

[RYouAskinMe]

Rutgers 12 ? Nebraska (15) manhandled them 22-10.

I guess it's gonna take at least the full season for things to shake out. From the results of the first month, RU may not even warrant top 25. Rinaldi, Zanetti and Smith are doin just fine, but once again Langel Melde, Mason and Winston are falling far short of expectations. The projected talent on this team could put it top 10, no doubt about it, but it's just not happening.

[pm-01]

Minnesota is a bit underranked by the formula.

Minnesota is definitely under ranked and will likely beat Iowa. If Iowa doesn't settled in at some middle and upper weights they could have a frustrating season.

[SetonHallPirate]

Rutgers 12 ? Nebraska (15) manhandled them 22-10.

I guess it's gonna take at least the full season for things to shake out. From the results of the first month, RU may not even warrant top 25. Rinaldi, Zanetti and Smith are doin just fine, but once again Langel Melde, Mason and Winston are falling far short of expectations. The projected talent on this team could put it top 10, no doubt about it, but it's just not happening.

That's Rutgers' only loss, and the only dual they've won by less than that twelve-point margin is an eleven point win over Army. Further, while Rutgers' wins haven't exactly been quality, they also haven't been against horrible competition.

Also, the dual-results ranking only takes into account the scores of dual meets, and doesn't take into account who contributes to those scores. (the dual-component ranking, on the other hand, takes only the individual contributions into account)