Execution Cost

The term that was missing

When a player beats a press with a turn, the coaching conversation is always about the defender — he bit, he was wrong-footed, he lost a step. All of that is real, and I have written about how much of it survives being scouted. But the attacker turned too. He also had to kill his own momentum, plant a foot and rebuild speed in a new direction. Nobody counts that. I call it the execution cost: the time the escape move costs the player performing it. Once you put that term into the equation, a coaching argument falls out of it that I did not expect to find.

The escape window, written out

Narizuka and colleagues (2026) analysed all 306 matches of the 2023 J1 League with synchronised event and tracking data and defined pressure as the minimum arrival time: how long the fastest opponent needs to reach the player on the ball. Ball loss probability rose as that time fell. So pressing is a budget problem, and a Signature Move is an attempt to buy more budget. It buys in two currencies. The perceptual one is the defender reading the wrong signal — Brault and colleagues (2012) measured a 167 millisecond gap between how early an expert and a novice pick up the honest centre-of-mass cue. The mechanical one is the defender reversing committed momentum — van den Tillaar and colleagues (2020) put the plant-step ground contact at roughly 0.15 seconds for a 45-degree cut and 0.30 seconds for a 180-degree one, with deceleration steps rising from about 0.4 to 3.3. Subtract what the attacker himself spends and you get the net escape window: perceptual gain plus mechanical gain, minus execution cost.

Why this decides things at the top and not at the bottom

The two gains behave completely differently as opponents improve. The perceptual one shrinks toward nothing, because a better defender reads the centre of mass instead of the shoulders. The mechanical one does not shrink, because physics is not persuaded by video analysis. So against a weak defender the net window is roughly the perceptual gain plus the mechanical gain minus whatever the move costs you, and it stays positive even if your technique is expensive. Against an elite defender the perceptual term is gone and you are left with the mechanical gain minus the execution cost — two quantities in the same order of magnitude. At that point the execution cost decides the sign. I want to be careful here: these are magnitude bands from different studies, not measurements of one duel, and the mechanical figures come from a change-of-direction task with no ball. The claim is not a specific number. The claim is structural: as the level rises, beating a press stops being about how well you deceive and starts being about how cheap the movement is for you.

What makes a movement cheap

This is where my argument about the body comes in, and it turns out to have a proper name. Nigg and colleagues (2017) filmed 35 runners at 240 Hz across three shoe types and barefoot and found that each runner holds a preferred movement path — defined as the joint trajectory demanding the least energy. Change their shoes and 80 to 100 per cent of them keep that path within three degrees; only barefoot running pushed them meaningfully off it. Selinger and colleagues (2015) went further: using an exoskeleton to move where the energetic optimum sits, they showed people converge on a new optimum within minutes, re-converge within seconds after a disturbance, and will chase savings of under five per cent. So when I say a player has a movement that feels comfortable to his body, that is not a figure of speech. It is an energetic minimum, the nervous system hunts for it without being asked, and it defends it against interference.

The move you were given fails twice, and the second time is the one that matters

A movement imposed from outside is not a player's own minimum, so its execution cost starts higher. That is the first loss. The second is worse. Masters (1992) showed that skills acquired through explicit instruction collapse under pressure as the performer reinvests conscious control into them, and Mullen, Hardy and Oldham (2007) replicated it after two earlier studies had muddied the finding. So the coached-in move gets more expensive exactly when the player is rushed — which is precisely the situation where the arrival time is short and the escape is needed. The technique fails hardest at the moment it is required. That, I think, is the real content of the coaching cliche that under pressure you fall back on habit. What surfaces is not merely what was drilled most. It is what the body found cheapest.

So where does standardisation stop

Not at zero, and I want to be clear about that because the opposite reading is the easy one. Four layers sit between a coaching instruction and a movement. What the player looks at and when. Which way he opens his body to receive. Where the ball has to end up. And the joint path that produces all of it. The first three are information and constraint, and they are common language — they should be taught the same way to everyone, and a coach who does that well is doing the job. Only the fourth is the player's own, and that is the one the biomechanics says you cannot install from outside. This also settles a disagreement I had with myself. Coaching literature is right that receiving shape can be trained and becomes more natural the earlier it is learned. That is layer two. What I have argued cannot be taught is layer four. They were never the same claim. The practical version is a single diagnostic question: am I correcting where the ball has to go, or how the player gets it there? If it is the second, the correction is making the movement more expensive for him, and it will cost him at the exact moment a press arrives.

The weakest link, stated plainly

There is a hole in this and I would rather name it than let someone else find it. Comfortable is not the same as effective. What the research establishes is an energetic minimum, and what beating a defender requires is an advantage over another person. Nothing guarantees those coincide, and it is easy to imagine the opposite — that a player's most comfortable turn is also his most predictable one. The partial answer is that at the top level the perceptual term has gone to zero anyway, so predictability has already been paid for and lowering the execution cost is the only improvement left on the table. But that is my inference, not a finding. And the execution cost itself has never been measured: of the three terms in the equation, it is the only one for which I cannot even quote a range. I would rather publish the equation with an unmeasured term in it than quietly drop the term.