The Odds Are Good, But the Goods Are Odd Han Solo hated being told the odds. But that was a long time ago…. Today’s sports fans are constantly bombarded with data and information, even in a simple and straightforward sport like MMA. As any sport grows, the metrics that measure it and the statistics that report it all evolve and advance. But there’s one set of numbers that are omnipresent from the inception of almost any game, from the back alley to the big leagues: the betting odds. In MMA, the Tale of the Tape summarizes the basic physique of each fighter, while their records summarize their performance history within the sport. But it’s the betting line that is the most direct and immediate hint to what’s about to happen when the cage door shuts on two fighters. So let’s take a closer look at what the odds can tell us about MMA, matchmaking, and upsets. Hey Han Solo, “earmuffs.” Putting the Extreme into Extreme Sports In an academic sense, betting lines are basically the market price for a certain event or outcome. These prices can move according to betting activity leading up to the event. And when a UFC fight begins, that betting line is the public’s final guess at the probability of each fighter winning, with roughly half of bettors picking each side of the line. Many experts make bold and confident predictions about fights, and they’re all wrong a good portion of the time. But what about the odds? How do we tell if they’re right? And what can we learn from looking at them in aggregate? The fact is that only a small portion of fights are truly evenly matched according to odds makers. So called “Pick ‘Em” fights made up only 12% of all matchups in the UFC since 2007, with the rest of fights having a clear favorite and “underdog.” UFC President Dana White mentions these betting lines to help build the story around matchups, often to point out why a particular fighter might be a “live dog.” White’s correct to play up that possibility, because upsets occur in roughly 30% of all fights where there’s a clear favorite and underdog. So next time you look at a fight card expecting no surprises, just remember that on average there will be two or three upsets on any given night. What Do Odds Makers Know? In a macro sense, cage fighting is inherently difficult to predict for a variety of reasons. The young sport is competed by individuals, and there are no teammates in the cage to pick up slack or help cover for mistakes. Individual competitors only fight mere minutes per outing, and, if they’re lucky, only a few times per year. And let’s not forget the raw and primal forces at work in the cage, where a single strike or mistake of position can end the fight in seconds. The volatility of these factors means there is absolutely no such thing as a guaranteed win when you’re allowing one trained competitor unmitigated access to do violence on another. The sport is completely dynamic, often intense, and with only a few round breaks to reset the action. These are also the reasons we watch and love the sport: it’s fast, furious, and anything can happen. It’s the polar opposite of the true statistician’s sport, baseball. Given how difficult MMA might be to predict and quantify compared to other mainstream team sports, it could be very revealing to look at how accurate odds makers have been over the last few years of UFC events. For this analysis I’ve translated the odds from the standard American odds format. For example, -200 indicates a two-to-one favorite, and +300 a three-to-one underdog. We can convert these somewhat confusing numbers to a percentage that is the implied win probability based on the odds at fight time (or 67% and 25% for these two examples, respectively). If 10 fighters were each listed as slight underdogs with a 40% implied win probability, we should expect 4 of them to win, and 6 to lose, if the odds were indeed accurately set. So are they? 
It turns out that odds makers have been pretty good at assessing the likelihood of fight outcomes. Their odds-based predictions closely match the actual outcomes of the fights. The exception to this occurs at the extreme end of the spectrum where very heavy underdogs have been overperforming by winning more fights than expected over heavy favorites – but the sample size for this group is the smallest one of all. Don’t go betting your life on long shots just yet, but never count anyone out either. We also see a slight drop in the middle of the betting spread. Slight favorites could be losing at a higher than predicted rate due to some basic bias of the betting public. When two guys face off that appear to be very evenly matched, most pick the more famous name. This is one possible explanation for favorites underperforming against the implied win probability of the betting lines. Patterns in the Chaos The betting line allows us to view fights through the lens of how evenly matched the fighters are. In theory, if betting lines are accurate (and they seem to be), then more extreme spreads mean the fighters are mismatched against each other. Conversely, when two fighters are very evenly matched, it’s rare to see one then dominate the other. What does the evenness of the matchup mean for how fights play out? 
One clear pattern that emerges is when there is a significant mismatch between fighters. When fighters are mismatched, finish rates go up. This makes sense, because either the favorite truly is much better than their opponent, and hence more likely to finish them, or the underdog’s only hope for victory is a miraculous move that finishes the fight. At face value, there’s no particular betting line where outcomes are drastically different from the expectation set by the odds. But we also know that generally speaking, almost a third of fights will not end as the masses predict. Heavier fighters tend to have slightly higher upset rates, perhaps because the “puncher’s chance” means more when larger fighters land that one fight-ending strike. There are also almost twice as many upsets in co-main events than in main events. Either way, there are no guaranteed outcomes in MMA, with potential for a spectacular finish to every fight. That’s what makes for an exciting sport. And if you don’t believe me, here are 10 reminders never to count anyone out of a fight. Top 10 Biggest Upsets in the UFC 
What Did We Learn?
- Betting lines are generally an accurate predictor of who will win a fight
- Still, 30% of the fights with a clear favorite and underdog result in an upset
- Only 12% of UFC fights are an even “Pick ‘Em”
- The more extreme the betting line, the more likely there will be a finish
- Co-Main events see the highest rate of upsets at 40%, Main Events the lowest at 23%
- Anything can happen in a fight, and that’s why we watch
– The Fight Scientist Like Fightnomics on Facebook, or follow on Twitter @Fightnomics to hear when new research and blog posts are available. Raw data for the analysis was provided by, and in partnership with FightMetric, and all analysis performed by Mr. Kuhn. Fightnomics, FightMetric and Sherdog assume no responsibility for bets placed on fights, financial or otherwise.