Small Cage Analysis Part 2: the What and the How of the Finisher’s Cage
Analysis of cage size effects was presented in the “Fightnomics” book, this analysis is more recent and focuses on UFC competition only.
In Part 1 of the analysis of Cage Size in the UFC we found that finish rates in the smaller Octagon were substantially higher than those in the normal, full-size Octagon. It’s not a perfect experiment given the sample sizes involved, but they were found to be statistically significant beyond a reasonable doubt. So the phenomenon of the “Finisher’s Cage” is, therefore, real. We just don’t how, or why.
There are a lot of potential drivers for finish rates. More action could lead to more finishes, or perhaps matchmaking or fighter skill level is fueling this result. In order to test these ideas, we first need to explore what’s different about a fight in a small cage versus a fight in a larger one.
Given prior analysis that has essentially shown that finish rates in the lower weight classes (men’s Flyweight through Welterweight) are basically the same (46%), performance metrics from these divisions were isolated for further analysis. Separating fighter metrics by cage size for this pool of fights revealed some changes in how fighters performed, but not all changes were equal. Some were small and some were big, and then there were some metrics that increased and others that decreased.
Here’s how it all shakes out in one view.
For detailed explanations and historical UFC benchmarks for these and other many other MMA statistics, check out the book: “Fightnomics,” available at Amazon.
There’s a lot going on here, and not all of it is obvious. From a fighting position standpoint, the confines of the smaller cage mean that fighters are standing and trading from a distance for a little less time. That means more time in the clinch (slightly, 4% more) and more time on the ground (a 12% increase) than they spend in the large cage. This makes sense because hitting the fence of the cage makes takedowns a little easier to land, and there’s less open ground to cover before running into the fence. So if fighters have less time to trade leather from a distance overall, how is it that the most pronounced changes in fighting metrics are dramatically increased knockdowns?
To generalize the requirements for a knockdown to occur, a fighter must attempt a strike, target their opponent (usually their head), land the strike, and then have it be powerful enough to cause a knockdown. But the metrics for overall Standup Striking Rate (total attempts), Share of Standing Head Strikes (target selection most likely to cause a knockdown), and Distance Power Head Accuracy (landing success rate) all see either negative or miniscule changes when competing in the smaller cage. Absolute power is not something that can be measured quantitatively (yet!). Nevertheless, despite no change to the metrics just listed, fighters are still getting dropped at a much higher rate in the small cage. And those knockdowns are driving the increase in stoppages.
Here’s how just the Distance Knockdown Rate looks for each division in the small and large cage.
One hidden explanation could be that the Standup Head Power Ratio of the fighters changes in the smaller cage. For some reason, fighters in the smaller cage used more power head strikes compared to jabs than when they were in the larger cage. But given the drop in volume these factors offset, and ultimately the number of power head strikes that land per minute of fight time in this position was about equal across the two types of cages. Volume and pace are not the answer then.
The jump in Knockdown Rates could only come from a few remaining sources. The first is age. Readers of the “Fightnomics” book already know that age is a good proxy variable for experience in MMA, and there’s a strong relationship between getting older and becoming more knockout prone. In this case, the average fighter age in the large cage was 29.8 years old, while it was 29.7 years in the small cage. Age was a negligible difference, but if anything the larger cage had slightly older fighters meaning if this was a factor it should have led to more finishes in the larger cage. So that’s not it either.
One more thing we can measure is the competitiveness of the fight. Historically, betting odds have been quite accurate in assessing each fighter’s probability of victory, and mismatches do lead to more finishes (see chapter 9 of the” Fightnomics” book). So if more mismatches were occurring in the small cage due to less experienced fighters or less visibility into their relative skill (or even intentional mismatches for the purpose of supplying exciting finishes to a new market), the wisdom of crowds would account for this in the betting odds. As it turns out, we don’t see evidence for that when it comes to fights taking place in different cage sizes.
Again, if anything the fights taking place in the larger cage saw more matchups with Long Odds, suggesting more extreme mismatches in those events. But overall, the average odds spread was nearly identical in both cage sizes. So this is not likely to be the cause of the increased finish rate.
Another hypothesis is that Fight Night cards have fighters with less skill or experience, who are therefore more prone to being finished. It’s a reasonable idea, as UFC matchmaker Joe Silva once pointed out to me: anyone who wants to be a fighter will focus on offense first, and defense last. So, anyone can throw a punch, but it takes some practice to avoid one. Unfortunately, we don’t have a perfect variable for measuring the skill of a fighter, although we can group fighters by their placement on UFC fight cards. Fighters deep on the undercard are either new to the UFC, or have lost fights and are therefore less skilled. Fighters higher up on the main card are either more skilled, more experienced, or possibly both. So does this measure affect finishes?
It turns out that undercard fighters see the smallest effect in the Finisher’s Cage, while moving up the fight card sees an increasingly pronounced effect. The main and co-main events where fighter skill and experience is presumably highest saw the largest effect, albeit with the smallest sample size. While I wouldn’t close the door on this theory, the evidence so far doesn’t support it.
