Share This

Share on facebook
Share on twitter
Share on google

Speed Parkour Competition Stats | Pilot Study

Disclaimer: I’m just a data nerd who’s curious to learn more about parkour speed training. There’s a lot I don’t know but I find this measurable movement style to be one of the most innovative, unexplored areas in all of athletics. No other jump/land/sprint type sport has gone as far as parkour has in adapting fundamental movement skills to so many diverse environments and complex challenges. On top of that, speed parkour competitions put it all to the test against an indisputable, unwavering judge known as TIME.

Much of this pilot study may not be useful in our goal to get faster but we must start somewhere. I hope this is the beginning of a more thorough and thoughtful meta-analysis of many parkour speed competitions, preferably with help from other expert athletes, coaches, and researchers too. While I took some stats in college, it’s been a few years since I used any advanced stuff. If you know more about the MATH relevant to this project (stats, data science, sports analytics, etc.), hit us up! If you’re good at generating creative connections and ideas, let us know! If you’re a relentless number-crunching machine, we need lots of help with tedious video edits and calculations haha

If you want to help grow this project, be sure to also check out the raw data & charts, and please drop a comment or send an email/DM to discuss more πŸ™‚

P.S. Special thanks to Amos Rendao, Julian Frazier, Brandon Douglass, and Taylor Carpenter for giving key initial feedback to help improve this pilot study.

🔥 Why speed parkour

Many old-school jump/land sports are extremely specialized. How far can you jump? How high? In parkour, it’s can you do that AND land on a rail? While 3 meters up? Above hard ground? Oh and for speed parkour, do all that and more as fast as you can, and for a full 10-30 second run. Minimal practice time. No restarts.

As an underrated, undervalued training style, parkour speed training and competitions are also getting crazy competitive and insightful. It’s amazing to see such power, skill, and confidence from elite speed athletes and I’m excited for this style of training to keep growing. Recently, I’ve enjoyed a speed training focus myself. Which got me thinking…

What does the data say about speed parkour competitions? What’s the data we should pay most attention to? How can we use the numbers and insights to become faster, better athletes?

Over the past year, I fired up a camera/gimbal and filmed 300+ runs from the last 9 major speed parkour competitions at Apex Denver, Apex Louisville, and Apex Fort Collins. Much respect and thanks to all the athletes who tested themselves, and all the organizers who helped make this happen. I’ve watched many of these runs more times than I can count, in slow-motion, real-time, and as side-by-side comparisons. Studying the film has taught me so much and so I recently decided to progress our understanding by calculating stats like contact times and number/types of touches.

For this pilot study, I started with a recent RMPC competition held at Apex Denver in April 2019.Β I want to start a conversation with an initial analysis of the top 8 competitors on each course. In this study, I focus on what is the data, but not what the data means. As we increase our sample size and number of expert eyes on this research, we will slowly decode a formula for hacking our times on a parkour speed course.

This is a work-in-progress so please let me know if you’re seeing any interesting patterns, or if you think there are any key metrics that have not yet been calculated. In the future, perhaps we can hook up course runners with accelerometers or other tech to track more stats like acceleration rates, top speeds, elevation changes, distance traveled, and peak forces. Keep reading for some insights into speed parkour competitions:

Overall rankings

Top 3 men

1) Seth Wang, 31 points (Apex Louisville)
2) Michael Sliger, 28 points (Apex Louisville)
3) Jared Nahulu, 26 points (Apex Denver)

Top 3 women

1) Kasia Kilijanek, 19 points (Apex Denver)
2) Taylor Carpenter, 16 points (Apex Fort Collins / Apex Louisville)
3) Mikaila Quinn, 15 points (Move to Inspire)

Note: Total points = course 1 points + course 2 points

Data tracked

Most stats were determined by analyzing the slow-motion videos embedded in this post. I’m also the guy behind the camera, chasing each athlete with a high-speed camera/gimbal rig (Pansonic GH4 / DJI Ronin-M). Because I shot most runs in 96 frames per second (25% speed), I used frame-by-frame analysis to calculate accurate start/finish/contact times down to 1/100th of a second (1/96fps = 0.0104 sec). This is tedious work because it involves calculating many of the 1200+ data points and charts one at a time.

