Baker's Game: History, Records and Competitions

FreeCell was born when someone changed one line of this game's rules. Yet the man it is named after, C. L. Baker, is known by little more than his initials — and according to Gardner he didn't invent it either: he learned it from his father. Here is what the record shows, and what it doesn't.

September 23, 2026 · 8-minute read

Name and origin

The earliest printed description of Baker's Game that can be found is in the June 1968 issue of Scientific American. In his "Mathematical Games" column (vol. 218, no. 6, from page 112), Martin Gardner described a game by C. L. Baker — one item in a column about the combinatorics of a shuffled pack.

Gardner also recorded where it came from. As quoted by Wikipedia, "the game was taught to Baker by his father, who in turn learned it from an Englishman during the 1920s." So Baker passed the game on; he didn't invent it. The inventor is unknown, and even the 1920s date rests on this one family story — no printed rules from that decade have been found. Of Baker himself we know almost nothing: solitaire sites call him a "mathematician" without citing a source, and his full name and dates are undocumented.

A variant of Eight Off

Michael Keller's FreeCell FAQ treats Baker's Game as a variant of Eight Off, which he calls "one of the oldest games of this type". Eight Off has eight free cells (four filled at the start; older books call them "depots") and allows only Kings in empty columns. Baker's Game cuts the cells to four but lets any card fill an empty column, turning empty columns into storage so long runs can move — "a better, more interesting game", says Keller, "though it is harder". When Eight Off first appeared in print hasn't been established, so it can't be shown to predate the Baker family's 1920s story.

From Gardner's column to FreeCell

A game without a name

Gardner reprinted the column in his 1977 book Mathematical Magic Show, where, according to Keller's Solitaire Bibliography, the game is described without a name. In Lincoln David Stein's Family Games (1979) it appears as "No Name"; Keller identified it as Gardner's game. The name "Baker's Game" seems to have settled later; who first used it is not known.

One rule changed, and FreeCell appeared

Meanwhile, on the University of Illinois's PLATO educational computer system, medical student Paul Alfille wrote the first computer FreeCell — in 1978, according to Keller. In Keller's account he changed Baker's Game in just one respect, packing the tableau in alternating colours instead of in suit, which had "the happy effect of making nearly every deal winnable". Jim Horne's Windows version first appeared on Microsoft Entertainment Pack 2 in 1991 (Keller's FAQ says 1992).

Nor was alternating colour new to this family. Einar Werner's Swedish book of 1945 describes "Napoleon på S:t Helena", a very close relative of FreeCell that deals its last four cards into the cells and takes only Kings in empty columns. On the same-suit side, Art Cabral of Longwood released Seahaven Towers for the Macintosh in 1988: ten columns of five, the last two cards in the cells, Kings only in empty columns. Longwood did not claim to have invented the game; its website called the program "the original computer version of the solitaire game that later came to bear its name".

The computer and internet era

Brain Jam and Games magazine

Baker's Game got its own Windows program in Brain Jam, by Peter Liepa, the Canadian programmer who co-created Boulder Dash; version 1.5 went out as Windows 3.1 shareware in 1993. Instead of random shuffles it ships a fixed set of deals, all guaranteed winnable, and its official site also offers a browser version. In June 1995 Keller's "Big Deal" in Games magazine — possibly, he notes, the first published article on FreeCell — also covered Baker's Game (as Brain Jam), with three sample deals of each.

One deal in four can't be won

The earliest computer estimate we could find came from Don Woods in 1995. Reporting on Usenet that his solver had cracked all but 14 of a million random FreeCell deals, he set that against 75% for Baker's Game and 89% for Seahaven Towers. Later work agrees. Theodore Pringle's analysis found 7,501,119 of 10 million random deals winnable (75.01%), and Charlie Blake and Ian P. Gent's solver Solvitaire measured 75.053% ± 0.028% on its own 10 million (Journal of Artificial Intelligence Research, 2026). Even with perfect play about one deal in four is lost; for FreeCell, an exhaustive check of about 8.6 billion deals found one in 84,000. The 74.3% on Wikipedia comes from an older run that, as Blake and Gent point out, used a configuration that could miss solvable deals, so it is too low.

