Speed, deceleration, acceleration and shot quality characteristics during Australian Open 2021-2023

/in /by

Dear Reader,

 

I’ll admit it that the last few posts have been a bit nerdy, focusing on average data for speed, acceleration and deceleration in tennis matches, followed by the last blog looking at peak data for speed, acceleration and deceleration during popular speed and change of direction tests.  But stay with me here, as there is a good reason to go granular with the detail.

 

I have to thank Joey Guarascio for his insights on a Pacey Performance Podcast Episode 423, where he was discussing all things deceleration training and its application to college football.  I’m a bit shocked to see that I didn’t actually get around to featuring it in one of my 40+ PPP review blogs.  I definitely need to do that!

 

Anyway, he had a really neat way of laying out his philosophy where he spelt out that “our capacities are going to feed our general skills, and our general skills are going to feed our sport skills.

 

Capacities

 

  • Peak Force – impulse
  • Rate of Force Development (RFD) – time to express force
  • Stretch shortening cycle (SSC) – how elastic are you?
  • Endurance – repeatability of an action

 

General Skills

 

  • Acceleration
  • Deceleration
  • Change of Direction (COD)
  • Maximum Velocity

 

Joey went on to say that like a lot of coaches, he thought the team were getting a lot of high speed deceleration from their COD work, and that wasn’t the truth.  When he looked at the catapult data and looked at the highest peak deceleration, what he found was that they were not maximal when compared to a 10 Yard deceleration to a full stop.  He saw -5 to -6 ms/s on those decelerations versus in the COD drills they were around -3-5 ms/s.

 

“The way I look at is you want to create thresholds and you want to create RESERVES.  Just like we have a speed reserve, I also want a deceleration reserve so when they have to do -3.5 ms/s decel and they have to do it 30 times in a match that’s nothing to them as it’s only 50% of what their maximum is.”

 

 

This got me thinking immediately, and right off the bat I had a few questions:

 

  1. What is the peak rate of deceleration during our most demanding COD test – the m505?
  2. What is the peak rate of deceleration in a tennis match?

 

Well, in the last blog I  just shared with you the peak data for the m505 – see below:

 

The group average peak rate of deceleration was -9.78 ms/s, which typically occurred during the penultimate step (10/16 athletes).  By comparison, it was -10.28 ms/s for the males.

 

So let’s get to point 2.   What do we see in a tennis match?  I didn’t have anything to use a reference until I came across a journal article – Lateral End-Range Movement Profile and Shot Effectiveness During Grand Slam Tennis Match-Play (2025).

 

Any research paper that is based on Hawk-Eye match data taken from a Grand Slam professional tournament will always grab my attention. Sadly the data reported was “average deceleration” and “average re-acceleration,” so it didn’t quite give me a definitive comparison for “peak” accel and decel data.  It’s always a bit tricky for me to compare the kinematic data when I don’t know the ins and outs of the data analysis process.  It stated on the Data Collection selection of the paper, “peak speed was identified as the starting point in each movement cycle.  The minimum speed around the COD was identified, and the average deceleration between peak speed and minimum speed was calculated.  Furthermore, average acceleration from minimum speed to peak speed post COD was also completed.”

 

Just before I get to the data, it is worth pointing out that:

 

  • Top 10 ranked males are faster and accelerate harder during lateral end-range COD in match-play compared to those ranked > 10.
  • Top 50 ranked males are able to decelerate harder during lateral end-range COD in match-play compared to those ranked > 50.
  • Peak speed may be influenced by deceleration ability, where poorer deceleration capability results in athletes self-regulating their peak speed in anticipation of stopping or changing direction
  • The enhanced deceleration ability of better ranked players may help explain the observed differences in peak speed.

 

The Data

 

Results show that tennis athletes record average deceleration intensities of -9.0 ms/s and acceleration intensities of 9.9 ms/s during COD tasks.  [This is a lot more than the average deceleration of -4.7 ms/s I recorded for our cohort of females for the m505 and the -4.2 ms/s reported for Wimbledon main draw matches from in house data shared by LTA].  The value of -9.0 ms/s is almost as high as the peak deceleration of -9.78 ms/s recorded on the m505.

 

Of the limited comparable data available in the literature, elite soccer players have been shown to produce average deceleration intensities of -4.99 ms/s in testing-based environments, with a theoretical maximal acceleration of ~7.7 ms/s.  This highlights the physicality of professional tennis and underlies why COD performance is a large focus of tennis training and conditioning programmes.

 

My thoughts?

 

Well honestly I’m still a little stuck….from the data I have seen on tagged matches played using similar ball and player tracking technology at the National Tennis Centre in pro males, the range for deceleration on the tagged clipped shared was -4.7 ms/s to – 7.2 ms/s.  I’ll need to check if it is average or peak data.

 

The specific COD tasks that were analysed in this paper were referred to as “Lateral end range movements.”   These are thought to be some of the most physically demanding COD tasks on a tennis court.  If you want more info you can download a previous paper An application of clustering to classify movement patterns in mens professional grand slam hard court tennis which goes into more detail.

