Deciphering the Fightnomics Uber Tale of the Tape
During one of my first interviews with MMA media I was unexpectedly challenged on what variables I would suggest including on some sort of “new and improved Tale of the Tape.” Thanks to the FightMetric stat-keeping of all UFC fights, there is now a great deal of information on each fighter in surprising detail. Unfortunately, very little of it is ever harnessed to analyze upcoming matchups. So I set out to assemble something that would analyze this information on prior performance and package it in (nearly) as small a Tale of the Tape-type graphic.
The tool I have created to help me understand UFC matchups more succinctly, yet also more informatively, is what I call the “Uber Tale of the Tape.” While the traditional Tale of the Tape can offer some clues into the context of a certain matchup like height, reach (both literally assessed with measuring “tape”) or even age differentials, my Uber Tape was a satisfying middle ground of additional performance metrics boiled down from a much larger list of variables that I crunch prior to every UFC event. Über is a German prefix meaning a superlative example of something. We can attach this prefix to a word or phrase to mean a version that is better, the best, or above and beyond the norm. Those of you following @fightnomics on Twitter or who read my matchup analysis articles at MMA Oddsbreaker will recognize this graphic immediately, but you may still have questions about some of the variables that I show. So here’s an example that I will walk through line-by-line, one that had real-world implications.
Analysis revealed lots of performance advantages for Michael McDonald over Brad Pickett when they fought at the debut of the UFC on FOX Sports 1 in August of 2013. McDonald won by submission after dominating the standup striking and knocking Pickett down twice.
The “Bio” portion of the Uber Tape supplements some of the traditional information with the addition of the fighting stance. Since the Southpaw Advantage is real, it’s important to note in a matchup especially if an orthodox stance fighter is inexperienced. The bottom of the Bio section shows how many minutes of analysis went into these metrics. Data usually includes all UFC, WEC and Strikeforce fighter appearances. Although FightMetric quantified some additional fights from other promotions in their database, I exclude these due to less consistent matchmaking, lesser competition, occasionally differing rules, and usually the fact that they took place a long time ago. The Analyzed Minutes metric therefore demonstrates two important things. First, it shows how much experience the fighter has competing at a high level, and second, it shows how big the sample size is for the rest of the Uber Tape. The more minutes of data we have, the more confidently we can draw conclusions about the relative strengths and weaknesses of each fighter. Small sample sizes mean the data can be skewed heavily by a single fight, which may have been against a much tougher or easier opponent, causing the metrics to sway accordingly. Usually, a few recent fights are all we need to understand the tendencies and relative skills of a fighter.
In the Bio section, I only marked two variables to demonstrate advantages: age and reach. You’ll have to read my book to understand why height alone does not carry much of an advantage and how critical age is in MMA. So these two metrics have been scored with a check mark or “X” to indicate which fighter has an inherent advantage or disadvantage, respectively. We’re currently limited to just the biographical data that the UFC reports, although it’s certainly possible that new metrics like “leg reach” could come to market soon. My spreadsheets are anxiously awaiting testing hypotheses about that particular metric. But for now, this is it.
The next level on the Uber Tape is Standup Striking Offense, and these five metrics summarize the key performance variables that allow you to assess a fighter’s offensive striking abilities. The first value is a ratio of the total number of knockdowns scored to the number of knockdowns received, regardless of whether they occurred in the clinch or at a distance. Great strikers have a much higher first number, and a minimal second number. This is also a hint to identifying fighters who have suffered some knockdowns, and who are therefore at increased risk of going down early in the future. The next value is the Distance Knockdown Rate, which normalizes the punch-for-punch power for each fighter. The Knockdown Rate measures how many distance knockdowns were scored divided by how many landed power head strikes it took to cause them. It tracks who does more damage per strike. It’s very telling when one fighter shows up as well above average while the other is below, as is the case in this matchup example. That means if these two stand and bang, one is more likely to go down than the other.
Further down we assess the accuracy of each fighter in terms of distance head striking, which comprise the vast majority of strikes thrown from a distance. I’ve broken out jabs and power strikes separately because many fighters will perform differently depending on the strike’s strength. This allows us to differentiate fighters with a great jab but poor power striking from others who use a lazy jab only to set up their precise power hand. The overall averages for these metrics are 28% for jabs and 24% for power head strikes, but they do vary slightly by weight class.
Last in the category is the Standup Strike Ratio based on attempts, not strikes landed. Regardless of what type of strike is thrown at whatever target, I’m interested in knowing the total volume of standup strikes a fighter throws compared to his opponents on average. A number of 1.0 indicates a completely even ratio and represents a fighter who exactly matches the pace of his opponent. Ratios higher than 1.0 identify fighters who tend to outwork their opponents and likely have more effective cage control. Fighters with a ratio of less than 1.0 tend to get outworked by their opponents, and are on the receiving end of more strikes than they are able to throw themselves.
The Striking Defense metrics combine all distance head strikes to show what percentage is avoided by the fighter. A higher number means fewer strikes get through. Defense is calculated as one minus the accuracy of opponent head strikes. The same goes for the “Chin” calculation, which is one minus the Knockdown Rate of opponents for a certain fighter. Generally, you should see good fighters showing better accuracy and power than their opponents while fighters with poor accuracy and low defense probably won’t last long in the UFC.
