If you've ever played Klondike Solitaire — the standard "Windows Solitaire" — and gotten frustrated that some deals seem impossible while others come together effortlessly, your instinct is right. Mathematicians have spent decades studying exactly this question, and the answer turns out to be surprisingly precise.

The short version: with optimal play (where you make every decision perfectly), approximately 79% of standard Klondike deals are mathematically winnable. With realistic human play, that number drops to around 43%. Most of the puzzles you give up on weren't bad luck — they were genuinely lost, often from your first or second move.

Why this question is hard

You'd think computing the win rate would be trivial — just simulate billions of games and count. The problem is that Klondike has imperfect information. When you turn over a face-down card in the tableau, you don't know what's underneath. That means a decision early in the game might depend on what cards you'll find later, which you can't yet see.

This makes Klondike fundamentally different from Chess or Tic-Tac-Toe. In Chess, all information is visible — minimax works in theory if not in practice. In Klondike, the optimal strategy requires reasoning about hidden state, which makes brute-force computation intractable. You can't just expand the game tree, because the "tree" branches in ways that depend on randomness you haven't observed yet.

Two different "winnable" rates

Researchers have distinguished between two interesting numbers:

1. Thoughtful Solitaire (~79% winnable)

"Thoughtful" Solitaire is a hypothetical version where, before you make any move, all face-down cards are flipped face-up so you can see the entire deal. This eliminates the imperfect-information problem entirely. Researchers at the University of Alberta published a paper in 2008 (Yan, Diaconis, Rusmevichientong, and Van Roy) computing this number through exhaustive search. They found that between 79.4% and 81.9% of deals can be won when you have full information from the start.

This number tells us something important: at least 4 out of 5 Klondike deals are theoretically winnable. The deck is, in some sense, fair.

2. Realistic Klondike (~43% winnable for skilled players)

But humans don't play with full information. We see what cards are face-up and have to guess about the rest. When researchers simulate skilled but imperfect play — using heuristics like "always move Aces to the foundation immediately, build long descending sequences on Kings" — the win rate drops to around 43%.

That gap, between 79% and 43%, is the cost of not knowing what's underneath the face-down cards. Every decision you make in the first few moves is based on incomplete information, and many of those decisions are wrong. Worse, you usually don't know they were wrong until much later, when you're stuck.

Why the popular "70%" figure

When people casually mention that Klondike is "70% winnable," they usually mean somewhere between the two numbers above — averaging thoughtful play (which is unrealistic) with skilled human play (which is what people actually do). Microsoft Solitaire's own documentation has at various points cited figures between 60% and 80%, depending on how they defined "win" and which scoring variant they meant.

The honest answer is: if you played perfectly, with hindsight, about 79% of deals you've seen were winnable. If you played at a normal skill level, you probably won about 30-45%. The gap is information, not luck.

Which decisions matter most

The University of Alberta paper analyzed which strategic choices in the opening phase predict eventual victory. Three findings stood out:

Move Aces and 2s to the foundation as soon as possible. This sounds obvious, but many players hesitate, thinking they'll need an Ace to support an upcoming sequence. They don't. Move the Ace.

Don't move cards to the foundations too eagerly past 2. Once you've moved an Ace and a 2 of a suit, holding off on moving the 3, 4, 5 can be valuable because those cards might be useful in tableau columns. The exception: if you have no use for them, move them.

Empty columns are gold. An empty tableau column is the single most valuable resource in Klondike. Use it for Kings or for temporary parking. Players who refuse to empty columns "in case they need them later" tend to lose. Players who aggressively empty columns and re-use them tend to win.

The unwinnable 20-25%

So what makes a deal unwinnable, even with perfect play? Two structural patterns emerge from the research:

Buried Aces. If a critical Ace is buried under multiple face-down cards that themselves are blocked by other cards, you may never be able to free it. The Ace can't go to the foundation, so the suit can never be cleared, so you lose.

Color-locked columns. In Klondike, tableau columns alternate colors. If a critical card ends up in a position where the colors around it can never alternate correctly (e.g., a red Queen needs a black King above it, but both black Kings are buried), the column locks up.

Both of these are determined by the initial deal. There's no decision you can make to fix them. They're roughly 20-25% of all deals.

Implications for our game

The Solitaire on GerGame uses a standard random deck shuffle. We don't filter for "winnable" deals or stack the deck — that would defeat the purpose. The math suggests that if you play 10 games on our site, statistically:

So if you lose a game and feel like you must have made a mistake somewhere — about a quarter of the time, you didn't. The deal just was what it was. The other three-quarters, there was a path forward you missed.

The deeper point

Klondike isn't usually thought of as a game with strategic depth. It's a coffee-break activity, something to play one-handed while waiting for a meeting to start. But underneath the simplicity is a real combinatorial puzzle, and the difference between a 40% winner and a 60% winner is genuine skill — not luck, not memorization, but pattern recognition and information management.

The next time you start a deal, try playing the first three moves more deliberately than usual. The opening phase, where you're committing to a sequence based on partial information, determines the outcome far more than the late game. Most lost games are lost in moves 1-5. Most won games are won there too.

The mathematics says: 79% of the time, the deal you got was winnable. The question is whether you played it well.