How to Use Data to Create Realistic Training Sessions

The development of realistic, task-specific training sessions is hugely important when it comes to preparing elite athletes for the demands of competition. As a means of enhancing athlete performance, I often find myself thinking about how to design more realistic training sessions that are more representative of competition. I’ve written about this previously.

Training aims to prepare athletes to compete. Yet, quite often, a session focuses on improving an individual skill or trait in isolation—frequently in a way that does not match the demands of competition. Of course, for the majority of sessions, training does not have to represent competition; to enhance performance, we often have to isolate either a key skill or a physiological or psychological adaptation. But, by failing to routinely shift the training focus from individual skills to a practice that better represents competition, I wonder if we’re preventing our athletes from reaching their true potential.

Practice as Rehearsal

Looking back on my athletic career, I’ve written quite widely about my disaster in the 4x100m relay at the 2008 Olympic Games, where I was responsible for the disqualification of the Great Britain men’s team. But I’m not the lone athlete to experience this exact failure—the Great Britain men’s relay team was disqualified (or failed to finish or messed up a changeover) at:

• The 2000, 2008, and 2012 Olympic Games
• The 2001, 2011, 2013, and 2015 World Championships
• The 2010 and 2012 European Championships

That’s a lot of missed medal opportunities.

The reasons for these competition errors are potentially varied, but I recall doing very little competitive changeovers. In general, we were very good at passing the baton in isolation; but in competition, we lost this ability—possibly because we had not practiced under competitive conditions. Had we practiced passing the baton under competitive—and therefore truly representative practice conditions—perhaps our flaws would have become more obvious, and we might have rectified them sooner.

Game Scenario Training Drills and the Latest Research

Sports scientists, researchers, and coaches are becoming more interested in how to design and develop representative training sessions. The concept of tactical periodization used most widely in soccer allows for integrated training sessions during which many component skills and abilities are honed as a whole, often under real-world conditions.

Tactical periodization sharpens component skills & abilities as a whole under real-world conditions and may improve performance, says @craig100m. Share on X

Similarly, performance analysis across a range of other sports highlights areas where we might want to focus our efforts on increasing the real-world representations of training. For example, a recent paper investigating skill patterns in Australian football revealed some important gaps in best practices within AFL training session development. Here, the authors looked at game scenario training drills and compared performance during game scenario sessions to actual match performance.

Direct physical contact from an opposition player occurred more often in matches as opposed to training, says @craig100m. #TrainingForCompetition Share on X

The results showed that physical pressure situations (i.e., direct physical contact from an opposition player) occurred more often in matches as opposed to training. Conversely, low-intensity defensive pressure—defined as when opponents were more than five meters away—and situations with no defensive pressure were less common in matches than in training.

There were also differences in kicking. In match play, kicks performed after 0-1s possession time were more common than they were during training, meaning players had less time on the ball. Furthermore, longer kicks and kicks made after collecting the ball from the ground, or when deflected from a player, occurred more often in matches than training sessions. Overall, the majority (~54%) of kicks in a match were executed in less than two seconds, and the number of kicks made while under a tackle doubled the number in training. Also, handballs—the passing action used in the AFL—were more frequent in training than in matches.

Team training sessions did not accurately represent match performance and the skills required to be successful in a game, says @craig100m. Share on X

These results suggest that the AFL team training sessions often did not accurately represent match performance and the skills required to be successful in a game. For example, in one match, players completed twice as many kicks in less than one second than in training. And the kicks occurring in the match were twice as likely to come under competitive training.

This means that AFL clubs may want to include more training scenarios where athletes are forced to kick both in a short amount of time and under competitive pressure. This could be drill-based, or perhaps more realistically, accomplished by modifying the constraints of a small-sided game to drive these conditions (most likely by increasing the number of players per team and reducing the pitch size).

Also, there were more kicks from the mark in game situations; in the AFL, when a player catches the ball from a kick without it bouncing, they can take an unopposed kick. Again, this suggests increasing set kicks in training to mimic the demands of a match better. These results were predominately mirrored in another recent study, again in the AFL.

Training Meet Conditions and the Long Jump

Track and field researchers have used a similar approach, specifically with the long jump. In a paper published earlier this year, a group of authors analyzed male and female long jump performance from 108 different competitions during 1999-2016. Researchers aimed to understand how different situations affected performance and then explore how this information could impact training session design.

The results suggested an interesting global trend: the mean elite long jump performance is decreasing by just over 1cm per year. This means performance has stagnated and is perhaps beginning to regress. Additionally, around 30% of all jumps were fouls. These results tended to be different between rounds: in round one, the average jump distance was lower than in all other rounds, and the total percentage of foul jumps was much lower. This suggests that athletes are typically performing reasonably safe jumps in the first round, most likely because after three rounds those who are beyond a given cut-off are eliminated from the competition.