Unfortunately, I had minor camera issues at the end of this competition and had to film a few final runs (Nahulu, Sliger, Frazier) with an iPhone at 24 frames per second (1/24 fps = 0.04 sec). Also, Boden’s run video cut out halfway through (memory card full) and so I guesstimated his second-half stats by comparing/contrasting it to other runs and asking him to recount how it went.

Keep in mind that there are likely some small errors in the stats that I have calculated so far. However, any errors are most likely less than the time of each video frame (.01β€”.04 sec).

For now, I’m not sharing my opinions or conclusions from the data. It’s too early to definitively say much and if we want to learn more, I still have hundreds of recent runs to analyze. As an initial exploration into some of the stats behind parkour speed training, my main goal is to see if there is value in digging deeper.

Please share your ideas regarding how we should progress this analysis in order to better understand the limits of human-powered jumps, landings, sprints, and climbs. Drop a comment below, or send me an email/DM and let’s talk πŸ™‚

See full stats & charts from the event

Footwork

β€” Total left foot touches = total number of left foot touches on the course
β€” Total right foot touches = total number of right foot touches on the course
β€” Total foot touches = total left foot touches + total right foot touches
β€” Foot touches per second = total foot touches / time (sec)
β€” Total foot contact time (sec) = total amount of foot contact time
β€” Average contact time per foot touch (sec) = total amount of foot contact time / total foot touches

Handwork

β€” Total left hand touches = total number of left hand touches on the course
β€” Total right hand touches = total number of right hand touches on the course
β€” Total hand touches = total left hand touches + total right hand touches
β€” Hand touches per second = total hand touches / time (sec)
β€” Total hand contact time (sec) = total amount of hand contact time
β€” Average contact time per hand touch (sec) = total amount of hand contact time / total hand touches

Total work

β€” Total touches = total hand touches + total foot touches
β€” Total touches per second = total touches / time (sec)
β€” Total contact time (sec) = total contact time of hands, feet, and any other body part (sec)
β€” Average contact time per touch (sec) = total amount of contact time / total touches
β€” Total air time (sec) = total time (sec) – total contact time (sec)
β€” Foot dominance = total touches / total foot touches

Segments & segment times

β€” Starting segment times = total time of start to first touch of first obstacle
β€” Segments are categorized as ascent (going up 1 body height or more), descent (going down 1 body height or more), or sprint (going across things without going up or down 1 body height)
β€” Segment times are calculated on an individual basis depending on the locomotive strategy used at what time, not necessarily based on when a certain physical checkpoint is hit (this is due to some athletes using different routes and/or strategies)
β€” Segments depend on course/division, see below for more details
β€” Total descent time = total time of all descent segments
β€” Total ascent time = total time of all ascent segments
β€” Total sprint time = total time of all sprint segments
β€” Sprint dominance = total sprint time / time

Charts

β€” Time (sec) vs. total foot contact time (sec)
β€” Time (sec) vs. total hand contact time (sec)
β€” Time (sec) vs. total contact time (sec)
β€” Time (sec) vs. total air time (sec)
β€” Time (sec) vs. foot dominance
β€” Time (sec) vs. total foot touches
β€” Time (sec) vs. total hand touches
β€” Time (sec) vs. total touches
β€” Time (sec) vs. total foot touches per second
β€” Time (sec) vs. total hand touches per second
β€” Time (sec) vs. total touches per second
β€” Time (sec) vs. avg. contact time per foot touch (sec)
β€” Time (sec) vs. avg. contact time per hand touch (sec)
β€” Time (sec) vs. avg. contact time per touch (sec)

Course 1 men

1) Seth Wang, 10.17 sec (Apex Louisville)
2) Julian Frazier, 10.53 sec (Apex Denver)
3) Michael Sliger, 10.66 sec (Apex Louisville)

Stats from course 1 men:

  • 23 athletes from several states ran course 1 two times each. Each athlete’s fastest time was used to determine the top 8 moving on to finals.
  • The average foot dominance of the top 3 was 75.33%
  • The average sprint dominance of the top 3 was 62.99%
  • The top 8 all finished within 3.07 seconds of each other
  • The top 5 all used the same number of R hand touches (5), L hand touches (5), and total hand touches (10)

Segments from course 1 men:

  • Segment 1 (sprint) β€” time from start to first touch on handrail
  • Segment 2 (sprint) β€” time from first touch of handrail to first touch on incline wall and/or high bar
  • Segment 3 (ascent) β€” time from first touch of incline wall and/or high bar to fully above gray wall
  • Segment 4 (sprint) β€” time from fully above gray wall to first touch of tallest blue tower
  • Segment 5 (descent) β€” time from tallest blue tower to finish

See full stats & charts from the event

🥇 Seth Wang, 10.17 sec

Stats from Seth’s run (compared to top 8):

  • Tied Julian for fewest amount of foot touches (29), tied 5 others for fewest amount of hand touches (10), tied Julian for fewest amount of total touches (39)
  • Least amount of total foot contact time (5.79 sec), total contact time (7.25 sec), & total sprint time (6.29 sec)
  • Second highest amount of air time (2.92 sec)
  • Fastest time on segment 2 (3.17 sec) segment 4 (2.54 sec), & segment 5 (2.08 sec)
  • Tied Michael for lowest average contact time per foot touch (0.20 sec), tied Michael for lowest average contact time per touch (0.19 sec)

🥈 Julian Frazier, 10.53 sec

Stats from Julian’s run (compared to top 8):

  • Tied Seth for fewest amount of foot touches (29), tied 5 others for fewest amount of hand touches (10), tied Seth for fewest amount of total touches (39)
  • Used 13 L-foot touches and 16 R-foot touches, compared to Seth’s 14 & 15
  • Second least amount of total foot contact time (6.63 sec) & total contact time (8.00 sec)
  • Fastest time on segment 1 (0.46 sec)
  • Second least amount of total sprint time (6.51 sec)

🥉 Michael Sliger, 10.66 sec

Stats from Michael’s run (compared to top 8):

  • Highest amount of foot touches/sec (3.19/sec)
  • Tied 5 others for fewest amount of hand touches (10)
  • 2nd least amount of total foot contact time (6.67 sec), total hand contact time (2.58 sec), & total contact time (8.16 sec)
  • Fastest time on segment 3 (1.58 sec)
  • Tied Seth for fastest time on segment 5 (2.08 sec)

📈 Trends

Course 1 men, top 8

Note: R-squared (RΒ²) is a statistical measure that represents the proportion of the variance for a dependent variable that’s explained by an independent variable or variables in a regression model. Whereas correlation explains the strength of the relationship between an independent and dependent variable, RΒ² explains to what extent the variance of one variable explains the variance of the second variable. E.g. if the RΒ² of a model is 0.50, then approximately half of the observed variation can be explained by the model’s inputs.

See full stats & charts from the event

Course 1 women

1) Kasia Kilijanek, 15.66 sec (Apex Denver)
2) Michaela Hoelldobler, 16.61 sec (Apex Denver)
3) Taylor Carpenter, 17.09 sec (Apex Fort Collins / Apex Louisville)

Stats from course 1 women:

  • 12 athletes from several states ran course 1 two times each. Each athlete’s fastest time was used to determine the top 8 moving on to finals.
  • The average foot dominance of the top 3 was 74.45%
  • The average sprint dominance of the top 3 was 58.08%

Segments from course 1 women:

  • Segment 1 (sprint) β€” time from start to first touch on handrail
  • Segment 2 (sprint) β€” time from first touch of handrail to first touch on incline wall and/or high bar
  • Segment 3 (ascent) β€” time from first touch of incline wall and/or high bar to fully above gray wall
  • Segment 4 (sprint) β€” time from fully above gray wall to first touch of tallest blue tower
  • Segment 5 (descent) β€” time from tallest blue tower to finish

See full stats & charts from the event

🥇 Kasia Kilijanek, 15.66 sec

Stats from Kasia’s run (compared to top 8):

  • Lowest average contact time per touch (0.23 sec)
  • Highest amount of hand touches per second (1.02/sec)
  • Least amount of total contact time (12.88 sec)
  • Fastest time on segment 3 (3.00 sec) & segment 5 (2.79 sec)
  • Least amount of total ascent time (3.00 sec) & total descent time (2.79 sec)

🥈 Michaela Hoelldobler, 16.61 sec

Stats from Michaela’s run (compared to top 8):

