There is an old line, borrowed from sport and the military, that people do not rise to the occasion under pressure, they fall to the level of their training. It is a hard thing to hear if your leadership development has been a series of models on slides, because a model on a slide is not training in any sense that survives contact with a real, high-stakes, fast-moving decision. When the pressure arrives, you do not perform what you read. You perform what you rehearsed.
Under pressure, people default to what they have rehearsed, not what they have read, because stress narrows working memory and triggers a threat response. Roediger and Karpicke's research shows active retrieval builds roughly 50% better long-term retention than re-reading. So decision-making under pressure is trained by rehearsing the real moment in realistic simulation, not by studying a framework, which is recalled too slowly to help when it counts.
Why the framework fails at the worst possible time
A decision-making framework learned passively lives in the conscious, deliberate part of the mind: the part that needs time, calm and spare attention to work. A genuine high-pressure moment removes all three. The stakes spike, the clock runs, and the body treats the situation as a threat, flooding the system and narrowing working memory to a slit. In that state, the carefully memorised five-step model is not available. What is available is whatever the person has done so often that it runs without conscious effort.
This is why so much leadership training disappoints exactly when it is most needed. People leave the course able to describe good decision-making and no more able to perform it under fire, because describing and performing draw on different systems. As we have written about leadership under pressure, the gap between the two is not a knowledge gap that more content can close. It is a rehearsal gap.
What the evidence actually says about how skills stick
The learning science here is unusually clear. Roediger and Karpicke's 2006 work in Psychological Science established what is now called the testing effect: actively retrieving and using knowledge produces dramatically stronger long-term retention than passively re-reading it, on the order of 50% better durability. The act of pulling something out and using it, under conditions that demand it, is what embeds it. Reading it again, however carefully, does not.
Sidestream's own roots are in exactly this question. Building on academic behaviour-change work from UCL, Cambridge and Bocconi, the finding that shaped how we work was that immersive role-play was roughly 20% more effective than passive modalities, slide-shows and video e-learning, at building genuine communication and decision skill. Just as telling was the confidence problem underneath it. When people rate their own skill after passive learning, they tend to be sure they are good and wrong about it, a live case of the Dunning-Kruger effect. The study replaced self-report with behavioural measurement, watching what people actually did rather than trusting what they claimed, and the picture changed completely. Rehearsal does not just build the skill. It reveals the truth about it.
Rehearsal is uncomfortable, and that is the point
Some things cannot be taught, they have to be felt. Deciding under pressure is one of them, because the pressure is not incidental to the skill, it is the skill. Rehearse the same decision in a calm, low-stakes classroom and you have practised the wrong thing. The rehearsal has to carry enough of the genuine article, the ambiguity, the stakes, the discomfort, that the response you build is the one that will actually turn up on the day.
That is the design behind our actor-based simulation and immersive simulation training. Trained actors recreate the real moment, the crisis with incomplete information, the team looking to you for a call you are not sure of, the decision with no clean answer, and leaders practise deciding, then get feedback on what they actually did under that load. It is deliberate practice in Ericsson's sense: repeated, specific, feedback-rich reps of the exact thing that matters. And yes, it is uncomfortable, which is precisely why it works. As we have explored in why role-play feels uncomfortable, the discomfort is the pressure doing its job, building a response robust enough to hold when the stakes are real.
Under pressure you do not perform what you learned. You perform what you rehearsed. Everything else is a slide you once read.
The practical implication for anyone designing leadership development is uncomfortable but freeing. If the goal is better decisions when it counts, the measure of a programme is not how much content it delivered or how well people scored on a quiz. It is whether they can actually do the thing under pressure, and the only way to build that, and the only honest way to know you have, is to have them do it, repeatedly, in conditions close enough to the real one that the response transfers. Rehearse the moment, and the moment will find you ready.
In Summary: Rehearsing Decisions Under Pressure
Under pressure people default to what they have rehearsed, not what they have read, because stress narrows working memory and triggers a threat response. Roediger and Karpicke's testing effect shows active retrieval builds roughly 50% better long-term retention than re-reading, and Sidestream's UCL-rooted work found immersive role-play about 20% more effective than passive modalities, while exposing how badly people judge their own skill. Decision-making under pressure is therefore trained by rehearsing the real moment, not by studying a framework.
- A framework learned passively lives in the slow, conscious mind, which a high-pressure moment shuts down; rehearsed behaviour runs without it.
- The testing effect shows retrieval and use build far more durable capability, around 50% better retention, than re-reading.
- Sidestream's UCL, Cambridge and Bocconi roots found immersive role-play roughly 20% more effective than passive learning, and behavioural measurement more honest than self-report.
- Realistic, actor-based simulation is deliberate practice: repeated, feedback-rich reps of the exact decision, and the discomfort is the pressure doing its work.
Frequently Asked Questions
Why do leaders freeze under pressure despite training?
Because pressure narrows working memory and triggers a threat response, so people default to whatever they have already practised, not what they have merely read. A framework learned on a slide is recalled slowly and consciously, which is exactly what a high-pressure moment does not allow. Rehearsed behaviour is faster and holds when it counts.
What is the testing effect, and why does it matter for leaders?
Roediger and Karpicke's 2006 research showed that actively retrieving and using knowledge produces far stronger long-term retention, roughly 50% better, than re-reading it. For leaders, this means practising a decision under realistic conditions embeds it far more durably than absorbing a model passively, because retrieval under pressure is the skill itself.
How do you train decision-making under pressure?
By rehearsing the real moment, not the theory. Realistic simulation with trained actors recreates the ambiguity, stakes and time pressure of a genuine decision, and participants practise deciding and get feedback on what they actually did. That behavioural rehearsal builds the judgement that reading a framework never reaches.