Another interesting finding is that jump performance across rounds is connected. Instead of viewing a competition as six (or three) discrete jumps, we have to consider it in its entirety. Performance in a preceding jump appears to influence subsequent jump performance; as such, the first round jump becomes increasingly important, potentially setting the tone for the competition as a whole. If the first round jump was a foul, for example, the odds of the next jump being a foul increased by two-thirds. In females, foul jumps were about 50% more likely in rounds four and five than in round one.

In a meet, the outcome of each long jump influences the constraints of the next jump, says @craig100m. #TrainingForCompetition Share on X

Why this inter-connection? It’s because the outcome of each jump influences the constraints of the next. For example, a foul in the first round increases the pressure on the long jumper to achieve two outcomes in the next round—not producing a foul while jumping far enough to qualify for jumps 4-6. The pressure increases because the jumper knows that failure to register a valid jump in round two—having fouled in round one—leaves them in the highly stressful position of pulling off a long enough legal jump with their final, third round effort.

Coaches can use this information to inform specific long jump training sessions. Coaches may wish to prepare athletes for the high-pressure scenario of nailing a second-round jump following a first-round fail in during competitive training sessions. This adds to the training complexity by moving a specific jump session from a series of individual, discrete jumps—which does not represent a long jump competition—into a more real-world scenario. Coaches can design mini-stories or scenarios for training, which break up the repetitiveness of training and provides variation that may enhance learning.

Athletes may handle environmental variations in competition better if they're exposed to different environmental conditions when training. Share on X

Environmental variation also has significant effects on long jump performance. The study mentioned above reported that, for every 1 m/s of wind speed, long jump performance was affected by ~4cm. Wind affects running speed, which requires modifying the long jump run-up. If an athlete is continually exposed to different environmental conditions (typically by training outdoors), then they may be better able to handle environmental variations in competition.

Conversely, if an athlete always trains indoors—a relatively benign, unchanging environment—they will be less likely to tolerate variations during competition. While we might often think that, as coaches, we dictate the constraints of the performance, the ambient environment does too. And we should take advantage of that.

Considerations On the Field and On the Track

There are times when designing representative training is not feasible. Returning to the example from the AFL paper, physical contact was much more frequent in games compared to training. This might be by design—increasing physical contact during training may harm recovery from the previous game and increase the risk of contact-based injuries. Running full-contact sessions during the week between games—which would be more representative of competition—would likely compromise recovery, and contact-induced injuries and niggles would take longer to heal. Clearly, a pragmatic approach is required.

There are some further caveats to designing highly representative training sessions. First, most of the research has focused on elite adult athletes. Using this approach with developing or non-elite athletes may not be optimal because it’s not clear whether the constraints and performance dynamics are the same at these levels as they are at the elite level.

In the AFL, for example, we don’t know whether match dynamics at the pro level are equivalent to those at the youth level (i.e., age 16-18 years). Do the lower physiological capacities of players influence the match demands? For example, if the players have lower levels of physical fitness, they are less likely to close down the opposition kicker promptly, which may decrease the proportion of kicks occurring in less than one second during a match.

Also, competition rules may be different at the developmental level. In track and field, for example, the height of the hurdles is lower, which may alter certain performance constraints. In team sports, the pitch size or number of allowed substitutions may affect match play.

When it comes to sprinting, we know that stride length and frequency and their related constructs (e.g., ground contact time, flight time, etc.) dictate sprint performance. We have a pretty good idea what the optimal average values are for these—Ralph Mann details them extensively in his book. We’re less clear, however, on what these variables should look like in a developing athlete.

So, if an elite male sprinter has a 2.4m stride length, what should a top under-18 sprinter’s stride length be? Also, how do factors such as training age and maturation affect this? Of course, the solution is for teams and governing bodies to analyze players and athletes at this level, and then use this to inform optimal athlete development protocols (an approach that would be expensive in both costs and resources).

Finally, when attempting to apply this approach to athletics, we need to consider individual variation. While we know what elite athletes achieve on a certain measure on average, there is often a considerable acceptable bandwidth around this average. What comprises this bandwidth isn’t always clear, and we run the risk of becoming over-prescriptive with this data. Instead, we should use this information as a guide—where the athlete is now and where should they progress to—and then use the coach’s knowledge of the individual to understand how the athlete’s nuances may affect this “optimal” training.

Since you’re here…
…we have a small favor to ask. More people are reading SimpliFaster than ever, and each week we bring you compelling content from coaches, sport scientists, and physiotherapists who are devoted to building better athletes. 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. — SF