You've probably coached that player. Technically brilliant in training — cleaner pass, sharper read, faster execution than anyone else in the squad. Then they step into a real match and it falls apart. The decisions come too late. The ball goes to the wrong place. The moment is gone.

Professor Michael Richardson has a precise explanation for why that happens — and a research programme built to fix it.

Richardson is a Professor at Macquarie University's Performance and Expertise Research Centre, one of the world's foremost ecological psychologists, and a man who has been coding since the age of 10. His work sits at an unusual crossroads: deep ecological theory on one side, cutting-edge AI and LiDAR hardware on the other. When Ian Renshaw and Keith Davids sat down with him for Episode 94 of The Constraints Collective Podcast, the result was one of the most practically provocative conversations the show has produced.

Give Someone a Ball and an Open Field.
Nothing Happens.

Richardson opens with an image that should reframe how coaches think about practice design. Put an athlete in an empty field with a ball and nothing else. Within seconds, they're bored. Nothing interesting emerges. Then add lines. Add goals. Add rules. Add teammates and opponents.

"All of a sudden, you've got a beautiful game. It's different every single time."

This isn't a metaphor. It's the mechanism. Constraints don't limit behaviour — they generate it. The state space of what is possible shrinks, but within that space, there is always an infinite number of solutions. The game's complexity is not designed play-by-play. It emerges from the interaction between simple boundary conditions and human movers.

The same process drives flocks of birds, schools of fish, and swarms: very simple formal rules — sometimes as few as three equations — and yet you get highly dynamic, context-sensitive behaviours. Complex order from minimal constraint.

The practical implication: coaches who design elaborate, choreographed drills may be doing less than coaches who choose constraints well and let the game do the teaching.

Coordination Isn't
Always Better

Richardson's early career work produced a finding that should sit uncomfortably with coaches who prize synchronised team play above all else.

In a now-famous series of experiments, people sitting near each other in rocking chairs — with no instruction and no physical contact — spontaneously synchronised their rocking rhythms purely through visual coupling. Perceptual information fields, he argues, are just as powerful as physical forces in shaping the patterns that emerge between people.

"These kind of perceptual fields — these informational fields — are just as powerful as physical fields in terms of constraining the patterns that can emerge."

But here's the uncomfortable part: constant coordination is a liability. "If you're always coordinated, then you're going to fail a lot." What distinguishes elite teams isn't seamless synchrony — it's context-sensitive, intermittent coordination. The ability to move in and out of coupled states fluidly, depending on what the game demands in this moment.

Applied to the relational play versus positional play debate that recurs across coaching circles, Richardson is precise: by definition, everything is relational. Performers don't perceive in terms of absolute position. Action possibilities are all relational. You can use positional structure to set initial conditions. What happens once play begins is always relational, regardless of the system.

Movement Is the Best Measure.
Always.

A strong position runs through Richardson's work: if you want to know what an athlete is actually doing — cognitively, perceptually, under pressure — measure their movement.

"Heart rate, skin conductance, EEG — they're so confounded with other variables. But action — that is a direct window into what they're doing. If you want to measure if someone is performing well, under cognitive load, situational awareness — measure their movements. Those will always give you the best way of classifying performance. Always. Every single time."

One of his current PhD students is classifying cognitive load purely from movement fluctuations, without any physiological signal. The logic: every cognitive task is fundamentally a coordination problem. Perception and action cannot be separated. The movement signature of a struggling performer is always there — if you know what to look for.

Predicting Decisions
Before They Happen

Which brings us back to the player who looks spectacular in training and falls apart in a game.

Richardson's lab is building AI models that predict player decisions one to two seconds before they occur. Using fine-tuned computer vision trained on AFL and football footage, the models track full-body pose and 3D game states from tactical camera angles. For binary events — will this player pass? — accuracy runs at 70–80% for players with sufficient data in the model.

The goal is not just prediction. It's explanation. What information was this player accessing — or failing to access — in the one to two seconds before the decision?

Elite performers don't see the present moment. They see four or five steps ahead, reading the information available now to anticipate what will be required then. Richardson reframes this not as "situational awareness" but as perceptual attunement — the calibrated ability to couple with the informational structure of the environment ahead of time. The training environment player is technically excellent. They just haven't developed the attunement. And if your practice environment doesn't require early information pick-up, it will never develop.

The Next
Three Years

Richardson's team is building hardware: solid-state LiDAR sensors combined with RGB cameras — the same technology used in autonomous vehicles — capable of real-time 3D tracking of every player across 120 degrees of field at 50–100-metre range. No post-processing lag. Half-time data available.

The coaching application is striking. Imagine the scenario: a VR helmet lets you zoom into any location on the field in real time, watch from any angle, stand in any location. Not a video review. Standing in the game.

And instead of generic analytics software, coaches and analysts will soon be able to ask their own questions in natural language, get bespoke code scripts written automatically, and regenerate analyses on demand.

"You're no longer trapped to what someone builds you. The analyst's domain expertise becomes the driving force."

His timeline: within the next three years, this technology will start to change how people coach, train, and how teams play.

The ecological framework and AI are not opposites. They are, in Richardson's hands, the same inquiry: how does a performer couple with their environment, and what information are they using to do it? The rocking chairs from his PhD and the LiDAR sensors he is building now are answering the same question. The difference is resolution.

Listen to the Full Podcast

Prof. Michael Richardson in conversation with Ian Renshaw and Keith Davids — ecological psychology, AI decision prediction, LiDAR hardware, and the future of performance analysis.