The last striking metrics show Significant Strikes Attempted per Minute and the overall Significant Strike Accuracy. This metric is an aggregated metric for strikes that are potentially important in a fight, rather than just superficial. This metric is heavily quoted by UFC media, as it is a basic stat captured by the FightMetric scoring system. The attempts per minute metric is very informative in the context of what I call the “Pace Advantage” in MMA. While it’s true that great strikers will tend to have a good Significant Strike Accuracy, that number is heavily influenced by how much time the fighter spends in positions with higher success rates for strikes, or for fighters who mix in a lot of body and leg strikes, which have higher accuracy than head strikes alone. The average for the UFC is around seven attempts per minute with a success rate of 41%. Generally, I look at pace more than accuracy to find high-output fighters, but because accuracy is often quoted in UFC broadcasts I keep it on the Uber Tape for reference.
It was clear when looking at this matchup that McDonald (on the left) has some advantages in the standup game over Pickett (on the right). Most importantly, McDonald’s Knockdown Rate is over three times the Bantamweight average. Combined with Pickett’s weaker chin, McDonald went on to score two different knockdowns in the first round of this fight. The one metric that favored Pickett was still revealing in this instance. His high average rate of Significant Strikes Attempted per Minute indicated he is a fighter who likes to press the action. And despite being on the losing end of most exchanges, Pickett fought with admirable bravery and a frenetic pace, which helped earn both fighters the Fight of the Night Bonus. Remember the rule about the Pace Advantage: having high pace is good, but having low pace is terrible. In this matchup Pickett did have the higher pace, but McDonald was safely in the middle group in terms of his own pace, and so was not at too much risk.
The last category on the Uber Tape encompasses a lot of variables that are key components of the Ground Game. The first area is takedowns. The Takedown Attempt rate shows how often a fighter attempts takedowns, while the success rate indicates how often they achieve them. That’s fairly straightforward. In the UFC the average number of takedowns attempted by a fighter per five-minute round is 1.5, with an overall success rate of 40%. Looking at the Uber Tape can tell us which fighter is more likely to try to get the fight to the ground, and also who is more likely to succeed.
Below that, Takedown Defense is calculated as one minus the success rate for opponent takedowns, or the percentage that a fighter has successfully defended when facing takedowns. Experienced wrestlers don’t always have high offensive takedown success rates due to the frequency with which they make attempts and the fact that their opponents are expecting them. But the best wrestlers in MMA often show very high Takedown Success values. Because the average success rate for takedowns is 40%, that means 60% is the UFC benchmark for takedown defense. The other metric shows how many actual takedowns a fighter has faced inside the Octagon to give some level of fidelity to their takedown metric. Seeing a Defense metric of 100% is less meaningful if they’ve only faced two takedowns. That’s like emphasizing that a pitcher is undefeated after only one strikeout. Similarly, seasoned veterans may learn from experience and improve their defense over time.
Second from the bottom is a critical metric for understanding the grappling tendencies of a fighter: the Share of Ground Time in Control. Using extremely detailed position data from FightMetric, I rolled up aggregated data for every UFC fighter and calculated what percentage of time spent on the ground was in a position of dominance (top or back control). The UFC average must less than 50% (given that some ground time is neutral), so any fighter above that threshold has been more likely than not to be in control while on the ground.
That leads to the final metric, Submission Attempts per Trip to Ground. I know, I know, fighters can attempt submissions while standing too, but most either occur on the ground, or lead to a grounded position eventually. The alternative metric was submission attempts per minute of fight time, but the value would be skewed by how much time was spent in various positions. Many fights don’t see any submission attempts because the two fighters stand and trade for the entire three rounds. Generally, active submission artists will show up either way and so I’ve included the metric regardless of its imperfection. In this particular matchup, the metric still proved to be prophetic.
Despite Pickett’s higher pace of takedown attempts, McDonald was evenly matched with slightly better than average takedown defense, and good offense of his own. One differentiator was the ground control stat, which suggested Pickett was much more likely to end up in control and work from the top. On the flipside, it was McDonald who was more than twice as likely to attempt a submission if they ended up on the ground. After dominating the standup game McDonald did in fact end up on his back, but then used the position to sink in a triangle choke that ended the fight in the second round and earned him a Submission of the Night bonus.
The numbers are certainly an important part of all this, but usually I’ll save those until Fight Night or save them for the MMA Oddsbreaker Premium members. More commonly I’ll release a simpler version of the Uber tape for all to see in matchup reports earlier in Fight Week. In the case of McDonald versus Pickett, the story was still the same.
So that’s the Uber Tale of the Tape. In this case, my Uber Tape revealed overall anthropometric and performance advantages in the McDonald-Pickett matchup, enough to prompt me to write a short matchup analysis and prediction of McDonald winning inside the distance, all the way down to foreseeing a Fight of the Night bonus (cha-ching). That was all posted right here.
The betting lines at 5Dimes.com opened at -215 for McDonald, making him a clear, but not huge favorite in the matchup. By the time the fight began “the market” had pushed the moneyline odds to -320, boosting his implied win probability from 68% to 76%. The market had the right hunch in this case, but understanding the big advantages that McDonald had in the fight would have presented a lot of value for bettors when the moneyline was released, and in my opinion even on fight night. Perhaps the readers of this blog or of MMA Oddsbreaker had something to do with that line movement.
The “Tale” isn’t always so compelling a story. Often fighters are very evenly matched, or advantages in one area are offset by advantages in another. Such is the nature of a competitive environment. But analysis like this can help justify a pick on a favorite, or even better, identify an underdog or pick ‘em who matches up well against a better known opponent, as recently happened when Tyron Woodley faced Josh Koscheck at UFC 167. Despite the seemingly lopsided nature of the matchup covered here, Pickett was by no means out of the fight. Any fighter worth a roster spot in the UFC is a highly trained athlete capable of hurting his or her opponent in a flash. That’s why we watch – anything can happen on any given fight night. Just don’t go on assuming that every fight’s a “fair” one.