The following article has been provided by Bob Smith, physical trainer at TeamBath-MCTA

In the majority of sports the competition and training schedule is known in advance. If you have good methods of objectifying an athlete’s current state of adaptive health, you can adjust training and recovery methods to manipulate the athlete’s biological state. Easier said than done but achievable all the same. Tennis tournaments are a totally unknown with uncontrollable and variable daily on-court loads which can anything from 60 minutes to 5 hours (2 matches in same day x 2.5 hours). Tournaments can be 5 days of anything from 60-250 minutes of tennis, the intensity of which is variable. So the more traditional methods of collecting and analysing data in team or linear endurance sports are of little help to directing training outside of competition. The goal of training load data collection is to provide information pertinent to reviewing for planning and training. The decision on when to train, how long to train and how intense it should be on any given day or week needs to take into consideration 3 broad questions:

Q1. What are we preparing the player to be able to handle?

• Needs analysis of the sport (singles/ doubles/ both)
• Gap analysis of the player vs the competition
• Tournament schedule and upcoming events

This provides context of what needs to be done to be successful at a specific level. It starts with a detailed needs analysis of that player. The picture is built through retrospective analysis of subjective and objective data collected for that individual. A gap analysis then identifies what needs to be done to get to where we want to go. This underpins every decision that we make regarding training and schedule.

Q2. How physically fit is a player at that moment:

• Tennis fitness
• Physical fitness

Q3. How fatigued is the player at that moment:

• Fatigue is multifaceted and therefore our monitoring system must be sensitive to different types of fatigue.

An effective load monitoring system needs to answer these questions if it is to be an applicable tool to aid in planning and reviewing training. Using data to prescribe training is by no means a hard science, especially when assessment tools are limited. With a subject number of 1, there is no way of knowing if our intervention is optimal or has even had a positive effect. This article outlines how we are developing our load monitoring systems for touring tennis players.

Part 1: Data Collection

Training Load

In the absence of GPS for every training session, training impulse (TRIMP) is an effective way to uniformly characterise training loads across different modes of exercise. For this reason it can be easily applied to quantify training in tennis which has a broad range of training modalities. The data that we collect training load data on is:

  • Tennis Training 1
  • Tennis Training 2
  • Singles Matchplay
  • Doubles Matchplay
  • Speed / Plyometrics
    Strength Training
  • Conditioning
  • Prehab / Rehab
  • Recovery

For each of these we use Borg’s Rate of Perceived Exertion 1-10 Scale and duration (minutes) to characterise intensity and volume respectively. RPE x duration gives us training load for that session. This data allows us to see how training loads are distributed across types of training and how frequently certain sessions are done. When training players remotely, it also helps with advising players on the decisions they make on the road.


Subjective Ratings

Subjective wellness ratings are critical if we are to know how the player reacted to the loads they reported and gives an indication of fatigue. Some empirically sound scales that we use are:

  • Energy Levels
  • Muscle Soreness
  • Mood
  • Quality of Sleep
  • Hours of Sleep
  • Readiness to Perform (an overall feeling and something I’ve added in)

Injury Data

We want to know if a player is carrying an injury, however big or small. Injuries in tennis are predominantly non-contact overuse injury and in that respect accumulate over time. Therefore with the correct intervention, we should in theory, be able to prevent niggles reaching a critical threshold where time off court is required. More easily said than done but an aspiration all the same. In a world of indirect measurements and estimations, niggles and injuries provide us with definitive information that tolerable load has been exceeded either from a local (e.g., muscle not strong enough) or systemic (whole body is not able to recovery) point of view. This valuable information helps to refine the player’s physical preparation programme to address these weaknesses.

Part 2: Basic Analysis

Collection of accurate data is one thing, but the strength of load monitoring is in asking the right questions, analysing the data to give you the answers and then modifying training or schedule accordingly. The next few paragraphs outline how we have tried to develop a useful tool for achieving these goals for tennis.

