UFC Daily Fantasy: Can Striking Data Help Us Build Better FanDuel Lineups?
Fantasy MMA is truly one of one.
With most stick-and-ball sports, the name of the game in fantasy is opportunity. In football, how many times will this player get an opportunity to touch the ball? In basketball, how many minutes will they play to collect points, rebounds, or assists? In hockey, how much are they on the ice -- especially on the power play? In baseball, how many times will they get to bat?
Efficiency matters, but overwhelmingly, players who get opportunities will be relevant in those fantasy sports.
Then, there are other individual sports in which might be a better comparison. Golf comes to mind; you get at least two rounds of driving, iron play, chipping, and putting. Those are guaranteed should a player not withdraw. Also, the PGA Tour keeps detailed, advanced statistics on how every single player performs compared to average in each one of those four categories.
UFC is very different. In each fight, two fighters are locked into a 25-or-30-foot octagon. It can end as soon as five seconds, or it can go the entire scheduled duration -- up to 25 minutes in championship fights. Time and opportunity are no guarantee should a single mistake happen.
Unfortunately, UFC is in its infancy as a sport founded in 1993. Baseline statistical recordkeeping did not enter the fray until years later. To this date, in a world where we can quite literally measure how fast a player on TV is running on a football field, UFC officially keeps just eight recorded statistics. They can be found on any fighter's page on UFCStats.com -- such as UFC global star Conor McGregor's here.
When UFC launched in March of 2020 at the beginning of the COVID-19 pandemic, I was tasked to formulate a strategy for daily fantasy MMA. Like most who watch the sport, I had never really considered using statistics as a means to an end in a sport with the possibility of such violent, wild endings. I fell back on plenty of "old school" handicapping tricks such as a fighter's reach and their career wins by type as well as other film-based analyses.
However, previewing fights very early on, I noticed a trend amongst the fighters' profiles that I gathered who ended up performing well. And, after a full year of FanDuel action to study in 2021, there at least seems to be some correlation on how to quickly identify fighters worth targeting in daily fantasy.
Very early on, something became very apparent when evaluating several of the fighters I knew were solid fighters -- such as Dustin Poirier and Curtis Blaydes. They all shared a similar trait.
Some fighters don't average a lot of significant strikes landed per minute (SSLPM). You get points on FanDuel for significant strikes, so in a perfect world, you would obviously want to see that number higher.
But, firing offense, as you can observe in a fight, can also drain energy and open a fighter to a counter punch. Both of those things may eventually result in a knockout loss that's terrible for fantasy.
However, some fighters are still quality FanDuel options -- if not world-class ones -- without high striking volume. They may prefer to wrestle or grapple, and their significant strikes absorbed per minute (SSAPM) is very low. The two most notable examples that came to mind were former undefeated champion Khabib Nurmagomedov (1.75 SSAPM) and his protégé Islam Makhachev (0.79 SSAPM).
Therefore, I coined a metric that isn't unique, but it's often not referenced or titled as notably as it should be. Some may refer to it as "significant striking differential per minute" or "significant strikes landed less absorbed per minute", but I refer to this mark as striking success rate (SSR):
Striking Success Rate (SSR) = Significant Strikes Landed Per Minute (SSLPM) -
Significant Strikes Absorbed Per Minute (SSAPM)
I collected pre-fight data from virtually every one of the 43 official UFC events in 2021. Each of the cards had a daily fantasy main slate on FanDuel.
In total, 936 fighters were included in this sample, and notably, 887 had pre-fight data of any kind, leaving roughly 5% of the fighters in the sample without any UFC data entering their fight.
The pre-fight data taken included the fighter's FanDuel salary, pre-fight significant strikes landed per minute, significant strikes absorbed per minute, striking accuracy percentage, striking defense percentage, takedowns per 15 minutes, takedown accuracy, takedown defense, and submission attempts per 15 minutes.
The top-shelf, ranked fighters in UFC sparked this hypothesis. Therefore, evaluating the champions of each division made sense.
One of the issues with UFC-specific data is that the samples can be minuscule. At the championship level, though, we have a good understanding of fighters analytically. The fewest number of career fights for a UFC champion at the end of 2021 was nine.
As demonstrated below, many champions have a striking success rate that is above the 70th percentile of fighters that made a 2021 appearance in my sample (1.21). Every single champion that has defended their current title at least twice is in the 70th percentile or better.
Two of the bottom-four divisions included are heavyweight and light heavyweight. Obviously, the additional punching power at the higher weight classes can erase plenty of per-minute efficiency gaps, but it's worth noting that both of those champions just did not have an opportunity to defend their belt yet:
As evidenced by just one UFC champion with a slightly negative striking success rate, it appears -- at least on the surface -- that a positive striking success rate is a key ingredient of becoming a UFC champion.
