MMA FIght Strategy : Developing Tactics Using UFC Round By Round Data

If you’re going to work on a successful fight strategy then a great place to start is with an understanding of what actually happens in terms of risk during the progress of a UFC fight. If you know the numbers then you know the chances of potential risks and rewards throughout a fight and consequently you start to understand when the best time is to focus on attacking, when to go after submissions, when to be more defensive and conservative and more. If you have a tactical game based on the evidence of the likely outcomes of rounds and fights then you have a distinct tactical advantage.

Play Video

Round By Round

This first set of numbers gives us an overall context to all the subsequent analysis, this helps us understand how the volume of finishes change throughout a fight. The chart is broken down round by round and shows us the total volume of submissions (blue) and the total volume of KO/TKOs (red) per round. This data is taken from the entire 25+ years of UFC fights – over 5,000 fights.

As you would expect the volume naturally drops off through each round, the volume in round two drops by around 40% from round one. Of course as a percentage of the fights get finished in the first round that same amount of fights don’t ever get to round two, so while the volume graph is interesting it doesn’t tell us that much in terms of the amount of risk that each round represents.

To help us to visualise the risk each round represents more clearly we need to look at the percentage of fights that get finished in each round if the fight makes it to that round. This means that if twenty fights started in round 1 and ten fights got finished then round 1 would have a 50% finishing rate. If six of the remaining ten fights got finished in round 2 then that round would have a finishing rate of 60%. So with this model it doesn’t necessarily follow that the rate should always decrease as we go through the rounds.

The following graph shows three bars per round, the first shows the overall percentage chance of getting finished in that round, the second shows the chance of you getting KO/TKO’d in that round and the third shows the chance of you getting submitted in that round.

By Round 3 The Chance Of A Finish Is Cut In Half

You can see that the chances of you getting finished in any way does decrease as you go through the rounds but not at the same rate. The overall chance of you getting finished in round 2 is around 20% less than that of round 1, but by the time you get to round 3 the risk reduces by 50% – so you’re twice as likely to get finished in round 1 as compared to round 3.

Looking at the KO/TKO bar you can see that drop off is a bit more uniform dropping by around 25% each round so that again it’s around half by the time it gets to round 3. It’s the submission rate that presents the anomaly here, instead of a steady drop off through the rounds there is only a 10% drop off between round 1 and 2 but there is still a drop off to around 50% by round 3. So in terms of submissions, both rounds 1 and 2 are almost equally dangerous but round 3 is far safer. With striking there is a more steady decrease in danger as the fight goes through the round.

Finally for this section we present a slightly different view of the same data, here each round is represented by a full bar which is broken down into the three main possible outcomes – survival, TKO/KO or submission.

Play Video

The Beginning, The Middle Or The End?

All of these initial pieces of data give us a pretty good overview of what happens through the rounds but now it’s time for a bit more focus. So now we’re just going to focus in on the first three rounds as the idea of this data is really to bring you information that might help you to adjust and strategise your own Mixed Martial Arts game. Given that most of us aren’t about to main event or fight for a UFC Championship then the fights we have are likely to be over 3 rounds. Also, given that 5 round fights represent only a small percentage of UFC fights the amount of data that we have on them isn’t anywhere near as substantial and so it is subject to a far greater margin of error. However, just seeing the three rounds as the data points doesn’t give us quite the level of insight we might need and so we’re going to break each round down even further.

We’ve split this data up in many different ways as we’re analysed it and we’ve found that one of the most informative views occurs when we divide the round up into thirds:

  • The beginning of the round 0:00 – 1:40
  • The middle of the round 1:41 – 3:20
  • The end of the round 3:21 – 5:00

This split gives us the easiest way of starting to develop a strategy for fights as we can divide up our approach into what we might do at the start, the middle and the end of each round. So let’s take a look at the overall finishes for each of the rounds from 1 to 3 when broken down into thirds, it looks like this:

The Round 2 Anomaly

So all three rounds have a similar pattern, they likelihood of a finish generally increases as the round progresses. Even this may be contrary to what you may have thought, the line of thinking that’s most often expressed is that the danger is at the start of the round and then as the competitors get exhausted through the round the threat decreases. As we can see this is not strictly the case.