Much like other sports, we collect data in calendar weeks for the players. Tournaments are nearly always wrapped up by the Sunday of each week and our training programme at Bath often starts on a Monday so it makes sense to apportion the data collection into 7 day microcyles from Monday to Sunday. Weekly load is calculated as the summation of all session loads in that calendar week. Training monotony is a measure of variability of training (average daily load divided by standard deviation Monday-Sunday). A low training monotony with high day-to-day variation is good. We do plan training weeks to minimise load monotony but it is actually a strong characteristic of tour tennis. As such being able to handle load monotony is critical to being a successful tennis player.

Using basic analyses we can give a snapshot as to what occurred in that training week by splitting total load into different types of training (e.g – tennis, strength, conditioning, pre-habilitation work). Thorough data collection over a long period of time builds the picture of what we are preparing the player to handle physically (Q1). This is extremely useful when setting out a training philosophy but we need to further analyse the data to inform planning of training.

Part 3: Making the data useful for tennis

Assuming our needs analysis for that player is set and we can see their competitive schedule in front of us, the two questions at any given time that should dictate how we plan training are:

  • How fit is that player at that moment
  • How fatigued is the player at that moment

Q3) is more easily addressed so we’ll do that first. In our load monitoring system, fatigue is characterised by our subjective wellness ratings. There are a mass of tools available that can be used to characterise fatigue more accurately with players (e.g., heart rate variability, submaximal heart rate tests) but the practicality of using these methods for a player travelling without coaches makes the data inconsistent and therefore useless or worse, misleading.

While there are many wellness questions available, we’ve abbreviated our questions to some commonly used and validated scales. Healthy players will feel soreness, have bad moods, poor sleep and generally feel bad from time to time. Variations in these values give valuable information as to what patterns of training make a player feel at their best and on the flipside what makes them feel poor. If more than 3 of these factors drop sharply this is alarm bells that the player is overly fatigued and not responding well to the imposed demands. We have additional, player specific testing and monitoring batteries for when players return to Bath to train. The goal of all of these tests is to highlight any physical or psychological issue that the player may be managing that as a coaching team we be able to help them work through.

Establishing analyses to answer the final question…

How fit is that player at that moment

…is a little more complex but very important in managing a player. We can measure different types of physical capacity fitness with timing gates, yo-yo and repetition maximum tests. They certainly have an influence on training capacity but what we are more interested in here is the amount of total load that a player is used to, specifically how much tennis load they are accustomed to handling. To do this we have made adaptations to a well-known principle called training-stress balance. Training stress balance takes the last 7 days of physical loading and compares it to the 21 days preceding it. If the most recent week’s training load exceeds 200% of the average of the 3 weeks preceding it, the most recent week is highlighted as a high risk of injury or illness.

What I like about the training stress balance analysis is that it makes a prediction of the athlete’s training tolerance. It’s a logical assumption that I deem necessary if we are to be able to subsequently prescribe training loads. If my remit as a tennis sports science service is to provide information to keep the player healthy then I believe there is mileage in preventing training stress balance exceeding 200% in training. If it occurs in competition then I should be able to highlight to the coaching staff when this is occurring so we are all aware and can act accordingly around matches and post-tournament. So we took the training stress balance principle and modified it.

Modification 1: Rolling training stress balance

Training stress balance is normally measured in calendar weeks, but tour tennis doesn’t work like this. We need to be able to highlight intensive periods of training or competition that span calendar weeks. What I now use is a rolling training stress balance to highlight large loads: Instead of calendar weeks I use the average of the previous 7 days compared to the average of the 21 days preceding it. We can apply the same 200% rule to see if any 7 day period has exceed the assumed tolerable load of the athlete.

So I now have a more accurate retrospective analysis of what has been done adapted from an empirically sound marker to show me when excessive loads may make a player susceptible to injury. However, a player and coach are not going to appreciate knowing that they’ve already made an error and training has to be massively reduced in the following week to compensate. We need to be able to plan proactively. In theory, every athlete at any moment in time has a minimal effective load (below which he/she will undertrain). We then have a theoretical optimal load and a maximal tolerable load (between which we can refer to as functional over-reaching). Training loads in excess of maximal tolerable load are known to induce non-functional overreaching characterised by chronic under-recovery, underperformance and injury. Our goal is to identify and design training periods that fall between the minimum effective load and maximum tolerable load to induce adaptation. This has led us to a concept of using data to establish tolerable load.

