Study Sample
The sample included UFC Bouts from 2020 to 2022 scheduled for five rounds across Featherweight (FTW - 145lbs, n = 13), Middleweight (MW - 185lbs, n = 16) and Heavyweight (HW - 206-265lbs, n = 18) divisions, totalling 47 bouts and 183 rounds. These included title fights (n = 14), which are all scheduled for five rounds, and other five round scheduled fights (n = 33). Fights lasting less than 30 seconds were removed, alongside those ending in no contests, or draws. This defined sample was selected to ensure the analysis was focusing on the very pinnacle of elite MMA athletes competing in the UFC.
Operational Definitions
The following indicators were collected and analysed for winners and losers:
- Striking:
- Total Strikes: Any attempt to carry out an attack with striking techniques (punches, kicks, knees, elbows) on the opponent.
- Strikes Landed: Any striking attempt which lands on the opponent, regardless of location or impact.
- Significant Strikes Landed: All strikes at a distance and power strikes in the clinch and on the ground.
- Strike Type:
- Punch: Any significant striking technique landed using punches.
- Kick: Any significant striking technique landed using kicks.
- Knee: Any significant striking technique landed using either knee.
- Elbow: Any significant striking technique landed using either elbow.
- Strike Location:
- Head: Any significant striking technique (punches, kicks, knees, elbows) on the opponent’s head.
- Body: Any significant striking technique (punches, kicks, knees, elbows) on the opponent’s body.
- Leg: Any significant striking technique (punches, kicks, knees, elbows) on the opponent’s leg.
- Strike Origin:
- Standing Strikes: Any significant striking technique on the opponent from a standing position.
- Clinch Strikes: Any significant striking technique on the opponent from the clinch. The clinch is when fighters are standing up in the grappling phase, when either fighter has direct contact and/or control of the other.
- Ground Strikes: Any significant striking technique on the opponent from the ground. This includes strikes from any ground position (mount, side control, guard, etc.).
- Takedowns:
- Total Takedown Actions: A technique that involves off-balancing an opponent trying to bring him/her to the ground, typically with the attacker on top.
- Takedowns Landed: Successful grounding of an opponent, with a level of control required by the TD initiator once the fighter is taken down.
- Submissions:
- Submission Attempts: A technique when one fighter tries to lock up their opponent causing one of their body parts to be compressed or stretched. If successful, the fight will end via submission or a technical submission. It generally comprises of either chokes or locks.
- Fight Outcome:
- Outcome Detailed: The method by which the fight ended with possible values of KO/TKO, Submission or Decision.
- Outcome: The method by which the fight ended with possible values of Finish (which includes KO/TKO and Submission) and Decision.
- Division: the weight division of the fighters with possible values of Featherweight (145lbs), Middleweight (185lbs) and Heavyweight (206-265lbs).
Reliability Study
Inter- and intra-rater reliability assessments were conducted with variance calculated using the percentage error method (< 5% percentage error was deemed acceptable)1. The inter-reliability assessment conducted by the author (> 4 years’ performance analysis experience) and a skilled analyst with extensive knowledge of MMA (> 15 years coaching experience), reported an overall percentage error of 2.0%. The intra-reliability assessment was conducted by the author with a period of three weeks given between the initial observation and the test comparison to reduce the effects of memory bias2, resulting in a percentage error of 4.7%.
