Early in my consulting career, I worked with a company trying to improve their sales team’s performance.
The approach they’d taken was intuitive and reasonable: identify the highest performers, figure out what they do differently, train everyone else to do the same things. Pattern recognition and replication. It seems like it should work.
It mostly didn’t. The high performers’ patterns were real — they did do certain things differently. But the things they did weren’t reliably predictive when adopted by others. The feedback loop was too noisy. Too many other variables. What worked for one person in one context failed for another person in a different context, even though the behavior looked identical.
Years later, I found a framework that explained this. David Epstein, in Range, distinguishes between “kind” and “wicked” learning environments. The moment I read it, I thought: this is exactly what I’ve been experiencing in magic.
Kind Learning Environments
In a kind learning environment, the rules are stable, feedback is immediate and accurate, and patterns repeat reliably.
Chess is the canonical example. The rules don’t change. Every move produces a clear result. Cause and effect are transparent. Feedback is rapid and accurate — you know immediately whether your move improved or worsened your position. Practicing in a kind environment builds real expertise, because your pattern-recognition is trained on real signal rather than noise.
Golf is another example. The rules of physics are stable. The relationship between your swing and the ball’s trajectory is lawful and consistent. You get immediate feedback from where the ball goes. Deliberate practice in golf builds real skill because the environment responds honestly.
In kind environments, the standard advice works: practice deliberately, do it a lot, get feedback, improve. The learning curves are steep and reliable. The wisdom of the domain transfers usefully from master to student.
Wicked Learning Environments
Wicked environments are different in several ways.
The rules are unclear or change. Feedback is delayed, incomplete, or misleading. Patterns from one situation don’t transfer reliably to other situations. Experience can build confidence without building genuine expertise, because what you’re learning from might be noise rather than signal.
Financial markets are wicked. You can make the right decision and have it turn out badly. You can make the wrong decision and have it turn out well. The feedback doesn’t reliably track the quality of your decision. Experienced traders can develop strong convictions about patterns that are actually statistical noise. Their experience feels genuine. Their expertise is partly illusory.
Medicine is partly wicked. Doctors who treated patients in certain eras learned patterns from their outcomes — but the feedback was confounded by placebo effects, natural disease progression, and countless uncontrolled variables. Bloodletting was practiced with confident expertise for centuries by people who were genuinely trying to help and genuinely observing results.
Why Magic Performance Is Wicked
Magic performance has all the characteristics of a wicked learning environment.
The feedback is delayed and confounded. When an effect lands well, you can rarely isolate which element caused the success. Was it the method? The presentation? The timing? The specific audience? The warm-up material that preceded it? The energy in the room from something that happened before you started? All of these factors interact, and the outcome doesn’t tell you which mattered.
The rules change. What works for one audience fails for another. What works in a corporate boardroom fails in a late-night bar. What works for this demographic fails for that one. What works when you’re in excellent form and feeling sharp fails when you’re tired or distracted. The same material gets completely different responses from different groups, sometimes with no identifiable reason.
Patterns mislead. You do something, it works three times in a row, you conclude the pattern is real and the approach is sound. Then it fails twice. Was the failure the exception? Or were the three successes the exception? In a kind environment you’d know. In a wicked environment, small samples routinely mislead.
Experience can build false confidence. A performer who has done five hundred shows has vastly more experience than one who has done fifty. But if that experience was processed as “what I do works,” without careful analysis of when it failed and why, the five hundred shows built confidence more reliably than expertise. The expert feels certain. The certainty is partially earned and partially artifact.
The Specific Way This Burned Me
I want to be concrete about how wicked environment dynamics played out in my own early performing.
I did some informal showing in the period before I’d structured anything as a real performance — at conferences, in hotel bars, with colleagues and strangers. Several effects worked consistently well. I reached confident conclusions: these effects are strong, I handle them well, the approach works.
Then I did a structured performance for a larger group in a more formal setting. Several of those same effects, with essentially the same handling, landed noticeably less well. I was confused. I’d “tested” the material. It had “worked.”
What I hadn’t accounted for: the informal settings had created conditions that were enormously favorable for those particular effects. The small group size. The conversational context. The sense of discovery rather than performance. The complete absence of expectation. My nervousness hadn’t been visible because there was no formal context to reveal it.
I’d run my tests in a kind environment (informally, conditions self-selected for my material) and tried to transfer the results to a wicked one (formal performance, controlled conditions, more evaluative audience). The feedback I’d gotten was real — those effects really did work in informal settings. But the signal didn’t transfer the way I thought it would.
What Wickedness Demands
If a kind learning environment rewards pattern-matching and replication, a wicked learning environment demands something different.
Epstein’s argument is that wicked environments reward the ability to recognize when a pattern is and isn’t applicable — to understand the conditions under which a pattern holds, not just to recognize that a pattern exists. This requires a kind of abstract thinking that pure repetition doesn’t build.
For magic: it’s not enough to know that a particular effect works. You need to understand why it works — what conditions enable it, what audience states make it land, what contextual factors support or undermine it. With that understanding, you can predict when it will transfer and when it won’t.
This is why intellectual engagement with magic theory matters so much for wicked-environment skill. Reading deeply — studying why things work, not just what to do — builds the conditional understanding that lets you navigate a wicked performance environment accurately.
The performer who has practiced an effect ten thousand times and understands only that it works is less equipped for wicked environments than the performer who has practiced it a thousand times and deeply understands the conditions under which it works. In a kind environment, more repetition wins. In a wicked environment, more understanding wins.
The Honest Implication
There’s an uncomfortable implication here that I’ve had to sit with.
In wicked environments, expertise is genuinely harder to develop and harder to verify. The feedback you get doesn’t reliably track your skill. Your intuitions about what works can be confidently wrong. Your experience can be extensive without being deep.
This means the advice that circulates in magic communities — advice from experienced performers about what works, what doesn’t, what audiences respond to — should be held more lightly than it’s often offered.
The experienced performer is describing patterns from their experience. Those patterns are real. But they’re products of a wicked environment, which means they’re confounded and conditional in ways neither the advice-giver nor the advice-receiver may fully appreciate.
Take the advice. Engage with it seriously. But don’t treat it as transferable fact. Test it in your own context. Build your own understanding of the conditions under which it holds.
That’s what wicked environments demand: not deference to experience, but engagement with principles. Not pattern-matching, but genuine understanding.
Magic performance is wicked. Learning it requires wicked-environment strategies.
And the first step is knowing you’re in a wicked environment at all.