Welcome to a new series at Bloody Elbow. Regular readers have seen bits and pieces of RoboJudge before, but not with his new name - as dubbed by Rami Genauer of FightMetric - and not in a regular series. Since I receive monthly updates of FightMetric data, RoboJudge will be a monthly series reviewing the scoring of each and every UFC fight that goes to a decision. Basic RoboJudge numbers will be provided for all fights while controversial or otherwise interesting decisions will get more detailed breakdowns.
In January, I'll do an analysis of every UFC decision in 2014 and quantify the unusual ones. [Teaser: Diego Sanchez had less than a 5% chance of getting the nod over Ross Pearson] After that, new pieces will go up monthly. The most recent event won't always be included as its finalized data may not yet be available. But if that happens, the event in question will be covered the following month.
Today we're focusing on Johny Hendricks and his UFC title fight sequence - St-Pierre vs. Hendricks at UFC 167, Hendricks vs. Lawler 1 at UFC 171 and Hendricks vs. Lawler 2 at UFC 181. Many would argue that two, possibly even all three, of these decisions were controversial. Personally, I had Hendricks winning all three bouts and therefore being the two-time defending UFC welterweight champion in Gifted Fantasyland. But how do the decisions stack up to the way a typical MMA judge likely would've leaned? That's where RoboJudge comes into play.
RoboJudge is a variant of a scientific model calibrated to the round-by-round scoring decisions of real-life MMA judges and FightMetric's detailed fighter performance statistics. It crunches all the fight numbers and provides odds that an average judge would score each fighter the winner of a particular round. Going with or against these odds doesn't necessarily mean the actual decision is good or bad, but the more the odds are stacked against a particular fighter, the more anomalous a decision for said fighter becomes.
For example, two-and-a-half weeks ago Johny Hendricks had a 98.9 percent chance of being scored the winner of the 2nd round against Robbie Lawler at UFC 181. Put another way, Robbie Lawler had a teensy-tiny 1.1 percent chance of getting the nod. Sal D'Amato and Marcos Rosales both gave the round to Hendricks while Glenn Trowbridge went with Lawler. Was Trowbridge's decision bad? Not necessarily. It was just strange, anomalous, unusual...an outlier.
Oh what the hell, it was terrible! But who knows what happened? Perhaps Trowbridge missed a few key sequences, had a bad angle at times, wasn't looking at his monitor enough or got distracted by the Octagon Girls. Perhaps he put more weight on a few key Lawler strikes. And my personal favorite, perhaps he forgot which fighter was in which column when writing his scores down. We don't know exactly what was going on in Trowbridge's head, but we do know that a typical judge would've had a better chance of hitting the symbol bet on the Big Wheel than scoring Lawler the winner of round 2. Anomalous.
In what follows, I'll try not to say a particular decision was good or bad. That's up to you to decide. The point of the exercise is to see which decisions are relatively consistent with an average MMA judge and which ones make you go, "Hmmmm..."
For those who are new to this or have forgotten, a jab is any non-power strike. In their review of the tape, FightMetric uses various physical cues to determine if a strike has power. Since significant strikes include all strikes at distance plus power strikes in the clinch and on the ground, the jab/power distinction is a bit more finely tuned than the stats we see on TV. A common criticism is that classifying strikes into two categories (jab/power) doesn't account for possible nuanced effects, but there are other advantages. Unlike a regular judge, FightMetric - and indirectly RoboJudge - can view multiple camera angles, rewind and slow things down to frame-by-frame if need be to see what connects and how.
Georges St-Pierre vs. Johny Hendricks
Nov. 16, 2013
Nov. 16, 2013 was the last time my wife really enjoyed watching MMA which indirectly affects my viewership. Please come back, GSP!
Here's the actual decision. The critical round was the 1st. Sal D'Amato and Tony Weeks saw it for GSP while Glenn Trowbridge gave it to Hendricks. Here's what happened:
While everything is analyzed behind-the-scenes in the model, I include most, but not all, of the key elements judges tend to use when evaluating a round in the table. Missed strikes sometimes have value and sweeps matter, as do the different types of ground control. Judges also don't tend to value jabs to the body.
GSP had an early takedown with a quick guillotine attempt culminating in the "phantom tap" during Hendricks' escape. The round was largely contested on the feet with over 2 ½ minutes taking place in the clinch. Most of the clinch time saw GSP and Hendricks locked up off the cage where GSP landed knee strikes to the body and Hendricks attacked with knees to the thigh. They each threw 26 power strikes while GSP threw 9 more jabs than Hendricks, but GSP was also more inaccurate, missing or having 19 strikes blocked to Hendricks' 9.
