Baker's Solitaire: FreeCell with Spider's discipline
Baker's Solitaire looks exactly like FreeCell: 52 cards face up in eight columns, four free cells, four foundations. The difference is in two rules, and both will feel familiar to Spider players. You build down in the same suit, not alternating colors, and on this site only a King may go into an empty column.
Those two rules make Baker's the least winnable game on our ladder that we could measure exactly. Our solver settled every one of 10,000 deals, and only 31.8% can be won at all. The other 68.2% are proven impossible from the moment they are dealt. For comparison, its near-twin FreeCell came out at 99.68% solved with no impossible deals.
This page explains why Baker's sits where it does, what the Kings-only rule costs you, how the game works here, and how we arrived at every figure.
Where Baker's ranks
| Game | Deals that can be won | How we know |
|---|---|---|
| FreeCell | 99.68% solved, none impossible | Our solver, 10,000 deals |
| Spider, 4 suits, every card known | 98.5% | Blake and Gent, 2026 |
| Golf (as played here) | 92.7% | Our exhaustive solver |
| Baker's, classic rule (any card in an empty column) | 76.0% | Our solver, 2,000 deals |
| Baker's, as played here (Kings only) | 31.8% (30.9–32.8%) | Our solver, 10,000 deals, every deal settled |
Because every card is face up, the solver sees exactly what you see, so this is not an unfair "perfect knowledge" figure. It is simply how many deals have a way through.
What the Kings-only rule costs
Most descriptions of Baker's Game let any card fill an empty column. We ran that version too, on 2,000 deals, and 76.0% were winnable. That matches the published figure of 75.05% from Blake and Gent, which also confirms our solver is doing its job.
Switch to our site's rule, where only a King can fill an empty column, and the winnable share falls to 31.8%. That one rule removes more than half of all winnable deals. On top of that, empty columns here do not help you move a run of cards as a group. A cleared column is still worth having, but it is a home for a King, not a general-purpose workspace.
Spider players will recognize the feeling. In two-suit Spider, the rule that only same-suit runs move turns small build choices into big problems. In Baker's, same-suit building plus Kings-only spaces does the same.
Baker's rules on spidersolitaire.us
- Layout: eight columns of seven and six cards, all face up.
- Free cells: four, one card each.
- Foundations: four, built up by suit from Ace to King.
- Columns: build down in the same suit: the 8 of clubs goes on the 9 of clubs, nothing else.
- Empty columns: only a King, or a run headed by a King.
- Moving runs: sorted runs can be dragged together, limited by your empty free cells. Empty columns do not raise that limit.
Playing Baker's with the numbers in mind
Decide early whether the deal is live
With 68.2% of deals impossible, some games are lost before you start. Don't take a loss as proof you played badly. Restart once to try a different plan; if the same wall appears, deal a new game.
Locate all four Kings before moving
Since only Kings can use a cleared column, a King buried deep in a column you want to empty is a real problem. Count where the four Kings sit before your first move.
Don't clear a column without a King to put there
An empty column with no King available is dead space. Keep that column working until you can move a King into it.
Mind the Aces
49.7% of our deals had an Ace free to play on the first move. The other half start with digging, and because builds must follow suit, every card you lift needs a same-suit home or a free cell.
Solvable deals tend to be brisk
The first solutions our solver found averaged about 39 moves, compared with 62 in FreeCell. When a Baker's deal works, it tends to work quickly and cleanly.
Features on spidersolitaire.us
Baker's Solitaire is free and runs in any modern browser on computer, tablet or phone. There is nothing to download, and no account, sign-up or email address. Cards are large and clear with a zoom control. You get Hint, unlimited Undo, Restart, Auto Play, and drag-and-drop or tap to move. Wins, time and moves are tracked for the current game; sound can be switched off; fullscreen and table colors are available.
Behind the numbers: how we ran the study
In September 2026, the spidersolitaire.us team tested how many Baker's deals can be won on this site and how much our rule differs from the classic game.
- Started from the code that runs our game. Same layout as FreeCell, but builds in suit, Kings only in empty columns, and empty columns not counted when moving runs. We flagged early that this is stricter than the version usually described.
- Rebuilt the game in code, sharing the FreeCell program with Baker's rules switched on, plus a second setting with the classic empty-column rule for comparison.
- Dealt numbered hands. Deals 1 to 10,000 on our rules and 1 to 2,000 on the classic rule, each from a fixed seed so the test can be repeated card for card.
- Chose the solver. With every card face up, the solver sees what you see. On Baker's it finished its search on every deal, so each one ends as either won or proven impossible, with nothing left undecided.
- Ran all 12,000 deals.
- Checked against published research. For the classic rule, Blake and Gent (Journal of Artificial Intelligence Research, 2026) give 75.05%; ours is 76.0%. There is no published figure for the Kings-only version we play here.
- Calculated the margin of error. For 31.8%, the 95% range is 30.9% to 32.8%.
- Cross-checked the solver. A separate brute-force program replayed Baker's deals 1 to 5 and reached the same verdicts. The one known limit: the site allows a card back down from a foundation, and our solver does not use that move, so if anything the true share is a little higher.
The code behind the ladder
The simulations are written in Python 3.11 with only the standard library. Baker's runs in freecell.py alongside FreeCell. The solver is a best-first search that keeps a record of every position it has already visited, stored in a standard form so that equivalent positions are only explored once, and it plays cards to the foundations automatically only when that is provably safe. On Baker's the search space is small enough that it ran to completion on every deal. The 10,000 deals on our rules took about 16 minutes on 19 worker processes on a 20-core machine; the classic-rule check took about 3 minutes.
We report each percentage with a Wilson score interval as its 95% range. An AI assistant, Claude by Anthropic, helped the team write and review the code and draft this page. It produced no figures: every number comes from running the program, checked against published research before it went on the site.
Ladder questions, answered with data
Why do I lose so many Baker's games?
Because most deals cannot be won on these rules. Only 31.8% have a solution.
Is Baker's the same as FreeCell?
Same layout, different rules. FreeCell builds in alternating colors and lets any card into an empty column; almost every deal can be won there.
Is the Kings-only rule standard?
No. Most rule books let any card fill a space. This site uses the stricter Kings-only version, which is why the winnable share is 31.8% rather than 76.0%.
Is Baker's harder than Spider?
Different kinds of hard. Four-suit Spider deals are almost all winnable in theory but tough to play; Baker's deals are often impossible, but the ones that work tend to be short.
Which game should I switch to after a tough Baker's run?
For an easy climb, try one-suit Spider. For a quick game with a high success rate, try Golf. The games page has the rest.
By Admin, spidersolitaire.us. Research run September 2026.