There’s a finding in Anders Ericsson’s research that I almost can’t believe is true every time I think about it, even though I’ve seen the evidence, even though I’ve experienced something like it myself.
Doctors with twenty years of clinical experience frequently perform worse on objective diagnostic measures than doctors who graduated five years ago.
Not marginally worse. Measurably, significantly worse.
This isn’t about burnout or declining cognitive function with age, though those are real phenomena. This is something more fundamental. Doctors in standard clinical practice get enormous experience — thousands of patients, thousands of decisions — but almost no structured feedback. You make a diagnosis. The patient goes home or to surgery or to another department. In most cases, you never find out if you were right.
Twenty years of experience with no feedback is not twenty years of learning. It’s twenty years of reinforcing whatever habits you started with.
I thought about this for a long time after I read it. Then I thought about the first two years of my own practice.
The Illusion of Accumulated Experience
When you’ve been doing something for a while, there’s a very convincing story you tell yourself about the value of that time. You’ve done it thousands of times. You’ve seen hundreds of different reactions. You’ve navigated unexpected situations. Surely all of that experience is making you better.
Sometimes it is. But experience only improves you if two conditions are met: you get feedback, and you use that feedback to adjust.
If the feedback is absent, ambiguous, or wrong, you don’t improve. You develop confidence in whatever you happen to be doing, whether it’s correct or not.
I was performing magic for friends, colleagues, and casual encounters for well over a year before I had any serious feedback mechanism. The feedback I was getting was mostly applause — social warmth, polite surprise, the general positive reaction that people give a person who just did something entertaining. That feedback told me almost nothing about what was working, what wasn’t, where I was losing people, or where my skill was genuinely lagging behind where I thought it was.
I thought I was getting better because I was doing it more. The doctor problem was already operating on me.
What Feedback Actually Means
The key word in Ericsson’s framework is “structured.” Not feedback in general — structured feedback. Feedback that tells you, specifically, what you did and whether it matched what you intended.
In surgery, the ideal feedback loop would be: you performed this procedure, here is exactly how it went relative to the optimal approach, here are the specific deviations. Almost no surgeon gets this. They get outcomes — patient survived, patient didn’t — but outcomes are many steps removed from the specific decisions and techniques that produced them.
In magic, the equivalent would be: here is exactly what the audience perceived at each moment of your routine, here is when they were genuinely watching you versus letting their attention drift, here is the moment they started to think analytically rather than experience wonder.
You can’t get that feedback in real time, in performance. You’re too inside it. And audience responses are notoriously unreliable as technical feedback — people will enthusiastically applaud a routine that a more experienced eye would immediately spot the problems with.
So I had to build a different mechanism.
The Camera
Recording myself performing was not a new idea. Every aspiring performer hears this advice. Most do it occasionally, watch the recording with general interest, wince at a few things, and don’t build any systematic practice around it.
I did the same thing for a long time.
What changed was treating the recording as diagnostic data rather than as a performance review.
The difference is significant. A performance review asks: was this good? A diagnostic review asks: what specifically was wrong, and at what moment, and what caused it?
I started watching recordings with the sound off first. Just movement, positioning, visual rhythm. Then with sound, watching only my hands. Then with sound, watching only my face. Then from the beginning as an audience member, trying to track what I was actually paying attention to at each moment.
Each pass revealed different things. The silent watch showed me that my body language in the first thirty seconds of a routine was subtly tense — I’d never noticed it while performing. The hand-only watch showed me a moment where my eyes went to my hands for a fraction of a second — so brief I would have sworn it didn’t happen, but there it was on video. The face watch showed me that my expression during what should have been a moment of genuine surprise read as slightly anticipatory — like I already knew what was coming, because I did.
These are not things an audience consciously notices. But they leak. They accumulate. They create a subtle, undefined feeling that something isn’t quite right, even when the rational mind can’t identify what.
Structured feedback makes the invisible visible.
Seeking External Eyes
Video is one feedback channel. It’s limited by the fact that you’re both the performer and the analyst — you bring your own blind spots to both roles.
The second phase of my feedback system was finding people who could watch and tell me things I couldn’t see myself. Not enthusiastic friends — they’re too kind. Not random audience members — they don’t have the vocabulary for what you need to hear.
I needed someone who knew what to look for and had no social incentive to spare my feelings.
Adam Wilber was ruthlessly useful here. When we started working together seriously on Vulpine Creations, one of the most valuable things about that collaboration was having someone with genuine expertise watch me perform and tell me specifically what wasn’t working. Not generally — specifically. Not “that was great” or “that felt a bit off” — but “you telegraphed the outcome there, you can see it in your posture” or “the audience lost you for about eight seconds in the middle and you didn’t notice.”
That kind of feedback is uncomfortable. It’s supposed to be. The discomfort is information.
The Degradation Dynamic
What happens without feedback isn’t just stagnation. The doctor problem isn’t just about plateauing at a competent level. In some cases, skill actively degrades.
Here’s why: without feedback, you optimize for what feels right rather than what is right. And what feels right is usually what reduces your own anxiety, not what serves the performance.
I noticed this in myself. Left to my own devices, without external correction, I was subtly reshaping certain routines to be more comfortable for me. Moments that were technically correct but required concentration that felt risky in performance — I was unconsciously sanding them down to something slightly easier, slightly safer. The routine felt smoother. From the outside, it was marginally less sharp.
I didn’t decide to do this. It happened automatically, through accumulated micro-decisions in practice that prioritized my comfort over the quality of the output.
This is the mechanism behind the doctor problem. Without feedback to anchor you to external standards, you drift toward internal comfort. And internal comfort and high performance are not the same place.
Building a Real Feedback System
What I use now is three-layered. Video review on a schedule — not every session, but regular enough that I have ongoing diagnostic data. Trusted external eyes when I’m working on something new or when something feels off but I can’t pinpoint why. And a specific practice habit of defining, before each session, exactly what I’m working on and what successful completion looks like — so I have a personal benchmark to evaluate against.
The last piece matters more than it seems. If you define in advance what you’re trying to accomplish, you have a reference point for evaluating whether you did it. Without that reference point, the session floats, you default to comfortable repetition, and you learn almost nothing.
Twenty years of experience can mean twenty years of improvement. But only if you’ve built the feedback architecture that converts experience into information, and information into genuine adjustment.
Without it, experience is just time. Time doesn’t automatically teach anything. It mostly just makes you more confident in whatever habits you started with.
If experience doesn’t automatically improve you, and feedback is the missing mechanism — the next question is what to do when you’re genuinely stuck. That’s next.