First Attempt at Pablo Wrestling Rankings

[Pablo]

I'm new here, but not new to wrestling.  I grew up in Iowa and am an high school official on leave while my kids are in high school.

But in another part of my life, I think about things like rankings.  Back 25 years now, I created a ranking system for NCAA volleyball, the Pablo Volleyball Rankings.  It ended up working out pretty well, and is now well-recognized in the college volleyball community (I'm like the Ken Pomm of volleyball, before Ken Pomm was famous).  At one point, the coaches tried to get the NCAA to use Pablo Rankings instead of RPI for making NCAA selections (yeah, that got nowhere)!  I've even presented it at the American Statistical Society.

Recently I've been thinking about how to do it with wrestling.  There are some fascinating aspects that come with wrestling rankings (I love the concept of ranking teams for dual meets - I've got a methodology for that), but the big challenge is finding the match outcomes.  Well, I finally found something that would work with the Flo weekly reports, and with the help of contributors here, I've gotten to the point where I can try.  The methodology at this point is very raw.  My match data set is ok, I think.  I have the results for most of the D1 dual meets this year, and I have the big tournaments (Cliff Keen and Midlands).  As I slog through all the teams on trac, I hope to have all the matches for all the wrestlers from all the D1 teams, but that will take me a while (I'm talking all the JV guys and everything - it's going to be 10s of thousands of matches and probably more than 10 000 wrestlers).

I won't go into too much detail about the methodology, but the standard Pablo approach is that it uses match performance in head-to-head matchups to predict future match outcomes.  I don't have enough data yet to have a very robust model, but it's based on point differential in matches up to and including tech falls, and pins are treated separately.  It goes beyond an ELO (Chess, win/loss) approach because I have found that, at least in volleyball, point differential matters when it comes to predicting future performance.  When I get more data with wrestling, hopefully I will be able to establish that more solidly, but for now I am just working from a mostly assumed model.  I've done some testing and have found that, for example, a tech fall has basically the same predictive value as a 15 point win (IOW, there is no premium for a tech fall over and above the points) and that a pin is worth basically twice that of a tech. 

But there is a long way to go to determine what is best.  At this point, I am just treating all the matches through the season as equal, and not giving more weight to more recent matches, but that is possible and likely come eventually.   

What I am going to to do in the next posts is share the results that I have.  Some kind of surprises in small ways, but looking at the big picture, I think things are reasonable.  A lot of Penn St guys in the top 5, for example.  I won't post the actual Pablo "rating" values for each guy because I don't quite know how to interpret them yet.  Ultimately, the goal of the ratings to be able to calculate a match probability based on the ratings difference, but I just don't have that model yet.  It requires the right raw data.  I'm working on it.  Results below, once I figure out how to format them.

 

ETA:  By the way, in order to be included in the rankings the wrestler has to have at least 5 matches in my dataset

Edited 21 hours ago by Pablo
added info about minimum matches

[Pablo]

1	.	125:West Virginia_Jett Strickenberger
2	.	125:Purdue_Matt Ramos
3	.	125:Arizona St_Richard Figueroa
4	.	125:Ohio St_Brendan McCrone
5	.	125:Oklahoma St_Troy Spratley
6	.	125:Penn St_Luke Lilledahl
7	.	125:Wisconsin_Nicolar Rivera
8	.	125:Virginia Tech_Eddie Ventresca
9	.	125:Penn_Max Gallagher
10	.	125:Lehigh_Sheldon Seymour
11	.	125:NC St_Vince Robinson
12	.	125:Princeton_Marc-Anthony McGowan
13	.	125:Nebraska_Caleb Smith
14	.	125:Northern Illinois_Blake West
15	.	125:Missouri_Gage Walker
16	.	125:Rutgers_Dean Peterson
17	.	133:American_Raymond Lopez
18	.	125:Oregon St_Maximo Renteria
19	.	133:Army_Ethan Berginc
20	.	125:Indiana_Jacob Moran
21	.	125:Oklahoma_Antonio Lorenzo
22	.	125:North Carolina_Spencer Moore
23	.	125:Harvard_Diego Sotelo
24	.	125:Minnesota_Cooper Flynn
25	.	125:Iowa St_Kysen Terukina
26	.	125:West Virginia_Jace Schafer
27	.	125:South Dakota St_Tanner Jordan
28	.	133:Appalachian St_Chad Bellis
29	.	125:Cal Poly_Koda Holeman
30	.	125:Iowa_Joey Cruz
31	.	133:Bucknell_Kurt Phipps
32	.	125:Northern Iowa_Trever Anderson
33	.	125:Utah Valley_Bridger Ricks
34	.	125:Cornell_Marcello Milani
35	.	125:Purdue_Isaiah Quintero
36	.	125:Northwestern_Dedrick Navarro
37	.	125:Army_Charlie Farmer
38	.	125:Binghamton_Carson Wagner
39	.	125:Princeton_Ethan Rivera
40	.	125:Campbell_Cooper Shore

[Pablo]