The anomaly here is round 2. In round 2 there is an increased threat in the middle of the round that we don’t see in the other rounds. With only this level of data it’s hard to make an assessment on why this might happen so we need some additional data. Next we split the same data into two further graphs, one for KO/TKOs and one for Submissions. Let’s take a look at the KO/TKO data first:

The Impact Of Making It 'To The Bell'

We can see here that there’s actually fairly little movement in the chances of a knockout happening between the different phases of the rounds. They’re surprisingly flat and consistent with around a 7% chance during each phase of round 1, around a 5% chance during each phase of round 2 and around a 3% chance during the phases of round 3. This is surprising because start of each round subsequent round does not simply pick up where the previous round left off as you might expect, making it to the bell and taking the break seems to have a significant impact of the chances of finishes when the next round begins. For example, there’s an almost 8% chance of you getting knocked out at the end of round one but if you make it to the break then the chances of you being knocked out at the start of round 2 drop to around 4.5% – you almost half your chance of getting knocked out just by making it to the end of the round.

But none of this accounts for the round 2 anomaly that we saw in the overall finishes so let’s take a look at how the submissions break down…

The Round 2 Submission Sweet Spot

As you can see the pattern looks markedly different from the KO/TKO data. For starters the first third of the every round is significantly less dangerous than the middle and end of the round in terms of submission threat. It’s reduced by around 50% in every round. This is understandable considering that acquiring a submission often takes far more time to set up than a KO and usually requires a prior wrestling exchange to get the opponent to the mat.

In round 3 you see a steady increase in the submission success rate as the round progresses. Round 1 shows the same increasing trend through the round but the middle of the round performs very strongly and almost matches the end of the round. The round 2 anomaly is clear though, we see that the middle of the round outperforms the beginning of the round by a significant amount AND it also outperforms the end of the round by a decent amount.

So looking at this graph it becomes clear that the anomaly that we see in the middle of round 2 with overall finishes is due to it being a high performing zone for submissions not KOs. We can broadly summarise that the middle of round 2 is the sweet spot for submissions, it would seem like this is one of your best opportunities to get a submission finish. Ultimately this data is only really useful if it can help to identify submission or KO opportunities that will help you to fight more strategically and tactically. If we start to analyse the data with this objective then it can be helpful to view submissions and KOs as different and somewhat opposing objectives. As in, we imagine that a fighter might broadly be seeking a KO and attempting to avoid a submission or seeking a submission and attempting to avoid the KO. So what we are looking for here is not only high volume in terms of our objective, but also low volumes in terms of the counter objective. In other words, if we’re looking for a KO then we’re looking for a time when the chances of a KO are high and the chances of a submission are low. If we are looking for a submission then we are looking for a time when the chances of a submission are high and the chances of a KO are low.

Identifying The Opportunities

To help us identify these sweet spots we are going to have to make calculations that take into account both the volume of the desired objective and the gap between the potentially opposing objective. To help us, let’s present the KO and submission data side-by-side, remember we’re looking for areas where there is high volume of the desired objective but also comparatively low volume of the counter objective.

For example, we can see that even though the beginning of round 1 isn’t the highest volume for KOs but given that the submissions are very low at this time it represents a sweet spot for KOs – a high chance of a knockout while there is a relatively low chance of a submission. To help us visualise this more clearly we’re going to overlay an additional layer that calculates both the volume of the objective and the gap. The base calculation is as follows:

For Knockouts

KO/TKO chance * gap between submission chance

For Submissions

Submission chance / gap between KO/TKO chance

The calculations for KOs and submissions are different because knockouts are always at a higher volume than submissions, this means that for knockouts we’re looking for the largest gap between knockouts and submissions – the further the knockout gets ahead of the submission the better. But submissions never get ahead of knockouts so instead of the largest gap we’re looking for the smallest gap – the closer the submission level gets to the knockouts the better.