  • Fastest time on segment 4 (4.38 sec)
  • Least amount of total hand contact time (5.21 sec)
  • Second least amount of total foot contact time (12.04 sec) & total contact time (14.50 sec)
  • Tied Kasia for lowest average contact time per hand touch (0.37 sec)
  • Second least amount of total sprint time (9.59 sec)

🥉 Taylor Carpenter, 17.09 sec

Stats from Taylor’s run (compared to top 8):

  • Tied Kasia for the second fewest amount of foot touches (40)
  • Fewest amount of hand touches (12) & total touches (52)
  • Least amount of total touches per second (3.04/sec)
  • Highest foot dominance (76.92%)
  • Fastest time on segment 1 (0.67 sec)

📈 Trends

Course 1 women, top 8

Note: R-squared (RΒ²) is a statistical measure that represents the proportion of the variance for a dependent variable that’s explained by an independent variable or variables in a regression model. Whereas correlation explains the strength of the relationship between an independent and dependent variable, RΒ² explains to what extent the variance of one variable explains the variance of the second variable. E.g. if the RΒ² of a model is 0.50, then approximately half of the observed variation can be explained by the model’s inputs.

See full stats & charts from the event

Course 2 men

1) Seth Wang, 12.25 sec (Apex Louisville)
2) Michael Sliger, 12.91 sec (Apex Louisville)
3) Jared Nahulu, 13.50 sec (Apex Denver)

Stats from course 2 men:

  • 8 athletes from several states ran course 2 one time each. Athletes earned points relative to their ranking on each course and total points determined the overall winner.
  • The average foot dominance of the top 3 was 68.39%
  • The average sprint dominance of the top 3 was 34.25%

Segments from course 2 men:

  • Segment 1 (sprint) β€” time from start to first touch of wood tower
  • Segment 2 (ascent) β€” time from first touch of wood tower to fully above wood tower
  • Segment 3 (descent) β€” time from fully above wood tower to first touch of blue spring floor
  • Segment 4 (sprint) β€” time from first touch of blue spring floor to first touch of incline wall
  • Segment 5 (ascent) β€” time from first touch of incline wall to fully above highest wood wall
  • Segment 6 (descent) β€” time from fully above highest wood wall to first touch of medium-high decks/platforms
  • Segment 7 (sprint) β€” time from first touch of medium-high decks/platforms to finish

See full stats & charts from the event

🥇 Seth Wang, 12.25 sec

Stats from Seth’s run (compared to top 8):

  • Fewest amount of total touches (52), tied Jared for fewest amount of total foot touches (35)
  • Least amount of total foot contact time (7.08 sec), total hand contact time (5.00 sec), and& total contact time (9.42 sec)
  • Lowest average contact time per foot touch (0.20 sec), average contact time per hand touch (0.29 sec), & average contact time per touch (0.18 sec)
  • Fastest time on segment 2 (1.79 sec), segment 4 (1.17 sec), & segment 7 (2.21 sec)
  • Least amount of total sprint time (4.05 sec) & total ascent time (4.83 sec)

🥈 Michael Sliger, 12.91 sec

Stats from Michael’s run (compared to top 8):

  • Fastest time on segment 1 (0.54 sec) & segment 3 (1.58 sec)
  • Least amount of total descent time (3.29 sec)
  • Second fewest amount of total foot touches (38)
  • Second least amount of total hand contact time (5.79 sec)
  • Second least amount of total sprint time (4.37 sec)

🥉 Jared Nahulu, 13.50 sec

Stats from Jared’s run (compared to top 8):

  • Tied Seth for fewest amount of total foot touches (35)
  • Second least amount of total foot contact time (7.83 sec)
  • Fewest amount of hand touches (15)
  • Least amount of total touches (50)
  • Fastest time on segment 5 (2.79 sec)

📈 Trends

Course 2 men, top 8

Note: R-squared (RΒ²) is a statistical measure that represents the proportion of the variance for a dependent variable that’s explained by an independent variable or variables in a regression model. Whereas correlation explains the strength of the relationship between an independent and dependent variable, RΒ² explains to what extent the variance of one variable explains the variance of the second variable. E.g. if the RΒ² of a model is 0.50, then approximately half of the observed variation can be explained by the model’s inputs.