1.	133:Little Rock_Nasir Bailey
2.	133:Cal Poly_Zeth Romney
3.	133:Ohio St_Nic Bouzakis
4.	133:Iowa_Drake Ayala
5.	133:Illinois_Lucas Byrd
6.	133:Northern Colorado_Dominick Serrano
7.	133:Stanford_Tyler Knox
8.	133:Nebraska_Van Dee
9.	133:Iowa St_Evan Frost
10.	133:Virginia Tech_Connor McGonagle
11.	133:Maryland_Braxton Brown
12.	133:Ohio St_Ben Davino
13.	133:ND St_Kyle Burwick
14.	133:Penn St_Braeden Davis
15.	133:Minnesota_Tyler Wells
16.	133:NC St_Troy Hohman
17.	133:Michigan_Dylan Ragusin
18.	133:Indiana_Angelo Rini
19.	133:Penn_Ryan Miller
20.	133:Wisconsin_Zan Fugitt
21.	133:Lock Haven_Anthony Noto
22.	133:Rutgers_Dylan Shawver
23.	133:Oklahoma St_Reece Witcraft
24.	133:Northern Illinois_Markel Baker
25.	133:Chattanooga_Blake Boarman
26.	141:Arizona St_Emilio YsaguirreJr
27.	133:Gardner-Webb_Takeo Davis
28.	133:North Carolina_Ethan Oakley
29.	133:Central Michigan_Vince Perez
30.	133:Edinboro_Colton Camacho
31.	141:Army_Braden Basile
32.	133:Drexel_Kyle Waterman
33.	133:Northern Iowa_Julian Farber
34.	133:Virginia_Gable Porter
35.	133:Penn St_Kyison Garcia
36.	133:South Dakota St_Derrick Cardinal
37.	133:Central Michigan_Sean Spidle
38.	133:Utah Valley_Kase Mauger
39.	133:SIU Edwardsville_Marcel Lopez
40.	133:Missouri_Kade Moore

[Pablo]

1.	141:Ohio St_Jesse Mendez
2.	141:Nebraska_Brock Hardy
3.	141:Penn St_Beau Bartlett
4.	141:Navy_Josh Koderhandt
5.	141:Rutgers_Joesph Oliveri
6.	141:Oklahoma St_Tagen Jamison
7.	141:Northern Colorado_Andrew Alirez
8.	141:Michigan_Sergio Lemley
9.	141:Minnesota_Vance Vombaur
10.	141:Penn_CJ Composto
11.	141:Rutgers_Joseph Olivieri
12.	141:Iowa St_Jacob Frost
13.	141:Virginia Tech_Sam Latona
14.	141:Lehigh_Carter Bailey
15.	141:Columbia_Kai Owen
16.	149:Army_Trae McDaniel
17.	141:Oklahoma_Mosha Schwartz
18.	141:Northern Iowa_Cael Happel
19.	141:Stanford_Jason Miranda
20.	141:Virginia_Dylan Cedeno
21.	141:Princeton_Eligh Rivera
22.	141:Missouri_Josh Edmond
23.	141:Pittsburgh_Briar Priest
24.	141:Iowa_Cullan Schriever
25.	141:Cornell_Joshua Saunders
26.	141:Drexel_Jordan Soriano
27.	141:Illinois_Danny Pucino
28.	141:SIU Edwardsville_Danny Martinez
29.	141:Iowa_Ryder Block
30.	149:Arizona St_Jesse Vasquez
31.	141:Chattanooga_Eli Knight
32.	141:North Carolina_Jayden Scott
33.	141:Purdue_Greyson Clark
34.	141:Brown_Ian Oswalt
35.	141:Utah Valley_Haiden Drury
36.	141:Wyoming_Cole Brooks
37.	141:Lock Haven_Wyatt Henson
38.	141:Gardner-Webb_Todd Carter
39.	141:Little Rock_Van Hoecke
40.	141:Oregon St_Nash Singleton

[Pablo]