Modification 2: Estimating tolerable load to plan loads

We are assuming that training tolerance on any given day in a trained athlete is the average daily load for the previous 21 days. Provided the athlete did not get injured in this period, I am assuming also that they have adapted to this chronic loading period and that the data will indicate to me their current level of total load tolerance. I can calculate this as tolerable daily tennis load, tennis time and also total load.

Prescribing Training Using Daily Tolerable Load

Depending on how long the player is back training for, we prescribe the allocated loads for that period of time. i.e., player has a 21-day average daily load of 800 units (e.g., 120 minutes of combined match play and practice per day at RPE 5- Hard plus 200 units of physical). Using this as a basis, a 7-day training week with 2 days off should have a total load of 4000 units. A conservative overload of this would be +10% so 4400 units. We then allocate 4400 units of work to tennis and various forms of physical based on the player’s needs analysis.

As an example, if we have a player for 3 weeks, a conservative training plan would look like: week 1: 4400 Units; week 2: 4840 units; week 3: 5280 units adding 10% load each week. The player would return on the road with an increase in training tolerance and well recovered. We could go more aggressive than this putting a player deep into functional overreaching but would have to allow for additional recovery time to compensate after the block. If we don’t we can expect the player to underperform until they have recovered. Neither of these scenarios are right or wrong, the needs analysis for the player will dictate the strategy. If load exceeds 8000 units in week 1 (200% rule) the risk of under-recovery and injury is high. Whilst it seems obvious to avoid such sharp increases in physical load, it is easily done. If you take a player returning from tournament tennis and you play 4 hours per day Mon-Fri at an RPE of 5-Hard, you will get a tennis load of 6,000 units. Add some physical and any hitting preparation for a tournament over the weekend, 8000 units is easily exceeded.

When we were getting it wrong, players would leave our training centre having done a lot of training but potentially over-trained and susceptible to injury during their return to competition. There’s always the temptation to do more with a healthy player. My belief is that a well recovered, highly adaptive player with high concentration, focus and motivation will train and progress better than a player on the edge of tolerable load but playing and training a lot more. This is a training philosophy that prioritises longevity over short-term gains.

Reviewing the block

Load monitoring is then used to measure the effect of the intervention. Wellness ratings will give an indication of how fatigued the player became and the rate at which they recovered following the block. Physical testing will tell us the impact of any physical training alongside technical practice

There are lots of ways to manipulate training loads to evoke different adaptations in players. For example we may want to expose a player to the loads expected from winning a tournament if they’ve had a string of poor results and need match fitness. We may plan to do this in week 3 of a physical block- therefore we need approximately 16-18hrs on court plus light physical around it. Or we may want to prioritise strength gains in a player as part of their training goals- therefore we need 3 x heavy strength training sessions 48-72 hours apart for this training period with reduced tennis intensity to allow positive adaption processes. The goal is dictated by the needs and gap analyses but the tolerable load is dictated by physiological principles that cannot be abused. I’ve heard and read coaches say that tennis is unique and training loads must be super-high all the time to mimic tournament tennis. Unfortunately tennis players are also human beings and their ability to recover and adapt to training adheres to the same principles as other athletes in other sports. The coaching staff and player needs to make detailed decisions as to the goals of the athlete and allocate the available training loads accordingly.

Future directions:

Future development of our system has two aspects. One aspect is to improve the tools that we have to predict biological status. Above all these methods have to be simple to administer, cost effective and where possible empirically validated if they are to be successfully used for tennis players. Secondly, as a coaching team, we need to continue to improve our interpretation of what the data is telling us and constantly develop our routines and practice to manage and advise the player. The limiting factor here is not the technology but our ability to ask the correct questions, interpret the answers accurately and modify training and recovery with the best empirically driven protocols available to us.