I Know Your Card

Pick a secret starting card. Follow one simple counting rule. I never see your choice — and I'll name where you end up about 85% of the time.

The deck — freshly shuffled

↓ pick any ONE of the first 10 cards (gold outline) as your secret start

your card will be

? ?

Press Reveal my prediction before you trace your path — that's the whole trick.

How to play

Secretly pick any of the first 10 cards — they're the ones outlined in gold.
Its value tells you how far to jump: A = 1, number cards = their number, J, Q, K = 5.
Count that many cards forward. Land on a new card. That's your new card.
Repeat until a jump would run off the end of the deck.
The last card you land on is your final card.

Reveal the prediction first, then click your starting card above and watch your path light up.

Does it actually work?

running…

Live simulation across thousands of random shuffles: how often two chains starting from different cards in the first ten end on the same final card.

The trick is that chains eat each other

Here's what's really happening. Every card in the deck is the start of its own chain of jumps. You picked one. The magician quietly picked another. Two different chains, wandering forward through the same 52 cards.

The moment those two chains ever land on the same card — even once, by accident — they are identical from that point onward. Same card, same value, same jump, forever. They have merged, and they end together.

So how likely is a collision?

Each chain makes a jump of average length about 3.8 cards (A through 10 average 5.5, but J/Q/K count as 5, dragging it down). So a chain touches roughly 1 in every 3.8 cards — call it 26% of the deck. Every step your chain takes is another chance to land on one of the magician's cards. Over the dozen or so jumps it takes to cross 52 cards, the chance of never colliding is small.

P(no merge) ≈ (1 − 1/3.8)^(number of the magician's cards passed) ≈ 0.15

That works out to roughly 85% success. It's the same phenomenon as the birthday paradox: with enough chances, coincidence becomes near-certainty. Press Show every path above and watch it happen — all ten starting points collapse into a handful of chains, and usually into one.

Named for Martin Kruskal

The physicist Martin Kruskal (1925–2006) described this in the 1970s. He is far better known for co-discovering the soliton — a solitary wave that keeps its shape forever — and for the Kruskal–Szekeres coordinates that let physicists see through the coordinate singularity at a black hole's event horizon. The card trick is his most-performed work.

Performing it on a real person

Deal the whole deck face up in a long row, left to right, in several rows.
Explain the counting rule and let them pick silently from the first ten cards.
Do the count yourself from your own secret start — you get the answer at the same time they do.
Write your prediction down before they finish. That's what sells it.
Around one time in seven you'll be wrong. Have a line ready: "let's try that again with a different shuffle" works fine, and the second attempt usually lands.

Counting J, Q and K as 5 rather than 11, 12, 13 is not a stylistic choice — shorter jumps mean more cards visited, more chances to collide, and a noticeably higher success rate.