Even the lone exception, Glover Teixeira, has a +0.07 striking success rate on the five-fight run in which he's secured his world title. It just seems improbable to make it to the apex of the sport while taking more damage than what's being dished.
Winner, Winner Chicken Dinner
You probably didn't need me to tell you fighters that strike their opponents more than they're struck per minute are good. That's the whole point of MMA.
However, the power of calculating and knowing this metric can be incredibly valuable when hunting for winners and high-upside plays in daily fantasy.
Although this isn't a betting piece, finding an outright winner is essentially mandatory in every lineup to win a tournament. Of the 228 fighters who made a perfect FanDuel lineup in our sample, 209 (91.7%) won their fight.
Worth noting as well: 144 (65.8%) of the 219 fighters that made a perfect lineup in 2021 with pre-fight data available had a positive striking success rate overall.
Striking success rate can -- at times -- guide us in the right direction. While it's not a guarantee, the results from just one year of analysis are pretty jarring.
Of the 887 fighters in the sample, 451 fighters had a higher striking success rate than their opponent. 251 of those fighters (55.7%) won their fight outright.
To my amazement, no two fighters had an identical striking success rate entering their fight. Therefore, of the remaining 436 fighters with a lower striking success rate than their opponent, only 191 of them (43.8%) won their fight outright.
Data Quality: The More (and the More Experienced), The Better
Personally, the only reason I feel that winning percentage isn't higher is that each striking success rate data point is weighted equally -- when really it shouldn't be.
Take Australia's Jakob Malkoun as an example. Malkoun lost his UFC debut in November to Phil Hawes via an 18-second knockout. He landed no significant strikes compared to seven for Hawes, and therefore, entering his second fight, he had a striking success rate of -23.33. He dominated his next fight.
That data point counted the exact same in this sample as either Dustin Poirier or Charles Oliveira in their December title bout that featured a combined 54 UFC appearances.
Therefore, I found a quick way to test my theory with a better sample -- UFC main events.
In order to be selected for a UFC main event, there is some pedigree required. Most of the fighters are ranked, and many are around their division's respective title picture. Therefore, 80 fighters were represented with good samples (certainly ones better than Malkoun's), and the results were even more drastic.
Of the 40 main event fighters that had a higher striking success rate than their opponent, 24 of them (61.5%) won their fight. For those so inclined, 17 of those 24 winners made the perfect lineup on FanDuel.
Given that 17 of 40 (42.5%) main event participants with a higher striking success than their opponent made the perfect lineup, there is an eye-popping trend to potentially identify one-sixth of the missing pieces in daily fantasy quickly. The extra two rounds in every UFC main event can likely explain that.
The logical takeaway overall here is that, as the sample gets better with higher-ranked, more experienced fighters, the more valuable that striking success rate becomes to finding the winner.
Winning Isn't Everything
UFC fighters don't care if they win ugly, but UFC daily fantasy players do.
A win on FanDuel with minimal volume doesn't mean very much.
At UFC Fight Night: Reyes vs. Prochazka, Luana Carolina scored just 35 FanDuel points in a win -- the lowest total in my sample. Conversely, at UFC 267: Blachowicz vs. Teixeira, Cory Sandhagen posted 107.4 FanDuel points in his title-fight loss to Petr Yan.
Therefore, I wanted to study if this general trend of winners could actually be categorized into FanDuel production. It was a messy road to get there, though. I needed to remove a series of outliers that were making the data incredibly noisy:
Outlier #1: Striking Success Rates in Small Samples
The Malkouns of the world -- and there were plenty of them from the one-step-above-amateur appearances on Dana White Contender Series -- made it incredibly difficult to find a correlated relationship.
First, I determined it was best to use a differential to measure production. There are still FanDuel points to be had in a fight between two negative striking success rates. A -1.00 striking success rate isn't very good, but it's a lot better than one that's -5.00. The advantage one fighter had in striking success rate over the other was the true value.
Then, I incorporated a maximum and minimum. Striking success rates flatten as fighters' data gets better. Heavyweight Chris Daukaus had the highest striking success rate of any fighter with at least three UFC appearances in my model (+8.63 entering his September fight). Even so, Daukaus' five-fight career sample still sits at just 16 minutes and 43 seconds to this day. He's barely over the equivalent of one full-length UFC fight
Outlier #2: Wild, Quick Finishes on FanDuel
However, I also couldn't just look at the striking success rate variance. Wild, splash finishes like Terrance McKinney's in June created an absurd FanDuel-per-minute output that wasn't reflective of long-term volume. McKinney's flash knockout in seven seconds earned him a rate of 877.7 FanDuel points per minute. He probably can't sustain that over 15 minutes.