1.	149:Penn St_Van ness	1.	157:Cornell_Meyer Shapiro	1.	165:Penn St_Mitchell Mesenbrink
2.	149:Virginia Tech_Caleb Henson	2.	157:Nebraska_Antrell Taylor	2.	165:Iowa_Mike Caliendo
3.	149:Nebraska_Ridge Lovett	3.	165:Arizona St_Nicco Ruiz	3.	165:West Virginia_Peyton Hall
4.	149:Illinois_Kannon Webster	4.	157:Penn St_Tyler Kasak	4.	165:Minnesota_Andrew Sparks
5.	149:Iowa_Kyle Parco	5.	157:Ohio St_Paddy Gallagher	5.	165:Missouri_Cameron Steed
6.	149:North Carolina_Lachlan McNeil	6.	157:Ohio St_Brandon Cannon	6.	165:Cornell_Julian Ramirez
7.	149:Iowa St_Anthony Echemendia	7.	157:Little Rock_Matty Bianchi	7.	165:Illinois_Braeden Scoles
8.	149:Oregon St_Ethan Stiles	8.	157:Iowa_Jacori Teemer	8.	165:Nebraska_Christopher Minto
9.	149:Northern Iowa_Colin Realbuto	9.	157:Northwestern_Trevor Chumbley	9.	165:Utah Valley_Terrell Barraclough
10.	149:Penn_Cross Wasilewski	10.	157:Northern Colorado_Vinny Zerban	10.	165:Oklahoma St_Cameron Amine
11.	149:Ohio St_Dylan D'Emilio	11.	157:Northern Iowa_Ryder Downey	11.	165:Michigan_Beau Mantanona
12.	149:Iowa St_Paniro Johnson	12.	157:Oregon St_CJ Hamblin	12.	174:Appalachian St_Lucas Uliano
13.	149:Stanford_Jaden Abas	13.	157:Purdue_Joey Blaze	13.	165:South Dakota St_Drake Rhodes
14.	149:Princeton_Ty Whalen	14.	157:Iowa St_Cody Chittum	14.	165:George Mason_Evan Maag
15.	149:Rider_Sammy Alvarez	15.	157:Ohio University_Peyten Kellar	15.	165:Northwestern_Maxx Mayfield
16.	149:Wyoming_Gabe Willochell	16.	157:Penn St_Alex Facundo	16.	165:Nebraska_Bubba Wilson
17.	149:Pittsburgh_Finn Solomon	17.	157:Michigan_Chase Saldate	17.	165:Cal Baptist_Drayden Morton
18.	149:Pittsburgh_Kade Brown	18.	157:Virginia Tech_Rafael Hipolito	18.	165:Columbia_Cesar Alvan
19.	149:Oklahoma_Willie McDougald	19.	157:Penn_Jude Swisher	19.	165:Little Rock_Joseph Bianchi
20.	149:NC St_Koy Buesgens	20.	157:Stanford_Grigor Cholakyan	20.	165:Stanford_Hunter Garvin
21.	149:Little Rock_Jordan Williams	21.	157:Maryland_Ethen Miller	21.	165:Rider_Enrique Munguia
22.	149:Northwestern_Sam Cartella	22.	165:Army_Gunner Filipowicz	22.	165:South Dakota St_Marcus Espinoza-Owens
23.	149:Michigan_Dylan Gilcher	23.	157:Lehigh_Logan Rozynski	23.	165:Northern Iowa_Jack Thomsen
24.	149:Oklahoma St_Carter Young	24.	157:NC St_Ed Scott	24.	165:Indiana_Tyler Lillard
25.	149:Virginia_Jack Gioffre	25.	157:Central Michigan_Johnny Lovett	25.	165:Ohio St_Sammy Sasso
26.	149:Missouri_Logan Gioffre	26.	157:Minnesota_Tommy Askey	26.	165:Bucknell_Noah Mulvaney
27.	149:Edinboro_Ryan Burgos	27.	157:Pittsburgh_Dylan Evans	27.	165:Hofstra_Kyle Mosher
28.	149:Lehigh_Malyke Hines	28.	157:Brown_Blake Saito	28.	165:Virginia_Nick Hamilton
29.	149:Columbia_Richard Fedalen	29.	157:North Carolina_Sonny Santiago	29.	165:Binghamton_Carter Baer
30.	149:Rutgers_Andrew Clark	30.	157:Navy_Jonathan Ley	30.	165:Rutgers_Anthony White
31.	149:Minnesota_Drew Roberts	31.	157:Northern Illinois_Landen Johnson	31.	165:Cal Poly_Luka Wick
32.	149:Kent St_Billy Meiszner	32.	157:Oklahoma St_Caleb Fish	32.	165:NC St_Derek Fields
33.	149:Central Michigan_Mason Shrader	33.	157:Rider_Colton Washleski	33.	165:Iowa St_Aiden Riggins
34.	149:West Virginia_Sam Hillegas	34.	157:LIU_Brayden Roberts	34.	165:Virginia Tech_Mac Church
35.	149:Cornell_Ethan Fernandez	35.	157:Wyoming_Jared Hill	35.	165:Rutgers_Andrew Barbosa
36.	149:VMI_Patrick Jordon	36.	157:South Dakota St_Cobe Siebrecht	36.	165:Oklahoma_Tate Picklo
37.	149:George Mason_Kaden Cassidy	37.	157:Campbell_Chris Earnest	37.	165:Navy_Dylan Elmore
38.	149:South Dakota St_Colin Dupill	38.	157:Chattanooga_Noah Castillo	38.	165:Pittsburgh_Jared Keslar
39.	149:The Citadel_Carson DesRosier	39.	157:Cal Poly_Legend Lamer	39.	174:American_Caleb Campos
40.	149:Cal Baptist_Paul Kelly	40.	157:Buffalo_Kaleb Burgess	40.	174:Air Force_Jack Ganos

[Pablo]