The Submission Multiplier

Doing this kind of calculation on the submission side gives us a bit of an issue as there is never a point throughout the rounds that submissions are at a greater volume than knockouts so there is never a point where submissions are the strictly speaking the best bet. 
This would mean that our overlay would never show any points where submissions present the best opportunity. This doesn’t help us much if we have a submission finish as our objective, if we go into a fight knowing that our best chance for victory is a submission then we still want to know where the sweet spots are even if knockouts still hold the advantage.

To help with this we are going to apply a multiplier to the submissions to help raise them up and expose the best opportunities for them. The multiplier is going to be the exact ratio that represents the difference between knockouts and submissions (currently at around 1.6). This has the effect of raising the volume of submissions up to a point where they are equal with the volume of knockouts. So with this amendment our calculation is as follows:

 

For Knockouts

KO/TKO chance * gap between submission chance

For Submissions

(Submission chance / gap between KO/TKO chance) * overall submission:KO/TKO ratio

So with this calculation applied to the additional overlay the graph looks like this

This is one of the most informative graphs that we have ever produced. This graph can produce the building blocks of a strategy where you are playing a role of striker vs grappler or grappler vs striker. It’s unlikely in modern MMA that you consider yourself to be a straight grappler or striker but you could easily find yourself in a position where you know that you or your opponent have a particular advantage, for example that you are up against someone who excels at striking but has limited skills on the ground.

Grappler vs Striker / Striker vs Grappler

So let’s take a look at the key areas of opportunity for either grapplers or strikers, here’s the list of times when you’re most likely to succeed in either achieving a knockout or a submission

Striker vs Grappler

Grappler vs Striker

The Rhythm Of The Round

As we’ve analysed this data we’ve broken it down in many different ways to find the views and visualisations that are most informative. Outside of the beginning/middle/end round breakdown the major other breakdown that reveals the most about the data is the most granular one, breaking down the round into ten 30 second chunks. With this view we really get to see the ebb and flow and rhythm of the round. Let’s first take a look at the overall finishes when we break the first 3 rounds down into 30 second pieces

The Double Push

Here we see an incredibly interesting flow through each of the rounds, there’s an increasing threat through the start and first half of the round and then there’s a drop off at around the 3 minute mark until the final minute when the threat significantly increases again. This would seem to suggest that the flow is due to the natural limits of physical output, you can push for the finish in the first half of the round when you’re most fresh but the more you do the more you are going to find your limits. Once you find this limit you’re going to have less ability to keep pushing for the finish and you’re going to need to take a break in the round. Then, once you see the end of the round is in sight you can push again as you know you only have to go so long before you get a break.

Now let’s take a look at how this differs between knockouts and submissions. First let’s look at the submission data

Here we see the similar rhythm that we’ve highlighted with the overall finishes, there’s the increasing threat for the first 3 minutes and then the drop off between 3-4 mins before the final push in the final minute. With the submission finishes you might expect that there is a more steady upward threat consistently through the round as progression is made to close the distance, achieve the takedown and then work towards a submission. Even though this is still generally the case it’s interesting that we still see the 3-4 min dip still remains, this seems to confirm that it’s generally down to energy levels rather than any particular technical element. Looking at the start of the round it’s completely logical that we would see low submission volumes within the first minute, as submissions do naturally require an amount of set up time and can’t often be achieved in one quick move like KOs can.

Let’s take a look at the rhythm of the knockout data to see if the pattern is different

Again we see a pattern similar to what we’ve already highlighted but once again we see that the knockout chances at the end of one round does not necessarily carry over into the start of the next, at the end of round 1 to the start of round 2 there is a dramatic drop off in the knock out rate. While you would absolutely expect this on submissions as they require a potentially lengthy set up, but again you would naturally assume that if the chances of a knockout are high at the end of a round that this would just carry over into the start of the next round – for example if there’s a 3% chance of you getting knocked out in the final 30 seconds of round 1 then you would assume that the level would remain at around 3% at the start of the next round and then slowly decrease. But as we’ve seen already this is not the case, at the end of round 1 the knockout chances are at 3% but at the start of round 2 it drops by half to 1.5%. This means that making it to the end of the round becomes absolutely critical, especially in round 1 – if you make it back to your stool and get to round 2 then you severely reduce the chances of getting knocked out. The break between rounds appears to have a greater impact than you would expect and the fight doesn’t naturally pick up where it left off.

