How to remember a magic trick.

Simply put, when you work hard to remember something, you remember it better and for a longer time. The effort of pulling a memory back is what inscribes it to long-term memory. Psychologists call this retrieval practice, and it's one of the most tested findings in the science of learning.

In 2006, Henry Roediger and Jeffrey Karpicke gave students a passage of text to learn. One group spent 4 study periods reading it, about 14 times through. Another group studied it for 1 period, then spent the other 3 writing down everything they could remember. After 5 minutes, the re-readers remembered more. A week later it had flipped: the recall group remembered 61% of the passage, and the re-readers 40%.

The re-readers were also the most confident that they would remember it. Repetition felt like it was working, and it wasn't.

Roediger went on to co-write Make It Stick, which makes the point plainly: learning that feels easy often doesn't last, and the strategies that feel most productive are often the least effective.

Mindless repetition (like endlessly shuffling cards) can feel like progress, but it isn't helping you learn or retain anything new. Running a trick 20 times in a row mostly tests whether you can remember something you did 30 seconds ago. Of course you can.

A performance asks for something different: the whole trick, cold, days or weeks after you last touched it. If you never practice that, you never find out whether you can do it.

Too much strain isn't helpful either. Aim to struggle with roughly 5% to 15% of what you learned. An easy comparison is weight lifting: enough effort to break the muscle fibers down, but not so much that you can't complete the exercise.

There's some math behind the top of that range. A 2019 paper in Nature Communications calculated that a broad class of learning algorithms learn fastest when they get things wrong about 15% of the time. It's a result about learning models, not a law of practice, but it lands in the same place: mostly right, with a real chance of being wrong.

A mistake in rehearsal is information. If you stop to check the instructions every time you blank, you're practicing the act of looking things up. Instead, push through, finish the trick, and check what you missed afterwards.

Make It Stick describes experiments where trying to answer first and getting it wrong, then seeing the right answer, beat being shown the answer up front. Scott Young gives the same advice in Ultralearning: test yourself before you feel ready, because the difficulty is doing the work.

Just remember that this only works with Immediate Feedback. An error you never catch is a bad habit in the making.

• Write the trick out from memory, somewhere only you can read it: the effect, the method, and every step in order. Then compare it with the source.
• Rehearse with the instructions closed. Paused tutorials and open books are for learning a trick, not for rehearsing one.
• Start cold. A warm-up run is where the easy recall gets used up.
• Leave a gap before you rehearse again (Spaced Intervals). A little forgetting is what makes the next recall count.
• Say the patter out loud. Words you only think are words you haven't tested.

A Mastery Journey begins with Strenuous Recollection. Before your first rehearsal, you document the trick's effect, its method, and its performance steps from memory.

After that, it's built into the schedule. A Cold Rehearsal asks for the full trick with no practice or warm-up first, and as the gaps between rehearsals grow, each one takes a little more effort to pull back. If your scores drop, the gaps shorten. If they hold, the gaps keep growing. The intent is to balance Strenuous Recollection with forward progress: struggling a little, and still finishing.