1.	174:Penn St_Levi Haines	1.	184:Penn St_Carter Starocci	1.	197:Iowa_Stephen Buchanan
2.	174:Missouri_Keegan O'Toole	2.	184:Northern Iowa_Parker Keckeisen	2.	197:Penn St_Josh Barr
3.	174:Oklahoma St_Dean Hamiti	3.	184:Oklahoma St_Dustin Plott	3.	285:Air Force_Antonio Ramos
4.	174:Ohio St_Carson Kharchla	4.	184:Minnesota_Max McEnelly	4.	197:Michigan_Jacob Cardenas
5.	174:Iowa_Patrick Kennedy	5.	184:Ohio St_Ryder Rogotzke	5.	197:Little Rock_Stephen Little
6.	174:Nebraska_Lenny Pinto	6.	184:Maryland_Jaxon Smith	6.	197:Lehigh_Michael Beard
7.	174:Oklahoma_Gaven Sax	7.	184:Cornell_Chris Foca	7.	197:Oklahoma St_Luke Surber
8.	174:Ohio University_Garrett Thompson	8.	184:Rutgers_Brian Soldano	8.	197:Illinois_Zac Braunagel
9.	174:Stanford_Lorenzo Norman	9.	184:Iowa_Gabe Arnold	9.	197:Nebraska_Camden McDanel
10.	174:North Carolina_Joshua Ogunsanya	10.	184:Nebraska_Silas Allred	10.	197:Missouri_Aeoden Sinclair
11.	174:South Dakota St_Cade DeVos	11.	184:South Dakota St_Bennett Berge	11.	197:Minnesota_Isaiah Salazar
12.	174:Illinois_Dan Braunagel	12.	184:Iowa_Angelo Ferrari	12.	197:Pittsburgh_Mac Stout
13.	174:Virginia Tech_Lennox Wolak	13.	184:NC St_Dylan Fishback	13.	197:Northern Iowa_Wyatt Voelker
14.	174:Navy_Danny Wask	14.	184:Pittsburgh_Reece Heller	14.	197:Appalachian St_Carson Floyd
15.	174:Bucknell_Myles Takats	15.	184:West Virginia_Dennis Robin	15.	285:American_Emmanuel Ulrich
16.	174:Drexel_Jasiah Queen	16.	184:Oklahoma_Deanthony ParkerJr	16.	197:Oregon St_Trey Munoz
17.	174:Iowa St_MJ Gaitan	17.	184:Wyoming_Eddie Neitenbach	17.	197:Wyoming_Joseph Novak
18.	174:NC St_Matthew Singleton	18.	184:Rider_Isaac Dean	18.	197:CSU Bakersfield_AJ Ferrari
19.	174:Northern Iowa_Jared Simma	19.	184:Iowa St_Evan Bockman	19.	197:Indiana_Gabe Sollars
20.	174:Iowa_Nelson Brands	20.	184:Columbia_Nick Fine	20.	197:Ohio St_Seth Shumate
21.	174:Cornell_Simon Ruiz	21.	184:Missouri_Colton Hawks	21.	197:Virginia Tech_Andy Smith
22.	174:Michigan_Joseph Walker	22.	184:Illinois_Edmond Ruth	22.	197:Iowa St_Christian Carroll
23.	174:Pittsburgh_Luca Augustine	23.	184:North Carolina_Gavin Kane	23.	197:Drexel_Mickey O'Malley
24.	174:Minnesota_Clayton Whiting	24.	184:Drexel_Giuseppe Hoose	24.	197:Princeton_Luke Stout
25.	174:Binghamton_Brevin Cassella	25.	184:Virginia Tech_Thomas StewartJr	25.	197:West Virginia_Rune Lawrence
26.	174:Rutgers_Jackson Turley	26.	184:Princeton_Kole Mulhauser	26.	197:Northwestern_Evan Bates
27.	174:Lock Haven_Avery Bassett	27.	184:Edinboro_Jared McGill	27.	197:Stanford_Nikolas Stemmet
28.	174:West Virginia_Brody Conley	28.	184:Indiana_Donnell Washington	28.	197:Rutgers_John Poznanski
29.	174:Wisconsin_Lucas Condon	29.	184:Purdue_James Rowley	29.	197:West Virginia_Ian Bush
30.	174:Purdue_Brody Baumann	30.	184:Hofstra_Ross McFarland	30.	197:South Dakota St_Zach Glazier
31.	174:Army_Dalton Harkins	31.	184:Chattanooga_Kamdyn Munro	31.	197:Bucknell_Dillon Bechtold
32.	174:Iowa St_Aiden Riggins	32.	184:ND St_Aidan Brenot	32.	197:Purdue_Ben Vanadia
33.	174:Columbia_Jack McGill	33.	184:Rutgers_Shane Cartagena-Walsh	33.	197:Army_Wolfgang Frable
34.	174:Utah Valley_Mark Takara	34.	184:CSU Bakersfield_Gerrit Nijenhuis	34.	197:Cornell_Michael Dellagatta
35.	174:Penn_Nick Incontrera	35.	184:Utah Valley_Caleb Uhlenhopp	35.	197:North Carolina_Cade Lautt
36.	174:Little Rock_Tyler Brennan	36.	184:Bellarmine_Devan Hendricks	36.	197:NC St_Christian Knop
37.	174:Central Michigan_Alex Cramer	37.	184:George Mason_Malachi Duvall	37.	197:Wisconsin_Wyatt Ingham
38.	174:Oregon St_Sean Harman	38.	184:Lehigh_Caden Rogers	38.	197:Iowa St_Nate Schon
39.	174:Wyoming_Riley Davis	39.	184:Buffalo_Marcus Petite	39.	197:Campbell_Levi Hopkins
40.	174:Maryland_Branson John	40.	184:Navy_Daniel Williams	40.	197:Navy_Payton Thomas

[Pablo]