Therefore, I limited McKinney and all the other wild finishers into a maximum of 15.00 FanDuel points per minute. It's not perfect, but the average duration of those bouts was just 3 minutes and 40 seconds. The maximum per-minute rate in a win that went to a decision was Max Holloway's January performance (in which he set the UFC record for significant strikes) where he recorded 11.83 FanDuel points per minute.
His 287 FanDuel points that afternoon were the top score of the year, and that's the 100th percentile outcome in a single UFC fight.
Outlier #3: Wrestling and Submission Aces
Of the 887 fighters, 701 fell into a "normal" category for a UFC fighter. Most fighters have a striking base, and therefore, they predominately score their per-minute base through striking. However, certain jiu-jitsu practitioners and wrestlers can drag the FanDuel scoring output down from the norm.
They're easy to spot analytically because of their abnormal volume in either takedowns or submission attempts per 15 minutes. In my sample, I have them classified separately, but you could also combine this group based on two key factors:
a) Takedown or Submission Attempt Volume
b) Lack of Striking Volume
Below is how I classified each in the sample:
|Grappler||Less than 3.50
(40th Percentile or Lower)
|Less than 2.50
(80th Percentile or Lower)
|More than 0.50
(65th Percentile or Higher)
|Wrestler||Less than 4.00
(55th Percentile or Lower)
|More than 2.00
(75th Percentile or Higher)
|Less than 1.00
(80th Percentile or Lower)
Admittedly, this was the hardest step. Finding defined guardrails was challenging because these fighters vary in volume just as strikers do. With the wrong infrastructure, I could have made Derrick Lewis a "Grappler" or made Petr Yan a "Wrestler." Diehard UFC fans likely had a good chuckle at both of those notions.
These fighters are different to handicap when selecting ones for FanDuel. Of the 186 that fell into these categories, 171 averaged fewer than 4.00 FanDuel points per minute, and many of the outliers came with small samples. If you solely judge by per-minute volume, you'll likely pass over the potential for a dominant win.
Discovering a Relationship
The remaining 701 fighters are classified as either a "Striker" or "Hybrid" fighter in this data set. Therefore, we can apply this relationship to potential fights that do not have participants meeting the "Grappler" or "Wrestler" criteria.
With that the case, there is something tangible to grasp. Below is a chart displaying the relationship between striking success rate differential and average FanDuel points per minute scored, based on the criteria defined above:
Now, I would be the first to admit -- especially given the small, one-year sample of data to work with -- this is far from gospel. But, in the primitive stage of UFC statistic gathering without any opponent adjustments or much granularity, it's fair to wonder if these stats have any input on predicting how UFC fighters will perform in the octagon.
However, with the tiniest bit of conviction, this appears to state the following: fighters, so long as they are not outliers in terms of wrestling or grappling, with a higher striking success rate than their opponent generally perform better on FanDuel on a per-minute basis.
The trend in the data shows that as well. Exactly 70 of the top-100 FanDuel scorers in 2021 had a positive striking success rate. It's not everything, and there are exceptions, but it's something to consider when building FanDuel lineups -- especially in fights with shorter odds to go the entire distance. Those fights are projected to be even more heavily reliant on per-minute efficiency and volume.
In conclusion, I wanted to provide some key takeaways to keep in mind when building UFC lineups on FanDuel:
-- "Striking success rate" is a quick calculation of a fighter's damage landed minus absorbed per minute from any position inside the UFC octagon.
-- Each UFC champion with multiple title defenses has a striking success rate in the 70th percentile or better amongst fighters with a 2021 appearance.
-- 55.7% of fighters with a higher pre-fight striking success rate won their fight in 2021. That number jumped to 61.5% in main events with better fighters and pre-fight samples.
-- 209 of the 228 fighters who made a perfect lineup on FanDuel in 2021 won their fight.
-- 63.2% of fighters that made a perfect lineup on FanDuel in 2021 had a positive striking success rate. Though, only 51.2% of fighters had a higher striking success rate than their opponent.
-- Because of the different styles of UFC fighters, it is important to identify when situations (like fighters who spend an abnormal amount of time wrestling or grappling) where striking success rate is less indicative.
-- When a striking environment can be projected, a fighter with a higher striking success rate generally posts more FanDuel points per minute.
-- 70 of the top-100 FanDuel performances in 2021 came from fighters with a positive striking success rate.