Alaska Solitaire: Yukon's cousin, measured from scratch

Alaska uses exactly the same deal as Yukon: all 52 cards spread over seven columns, no stock, and the freedom to move any face-up card along with everything stacked on it. The difference is the building rule. In Alaska, a card goes on another of the same suit, and it may be one rank lower or one rank higher. The 7 of diamonds can go on the 8 or the 6 of diamonds, but on nothing else.

Our aim across solitaire-free.uk is to set our own results beside the published research. For Alaska that is not possible: the 2026 study by Blake and Gent at the University of St Andrews, the most thorough recent work on patience games, does not include it, and we found no other published figure. What follows is our own measurement of this site's Alaska, with its limits spelled out.

The key result: a solver that could see every hidden card proved 75.6% of 500 deals winnable, proved 7.2% impossible, and could not decide the other 17.2%. So somewhere between 75.6% and 92.8% of deals can be won with perfect play.

Our numbers at a glance

Please read the 1.5% with care. Our simple player is poorly suited to Alaska: the up-or-down rule creates many possible moves, and a player that sees only part of the table and never takes a move back struggles to choose among them. It got an average of about five cards home per game. That figure says more about the player than about the game. The solver's result is the one that describes the deals.

How Alaska is played here

• Seven columns of 1, 6, 7, 8, 9, 10 and 11 cards; in each column after the first, the bottom five cards are face up.
• No stock and no waste: every card is dealt at the start.
• Build in the same suit, one rank up or down.
• Any face-up card can be moved together with all the cards on top of it.
• Only a King, with anything resting on it, may go into an empty column.
• Build the four foundations up by suit from Ace to King to win.

Getting more from the up-and-down rule

Two directions means more choices, and more traps

Because a card can go on the next rank up or down, you often have two homes for the same stack. Before choosing, ask which one leaves the cards underneath free to move later. The wrong one can bury a card you need.

Suits matter from the first move

Unlike Yukon's colour-based rule, Alaska's same-suit rule means each card has at most two possible landing places on the table. Keep a mental note of where each suit's cards sit, and plan to bring each suit together in one area.

Uncover the hidden cards

With about 1.6 Aces starting face down, turning over hidden cards early is how you find them. Group moves are most valuable when they uncover something.

Plan Kings into empty columns

An emptied column accepts only a King, so clear one only when a King (and its stack) is ready to move in.

Lean on Undo

The solver's advantage is seeing everything. Yours is Undo. Try a line, see what turns over, and step back if it closes more options than it opens. Our player's winning games averaged about 111 moves, so there is room to explore.

On solitaire-free.uk

Alaska costs nothing to play, needs no download or sign-up and asks for no email address. It runs in any current browser on computer, tablet or phone, with large cards and zoom, Hint, unlimited Undo, Restart, Auto Play, drag and drop or tap to move, and wins, time and moves recorded for the game you are playing. Switch sound off, go fullscreen or change the table colour whenever you like.

Prefer the colour-based version? Try Yukon, which uses the same deal.

Mistakes that cost Alaska games

• Treating it like Yukon. Colour means nothing here. A red 6 on a black 7 is not allowed; only the 6 or 8 of the same suit will take a 7.
• Always building downwards. The up-or-down rule is the game's biggest freedom. Placing the 7 of hearts on the 6 of hearts can be exactly the move that frees a column.
• Emptying a column with no King ready. Only a King can use the space, so an empty column with no King to fill it helps nobody.
• Moving big stacks for the sake of it. A group move is only worth making if it uncovers a card or brings a suit together. Otherwise it simply reshuffles the problem.
• Giving up on a deal early. Our solver proved only 7.2% of deals impossible. The odds are that the one in front of you can be won.

How we ran the research, step by step

This study was carried out by the solitaire-free.uk team in September 2026. As no published Alaska figure exists to check against, we have been especially careful to describe what we did.

1. Rules from our code. The Yukon deal and group moves, with building in the same suit, one rank up or down, and only Kings in empty columns.
2. A program version. We rebuilt Alaska in Python with a solver and a simple player, sharing code with our Yukon model.
3. Numbered deals. Deals 1 to 10,000 for the simple player and deals 1 to 500 for the solver, each shuffled from a numbered seed so they can be repeated.
4. Two kinds of player. The solver sees every hidden card and could search up to 50,000 positions per deal, giving a ceiling. The simple player sees only face-up cards, looks up to 300 positions ahead and never undoes, giving a floor.
5. Running it. 500 solver deals and 10,000 simple-player games.
6. Published comparison. We found no published winnability figure for Alaska, including in the St Andrews study.
7. The margin of error. A 95% range is where the true value very probably lies. For the solver's proven 75.6%, it is 71.6% to 79.2%. Because of undecided deals, the full answer is 75.6% to 92.8%.
8. What the numbers can't tell you. The simple player is weak here and is not a guide to how often people win. The solver's undecided deals may be winnable or not; we simply don't know.

Behind the numbers: the code

The study is in Python 3.11, standard library only. Alaska shares yukon.py with Yukon, with the building rule switched to same suit, up or down. A shared runner ran the games on 19 processes of a 20-core machine. The solver is a depth-first search limited to 50,000 positions per deal; for Alaska it makes no automatic foundation moves, so every move is chosen by the search. If it runs out of options, the deal is proven impossible. The simple player looks up to 300 positions ahead on visible cards and scores them by cards home, hidden cards left and how many cards are correctly linked. Its 10,000 games took about a minute; the 500 solver deals about three and a half minutes. Ranges use the Wilson score interval.

We wrote and reviewed the code, and drafted this page, with the help of an AI assistant: Claude, made by Anthropic. The AI provided none of the figures. They are all outputs of the simulation, checked by the team before publication.

Frequently asked questions

What share of Alaska deals can be won?

Between 75.6% and 92.8% with perfect knowledge of every card, in our study of 500 deals. We know of no published figure to compare with.

How is Alaska different from Yukon?

Yukon builds down in alternating colours. Alaska builds in the same suit, and a card may go one rank up or one rank down.

Why is the computer player's result so low?

It sees only face-up cards, never undoes and handles the up-or-down choice poorly. Its 1.5% is a floor, far below what a thoughtful player achieves.

Does starting with Aces face down hurt?

Not in our data. The simple player won 1.9% when two or more Aces started hidden and 1.1% when fewer did, a small difference that runs the opposite way to what you might expect.

Where should I go after Alaska?

Yukon for the colour-based version, or Baker's Solitaire for another same-suit challenge. All twelve games are on the games page.

By Admin, solitaire-free.uk. Study date: 18 September 2026.