1.	285:Oklahoma St_Wyatt Hendrickson
2.	285:Minnesota_Gable Steveson
3.	285:Virginia Tech_Jimmy Mullen
4.	285:Penn St_Greg Kerkvliet
5.	285:NC St_Isaac Trumble
6.	285:Arizona St_Cohlton Schultz
7.	285:Illinois_Luke Luffman
8.	285:Lehigh_Owen Trephan
9.	285:Ohio St_Nick Feldman
10.	285:Northern Iowa_Lance Runyon
11.	285:Iowa_Ben Kueter
12.	285:Maryland_Seth Nevills
13.	285:Michigan_Joshua Heindselman
14.	285:Rutgers_Yaraslau Slavikouski
15.	285:Missouri_Seth Nitzel
16.	285:Nebraska_Harley Andrews
17.	285:Morgan St_Xavier Doolin
18.	285:Pittsburgh_Dayton Pitzer
19.	285:Indiana_Jacob Bullock
20.	285:South Dakota St_Luke Rasmussen
21.	285:Lock Haven_Gavin Hoffman
22.	285:Iowa St_Daniel Herrera
23.	285:Minnesota_Bennett Tabor
24.	285:Cal Poly_Trevor Tinker
25.	285:Binghamton_Cory Day
26.	285:ND St_Andrew Blackburn-Forst
27.	285:Cleveland St_Daniel Bucknavich
28.	285:Oklahoma_Juan Mora
29.	285:Stanford_Peter Ming
30.	285:Purdue_Hayden Filipovich
31.	285:Ohio University_Jordan Greer
32.	285:Brown_Alex Semenenko
33.	285:Bellarmine_Daulton Mayer
34.	285:North Carolina_Nolan Neves
35.	285:Bucknell_Logan Shephard
36.	285:Northern Colorado_Remington Peterson
37.	285:Cornell_Ashton Davis
38.	285:Wyoming_Sam Mitchell
39.	285:Missouri_Jarrett Stoner
40.	285:American_William Jarrell

[Idaho]

17 minutes ago, Pablo said:

1	.	125:West Virginia_Jett Strickenberger
2	.	125:Purdue_Matt Ramos
3	.	125:Arizona St_Richard Figueroa
4	.	125:Ohio St_Brendan McCrone
5	.	125:Oklahoma St_Troy Spratley
6	.	125:Penn St_Luke Lilledahl
7	.	125:Wisconsin_Nicolar Rivera
8	.	125:Virginia Tech_Eddie Ventresca
9	.	125:Penn_Max Gallagher
10	.	125:Lehigh_Sheldon Seymour
11	.	125:NC St_Vince Robinson
12	.	125:Princeton_Marc-Anthony McGowan
13	.	125:Nebraska_Caleb Smith
14	.	125:Northern Illinois_Blake West
15	.	125:Missouri_Gage Walker
16	.	125:Rutgers_Dean Peterson
17	.	133:American_Raymond Lopez
18	.	125:Oregon St_Maximo Renteria
19	.	133:Army_Ethan Berginc
20	.	125:Indiana_Jacob Moran
21	.	125:Oklahoma_Antonio Lorenzo
22	.	125:North Carolina_Spencer Moore
23	.	125:Harvard_Diego Sotelo
24	.	125:Minnesota_Cooper Flynn
25	.	125:Iowa St_Kysen Terukina
26	.	125:West Virginia_Jace Schafer
27	.	125:South Dakota St_Tanner Jordan
28	.	133:Appalachian St_Chad Bellis
29	.	125:Cal Poly_Koda Holeman
30	.	125:Iowa_Joey Cruz
31	.	133:Bucknell_Kurt Phipps
32	.	125:Northern Iowa_Trever Anderson
33	.	125:Utah Valley_Bridger Ricks
34	.	125:Cornell_Marcello Milani
35	.	125:Purdue_Isaiah Quintero
36	.	125:Northwestern_Dedrick Navarro
37	.	125:Army_Charlie Farmer
38	.	125:Binghamton_Carson Wagner
39	.	125:Princeton_Ethan Rivera
40	.	125:Campbell_Cooper Shore

You lost me already with Jett being #1 with 6 losses and just took a loss by major to Figs. Not sure how that happens in any methodology or algorthym. 

[Pablo]

14 minutes ago, Idaho said:

You lost me already with Jett being #1 with 6 losses and just took a loss by major to Figs. Not sure how that happens in any methodology or algorthym. 

With the matches I have, most of his losses are at 133.  He is 6-1 at 125, with his only loss to #3 Figueroa.  He is ranked #44 at 133, with a record of 3-3.

For now, I am basically ranking them solely within the weight classes, so his matches against 133 guys don't affect his ranking within 125

I don't know what matches of his I am missing. I went through the results on the WV website to make sure I had everything I could find.  I have not gotten to WV in trac yet.

ETA:  I think a lot of it is arising from his pin over Spratley. 

I want to come back to this just to clarify.  This is the problem with criticisms like this.  Idaho says, "But he just lost to #3 Figueroa!" and completely didn't mention the pin over #4 Spratley.  What is the best way to assess those results?  As I mentioned in the preamble, I have found that pins have much more significance in predicting future performance than do decisions, and so clearly what is happening here is that Pablo is putting a lot of emphasis on that pin.  Now, with only 7 matches at 125 in my dataset, that is going to carry a lot of weight.  More results at 125 will be useful, but based on what he has actually done (at least what I have), it is completely reasonable to see him ranked where he is.

Edited 20 hours ago by Pablo

[Idaho]

15 minutes ago, Pablo said:

With the matches I have, most of his losses are at 133.  He is 6-1 at 125, with his only loss to #3 Figueroa.  He is ranked #44 at 133, with a record of 3-3.

