Does your shuffle pass the test?

Hello everyone,

Apologies for posting in English: our site has no Finnish version.

We’ve put an online test on our website to see whether your usual way of shuffling leaves traces of the deck’s starting order. The idea is simple: a well-shuffled deck keeps no trace of that order. Passing the test doesn’t prove that your method mixes the cards well, but failing is a strong sign that it doesn’t mix them enough.

Sort a deck into the ”By suit and rank” order shown on the page. Shuffle as you normally would, with riffles, a wash (mixing the cards face down on the table), or a machine.

Turn the whole deck face up. Click the card you see in the on-screen grid, then place it face up on a pile beside the deck to reveal the next one. The animation below shows how. Entering a deck takes about two minutes.

If you’re curious or just want to see how the test works, try it once. For a more reliable result, repeat the test ten times in a row with the same deck: shuffle it and enter the result each time. No need to sort the cards again between shuffles: turn the pile over as one block and shuffle it face down as usual.

One shuffle isn’t enough to draw a conclusion. After ten shuffles, the page does the calculations and tells you whether your method leaves traces of the starting order. When each deck is well shuffled, about 98.5% of series of ten decks pass the test. Failing is therefore a strong sign that the method doesn’t mix the cards enough.

Like any statistical test, this one can show that a method doesn’t mix the cards well, but it cannot prove that it does.

Try the test: https://leshuffler.com/q85j

How the test works: https://leshuffler.com/kgst

François

We’re a small team based in France. We created LeShuffler, an automatic card shuffler that, by design, guarantees no correlation between the starting order and the final order.