Closing the Gap to Klondike's 79 Percent Win Ceiling

Roughly 79 of every 100 Klondike deals can be won. That figure comes from computer studies that flip all 52 cards face up and search every line to the end, and it should sting a little, because nobody's personal win rate lives anywhere near it. Most of us lose deals that were winnable from the first move, repeatedly, and mostly for the same handful of reasons. The distance between 79 percent and your percent isn't luck — luck is already priced into the 79.

What the ceiling actually measures

Be clear about what those solvers had that you don't: information. In Klondike, 21 of the 28 tableau cards start face down and the stock's order is a mystery. The research programs study the deal with everything visible — a setup sometimes called thoughtful solitaire — and ask whether any winning line exists at all. About four times in five, one does. The exact figure drifts a little with the rule set — draw-3 and redeal limits shave off a few points — but it stays near 79, which is why that's the number everyone quotes. So the ceiling is real but not reachable, since some wins are only findable when you can see the buried cards that justify each move. What the number genuinely tells you is that the margin for skill is enormous: if perfect information wins 79 percent and a careless player wins a small fraction of that, then most losses are not the deck's fault, and the recoverable games live in the habits below.

Four habits that close most of the gap

Draw-1 and draw-3 are different games

Under draw-1 rules, every stock card offers itself once per pass, and the game is nearly pure tableau craft. Draw-3 adds machinery: only every third card is playable, and which cards those are depends on your position in the cycle. Here's the part most players never notice — playing a single card out of the waste shifts the entire sequence by one, changing which cards surface on every later pass. A cycle that looks dead can be brought back to life by deliberately playing one mid-stock card, even a card you don't especially want, purely to change the rhythm. If you've only ever cycled draw-3 passively, that one idea will win you more games than anything else on this page. It cuts the other way too: a careless waste play can bury a card you were counting on, so draw-3 rewards tracking what you've seen and thinking a full pass ahead. Draw-1, by contrast, forgives sloppy sequencing — it's the version to build tableau habits on before draw-3 starts charging for mistakes.

The FreeCell control group

For proof that hidden information is the villain here, look at FreeCell, Klondike's fully exposed cousin. Nothing is face down, and the effect on winnability is absurd: of the 32,000 numbered deals in the classic Microsoft release, exactly one — game 11982 — has no solution. Same deck, same goal, no secrets, and the ceiling jumps from 79 percent to a whisker under 100. Play a few FreeCell deals after a Klondike session and you can feel the difference in your shoulders. Every loss is visibly your own, which is oddly relaxing.

The rest of the family isolates other muscles. Spider plays two decks and turns suit management into the entire game; golf solitaire throws out foundations and rewards spotting long chains at speed. Rotate through them and Klondike's particular blend — part information game, part packing puzzle — becomes easier to see and easier to respect. Your win rate will never touch 79 percent, and nobody's does. Moving from one win in five toward one in three, though, is entirely a matter of the four habits above, and playing signed in on Inkplay means the practice also earns coins and a spot on the leaderboard while you're at it.

More from the blog

Or skip the reading and just play something — everything here is free, in your browser, no download.