For now, I am basically ranking them solely within the weight classes, so his matches against 133 guys don't affect his ranking within 125

I don't know what matches of his I am missing. I went through the results on the WV website to make sure I had everything I could find.  I have not gotten to WV in trac yet.

ETA:  I think a lot of it is arising from his pin over Spratley. 

I want to come back to this just to clarify.  This is the problem with criticisms like this.  Idaho says, "But he just lost to #3 Figueroa!" and completely didn't mention the pin over #4 Spratley.  What is the best way to assess those results?  As I mentioned in the preamble, I have found that pins have much more significance in predicting future performance than do decisions, and so clearly what is happening here is that Pablo is putting a lot of emphasis on that pin.  Now, with only 7 matches at 125 in my dataset, that is going to carry a lot of weight.  More results at 125 will be useful, but based on what he has actually done (at least what I have), it is completely reasonable to see him ranked where he is.

He lost head to head by major decision to an undefeated and returning NCAA champ at the same weight. 

Edited 20 hours ago by Idaho

[OMW]

A lot of work and effort, great to see. However - and not a criticism- those look more like predictions than rankings. 

Regardless, good to read.

 

[Pablo]

12 minutes ago, Idaho said:

He lost head to head by major decision to an undefeated and returning NCAA champ at the same weight. 

Yeah, lots of guys do that.  

He also pinned the #4 ranked wrestler in the weight class.

Figueroa's result from last year does not affect this year's rankings.

The bigger issue that I just found is his loss to Moore from NC in the Southeast Open.  I don't have that in my dataset yet (not a dual, not Midlands or Cliff Keene).  I'm guessing that will make a big difference.

[Wrestleknownothing]

1 hour ago, Pablo said:

I'm new here, but not new to wrestling.  I grew up in Iowa and am an high school official on leave while my kids are in high school.

But in another part of my life, I think about things like rankings.  Back 25 years now, I created a ranking system for NCAA volleyball, the Pablo Volleyball Rankings.  It ended up working out pretty well, and is now well-recognized in the college volleyball community (I'm like the Ken Pomm of volleyball, before Ken Pomm was famous).  At one point, the coaches tried to get the NCAA to use Pablo Rankings instead of RPI for making NCAA selections (yeah, that got nowhere)!  I've even presented it at the American Statistical Society.

Recently I've been thinking about how to do it with wrestling.  There are some fascinating aspects that come with wrestling rankings (I love the concept of ranking teams for dual meets - I've got a methodology for that), but the big challenge is finding the match outcomes.  Well, I finally found something that would work with the Flo weekly reports, and with the help of contributors here, I've gotten to the point where I can try.  The methodology at this point is very raw.  My match data set is ok, I think.  I have the results for most of the D1 dual meets this year, and I have the big tournaments (Cliff Keen and Midlands).  As I slog through all the teams on trac, I hope to have all the matches for all the wrestlers from all the D1 teams, but that will take me a while (I'm talking all the JV guys and everything - it's going to be 10s of thousands of matches and probably more than 10 000 wrestlers).

I won't go into too much detail about the methodology, but the standard Pablo approach is that it uses match performance in head-to-head matchups to predict future match outcomes.  I don't have enough data yet to have a very robust model, but it's based on point differential in matches up to and including tech falls, and pins are treated separately.  It goes beyond an ELO (Chess, win/loss) approach because I have found that, at least in volleyball, point differential matters when it comes to predicting future performance.  When I get more data with wrestling, hopefully I will be able to establish that more solidly, but for now I am just working from a mostly assumed model.  I've done some testing and have found that, for example, a tech fall has basically the same predictive value as a 15 point win (IOW, there is no premium for a tech fall over and above the points) and that a pin is worth basically twice that of a tech. 

But there is a long way to go to determine what is best.  At this point, I am just treating all the matches through the season as equal, and not giving more weight to more recent matches, but that is possible and likely come eventually.   

What I am going to to do in the next posts is share the results that I have.  Some kind of surprises in small ways, but looking at the big picture, I think things are reasonable.  A lot of Penn St guys in the top 5, for example.  I won't post the actual Pablo "rating" values for each guy because I don't quite know how to interpret them yet.  Ultimately, the goal of the ratings to be able to calculate a match probability based on the ratings difference, but I just don't have that model yet.  It requires the right raw data.  I'm working on it.  Results below, once I figure out how to format them.

 

ETA:  By the way, in order to be included in the rankings the wrestler has to have at least 5 matches in my dataset

My brother.

I was even Pablo in sixth grade Spanish class.

[Pablo]

2 minutes ago, OMW said:

A lot of work and effort, great to see. However - and not a criticism- those look more like predictions than rankings. 

Regardless, good to read.

 

It all depends on what is your goal with a ranking.

Yes, the objective of Pablo rankings is to predict outcomes of matches that are to come, and it gets optimized toward that.  That's why I focus on predictive value of events.  And I do that because, if I ask the question, "Who is the better wrestler?" what I want to know is, if I put them on the mat against each other, who is more likely to win?  That's about probability.

Here's the way I usually describe it:

Suppose I took Gable Stevenson and replicated him, made an exact duplicate.  And I put them on the mat, and gave one a green ankle band, and the other gets a red ankle band.  Who do you think is going to win?  Well, since they are by definition exactly equal, there is an equal chance that either red or green wins.  Neither one is better than the other, they are equal.  But one of them is going to win, we just don't know which.