Inter-Reliability Study Results
Results of Inter-Reliability Study by Indicator
| Fighter |
Category |
Indicator |
Observation 1 |
Observation 2 |
Absolute Differences |
Mean Observations |
% Error |
| A |
Strikes |
Attempted |
82 |
84 |
2 |
83.0 |
2.4% |
| A |
Strikes |
Landed |
272 |
276 |
4 |
274.0 |
1.5% |
| A |
Strikes |
Significant |
237 |
240 |
3 |
238.5 |
1.3% |
| A |
Strikes |
Punch |
167 |
169 |
2 |
168.0 |
1.2% |
| A |
Strikes |
Kick |
59 |
60 |
1 |
59.5 |
1.7% |
| A |
Strikes |
Knee |
0 |
0 |
0 |
0.0 |
0.0% |
| A |
Strikes |
Elbow |
11 |
11 |
0 |
11.0 |
0.0% |
| A |
Strikes |
Head |
166 |
170 |
4 |
168.0 |
2.4% |
| A |
Strikes |
Body |
27 |
25 |
2 |
26.0 |
7.7% |
| A |
Strikes |
Legs |
44 |
45 |
1 |
44.5 |
2.2% |
| A |
Strikes |
Standing |
161 |
163 |
2 |
162.0 |
1.2% |
| A |
Strikes |
Clinch |
5 |
5 |
0 |
5.0 |
0.0% |
| A |
Strikes |
Ground |
71 |
72 |
1 |
71.5 |
1.4% |
| B |
Strikes |
Attempted |
103 |
107 |
4 |
105.0 |
3.8% |
| B |
Strikes |
Landed |
111 |
114 |
3 |
112.5 |
2.7% |
| B |
Strikes |
Significant |
94 |
96 |
2 |
95.0 |
2.1% |
| B |
Strikes |
Punch |
70 |
73 |
3 |
71.5 |
4.2% |
| B |
Strikes |
Kick |
18 |
17 |
1 |
17.5 |
5.7% |
| B |
Strikes |
Knee |
1 |
1 |
0 |
1.0 |
0.0% |
| B |
Strikes |
Elbow |
5 |
5 |
0 |
5.0 |
0.0% |
| B |
Strikes |
Head |
68 |
69 |
1 |
68.5 |
1.5% |
| B |
Strikes |
Body |
18 |
19 |
1 |
18.5 |
5.4% |
| B |
Strikes |
Legs |
8 |
8 |
0 |
8.0 |
0.0% |
| B |
Strikes |
Standing |
82 |
84 |
2 |
83.0 |
2.4% |
| B |
Strikes |
Clinch |
8 |
8 |
0 |
8.0 |
0.0% |
| B |
Strikes |
Ground |
4 |
4 |
0 |
4.0 |
0.0% |
| A |
Takedown |
Attempted |
0 |
0 |
0 |
0.0 |
0.0% |
| A |
Takedown |
Landed |
0 |
0 |
0 |
0.0 |
0.0% |
| B |
Takedown |
Attempted |
0 |
0 |
0 |
0.0 |
0.0% |
| B |
Takedown |
Landed |
0 |
0 |
0 |
0.0 |
0.0% |
| A |
Submission |
Attempted |
0 |
0 |
0 |
0.0 |
0.0% |
| B |
Submission |
Attempted |
4 |
4 |
0 |
4.0 |
0.0% |
Overall Results of Inter-Reliability Test
| Total Observation 1 |
Total Observation 2 |
Total Absolute Differences |
Total Mean Observations |
% Error |
| 1896 |
1929 |
39 |
1912.5 |
2.0% |
Intra-Reliability Study Results
Results of Intra-Reliability Study by Indicator
| Fighter |
Category |
Indicator |
Observation 1 |
Observation 2 |
Absolute Differences |
Mean Observations |
% Error |
| A |
Strikes |
Attempted |
38 |
33 |
5 |
35.5 |
14.1% |
| A |
Strikes |
Landed |
103 |
105 |
2 |
104.0 |
1.9% |
| A |
Strikes |
Significant |
68 |
67 |
1 |
67.5 |
1.5% |
| A |
Strikes |
Punch |
45 |
43 |
2 |
44.0 |
4.5% |
| A |
Strikes |
Kick |
16 |
16 |
0 |
16.0 |
0.0% |
| A |
Strikes |
Knee |
6 |
7 |
1 |
6.5 |
15.4% |
| A |
Strikes |
Elbow |
1 |
1 |
0 |
1.0 |
0.0% |
| A |
Strikes |
Head |
29 |
30 |
1 |
29.5 |
3.4% |
| A |
Strikes |
Body |
27 |
24 |
3 |
25.5 |
11.8% |
| A |
Strikes |
Legs |
12 |
13 |
1 |
12.5 |
8.0% |
| A |
Strikes |
Standing |
45 |
42 |
3 |
43.5 |
6.9% |
| A |
Strikes |
Clinch |
7 |
8 |
1 |
7.5 |
13.3% |
| A |
Strikes |
Ground |
16 |
17 |
1 |
16.5 |
6.1% |
| B |
Strikes |
Attempted |
15 |
13 |
2 |
14.0 |
14.3% |
| B |
Strikes |
Landed |
93 |
91 |
2 |
92.0 |
2.2% |
| B |
Strikes |
Significant |
70 |
70 |
0 |
70.0 |
0.0% |
| B |