See full stats & charts from the event

Course 2 women

1) Mikaila Quinn, 19.89 sec (Move to Inspire)
2) Kasia Kilijanek, 20.23 sec (Apex Denver)
3) Taylor Carpenter, 21.78 sec (Apex Fort Collins / Apex Louisville)

Stats from course 2 women:

  • 8 athletes from several states ran course 2 one time each. Athletes earned points relative to their ranking on each course and total points determined the overall winner.
  • The average foot dominance of the top 3 was 66.63%
  • The average sprint dominance of the top 3 was 41.39%

Segments from course 2 women:

  • Segment 1 (sprint) β€” time from start to first touch of wood tower
  • Segment 2 (ascent) β€” time from first touch of wood tower to fully above wood tower
  • Segment 3 (descent) β€” time from fully above wood tower to first touch of blue spring floor
  • Segment 4 (sprint) β€” time from first touch of blue spring floor to first touch of incline wall
  • Segment 5 (ascent) β€” time from first touch of incline wall to fully above highest wood wall
  • Segment 6 (descent) β€” time from fully above highest wood wall to first touch of medium-high decks/platforms
  • Segment 7 (sprint) β€” time from first touch of medium-high decks/platforms to finish

See full stats & charts from the event

🥇 Mikaila Quinn, 19.89 sec

Stats from Mikaila’s run (compared to top 8):

  • Fastest time on segment 2 (3.17 sec), segment 5 (4.33 sec), segment 6 (1.71 sec), & segment 7 (5.75 sec)
  • Least amount of total foot contact time (14.42 sec), total hand contact time (10.79 sec), & total contact time (17.67 sec)
  • Least amount of total sprint time (8.42 sec) & total ascent time (7.50 sec)
  • Highest sprint dominance (42.33%)
  • Highest amount of foot touches per second (2.46/sec) & total touches per second (3.57/sec)

🥈 Kasia Kilijanek, 20.23 sec

Stats from Kasia’s run (compared to top 8):

  • Least amount of total foot touches (41)
  • Highest amount of hand touches per second (1.24/sec)
  • Least amount of air time (0.56 sec)
  • Fastest time on segment 3 (2.00 sec)
  • Least amount of total descent time (3.83 sec)

🥉 Taylor Carpenter, 21.78 sec

Stats from Taylor’s run (compared to top 8):

  • Fastest time on segment 4 (1.33 sec)
  • Highest amount of air time (2.95 sec)
  • Lowest sprint dominance (39.81%)
  • Second least amount of total contact time (18.83 sec)
  • Fewest total hand touches (20)

📈 Trends

Course 2 women, top 8

Note: R-squared (RΒ²) is a statistical measure that represents the proportion of the variance for a dependent variable that’s explained by an independent variable or variables in a regression model. Whereas correlation explains the strength of the relationship between an independent and dependent variable, RΒ² explains to what extent the variance of one variable explains the variance of the second variable. E.g. if the RΒ² of a model is 0.50, then approximately half of the observed variation can be explained by the model’s inputs.

See full stats & charts from the event

Get involved with this project

Got any good ideas and/or know more about MATH stuff (stats, data science, sports analytics, etc.)? Want to help improve the analysis and insights gleaned from speed competition stats? Please send an email/DM to discuss or collaborate more πŸ™‚

Learn more

Apex Speed Clash
Obstacle Course Competition Highlights
Parkour Randori by Amos Rendao
Sport Parkour League


Ryan FordΒ is the author ofΒ Parkour Strength TrainingΒ and founder ofΒ ParkourEDUΒ &Β APEX School of Movement.

Join the discussion!

  • I would agree that it’s too early to determine exactly how and what we can use to know effectively what makes a fast and powerful athlete when it comes to speed competitions/run in Parkour. The one thing I am noticing is that even though each athlete is performing at their best they’re also performing the moves differently, so what if there was a limited amount of moves or you were only limited to do these certain moves to complete the course at speed? Currently the freedom to move as you want and as fast as you want .. is almost as individual as a fingerprint… I do think my thoughts are limiting and it’s in our perspective to try to get everyone to move the same but it’s one way to possibly drill down the stats of how fast we can really go. ‍♂️

  • Archives
    >