 

 

 

One thing I always said in my discussion with Jonas Dodoo about high powered tennis actions, is that unlike the m505, in a “running forehand” out wide we should really refer to it as a “leaping forehand” as there is a clear take off, flight and landing phase (see above).  Given the time pressure that tennis players are under to brake, plant and re-accelerate from this leap, my hypothesis is that tennis does indeed involve considerable demands on a player’s deceleration ability.

 

In house data from the LTA suggested that as a worse case scenario, elite players have to stop within 2 metres out wide from a running (aka leaping) forehand with an approach velocity of 6 m/s.  This is typically over a distance from the middle of the baseline to the outside tramline (5.49m) which is comparable to the 5m distance we use for the m505.

 

If you compare my data for the female cohort below, you will see that the best distance to stop was 2.67 m, which was given a weighting of 92/100 from the Speedworks normative data.   This athlete had a peak velocity of 5.41 m/s coming into the cut.  The average for the group was 3.77m distance to stop.

 

 

All I can say is that if elite male players are getting an average deceleration of -9.0 ms/s during Grand slam matches then goodness knows what the peak data would be?  For now, until I see more data, I’ll keep my eye on it.  If my athletes are hitting an average deceleration of -4.47 ms/s on the m505 and a peak deceleration of -9.78 ms/s, I’m confident that they are being prepared for the deceleration demands of tennis match-play (from a running point of view).  But the loads on the limbs from a leaping and landing point of view – well that’s a discussion for another day.  But let’s just say it is a violent braking manoeuvre.

 

 

 

Hope you have found this article useful.

 

Remember:

  • If you’re not subscribed yet, click here to get free email updates, so we can stay in touch.
  • Share this post using the buttons on the top and bottom of the post. As one of this blog’s first readers, I’m not just hoping you’ll tell your friends about it. I’m counting on it.
  • Leave a comment, telling me where you’re struggling and how I can help

 

Since you’re here…

…we have a small favour to ask.  APA aim to bring you compelling content from the world of sports science and coaching.  We are devoted to making athletes fitter, faster and stronger so they can excel in sport. Please take a moment to share the articles on social media, engage the authors with questions and comments below, and link to articles when appropriate if you have a blog or participate on forums of related topics. — APA TEAM

 

=> Follow us on Facebook

=> Follow us on Instagram

Deeper dives into the Speedworks PSR reports

/in /by

In case you missed it

 

At the start of this 2025 I wrote a two part blog on the Speedworks Linear speed and COD reports.  You can read the full blog articles here and here. I wrapped up the year with some insights on the kind of kinematic data I have been able to access on tennis matchplay.

 

If you read my last blog, you will recall that I shared some typical “average” data for accels and decels in terms of the speeds reached and also the frequency during competitive matches.

 

To repeat it here – in a 2-set match we will typically see:

 

  • High intensity tennis movements (>2.5ms/s) – e.g., 30 accels / 17 decels
  • Very high intensity tennis movements (>3.5ms/s) – e.g., 7 accels / 11 decels
  • Total high intensity efforts (average 59 and maximum 90)
  • Max speed 6.1 m/s – but 66% of match is played at speeds below 1 m/s
  • Average speed 2.9 m/s
  • Average accel 4.0 ms/s
  • Average decel -4.2 ms/s

 

Now this is certainly very useful to know.  25-30 high intensity tennis movements per set.

 

 

And certainly, this enables a coach to start comparing match demands to training sessions – if you are fortunate enough to have some sort of camera ball and player tracking system (see above).

 

Don’t let that be an excuse though, if you can categorise the game scenarios by those that have the highest physical demand you can start to have conversations with tennis coaches to ensure you plan for a few days in the week that have a higher running demand.

 

I’m not interested in average

 

I wanted to go a stage further to know what the “peak” rates of acceleration and deceleration in tennis matches are.  It’s good to know what is going on at the “average” level but from a physical preparation standpoint I want to know what the peaks are, or “worse case scenarios.”   And I want to know how stressful moving around the court at “Gamespeed” is compared to a traditional linear sprint and a 180-degree change of direction test.

 

I’m making an assumption that a 10m sprint and a m505 change of direction test are appropriate key performance indicators of physical preparedness to meet the game demands.  Put another way, I’m saying that if someone is really fast on these tests, I’m more confident in their ability to handle the demands of moving around the court and certainly at a lesser (energy) cost to them.

 

 

So before I go into the match data for accels and decels I thought it was important to look at a typical 10m sprint and m505 in this blog to know what the maximum or “peak” outputs are during these tests.  At APA we work with Speedworks who provide reports to interpret the kinematic data captured on video for these tests.

 

10m sprint

 

For the 10m sprint Speedworks analyse the first six steps after the initial starting motion.  For elite sprinters they may cover the 10m with six steps but for our tennis athletes they were mostly short of 10m by the time they had carried out six steps.

 

 

In the cohort of athletes we tested I was interested to see what the peak rate of acceleration was.  This always occurs at the “start.” The Speedworks protocol has the athlete start in a staggered stance, so the start usually involves some sort of unweighting of the front foot followed by a push of it into the ground.  Step 1 is then counted as the step that hits the ground inside the 10m designated area indicated by the yellow cones between the timing gates.