While Hendricks had more control, the major performance discrepancies were Hendricks' 8 power head shots landed (including 5 Travis Browne-ian elbows) to GSP's 1, Hendricks' 8 power leg strikes landed to GSP's 2, and GSP's 10 power body strikes landed to Hendricks' 1. Remember, landing a power shot to the opponent's dome tends to be 75 percent more valuable in judges' eyes than a power strike to the body or legs.
When 10-8's aren't involved, you win a title fight on one judge's scorecard by being selected the winner of at least three rounds. You win the overall fight when two or more judges score you the winner. In what follows, I give the odds of being scored the winner of each round, the winner of the fight by a single RoboJudge and the overall winner by a panel of three RoboJudges.
Consistent with the odds, Hendricks took rounds 2 and 4 with GSP taking 3 and 5. Round 1 was critical and Hendricks had a 66.7 percent chance of taking it on a typical judge's scorecard. Hendricks was 72.0 percent to win the fight on any one judge's scorecard and therefore 80.8 percent to win overall. Did he get robbed? That's for you to decide.
Johny Hendricks vs. Robbie Lawler 1
Mar. 15, 2014
Here's the actual decision. There weren't any incredibly controversial rounds in this fight, just a peculiar 10-10 round 5 from Doug Crosby.
Overall, here are the performance stats:
The first three minutes of the round were reasonably close while the last two minutes were all "Bigg Rigg" Hendricks. Don't you recall thinking at one point he just might be able to drop Lawler? RoboJudge clearly believes the quantitative version of that.
Both fighters had strong winning rounds with the first being the closest. Still, Hendricks had three solid rounds for an 80.2 percent chance of winning a typical judge's scorecard and an 89.8 percent chance of taking the overall fight.
Johny Hendricks vs. Robbie Lawler 2
Dec. 6, 2014
Here's the actual decision. The interesting rounds are the 2nd and 4th.
As we'll see below, round 2 was almost Hendricks' most dominant round, except to Glenn Trowbridge who saw things - real or otherwise - that made him say, "Lawler, he's my guy."
Round 4's the controversial one. The fighters spent 2:55 at distance, 1:22 in the clinch with Hendricks in control for 1:16 (often hanging his head super low for a controlling fighter) and 0:43 on the ground with Hendricks having a little more control and none of it taking place in an incredibly dominant position for either fighter.
I didn't count how many of Lawler's power shots came in the final 15 seconds (and have now deleted my recording) when Hendricks was on the ground seemingly riding out the round or gassed. That time was likely critical. Objectively, it doesn't make up for the rest of the round but may very well have played a significant role with Trowbridge and Marcos Rosales who each scored it for Lawler, while the majority of media members gave it to Hendricks.
With the exception of Trowbridge's round 2 brain fart, each fighter took their dominant rounds on the scorecards. At 51.7 percent for Hendricks, Round 4 was closer than I thought live but could definitely see more of the arguments for Lawler upon rewatching. Hendricks had a 56.4 percent chance of winning a typical judge's scorecard and a 59.5 percent chance of taking the overall fight.
So there you have it. Based on what actually went down inside the Octagon, Johny Hendricks was the odds-on favorite to get the decision nod in all three of this title shots. His best performance came with a busted bicep in his first encounter with Lawler and his most anomalous loss was the scrap with GSP. At 59.5 percent and thanks to a bad finish to round 4, his least controversial loss was the recent Lawler rematch.
Glenn Trowbridge put up a solid scorecard for GSP/Hendricks only to be a ginormous outlier for Hendricks/Lawler 2, albeit, in an insignificant way. Meanwhile Sal D'Amato was an outlier for GSP/Hendricks and then went consistent with the odds for Hendricks/Lawler 2. Judges, leaves, hands...Oh my.
What do you think? Were some of his losses legit or is Johny Hendricks the rightful UFC welterweight champion? Share your thoughts in the comments below.
RoboJudge will return in a few weeks with numbers from every UFC fight in 2014 and then every month after that. Happy holidays to all!
Paul is Bloody Elbow's analytics writer. All mistakes are his own and they've been known to happen sometimes. Follow him @MMAanalytics. Fight data provided by FightMetric.