Now, give me the green ankle band and let me wrestle Gable Stevenson.  He's a better wrestler than I am and so the probability of him winning is basically 100% (we shake hands, the whistle blows and a ceiling tile falls on his head knocking him out, and I win by injury default.  I just have to run fast enough to allow the 10 seconds for the ceiling tile to fall all the way to the floor; that's my only chance of winning, but it is non-zero probability!).  So Gable Stevenson over Pablo has a 100% probability. 

Somewhere between in-between me and Gable Stevenson there is the guy that is way better than me but not quite as good a Gable.  Call him "Gable with a toothache."  If I put Gable Stevenson up against Gable with a toothache, who is going to win?  Well, a healthy Gable is a better wrestler than Gable with a toothache, so he's more likely to win, but the difference between them isn't big, so Gable with a toothache will still win a lot of the time.  Maybe 45%.  Healthy Gable is better, sure, but there is still a chance that Gable with a toothache can beat him.

What Pablo rankings are trying to do is to figure out, what is that probability between them?  In this model, I am ranking who's the better wrestler, with "better" being defined as "who is more likely to win if they wrestle each other"

In this model, upsets will happen.  Actually, by definition, they must happen. There will be times when the better wrestler loses.  And this is true - we know that there are times when upsets happen.  Heck, we can see this in rematches (this is something I need to explore at the college level; I've looked at it at the Indiana state high school level and you can absolutely see the upset rate (for example, if two guys match up in the regional finals and then match up again in the semi-state finals, the guy that won the regionals only wins 75% of the time at semi-state - same two guys, a week apart)), and you can also see it in failed transitivity (A beats B who beats C who beats A).

So this is the way I think about it. In some respect yes it is a predictive model, but in the end, for me, when I asked the question "who is better" that's what we are talking about.  

I can create models that do things like, maximize the correct number of wins and losses for matches that have been wrestled, but if you think having Jett at #1 is screwy here, those things can go way off the rails.

I just think that Pablo does a good job of reflect what we really mean when we talk about who is the "better wrestler."  It's all about who going to win if they wrestle, right?  But it's just not that simple.  Pit replicated Gable against Gable, and someone is going to win.  That doesn't mean either is a better wrestler.

Sorry for the long screed....I like thinking about this stuff.

[Pablo]

22 minutes ago, Wrestleknownothing said:

My brother.

I was even Pablo in sixth grade Spanish class.

Over the years I've been known as Pablo, Agent Buchwald and lately I've settled on The Bofa on the Sofa in much of my life.

If you were The Bofa in 3rd grade, that would be amazing.

[Pablo]

So I added in the match against Moore from November (he had another match against some non_D1 guy but I didn't bother).  I'm always a fan of more data

1	.	125:Purdue_Matt Ramos
2	.	125:Ohio St_Brendan McCrone
3	.	125:Arizona St_Richard Figueroa
4	.	125:Oklahoma St_Troy Spratley
5	.	125:West Virginia_Jett Strickenberger
6	.	125:Penn St_Luke Lilledahl
7	.	125:Wisconsin_Nicolar Rivera
8	.	125:Virginia Tech_Eddie Ventresca
9	.	125:Penn_Max Gallagher
10	.	125:Lehigh_Sheldon Seymour
11	.	125:NC St_Vince Robinson
12	.	125:Princeton_Marc-Anthony McGowan
13	.	125:Nebraska_Caleb Smith
14	.	125:Northern Illinois_Blake West
15	.	125:North Carolina_Spencer Moore
16	.	125:Rutgers_Dean Peterson
17	.	125:Oregon St_Maximo Renteria
18	.	125:Missouri_Gage Walker
19	.	125:Iowa St_Kysen Terukina
20	.	125:Indiana_Jacob Moran
21	.	125:Harvard_Diego Sotelo
22	.	125:Minnesota_Cooper Flynn
23	.	125:West Virginia_Jace Schafer
24	.	125:Oklahoma_Antonio Lorenzo
25	.	125:South Dakota St_Tanner Jordan
26	.	125:Cal Poly_Koda Holeman
27	.	125:Iowa_Joey Cruz
28	.	125:Northern Iowa_Trever Anderson
29	.	125:Cornell_Marcello Milani
30	.	125:Northwestern_Dedrick Navarro
31	.	125:Utah Valley_Bridger Ricks
32	.	125:Purdue_Isaiah Quintero
33	.	125:Army_Charlie Farmer
34	.	125:Binghamton_Carson Wagner
35	.	125:Princeton_Ethan Rivera
36	.	125:Campbell_Cooper Shore
37	.	125:Iowa St_Adrian Meza
38	.	125:Northern Iowa_Kyle Gollhofer
39	.	125:Virginia_Keyveon Roller
40	.	125:Pittsburgh_Nick Babin

[Idaho]

43 minutes ago, Pablo said:

Yeah, lots of guys do that.  

He also pinned the #4 ranked wrestler in the weight class.

Figueroa's result from last year does not affect this year's rankings.

The bigger issue that I just found is his loss to Moore from NC in the Southeast Open.  I don't have that in my dataset yet (not a dual, not Midlands or Cliff Keene).  I'm guessing that will make a big difference.