Strikes |
Punch |
21 |
18 |
3 |
19.5 |
15.4% |
| B |
Strikes |
Kick |
41 |
42 |
1 |
41.5 |
2.4% |
| B |
Strikes |
Knee |
5 |
6 |
1 |
5.5 |
18.2% |
| B |
Strikes |
Elbow |
3 |
4 |
1 |
3.5 |
28.6% |
| B |
Strikes |
Head |
24 |
23 |
1 |
23.5 |
4.3% |
| B |
Strikes |
Body |
21 |
20 |
1 |
20.5 |
4.9% |
| B |
Strikes |
Legs |
26 |
27 |
1 |
26.5 |
3.8% |
| B |
Strikes |
Standing |
64 |
63 |
1 |
63.5 |
1.6% |
| B |
Strikes |
Clinch |
6 |
7 |
1 |
6.5 |
15.4% |
| B |
Strikes |
Ground |
0 |
0 |
0 |
0.0 |
0.0% |
| A |
Takedown |
Attempted |
3 |
2 |
1 |
2.5 |
40.0% |
| A |
Takedown |
Landed |
5 |
5 |
0 |
5.0 |
0.0% |
| B |
Takedown |
Attempted |
2 |
1 |
1 |
1.5 |
66.7% |
| B |
Takedown |
Landed |
1 |
1 |
0 |
1.0 |
0.0% |
| A |
Submission |
Attempted |
3 |
3 |
0 |
3.0 |
0.0% |
| B |
Submission |
Attempted |
4 |
4 |
0 |
4.0 |
0.0% |
Overall Results of Intra-Reliability Test
| Total Observation 1 |
Total Observation 2 |
Total Absolute Differences |
Total Mean Observations |
% Error |
| 820 |
806 |
38 |
813 |
4.7% |
Statistical Analysis
The data was tidied and transformed in Microsoft Excel, which included scaling the variables to actions per minute to ensure accurate comparisons across bouts3. Statistical analysis was conducted using R 4.4.04 and RStudio5. Descriptive statistics are presented as mean ± SD across all bouts, winning and losing performances, by weight and based on the method of fight outcome. The Shapiro-Wilk test was used to assess normality of samples and Levene’s Test was used to assess equal variances in independent samples.
Differences in indicators between winners and losers were examined using a paired samples t test for normally distributed samples, while the Wilcoxon signed rank test was used for non-normally distributed samples.
Relative indicators were calculated by subtracting the loser’s performance from the winner’s performance. Differences in these relative indicators between fight outcomes were examined using an independent samples t test for normally distributed samples (with equal variances either assumed or not assumed depending on the results of Levene’s test), while the Mann-Whitney U test was used for non-normally distributed samples.
The alpha level required for statistical significance was α = 0.05. The p values within each group of hypothesis tests were adjusted using the Bonferroni correction to ensure the cumulative Type 1 error remains below 0.05.
Cohen’s d was used to estimate effect sizes for normally distributed samples, while the Wilcoxon Effect Size r = z/√n was used to estimate effect sizes for non-normally distributed samples, where n is the number of observations and z is the z-score returned from either the Wilcoxon signed rank test or Mann-Whitney U test6.
Cohen’s d can be interpreted as follows7,8:
- 0.2 ≤ d < 0.5 indicates a small effect.
- 0.5 ≤ d < 0.8 indicates a moderate effect.
- d ≥ 0.8 indicates a large effect.
The Wilcoxon Effect Size r can be interpreted by taking the absolute value of r and using the following thresholds:
- 0.1 ≤ |r| < 0.3 indicates a small effect.
- 0.3 ≤ |r| < 0.5 indicates a moderate effect.
- |r| ≥ 0.5 indicates a large effect.
