I said something to a spectator after a card effect in Graz that I had not planned. It came out naturally, without thinking, and it changed the way I understood my own craft.
“So you shuffled the deck, you chose a card completely at random, and yet here it is — the only card facing the wrong way.”
She nodded. She agreed with every word. She remembered shuffling the deck.
She had not shuffled the deck. She had cut it. Once. A single, simple cut.
But the moment I said “you shuffled the deck” in my verbal recap of what had just happened, her memory reorganized around my words. The cut became a shuffle. A limited action became an unlimited one. The conditions of the effect, in her mind, became significantly more impossible than they had actually been.
I did not plan this. I was not trying to plant a false memory. I was casually summarizing the effect, and the imprecise language slipped out. But the result was unmistakable: my words had rewritten her memory.
That night, in my hotel room, I started researching what had just happened. And I fell into a rabbit hole that led me to one of the most powerful psychological principles I have ever encountered in the context of magic performance.
Elizabeth Loftus and the Misinformation Effect
The scientific foundation for what I experienced comes from the work of Elizabeth Loftus, one of the most cited psychologists in history. Loftus spent decades studying how post-event information — things people hear or are told after witnessing an event — can alter their memory of that event.
Her most famous finding: memory, as she puts it, “works like a Wikipedia page: you can go in there and change it, but so can other people.”
In her landmark experiments, Loftus showed participants a video of a car accident and then asked them questions about what they saw. When the question used the word “smashed” instead of “hit,” participants remembered the cars going faster and were more likely to report seeing broken glass — even though there was no broken glass in the video. The single word “smashed” inserted a false detail into their memory.
This is the misinformation effect: post-event information that is subtly different from what actually happened can overwrite the original memory. The person does not experience this as an update or a correction. They experience it as remembering. The false detail feels exactly as real as the accurate details surrounding it.
When I read about this in the context of Gustav Kuhn’s research at the MAGIC-lab at Goldsmiths, the connection to magic performance was immediate and obvious. Every time a performer summarizes what just happened — every verbal recap, every “so you chose a card and I never touched it” — they are in a position to shape the spectator’s memory of the event.
Not just to remind them. To rewrite them.
The Power of the Verbal Recap
Let me be precise about what I mean by a verbal recap, because it is something every performer does, usually without thinking about it.
A verbal recap is the moment after an effect when the performer summarizes the conditions. “You shuffled the deck.” “I never touched the card.” “You had a completely free choice.” “I was standing on the other side of the room.” These statements serve an obvious theatrical purpose: they emphasize the impossibility of what just happened. They are selling the effect.
But they are also doing something else. They are providing post-event information that integrates into the spectator’s memory of the procedure. And if the verbal recap is subtly different from what actually happened, the spectator’s memory shifts to match the recap.
This is not a trick. It is not deception in the traditional sense. It is a natural consequence of how human memory works. Post-event information and original memory merge in the brain. The spectator cannot distinguish between what they actually saw and what they were told they saw.
The Spectrum of Inaccuracy
Not all verbal recaps are equally effective at altering memory. The research suggests a spectrum of susceptibility.
At one end, major contradictions. If you tell a spectator they shuffled the deck when they clearly remember that you shuffled it, the contradiction may be too large. The spectator’s existing memory is strong enough to resist the overwrite. They might think, “Wait, I didn’t shuffle it, you did.” The false information is rejected because it conflicts too directly with a vivid, well-encoded memory.
At the other end, trivial substitutions. If you say “you shuffled the cards” when they actually cut the cards, the substitution is small. Cutting and shuffling are both forms of mixing. The spectator’s memory of the specific action is probably somewhat vague — did I shuffle or cut? — and the verbal recap tips the balance toward the more impressive version. The overwrite succeeds because the existing memory is not vivid enough to resist it.
The sweet spot for verbal recaps is in this second zone: small, plausible upgrades to the spectator’s memory of the procedure. Not lies. Not fabrications. Slight improvements. Generous interpretations. Verbal nudges that tip an ambiguous memory in the direction of impossibility.
“You shuffled the deck” instead of “you cut the deck.”
“You had a completely free choice” instead of “you chose from these options.”
“I never touched the card” instead of “I briefly handled the card at the beginning.”
Each of these is a small shift. Each is plausible enough that the spectator’s memory does not resist it. And each makes the reconstruction of the effect slightly more impossible.
The Timing Window
Research on the misinformation effect shows that timing matters. Post-event information is most effective when it is delivered during a specific window.
Too soon, and the original memory is still vivid. The spectator can compare the recap to their fresh memory and detect the discrepancy. This is why recapping immediately after the climax — while the audience is still processing what they just saw — is less effective than waiting a beat.
Too late, and the original memory has already been consolidated. Once a memory is firmly stored, it is harder (though not impossible) to alter with new information.
The ideal window is a few minutes after the event, when the original memory has begun to fade but has not yet been fully consolidated. In this window, the memory is malleable — soft, impressionable, receptive to shaping.
