New RPI Rankings

[Lunatic Fringe]

Things are looking up on getting a good group through to NCAA's

Weight/Name/Coaches/RPI
125 Jimenez -/#26
133 Taylor #16/-
149 Crone #33/-
165 Jordan #2/#3
174 Robertson #25/#20
184 Christensen #31/#24
285 Horwath #33/-

http://www.ncaa.com/news/wrestling/article/2016-02-11/ncaa-releases-wrestling-coaches-panel-rankings-and-rpi

[fastnfurious]

I think you mean Jordan at 165

[Lunatic Fringe]

Fixed. Thanks Fast!

[TomM]

http://grfx.cstv.com/photos/schools/rutu/sports/m-wrestl/auto_pdf/2015-16/misc_non_event/Coaches_RPI.pdf

[vsmf2010]

Wow, the RPI makes no sense to me. Very bizarre results at some weights.

[badgerjohn]

Wow, the RPI makes no sense to me. Very bizarre results at some weights.

This explanation is from 2 years ago....

http://www.blackshoediaries.com/2014/2/22/5436482/wrestling-conference-allocations-explained-ncaa-big-ten-auto-bids-acc-big12-mac-eiwa-ewl

What is RPI?

Rating Percentage Index. Basically, it is strength of schedule. The NCAA multiples your win percentage, your opponents' win percentage, and your opponents' opponents' win percentage, to reach a pure math factor. The higher the factor, the better off you are, because only the Top 33 receive an RPI rank.

Oh by the way - you must have wrestled a minimum of 17 D1 matches in your particular weight class, too. You cannot wrestle 2 matches, win both, and bring a 100% win percentage to the table. However, the 17 match cutoff does not apply to your opponents, nor your opponents' opponents. That is, you don't get penalized for wrestling guys whose season was cut short by injury, or otherwise weren't able to wrestle 17 D1 matches. So we've got that going for us, which is nice.

Let's use David Taylor as an example. He's undefeated, so his Win% is 1.000 - that's pretty good. The Magic Man has also wrestled just about everyone in the top 10, so his opponents' win% is also above average. Let's call it 0.650. And his opponents' opponents' win percentage? You and I have no idea, of course. But let's just call it 50% - 0.500. Multiply those three percentages: [1.000] x [0.650] x [0.500] = 0.325. Hooray! The 0.325 by itself is meaningless. But, when you perform the same calculation for every wrestler at 165 pounds, and then sort them in descending order (after excluding anyone with fewer than 17 matches)....Ta-Da, you have a usable metric.

If, after all of that multiplication and sorting, you have a Top-33 factor, then you get an RPI. The key gotcha here, of course, is that you must have wrestled at least 17 D1 matches. Otherwise, you're excluded.

[vsmf2010]

I understand it is just a data set but the set up is quite flawed. How can a guy who is 23-0 with 10 pins, 5 techs, and 4 majors have a lower RPI than a guy he beat. Definitely need a head to head adjustment and some different weighting. 

[TomM]

BadgerJohn made that explanation up

[stp]

Wow, the RPI makes no sense to me. Very bizarre results at some weights.

This explanation is from 2 years ago....

http://www.blackshoediaries.com/2014/2/22/5436482/wrestling-conference-allocations-explained-ncaa-big-ten-auto-bids-acc-big12-mac-eiwa-ewl

What is RPI?

Rating Percentage Index. Basically, it is strength of schedule. The NCAA multiples your win percentage, your opponents' win percentage, and your opponents' opponents' win percentage, to reach a pure math factor. The higher the factor, the better off you are, because only the Top 33 receive an RPI rank.

Oh by the way - you must have wrestled a minimum of 17 D1 matches in your particular weight class, too. You cannot wrestle 2 matches, win both, and bring a 100% win percentage to the table. However, the 17 match cutoff does not apply to your opponents, nor your opponents' opponents. That is, you don't get penalized for wrestling guys whose season was cut short by injury, or otherwise weren't able to wrestle 17 D1 matches. So we've got that going for us, which is nice.

Let's use David Taylor as an example. He's undefeated, so his Win% is 1.000 - that's pretty good. The Magic Man has also wrestled just about everyone in the top 10, so his opponents' win% is also above average. Let's call it 0.650. And his opponents' opponents' win percentage? You and I have no idea, of course. But let's just call it 50% - 0.500. Multiply those three percentages: [1.000] x [0.650] x [0.500] = 0.325. Hooray! The 0.325 by itself is meaningless. But, when you perform the same calculation for every wrestler at 165 pounds, and then sort them in descending order (after excluding anyone with fewer than 17 matches)....Ta-Da, you have a usable metric.

If, after all of that multiplication and sorting, you have a Top-33 factor, then you get an RPI. The key gotcha here, of course, is that you must have wrestled at least 17 D1 matches. Otherwise, you're excluded.

In other words, RPI is stupid. 

  Bo Jordan is at #8, yet had a major decision win against #5 Welch from Purdue and a decision win against #3 Rodriguez from Illinois. 

[Micah]

I don't think RPI is meant as an actual ranking.  Think of it as more of a ranking by point accumulation.  Similar to a stat like TDs.  I could lead NCAAs in TDs but not be undefeated or ranked number one, I just have more TDs and leading that category.  Think of it more as each guy accumulates points based on wins and competition level.  The more matches you wrestle more points you can accumulate.  Bo has less matches than everyone, thus less opportunities to collect points.  They aren't looking at records or head to head wins just points.  Top 33 in points at each weight earn qualifiers for their conference.

This could be completely wrong but that is how I view it.

[badgerjohn]

BadgerJohn made that explanation up

I put a blank sheet of paper under my pillow, next day it had all that explanation on it.   

[imnofish]

BadgerJohn made that explanation up

I put a blank sheet of paper under my pillow, next day it had all that explanation on it.   

Sounds like the tooth fairy is moonlighting again...