Descriptive Statistics of All Fighters, By Division & By Fight Outcome
Descriptive Statistics for All Fighters, By Division & By Fight Outcome (n = # Fighters)
|
Division |
Fight Outcome |
| Indicator |
Overall
(n = 94) |
FTW
(n = 26) |
MW
(n = 32) |
HW
(n = 36) |
Finish
(n = 40) |
Decision
(n = 54) |
| Total Strikes Attempted |
9.49±4.46 |
12.43±4.79 |
9.61±3.95 |
7.27±3.34 |
8.18±4.16 |
10.47±4.46 |
| Total Strikes Landed |
7.32±3.95 |
9.13±4.52 |
7.46±3.60 |
5.88±3.28 |
6.43±3.94 |
7.98±3.86 |
| Sig. Strikes Landed |
5.27±2.98 |
7.03±3.30 |
5.03±2.50 |
4.21±2.59 |
4.71±2.75 |
5.68±3.09 |
| Sig. Punch Strikes |
3.65±2.56 |
4.93±2.72 |
3.62±2.54 |
2.76±2.10 |
3.15±2.15 |
4.02±2.79 |
| Sig. Kick Strikes |
1.21±0.88 |
1.56±0.91 |
1.12±0.87 |
1.03±0.82 |
1.10±0.93 |
1.29±0.85 |
| Sig. Knee Strikes |
0.22±0.38 |
0.31±0.59 |
0.14±0.19 |
0.23±0.30 |
0.28±0.52 |
0.18±0.23 |
| Sig. Elbow Strikes |
0.19±0.41 |
0.24±0.36 |
0.16±0.23 |
0.19±0.55 |
0.19±0.53 |
0.20±0.29 |
| Sig. Head Strikes |
3.55±2.49 |
4.87±2.51 |
3.36±2.39 |
2.77±2.24 |
3.27±2.34 |
3.76±2.61 |
| Sig. Body Strikes |
0.89±0.69 |
1.11±0.86 |
0.93±0.63 |
0.69±0.56 |
0.68±0.52 |
1.04±0.76 |
| Sig. Leg Strikes |
0.83±0.67 |
1.06±0.71 |
0.74±0.68 |
0.76±0.61 |
0.77±0.69 |
0.88±0.66 |
| Sig. Standing Strikes |
4.26±2.83 |
6.26±3.11 |
4.11±2.54 |
2.95±1.98 |
3.35±2.20 |
4.94±3.07 |
| Sig. Clinch Strikes |
0.32±0.44 |
0.40±0.58 |
0.31±0.32 |
0.28±0.41 |
0.35±0.60 |
0.30±0.27 |
| Sig. Ground Strikes |
1.09±2.22 |
1.84±3.08 |
0.62±1.27 |
0.98±2.06 |
1.22±2.19 |
1.00±2.25 |
| Takedowns Attempted |
0.17±0.23 |
0.16±0.22 |
0.19±0.21 |
0.16±0.26 |
0.15±0.26 |
0.18±0.20 |
| Takedowns Landed |
0.06±0.11 |
0.05±0.07 |
0.07±0.10 |
0.07±0.14 |
0.06±0.11 |
0.07±0.11 |
| Submission Attempts |
0.04±0.10 |
0.05±0.07 |
0.02±0.04 |
0.05±0.15 |
0.05±0.15 |
0.03±0.05 |
Comparing Relative Indicators by Fight Outcomes
For each fight outcome, relative indicators are calculated by subtracting the loser’s performance from the winner’s performance. For example, Rel. Total Strikes Attempted is calculated as total strikes for the winner minus total strikes for the loser. Therefore, a positive value for Rel. Total Strikes Attempted means the winner had more total strikes than the loser.
