MMA Sabermetrics: Strength of Record

This has been an ongoing project for quite some time, and a lot more articles will likely be driven under the "MMA Sabermetrics" banner in the future. The intent behind these articles is to give fans an objective backing to many claims that analysts, fans, and bloggers have made in the past. It'll also give us a statistical look at certain aspects of MMA from not only a current snapshot of the overall landscape of the sport, but it can also include historical data as well. It should be pretty informative and get the minds of fans churning.

The first in this series will revolve around a statistic that I believe is very vital to those of you in the betting arena of MMA and also breaking down fights. It is the strength of schedule statistic. One of the ways to define this is to ask how you can compare two fighters who have never fought each other before. Most fans would subjectively look down their record and see the quality of wins, performances in fights, and their style matchups. While these will heavily outweight this statistic as it has attributes that don't come into the calculation, Strength of Schedule or Record (SOS) can give us a snapshot of the quality of wins a fighter has. It can also show us the weakness in their quality of wins vs. their record.

The calculation, in this case, was a replica of what many Football statisticians have been using for quite awhile, and it is modeled after the BCS SOS calculation and NCAA Ratings Index (Nothing to do with the rankings!). The calculation takes the Fighter's opponents win percentage + fighter's opponent's opponents win percentage and takes that average. This ensures that a fighter's SOS isn't simply because he fought guys who were 5-0, 6-0, 7-0, but those opponents fought 0-5, 0-6 easier opponents. I will be revamping the algorithms and trying different SoS calculations in the future as well. Most notably, this formula does include Fighter A vs. Opponent in the Opponent's Winning Percentage calculation, which is an ongoing debate as to whether it should be included or not. RIght now, most SoS calculations do not include it. For our purposes, it wouldn't affect the ratings very much because fighters do not have "seasons" like Football does.

In order to put more value on a fighter's direct opponents, the opponent's winning percentage is multiplied by a factor of 2 while the opponent's opponents winning percentage stays the same. These are then divided by 3. This puts more weight on the fighter's actual wins that their opponent's opponents wins.

The statistics below take into account that each fighter on the listing has fought in 2008. I didn't want to include older fighters who have went on to retire. It also filters out fighters who have only fought  10 or more fights, and it's been filtered by fighters that are likely not known to fans. It also only lists those fighters with a winning percentage above .500 or more wins than losses. We'll focus on fighters with losing records with a strong SoS later (i.e. a guy like Dong Sik Yoon).

Take note that B.J. Penn tops the list. Penn has not only fought quality opponents over his career, but his opponents have also gained great records by fighting quality opponents as well. This easily puts Penn at the top of the list.

Some names you'll want to avoid in this listing as they aren't really relevant, and in later articles, we'll focus solely on UFC fighters and promotion's listings.

Obviously, some fighters are boosted by one opponent such as Rousimar Palhares. Dan Henderson's extensive MMA record and quality opponents he has faced has obviously boomed Palhares' statistic along with solid veterans like Negao and Acacio. Griffin has consistently taken on very good competition on his route to the top as well.

#53: Demian Maia - .6693 (10-0)

#55: Joachim Hansen  0.6680 19 - 7       

#56: Thiago Alves 0.6677 16 - 3     

#57: Patrick Cote 6612 0.6670 13 - 5                            

#63: Andrei Arlovski 270 0.6660 14 - 6     

66: Dan Henderson 195 0.6653 24 - 7                            

68: Wilson Gouveia 3070 0.6650 12 - 6                    

69: Martin Kampmann 5344 0.6647 15 - 2                            

71: Wagnney Fabiano 905 0.6643 12 - 1                                                       

Keep in mind, this is out of over 1300 active fighters who have fought in 2008. Here's some of the bottom trollers:

1311: Jason Reinhardt 0.4837 20 - 1                            

1301: Ronnie Mann 0.4907 16 - 1        

1288: Matt Veach 0.4990 11 - 0                            

1272: Ryan Thomas 0.5083 9 - 3                             

1202: Rich Crunkilton 0.5263 16 - 2                            

1196: John Hathaway 0.5267 10 - 0                            

1168: Paul Bradley 0.5347 10 - 0                            

1142: Stefan Struve 0.5390 16 - 3        
                    
1143: Dan Lauzon 0.5387 12 - 2         

1099: Tom Speer 0.5467 11 - 4

1094: Brock Larson 0.5477 252 

Some of the notables I wanted to point out were that of Brock Larson and Jason Reinhardt. This stat is a perfect example as to how it can expose a fighter's wins when coupled with the opponent's opponents winning percentage.

Larson may have a phenomenal record, but his opponents lack the quality wins to boost his SOS to even average levels. It's safe to say that Larson has had a much easier time building his record than most of the quality fighters that are in the UFC today. Reinhardt is another perfect example as well as he was hyped as this undefeated phenom before his loss to Joe Lauzon.

Even more impressive, even with weight on Larson and Reinhardt's actual wins, the opponent's opponents winning percentage drags their averages down, as it should.

I'll have a bell curve in place at some point, but many of the bigger names are in the mid-range. Fedor is roughly a .65 and Rashad Evans hovers at .6437. These are typical and don't really tell us much. They've fought some very great fighters and some very mediocre fighters in comparison to their record.

Overall Analysis

This is my first sampling of this data which spreads over nearly 60,000 - 80,000+ fights historically. The statistics displayed here are filtered currently, and while I tweak algorithms and create some more meaningful stats, this is a statistic that would have some meaning at a quick glance if compared to record.

If their SoS is high, but their record is mediocre, it could mean big things in terms of an upset if the style is right. If their SoS is high with a phenomenal record, they are likely one of the top fighters in the sport, and it's justified.

Low SoS with a great record normally means a stacked record full of opponents who either were terrible or fought opponents who were equally terrible. It basically means a low quality of fights.  Low SoS and a losing record obviously indicates some lesser skills.

Some of the next installments will elaborate on some other stats that have been put together here on BloodyElbow. I'll be looking at submissions percentages OVER the entire landscape and history of MMA vs. what Mike Fagan has done in getting as much as he could within his sample. He did a great job with the piece, but I'd love to see the real percentages when including an even broader base to see if Josh Gross's claims that submissions have dramatically went down is justified.

Any other stats that anyone wants to see? I'd love to hear some SoS stat suggestions as well. I'm going to be working on excluding fights in which the fighter the statistic is being calculated for is eliminated as it should be from other formulas, and I'm also looking at a hybrid formula.