In this post, I'll try to bring together two things I enjoy: rationality and magic. Like Hazard, I've also practiced close-up magic for a good amount of time now. After recently seeing Tyler Alterman make a Facebook post about estimations and System 1, it occurred to me that there are a few calibration exercises you can do with a deck of playing cards. The three exercises below are all variants of cutting/manipulating a deck of cards, and then trying to intuit something about the deck.

This serves three purposes:

  1. Get a feel for your System 1:
    1. The goal of the following three exercises is to see how good your gut is at estimating uncertainty (hint: probably better than you think!)
  2. Improve calibration:
    1. These exercises all allow for some room for error. You can set your confidence intervals and see how quickly you can get calibrated using first principles.
  3. Practice cool party tricks:
    1. While I don't intend for this to be a full-on magic tutorial, the exercises I outline are building blocks for magic tricks, and even demonstrating your super-calibration (after getting good) might be impressive.

Below are the three exercises. If you have a deck of cards handy, you can tag along!


Cut Estimation


The simplest exercise is as follows:

  1. Lift up a packet of cards.
  2. Estimate how many cards you've picked up.
  3. Count to check how many you actually picked up.

You have a few obvious reference points. The entire deck is 52 cards, and you can easily tell if you've lifted up more or less than half (and this recurses). With a little practice. I've found that my gut is pretty good at this sort of thing. I'll ask myself how many cards, and there will be a number that feels right. It's usually quite close.

Things to pay attention to:

  • When your System 2 estimate conflicts with your System 1 gut answer of how many cards you cut off.
  • Whether you are systematically over or underestimating the amount.

Repeating The Cut


This is similar to the first one:

  1. Cut to a card.
  2. Replace the pack on the deck.
  3. Cut to the same card again.

Things to pay attention to:

  • When you try to cut to the card a second time, how quickly does your gut know that you got it right or wrong?
  • What does it feel like to "know" that the pack of cards in your hand is not the same size as the first time?
  • If you know you got it wrong, do you know if you cut too much or too little? And by how much?

Riffle Peek

This one is something I've just started playing with recently, and it's a mildly superhuman feat to get down right.


  1. Name a card. Any playing card.
  2. Riffle through the cards, watching the corners (where the number and suits are) flip towards you, and look for the card you named.
  3. Using the information in 2, cut to where you saw the card.

This is difficult. It's partially an estimation task because you need to know approximately where through the deck you saw the card, i.e. half-way, at the end, etc. To start, you can go slow, such that you can see each card as it slips off your thumb.

This gets harder the faster you riffle through the cards. To ramp up the difficulty, riffle faster, such that you can only get a fractional peek at each card's corner.

Things to pay attention to:

  • How does your visual experience of watching the cards differ when you aren't looking for a particular card vs when you are? Does anything jump out at you? Are there false positives?
  • How many riffles through the deck does it take for you to glimpse the card? (I don't always see it on the first pass through the deck myself.)
  • What is the visual experience of trying to differentiate between two similar cards (e.g. Ace of Hearts vs Ace of Diamonds)?

I think the above three exercises are fun learning experiences and a way to check in with your gut feelings through a medium many people may not have tried before. With enough practice, you can hit very levels of accuracy on these tasks, despite them seeming just a little impossible.

If you do decide to give these a go, let me know how it turns out!

New to LessWrong?

New Comment
3 comments, sorted by Click to highlight new comments since: Today at 3:01 PM

My thoughts:

  • I tried these exercises for about 15-20 mins. In that time, I tried cutting-and-counting like 10 times, and I was off-by-one every time *curses*. I managed to repeat the cut a few times - I am unsure whether you're supposed to move the cards around a bit after cutting, or whether you're allowed to hold them identically when repeating; they're often a little bit uneven at the place I just cut, which feels like an unfair advantage. I did not manage to get the riffle-peek-then-cut right any times. I took a few goes at riffling really slowly, and failed often to see the card (because two cards would flip together), which I felt unhappy about until I saw you explain that sometimes you don't always see it on a single pass through, and then I felt like I wasn't failing too hard.
  • I think this will be really useful if I ever seriously try to learn close-up magic, and much better than other intros I've seen. I really appreciated being guided in what to notice. It reminds me of The Inner Game of Tennis, which talks a lot about using S2 to guide where your attention is as you do the task, and then letting S1 take in that data and use it better - as opposed to getting S2 to tell S1 what to do (or not do) directly. I feel like the OP doesn't split close-up card magic into tricks but atomic moves, and helps guide my attention to learning the key skills that make-up those moves.
  • I feel some motivation to try to write a similar guide for playing the classical guitar, if only for my own benefit of making the moves explicit, and running metacognition on what to notice.
[-][anonymous]4y20

Thanks for trying these out, Ben!

If you ever are interested in learning close-up magic some more, I have lots more thoughts on what good resources are for learning / have strong opinions on what makes a good magic effect. I haven't written about them for the LW audience, but maybe more of this hybrid stuff will manifest later on.

I liked this exercise, I think it was very well explained.