Descriptive Statistics
Descriptive Statistics for Relative Indicators By Fight Outcome (n = # Fights)
| Indicator |
Finish
(n = 20) |
Decision
(n = 27) |
| Rel. Total Strikes Attempted |
4.82±4.02 |
2.37±3.74 |
| Rel. Total Strikes Landed |
4.86±3.67 |
2.68±3.54 |
| Rel. Sig. Strikes Landed |
3.70±2.84 |
2.40±2.62 |
| Rel. Sig. Punch Strikes |
2.68±2.27 |
1.92±2.38 |
| Rel. Sig. Kick Strikes |
0.77±1.10 |
0.25±1.07 |
| Rel. Sig. Knee Strikes |
0.05±0.57 |
0.08±0.31 |
| Rel. Sig. Elbow Strikes |
0.22±0.76 |
0.13±0.37 |
| Rel. Sig. Head Strikes |
2.94±2.43 |
1.86±2.41 |
| Rel. Sig. Body Strikes |
0.28±0.50 |
0.29±1.02 |
| Rel. Sig. Leg Strikes |
0.50±0.92 |
0.24±0.84 |
| Rel. Sig. Standing Strikes |
1.89±2.54 |
1.73±2.57 |
| Rel. Sig. Clinch Strikes |
0.16±0.72 |
0.06±0.33 |
| Rel. Sig. Ground Strikes |
1.77±2.30 |
1.08±1.85 |
| Rel. Takedowns Attempted |
0.08±0.41 |
0.06±0.36 |
| Rel. Takedowns Landed |
0.06±0.16 |
0.07±0.16 |
| Rel. Submission Attempts |
0.02±0.22 |
0.00±0.06 |
Hypothesis Testing
The Shapiro-Wilk test was applied to each individual relative indicator for each fight outcome to assess if the samples followed the Normal distribution. Levene’s test was applied to each pair of corresponding relative indicators for each fight outcome to assess whether or not the samples had equal variances.
Depending on the results of the Shapiro-Wilk test & Levene’s test:
- Either an independent samples t test (without or without equal variances assumed, according to the results of Levene’s test) or the Mann-Whitney U test was applied to test for significant differences between each pair of corresponding relative indicators for each fight outcome.
- Cohen’s d or Wilcoxon’s effect size r were used to measure effect size.
All p values returned from the independent samples t test and Mann-Whitney U test were adjusted using the Bonferroni correction to ensure the cumulative Type 1 error remains below 0.05.
Shapiro-Wilk Test & Levene’s Test Results
Results of Shapiro-Wilk & Levene's Test for Finish & Decision Samples
| Indicators |
p - SW - Finish Sample |
p - SW - Decision Sample |
Normal Dist? |
LW |
Equal Variances? |
Correct Test |
| Rel. Total Strikes Attempted |
0.25 |
0.00 |
Non-Normal |
0.30 |
Equal Variances |
Mann-Whitney Test |
| Rel. Total Strikes Landed |
0.47 |
0.00 |
Non-Normal |
0.28 |
Equal Variances |
Mann-Whitney Test |
| Rel. Sig. Strikes Landed |
0.19 |
0.00 |
Non-Normal |
0.27 |
Equal Variances |
Mann-Whitney Test |
| Rel. Sig. Punch Strikes |
0.85 |
0.00 |
Non-Normal |
0.96 |
Equal Variances |
Mann-Whitney Test |
| Rel. Sig. Kick Strikes |
0.36 |
0.32 |
Normal |
0.98 |
Equal Variances |
Ind Samples T-test |
| Rel. Sig. Knee Strikes |
0.00 |
0.05 |
Non-Normal |
0.17 |
Equal Variances |
Mann-Whitney Test |
| Rel. Sig. Elbow Strikes |
0.00 |
0.00 |
Non-Normal |
0.26 |
Equal Variances |
Mann-Whitney Test |
| Rel. Sig. Head Strikes |
0.80 |
0.02 |
Non-Normal |
0.97 |
Equal Variances |
Mann-Whitney Test |
| Rel. Sig. Body Strikes |
0.00 |
0.00 |
Non-Normal |
0.08 |
Equal Variances |
Mann-Whitney Test |
| Rel. Sig. Leg Strikes |
0.06 |
0.28 |
Normal |
0.68 |
Equal Variances |
Ind Samples T-test |
| Rel. Sig. Standing Strikes |
0.59 |
0.00 |
Non-Normal |
0.55 |
Equal Variances |
Mann-Whitney Test |
| Rel. Sig. Clinch Strikes |
0.00 |
0.27 |
Non-Normal |
0.03 |
Non Equal Variances |
Mann-Whitney Test |
| Rel. Sig. Ground Strikes |
0.00 |
0.00 |