Round 3 KOs Break The Rules

The start of round 3 for KOs seems to break the general pattern, here it seems like that there is a final push to try and get the fight finished early in the round for around the first minute before the threat wanes somewhat and then it surges again in the middle of the round before settling down. You don’t see the same surge in the final minute like you do in rounds 1 and 2, perhaps the surge of energy is simply shifted to the start of the round as combatants try to finish the fight rather than risk it going to the judges.

Round By Round Flow

This can all be a little difficult to see in the views that we’ve presented so far so for clarity we’re going to break that data down into individual rounds and represent the data with curved lines so that you really get a sense of the flow. We’ll show the overall finishing rate, the KO rate and the submission rate in the same graph. Here’s round 1

This really gives you a sense of the rhythm of the round, with the KO push in the first minute the dip around 3-4 mins in both KOs and submissions and then the surge from both in the final minute. You can also see a potential submission sweet spot around 3:30 in the round as the submission chances close the gap on the knockout chances to the greatest degree. Here’s round 2

Round 2 does not have the same surge in KOs at the 1 minute mark that round 1 has but generally you get the same rise and fall before the final push. Submissions have a steady incline and peak at around 2 mins then fall away a little and peak again in the final minute. The gap between KOs and submissions is at some of its lowest levels right from minute 2 through to minute 4 so this whole middle of the round could be considered a submission sweet spot. Here’s round 3

The KOs in round 3 represent the biggest break in the patterns that we see, there’s a big surge at the beginning that we don’t see in other rounds followed by a big dip and then recovery and there isn’t the big surge at the end that we usually do see. This means that the submission opportunities in the final round are quite sporadic, there are key opportunities at 1:30 after the initial surge, again at 3:00 when they hit their initial peak and then again for the whole final one and a half minutes of the fight when the KO threat wanes. At 4:30 in the final round we see the only time in the entire fight when the submission chances actually match the knockout chances.

High Detail Opportunity Breakdown

We’re going to finish our analysis of the 30 second break down by showing a view that summarise all of the above. This graph shows the KO chances, the submission chances and we return to our earlier opportunity calculation and overlay the sweet spots in terms of KOs and submissions

Once again we’re looking for a combination of the best peaks combined with the best gaps. So from a striking perspective we want to identify red areas where the volume is high and the gap is large. For submissions it’s the blue areas.

Here’s our list of the top opportunities for both submissions and knockouts, along with a list of the times in a round to avoid. 

Striker vs Grappler

Grappler vs Striker

The Top 5 Submissions Round By Round Flow

In our final graph we look at the top 5 submissions:

  1. Rear Naked Choke
  2. Guillotine Choke
  3. Arm Bar
  4. Triangle Choke
  5. Arm Triangle Choke

And we plot their chance of finishing during each phase of the round. We return to breaking down the round into the start, middle and end here as there’s simply not enough volume to draw good conclusions from 30 second breakdown.

As usual the volume of Rear Naked Chokes dwarfs the other submissions but we can see that it naturally requires more set up time and so peaks at the middle to end of the rounds. 

This rhythm is far less pronounced with the Guillotine Choke and it actually has three points in the fight where it matches the effectiveness of the Rear Naked Choke – the start of Round 1 and then the start and the end of Round 3. 

The Arm Bar presents a very interesting pattern – There’s a surprisingly pronounced high point at the end of Round 1 where it’s more effective than a Guillotine Choke and comes quite close to the Rear Naked Choke. The end of Round 1 seems to be a real sweet spot for getting an Arm Bar.