The Hardest Thing to Coach Is Restraint

Most coaching is additive. Here is a new shot, a new movement, a new decision to add to your repertoire. Progress is measured in things acquired, and the entire structure of instruction, from the drills to the assessments, is organised around accumulation.

The interventions that have most improved players I have watched were nearly all the opposite. They involved removing something, and the thing removed was almost always an action the player was capable of performing and should not have been performing.

Golf Makes This Unavoidable

Golf is the clearest laboratory for this because the sport is unusually honest about the cost of ambition. Every shot presents a menu, and the menu always includes an option that is available, achievable, and wrong.

The player in the trees can reach the green through a two-metre gap. They can genuinely do it, perhaps three times in ten. The alternative is a wedge back to the fairway that costs a shot and guarantees a bogey at worst. The arithmetic on this is not close and has not been close since anyone bothered to calculate it, and players at every level continue to take the gap.

They do not take it because they miscalculated. They take it because they can see the shot, and seeing a shot creates a desire to play it that the calculation does not touch.

The problem is not that they cannot play the shot. The problem is that they can.

Why Restraint Resists Instruction

Here is what makes this so hard to coach. Skills are taught through repetition and feedback, and restraint has terrible feedback properties.

When a player lays up correctly, nothing happens. There is no moment of confirmation, no sensation of having done something well, just a routine shot to the fairway and a par. When the player takes the gap and pulls it off, the reward is immediate, vivid, and remembered for years.

The feedback structure is exactly backwards from what learning requires. The correct decision produces no signal and the incorrect one produces an intense signal roughly a third of the time, which is precisely the reinforcement schedule most resistant to extinction. We are asking players to learn against the strongest form of conditioning there is.

Memory Is Not on Your Side

Compounding this, the recall is asymmetric. Every player remembers the recovery shot that worked. Almost none of them can list the seven attempts that finished in a different set of trees, because failure at this is undramatic and gets absorbed into a general sense of a scrappy round.

Statistics help here, which is the strongest practical argument for shot-tracking. A player who can be shown that their attempted recoveries have cost them a defined number of strokes over a season is being handed a memory they do not naturally possess.

It helps. It does not solve it, because the number is abstract and the shot is right there in front of them.

What Actually Works

The most effective approach I have seen removes the decision from the moment entirely. The player commits in advance, off the course, to a rule: from this lie, in these conditions, the answer is always the wedge, regardless of what the gap looks like.

This works because it relocates the choice to a context where the emotion is absent. A player sitting calmly deciding a policy makes a different decision than the same player standing in the trees with adrenaline and an audience. Pre-commitment lets the calm version of the player overrule the excited one.

It is also, notably, not a skill. It is a structure imposed to compensate for the fact that the skill cannot reliably be trained.

The Broader Version

This generalises well beyond golf. The bowler who tries the delivery they have not landed in practice. The footballer who attempts the pass rather than the one that maintains possession. The coach who intervenes tactically when the shape was working. In each case the failure is not ignorance and not inability. It is the presence of a capacity that should not have been used.

I have come to think this is the central asymmetry of coaching. We can teach people to do things. We are considerably worse at teaching them to have things and leave them alone, and that second capacity is where a lot of the difference between good and very good actually sits.