 

As you can see, the fastest athlete who clocked 2.13 seconds for 10m also had the second highest rate of acceleration off the start at 10.1 ms/s.

 

 

The group average peak acceleration was 8.7 ms/s for the females.  By comparison, it was 9.3 ms/s for the males.  This is comparable to the 9.5 ms/s that elite sprinters are thought to accelerate at immediately out of the blocks.

 

I noticed that acceleration would then drop to 2-4 ms/s over those first few steps inside the yellow cones.  Please remember this last point, as this is typically what I believe other coaches are reporting when they describe the typical rates of acceleration to a ball during tennis matches (2-4 ms/s).

 

M505

 

The modified 505 (m505) consists of a 10m approach and 5m re-acceleration in the opposite direction using a 180-degree cut.  The time recorded is the time to complete the last 5m before cutting and the 5 metre re-acceleration out, 10m in total.

 

 

At APA we actually use a 10m approach distance (as technically speaking the full width of the doubles court is 10.97m).  But the protocol photo above gives you the idea.

 

I like to compare the 10m linear sprint to the m505 time and see the change of direction deficit score; the difference between the m505 (target 2.45sec for m505 and 1.85sec for 10m sprint females) – which would be a 0.60sec COD deficit in this example).

 

A good change of direction (COD) deficit score is a low one, ideally close to zero, meaning there is minimal difference between the time it takes for an athlete to run a linear sprint and the time it takes them to complete a change of direction test.

 

A high score indicates a greater loss of time during the turn, suggesting poor COD ability relative to linear speed.  I have found that a difference of less than 0.5 seconds is pretty good.  Our female cohort has COD deficit scores of 0.20-0.35s.  By comparison the range for the males was 0.10-0.40s).

 

For the m505 sprint Speedworks analyse all the steps after the initial starting motion.  The average number of steps prior to decelerating (indicated by the anti-penultimate step) was six steps, (6.44 going to the left, and 6.13 going to the right).   The range was four to nine steps prior to decelerating.  This data isn’t presented in the table below, I just took it from the raw data step-by-step parameters that I asked Speedworks for.  This is just to say that there are variations in individual strategy that result in a large range of acceleration steps prior to slowing down, with six being the average.

 

 

In the cohort of athletes we tested I was interested to see what the peak rates of acceleration and deceleration were.

 

As you can see, the fastest athlete who clocked 2.44 seconds for m505 also had the third highest rate of average deceleration (-5.07 ms/s).  This compared to the group average of -4.47 ms/s.   This average deceleration value is similar to what I previously reported for a tennis match (-4.2 ms/s).

 

The group average peak rate of deceleration was -9.78 ms/s, which typically occurred during the penultimate step (10/16 athletes).  By comparison, it was -10.28 ms/s for the males.

 

 

What was also interesting was to compare the acceleration data and re-acceleration data (as they come out of the cut).

 

The table above doesn’t show the acceleration data for the 10m approach.  But from analysis the highest acceleration typically occurred at the start (2.99 ms/s for females versus 3.52 ms/s for males).

 

In comparison, the peak rate of acceleration occurred during re-acceleration out of the cut (10.07 ms/s) during the concentric plant step in the majority of cases.

 

Summary of Peaks

 

Group average data:

 

10m sprint

 

Peak rate of acceleration: 8.7 ms/s

 

M505

 

Peak rate of acceleration: 2.99 m/s (start)

 

Peak rate of deceleration; -9.78 ms/s

 

Peak rate of reacceleration: 10.07 ms/s (concentric plant step)

 

One take away is that when an athlete knows they are performing a 10m sprint they seem to “get off the mark” with an initial rate of acceleration to overcome their inertia (8.7 ms/s) not unlike the rate of acceleration thought to occur immediately out of the blocks with sprinters running a 100m race (9.5 ms/s).

 

For some reason, on the m505 athletes don’t seem to have the same “pop” to their initial movement.  They accelerate towards the change of direction at only 2.99 ms/s (similar to what we see in terms of movement on the tennis court!).  They tend to peak in deceleration at around the penultimate step (-9.78 ms/s) and peak their acceleration during commencement of the re-acceleration out of the cut during the concentric part of the plant step (10.07 ms/s).

 

In the final blog on this series, I will share the peak data for tennis matches!!  Coming soon!

 

Hope you have found this article useful.

 

Remember:

  • If you’re not subscribed yet, click here to get free email updates, so we can stay in touch.
  • Share this post using the buttons on the top and bottom of the post. As one of this blog’s first readers, I’m not just hoping you’ll tell your friends about it. I’m counting on it.
  • Leave a comment, telling me where you’re struggling and how I can help

 

Since you’re here…

…we have a small favour to ask.  APA aim to bring you compelling content from the world of sports science and coaching.  We are devoted to making athletes fitter, faster and stronger so they can excel in sport. Please take a moment to share the articles on social media, engage the authors with questions and comments below, and link to articles when appropriate if you have a blog or participate on forums of related topics. — APA TEAM

 

=> Follow us on Facebook

=> Follow us on Instagram