In practical performance terms, this means the verbal recap is most powerful when it happens not at the climax, but slightly after — during the applause, during the transition to the next effect, during the casual conversation that follows. “Wasn’t that amazing? You shuffled the deck yourself, and yet…” The recap lands in the window of maximum malleability.
What I Changed in My Own Performance
After studying this, I went through every effect in my repertoire and identified the moments where I was giving verbal recaps. Then I asked myself: are these recaps accurately describing what happened, or are they subtly upgrading the spectator’s memory?
In most cases, my recaps were accurate. I was simply restating what had happened, which meant I was wasting an opportunity. I was not using the misinformation effect. I was just reminding people of what they already knew.
So I started making deliberate, minor adjustments to my recaps. Not fabrications. Not outright lies. Gentle shifts in language that nudged the spectator’s memory toward cleaner, more impossible conditions.
Where the spectator had cut the deck, I said “mixed.” Where they had chosen from a limited set, I said “free choice.” Where I had briefly handled an object, I omitted that detail entirely, letting the spectator’s memory fill in the gap with the assumption that I had never touched it.
The results were noticeable. Not in the moment — the moment was the same. But in the retelling. When spectators described the effect to others, they used my language. They said “I shuffled the deck” because I had said they shuffled the deck. They said “he never touched it” because I had framed it that way.
My verbal recap became the template for their memory. And the memory they constructed from that template was more impossible than the event itself.
The Ethics of Verbal Framing
I should address something that might be bothering you, because it bothered me for a while.
Is this manipulative? Is it dishonest to use verbal recaps to shift a spectator’s memory?
I wrestled with this question during a late night in a hotel room in Innsbruck, and here is where I landed.
All magic is, by definition, an act of deception in the service of wonder. The spectator has entered into an unspoken contract — they know you are going to fool them, and they want to be fooled. The deception is not a violation of trust. It is the entire point of the interaction.
Verbal recaps that subtly upgrade the spectator’s memory are no different, in principle, from any other form of misdirection. They redirect not the spectator’s attention but their memory. They shape not what the spectator sees but what the spectator remembers. They are misdirection applied to memory rather than perception.
And the result is the same: a stronger experience of impossibility. A better memory. A more magical story to tell.
I draw the line at outright fabrication — claiming something happened that was dramatically and obviously different from the actual event. But gentle framing? Generous interpretation? Using language that tips an ambiguous memory in the direction of wonder?
That is not manipulation. That is performance.
The “Again” Technique
One of the most elegant applications of this principle comes from Derren Brown, who describes it in Tricks of the Mind. Brown talks about inserting the word “again” into instructions — “shuffle the deck again” — when the spectator has never shuffled before.
The word “again” implies a prior action. The spectator, focused on the task of shuffling, absorbs the implication without questioning it. Later, they genuinely remember having shuffled the deck at the start, because the word “again” created a false memory of an earlier shuffle that never happened.
This is breathtaking in its simplicity. A single word — four letters — inserted casually into an instruction creates a false memory that survives reconstruction and strengthens the effect.
When I first read this, I sat staring at the page for a long time. Because it meant that every word I say during a performance is not just communication. It is memory architecture. Every word either reinforces the spectator’s accurate memory or gently reshapes it. And the performer who is aware of this has a tool that the performer who is not aware of it is using randomly and wastefully.
Building the Memory You Want
Here is the framework I use now. For every effect, I identify three things:
The ideal memory. What do I want the spectator to remember? What is the cleanest, most impossible version of the effect that they could carry away? This is the target.
The actual procedure. What actually happens during the performance? This is the reality.
The gap. Where does the actual procedure fall short of the ideal memory? What details, if remembered accurately, would weaken the impression of impossibility?
The verbal recap is how I bridge the gap. Not by lying about what happened, but by framing the recap in language that nudges the spectator’s memory toward the ideal version.
“You chose a card” becomes “you chose completely freely.” “I looked through the deck” becomes “I never saw the faces of the cards.” “You cut the deck” becomes “you mixed the deck yourself.”
Each nudge is small. Each is plausible. And each is landing in the window of memory malleability where the spectator’s recall is soft enough to be shaped.
The Long-Term Effect
The misinformation effect does not weaken over time. This is one of the most important findings from Loftus’s research. False memories, once implanted, become indistinguishable from genuine memories. The spectator does not eventually realize “wait, I didn’t actually shuffle the deck.” The false memory is permanent. It becomes part of their story.
This means that verbal recaps have a compounding effect. In the hours after the performance, the spectator retells the effect to friends using the language you provided. Each retelling reinforces the false memory. Each retelling makes the effect more impossible. And each retelling moves the spectator’s memory further from the actual procedure and closer to the ideal version.
By the time they describe the effect a week later, the verbal recap has fully replaced the original memory. The effect they remember is the effect you told them to remember — the one where they shuffled the deck, chose completely freely, and you never touched a thing.
You performed an effect. Then you told them what they saw. And now that is what they saw.
The words you choose after the trick are not an afterthought. They are the final, invisible layer of the effect itself. They are the last piece of design. And they may be the most important one, because they determine not what the audience experienced, but what the audience will remember experiencing.
Choose them carefully.