Large-Sided Games Don't Replicate Worst-Case Match Demands
What the worst-case scenario research means for how you structure training.
One of the most common assumptions in football training is that if a practice game looks similar to the match, then the physical demands will naturally take care of themselves.
If the pitch is large enough, the player numbers are high enough, and the area per player resembles competition, practitioners may assume the session is automatically preparing players for match demands.
But a new study from Lukas Herzig and colleagues challenges that assumption quite strongly.
The study investigated whether 9v9 and 10v10 large-sided games, using a similar area per player to official matches, could replicate the worst-case scenarios experienced during professional football matches.
Rather than analysing average match values, the researchers focused on the most physically demanding 1, 3, and 5 minute periods of play.
Large-Sided Games Fell Short of Match Demands
The researchers monitored 21 professional players from a professional German club across official matches and large-sided training games over a 19-week period. The games used approximately 320 m² per player, very similar to official match conditions.
Across every position and every time window tested, large-sided games underperformed official matches.
Total distance, medium-speed running, high-speed running, and sprint distance were all significantly lower during training games compared to worst-case match periods.
The effect sizes weren't borderline either, they were large to very large.
The gaps were most pronounced at higher speeds. During the worst-case 1-minute window, wide midfielders in official matches covered nearly 40 metres at sprint speed (>25 km/h).
In large-sided games, that figure dropped to around 20 metres. For high-speed running, the story was similar across all five positions studied.
Accelerations and decelerations told a similar story. The mechanical demands, the explosive stops and starts that stress muscles and tendons, were also consistently underrepresented in large-sided games across most positions and time windows.
Why Does This Happen?
Even with a comparable pitch size, several things conspire to reduce intensity in training games.
Competitive pressure is lower, and players pace themselves differently when nothing is at stake.
Knowing the bout duration in advance changes how hard people push. Real matches also generate unpredictable bursts of intensity when possession changes, and training games simply have fewer of those high-pressure transitions.
Underlying all of this is psychological context: the urgency of a counter-attack or the desperation of a last-minute press doesn’t replicate in a training environment, no matter how well the drill is designed.
Interestingly, increasing pitch size beyond match dimensions can push locomotor output higher. But doing so sacrifices technical-tactical specificity, which is arguably the main reason you’re running large-sided games in the first place.
So What Should Practitioners Do?
The authors suggest supplementing large-sided games with running-based drills to hit worst-case scenario demands.
To replicate the WCS1 demands of a wide midfielder, for example, a drill might involve a 50-metre sprint at full speed, followed by a 56-metre run at high speed, then an 80-metre run at medium speed, with around 10 seconds of active recovery between each effort.
The problem many coaches will immediately raise is that running drills strip away football context entirely. There is no ball, no decision-making, no tactical relevance.
For coaches who have built their entire training philosophy around game-based methods, prescribing isolated runs feels like a step backwards.
But the data forces an uncomfortable question. If large-sided games cannot reach the peak intensities of match play, and purely football-based methods cannot close that gap, then something has to give.
The choice is not between a perfect solution and an imperfect one. It is between acknowledging the physical preparation gap and addressing it pragmatically, or ignoring it in favour of a philosophy that leaves players underprepared for the hardest moments of a match.
For acceleration and deceleration demands, there is at least a more football-specific alternative. Smaller-sided games and medium-sided games, specifically 4v4 or 5v5 formats, have been shown to match or even exceed the mechanical demands of worst-case scenarios.
The inverse relationship between speed and acceleration intensity means lower-velocity contexts actually produce higher acceleration peaks, so coaches do not necessarily need to abandon game-based methods to address that part of the problem.
The Takeaway
Large-sided games are a valuable training tool. They develop tactical understanding, technical execution, and positional relationships in a realistic environment. Nobody is suggesting you scrap them.
But they are not necessarily a conditioning tool for worst-case scenarios. If you’re relying on them to prepare players for the most physically demanding moments of a match, there may be a gap in your programme.
Know what your large-sided games are doing well. Know what they’re leaving on the table.
Football Performance Network
The tension this research surfaces, between game-based philosophy and genuine physical preparation, is exactly the kind of problem that rarely has a clean answer in practice. How you navigate it depends on your club’s culture, your players’ individual demands, and how much you can push back on the coaching staff.
That conversation is what happens every day inside the Football Performance Network, where 70+ physical performance coaches and sport scientists working in professional football share ideas, challenge each other’s thinking, and work through the problems that research alone cannot solve.
The next intake opens in July 2026. Join the waitlist
Reference: Herzig L, Stöckert J, Lochmann M, Rumpf MC (2026). Large-sided games do not replicate the positional worst-case scenarios of official matches in professional football. Frontiers in Sports and Active Living, 8:1797322.