Non-Normal |
0.16 |
Equal Variances |
Mann-Whitney Test |
| Rel. Takedowns Attempted |
0.08 |
0.97 |
Normal |
0.91 |
Equal Variances |
Ind Samples T-test |
| Rel. Takedowns Landed |
0.00 |
0.00 |
Non-Normal |
0.91 |
Equal Variances |
Mann-Whitney Test |
| Rel. Submission Attempts |
0.00 |
0.00 |
Non-Normal |
0.06 |
Equal Variances |
Mann-Whitney Test |
Independent t Test/Mann Whitney U Test Results & Effect Sizes
Results of Independent Samples Testing for Finish vs Decision Fights
|
Non-adjusted |
Bonferroni adjusted |
|
| Indicator |
Test |
p |
Result |
p |
Result |
Effect Calc |
Effect Size |
| Rel. Total Strikes Attempted |
Mann Whitney |
0.006 |
Significant |
0.088 |
Non-Significant |
r |
-0.405 |
| Rel. Total Strikes Landed |
Mann Whitney |
0.005 |
Significant |
0.082 |
Non-Significant |
r |
-0.408 |
| Rel. Sig. Strikes Landed |
Mann Whitney |
0.023 |
Significant |
0.361 |
Non-Significant |
r |
-0.333 |
| Rel. Sig. Punch Strikes |
Mann Whitney |
0.116 |
Non-Significant |
1.000 |
Non-Significant |
r |
-0.229 |
| Rel. Sig. Kick Strikes |
Ind T-test |
0.113 |
Non-Significant |
1.000 |
Non-Significant |
d |
0.087 |
| Rel. Sig. Knee Strikes |
Mann Whitney |
0.674 |
Non-Significant |
1.000 |
Non-Significant |
r |
-0.061 |
| Rel. Sig. Elbow Strikes |
Mann Whitney |
0.769 |
Non-Significant |
1.000 |
Non-Significant |
r |
0.043 |
| Rel. Sig. Head Strikes |
Mann Whitney |
0.064 |
Non-Significant |
1.000 |
Non-Significant |
r |
-0.270 |
| Rel. Sig. Body Strikes |
Mann Whitney |
0.966 |
Non-Significant |
1.000 |
Non-Significant |
r |
-0.006 |
| Rel. Sig. Leg Strikes |
Ind T-test |
0.320 |
Non-Significant |
1.000 |
Non-Significant |
d |
0.297 |
| Rel. Sig. Standing Strikes |
Mann Whitney |
0.232 |
Non-Significant |
1.000 |
Non-Significant |
r |
-0.174 |
| Rel. Sig. Clinch Strikes |
Mann Whitney |
0.795 |
Non-Significant |
1.000 |
Non-Significant |
r |
-0.038 |
| Rel. Sig. Ground Strikes |
Mann Whitney |
0.133 |
Non-Significant |
1.000 |
Non-Significant |
r |
-0.219 |
| Rel. Takedowns Attempted |
Ind T-test |
0.841 |
Non-Significant |
1.000 |
Non-Significant |
d |
0.060 |
| Rel. Takedowns Landed |
Mann Whitney |
0.707 |
Non-Significant |
1.000 |
Non-Significant |
r |
-0.055 |
| Rel. Submission Attempts |
Mann Whitney |
0.241 |
Non-Significant |
1.000 |
Non-Significant |
r |
-0.171 |
References
Hughes, M. (2004). Performance analysis–a 2004 perspective. International Journal of Performance Analysis in Sport, 4(1), 103-109.
Cooper, S.-M., Hughes, M., O’Donoghue, P., & Nevill, M. A. (2007). A simple statistical method for assessing the reliability of data entered into sport performance analysis systems. International Journal of Performance Analysis in Sport, 7(1), 87–109.
James, L. P., Robertson, S., Haff, G. G., Beckman, E. M., & Kelly, V. G. (2017). Identifying the performance characteristics of a winning outcome in elite mixed martial arts competition. Journal of Science and Medicine in Sport, 20(3), 296–301.
R Core Team (2024). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/
Posit team (2023). RStudio: Integrated Development Environment for R. Posit Software, PBC, Boston, MA. URL http://www.posit.co/
Rosenthal, R. (1991). Meta-analytic procedures for social research (2nd ed.). Newbury Park, Ca: Sage.
Cohen J. (1992). A power primer. Psychological Bulletin, 112(1), 155-159.
Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (2nd ed.). New York: Academic Press.