The last factors to address are currently less quantifiable, and definitely open to interpretation. As we have run out of variables in our data to drive the higher knockout trend in the small cage, the last causes are simply that fighters in confined spaces are more aggressive, and/or that hurt fighters have nowhere to run. There have been some famously boring fights where a fighter refuses to engage his more aggressive opponent, leading to endless circling around the Octagon. Kalib Starnes ran away from Nate Quarry at UFC 83 to such an egregious extent that Starnes was mocked by Quarry in real-time towards the end of the fight. One judge scored the fight 30-24, giving Starnes the second lowest point total in UFC history.
Research into animal behavior in the confines of a cage or pen has shown that there is increased aggression between unknown members of the same species placed in smaller spaces. Close quarters aggression among natural competitors is a simple, yet primal behavioral force. And as technically proficient as humans have become, we are still slaves to our evolutionary past in the general sense. Placing two high testosterone males into a cage for the purpose a prize fight already has strong evolutionary forces harnessing hormonal drivers. The larger space of the full-size cage could certainly mitigate some of the natural urge to fight that these athletes are already experiencing. The small cage simply exacerbates that urge, and could be driving more aggressive striking and submission attempts. These fighters are certainly throwing more power strikes, so perhaps there’s also more power in those strikes, or at least more aggressive risk-taking in the striking selection. Perhaps, and we can’t rule this out, but we also don’t have the data to test this. Again, if only we had power sensors in the gloves we could test this idea – but we don’t, not yet.
The Finisher’s Cage
Lastly, there’s the tactical effect. Recently, some fighters have lamented the smaller cage claiming they have less room to set up their kicking attacks. However, the target selection metrics aren’t much different between cages, and body strikes (mostly kicks) are nearly equal in both sizes. Other fighters like Tim Kennedy have praised the smaller cage, suggesting they are better able to cut off their opponents in retreat. This comment hints at a key aspect of the small cage: there is less room to run when you’re hurt.
If strikes are landed at the exact same rate in the small cage but cause more knockdowns, this explanation is consistent with what we’re seeing. It’s not about a boost to offense; it’s about a decrease in defensive capability. We’re not seeing more volume or sloppier engagement, and were not seeing mismatches or older fighters. We’re just seeing what happens when two fighters with the same offensive weapons are put into an environment where one of their defensive outlets is reduced.
And it’s not just more knockouts in the small cage, as we also see an increase in submission finishes. Remember, we corrected for weight class already, so the same fighters are tapping out more frequently in the small cage. The performance metrics analysis reveals that fighters are spending more time past the position of full guard, and are also attempting submissions at a higher rate. So the increase in finish rate isn’t just due to the benefit given to aggressive strikers, fighters are also being more aggressive on the mat.
Part of this submission boost may dovetail the increase in knockdowns. When a fighter gets dropped, some opponents will go for a choke to finish the fight rather than finish with ground and pound. But given the increase in time on the mat and in the share of that time spent in advanced positions, the boost to submission attempts and finishes appears to be also reflect greater aggression by the fighters in small cages even after they hit the ground. It’s not a question of grappling expertise, as the submission success rate is not higher in the small cage (due to lower skilled fighters for example), but rather the offensive ground fighters are just being more aggressive in going for finishes. Specific grappling metrics show only slight differences, but the key metric of submission success rate is not one of the ones getting a boost. It’s aggression, not finesse.
This is all relevant to any upcoming cards using the small, aggression-boosting “Finisher’s Cage.” Unfortunately, it seems there’s some lingering unawareness of the effect, and cage size is rarely announced prior to events. But one of the takeaways from this analysis as that lower profile UFC cards in tiny venues may still provide plenty of highlight reel potential if they utilize the small cage.
Whether or not fight promotions use these findings to change or standardize the size of their cages remains to be seen, but the issue is still important to fighters so they can properly train for the right sized playing field. Many gyms may be unintentionally influencing the fighting style of their trainees by using a small practice cage. Fighters accustomed to the small confines of a cage that allows them to cut off opponents will be disadvantaged when competing on the larger stage of the UFC’s full-size Octagon. There are, of course, other variables affected by the cage size, like sponsor logos and camera angles. But each of these can be handled and overcome, in some cases with improved technology like green screen mats for digital sponsor logos.
So there it is. Since implementing parallel cage sizes in the UFC we have seen an underlying factor affect how fights play out. We now know how big the effect is, and even some of the mechanics in how it influences fighter performance. We have a reasonable working theory as to why this is occurring, but as with any worthwhile experiment, more data would help. When the sport of MMA catches up with more established and mainstream sports like basketball, football and hockey, we should see the advent of fighter position data-capture and even real-time smart-glove enabled power measurements. I’ve always believed that the sport has great potential, and that goes for how we view and understand the sport as well.
Raw data was provided by, and in partnership with FightMetric, with additional research by Eli Langbaum, Columbia University.