The  bigger issue is you have Jett ahead of an undefeated wrestler who majored him head to head 11-2 in a recent match Sorry, his pin of #4 does not trump his head to head loss to Figs. I don't know if you will find that logic anywhere. Even the Big 12 seedings have Figs 1 and Jett 3.   Beau Bartlett undefeated with head to head wins over both Mendez and Hardy, but he's #3 behind both of those guys??????

[Idaho]

7 minutes ago, Pablo said:

So I added in the match against Moore from November (he had another match against some non_D1 guy but I didn't bother).  I'm always a fan of more data

1	.	125:Purdue_Matt Ramos
2	.	125:Ohio St_Brendan McCrone
3	.	125:Arizona St_Richard Figueroa
4	.	125:Oklahoma St_Troy Spratley
5	.	125:West Virginia_Jett Strickenberger
6	.	125:Penn St_Luke Lilledahl
7	.	125:Wisconsin_Nicolar Rivera
8	.	125:Virginia Tech_Eddie Ventresca
9	.	125:Penn_Max Gallagher
10	.	125:Lehigh_Sheldon Seymour
11	.	125:NC St_Vince Robinson
12	.	125:Princeton_Marc-Anthony McGowan
13	.	125:Nebraska_Caleb Smith
14	.	125:Northern Illinois_Blake West
15	.	125:North Carolina_Spencer Moore
16	.	125:Rutgers_Dean Peterson
17	.	125:Oregon St_Maximo Renteria
18	.	125:Missouri_Gage Walker
19	.	125:Iowa St_Kysen Terukina
20	.	125:Indiana_Jacob Moran
21	.	125:Harvard_Diego Sotelo
22	.	125:Minnesota_Cooper Flynn
23	.	125:West Virginia_Jace Schafer
24	.	125:Oklahoma_Antonio Lorenzo
25	.	125:South Dakota St_Tanner Jordan
26	.	125:Cal Poly_Koda Holeman
27	.	125:Iowa_Joey Cruz
28	.	125:Northern Iowa_Trever Anderson
29	.	125:Cornell_Marcello Milani
30	.	125:Northwestern_Dedrick Navarro
31	.	125:Utah Valley_Bridger Ricks
32	.	125:Purdue_Isaiah Quintero
33	.	125:Army_Charlie Farmer
34	.	125:Binghamton_Carson Wagner
35	.	125:Princeton_Ethan Rivera
36	.	125:Campbell_Cooper Shore
37	.	125:Iowa St_Adrian Meza
38	.	125:Northern Iowa_Kyle Gollhofer
39	.	125:Virginia_Keyveon Roller
40	.	125:Pittsburgh_Nick Babin

So now the updated rankings have a 7 loss McCrone somehow passing undefeated Figs into #2 because you added a match to Jett?

 

[Gene Mills Fan]

McCrone too high got beat by 27 Cruz  

[Pablo]

1 minute ago, Idaho said:

Sorry, his pin of #4 does not trump his head to head loss to Figs. 

How did you determine that?

I have data that informs the value of a pin.  You are making things up.

[Pablo]

6 minutes ago, Idaho said:

The  bigger issue is you have Jett ahead of an undefeated wrestler who majored him head to head 11-2 in a recent match Sorry, his pin of #4 does not trump his head to head loss to Figs. I don't know if you will find that logic anywhere. Even the Big 12 seedings have Figs 1 and Jett 3.   Beau Bartlett undefeated with head to head wins over both Mendez and Hardy, but he's #3 behind both of those guys??????

Bartlett also lost to Mendez in the All-star classic.

You can't just ignore results you don't like.

 

[TylerDurden]

18 minutes ago, Pablo said:

How did you determine that?

I have data that informs the value of a pin.  You are making things up.

Counterpoint: you decided that a pin is more valuable in every situation for your calculation. 

Wrestling isn't the same as volleyball. A pin can happen for reasons other than domination - people get caught while leading, etc. 

In volleyball, you can't get caught. If you lose a set 25-14, you got smacked. 

I think that's where your system will fall apart when trying to make it work for wrestling. 

Edited 18 hours ago by TylerDurden

[Gene Mills Fan]

19 minutes ago, Pablo said:

How did you determine that?

I have data that informs the value of a pin.  You are making things up.

So he gets beat by a guy but he pinned someone else so he's better than the guy???

[PortaJohn]

16 minutes ago, Pablo said:

Bartlett also lost to Mendez in the All-star classic.

You can't just ignore results you don't like.

 

That match doesn't count.  It's an exhibition.  Kudos for the work you put into this but your top 3 at 141 is incorrect and Idaho is the furthest thing from a PSU homer

[jchapman]

6 minutes ago, TylerDurden said:

Counterpoint: you decided that a pin is more valuable in every situation for your calculation. 

Wrestling isn't the same as volleyball. A pin can happen for reasons other than domination - people get caught while leading, etc. 

In volleyball, you can't get caught. If you lose a set 25-14, you got smacked. 

I think that's where your system will fall apart when trying to make it work for wrestling. 

Plus, wrestling is a timed event, whereas volleyball is a first to so many points event.  So matches that are close in score at the end can reflect a strategy of the leading wrestler allowing the losing wrestler to score meaningless points as time runs out.  There is no such strategy in volleyball.