Building in suit is expensive, though with eight cells, as in Eight Off, it hardly matters. Pringle's re-runs of the same 10 million deals with different numbers of cells show how much each cell is worth:

People win far less often: Goodsol estimates that "average players win it about 35% of the time". Our own check agrees with the research: of 5,000 random deals, our solver could not solve 1,251 (25.0%).

Curious stories and common myths

FreeCell's "impossible" deal, solved with the wrong rules

Of the original 32,000 Microsoft FreeCell deals, only #11982 is unsolvable. A report that someone had finally solved it turned out to be a win under Baker's Game rules, played by hand. As Blake and Gent show, the two building rules are incomparable — neither is strictly easier. ♠5 on ♠6 is legal only in Baker's Game, ♥5 on ♠6 only in FreeCell, so a deal that beats FreeCell can still fall to Baker's Game.

What people say, and what the record says

• "C. L. Baker invented it." Gardner's own account has Baker learning it from his father. The inventor is unknown.
• "Gardner named it." His book describes it without a name, and a 1979 book printed it as "No Name".
• "Alfille created FreeCell after reading Gardner's column." Plausible but unconfirmed. In a 2000 freecell.net interview Alfille wasn't sure where the game came from, recalled a solitaire book read in childhood, thought he had made up the name, and didn't mention Gardner. The 2002 New York Times article Wikipedia cites for the claim doesn't mention Gardner either.
• "Every deal can be won." About 25% of random deals can't be, even with perfect play. Brain Jam's "all hands are winnable" refers to its hand-picked set.
• "You win about 1 game in 10." That is SolSuite's figure for a Kings-only empty-column rule — a strictly harder game, so its win rates can't be compared. Under the standard any-card rule (the one used here) about 75% of deals are winnable.
• "It's the same as Baker's Dozen." Baker's Dozen is an unrelated patience first published by Dick in 1883, with 13 columns of four. The "Bakers Dozen" on freecell.net is different again: a 12-column, one-cell FreeCell.

Competitions, records and community

Offline

We found no tournaments, championships or official records for Baker's Game, and no Guinness World Records entry.

Online

Competitive play in the FreeCell family centres on freecell.net, whose game is based on Alfille's PLATO original. It has regular tournaments, streaks and Elo ratings, and mixes in variants such as 13x0 and 12x1 — but all of them are alternating-colour FreeCell, and there is no Baker's Game category. We found none on speedrun.com either; beyond that, casual sites keep their own high-score tables. Tools for checking deals are free, though: Shlomi Fish's open-source Freecell Solver can solve Baker's Game under both the standard and the Kings-only rule.

Why contests are hard to find

Competing at a one-player game usually means several people playing the same deal and comparing time or moves. Baker's Game adds a hurdle: with a quarter of random deals unwinnable, every deal must be checked before anyone races on it. For solitaire contests in general, see Solitaire Competitions and Communities; for where these games came from, A History of Solitaire.

How to compete here

FreeCell Mania runs "same deal" competition through deal numbers: the same number opens the same deal anywhere, any time. Jim Horne numbered the Windows FreeCell deals for a similar reason — so people could swap the numbers of difficult or interesting deals with friends.

• No deal is unsolvable. Our solver played every number first, and the ones that failed were given a different layout.
• Each number has two leaderboards, fastest time and fewest moves. Your two best results can come from different plays.
• A multi-card move counts as one move and automatic moves count too, so no deal takes fewer than 52 moves. Undone moves drop out of the count, but the clock keeps running.
• Deal #617 in Baker's Game is a different layout from #617 in FreeCell, with its own leaderboards.

Tips for climbing the leaderboards are in the strategy guide.

Sources

1. Mathematical Games (June 1968) — Martin Gardner, Scientific American vol. 218 no. 6, 1968
2. List of Martin Gardner Mathematical Games columns — Wikipedia
3. FreeCell FAQ — Michael Keller, Solitaire Laboratory
4. Solitaire Bibliography — Michael Keller, Solitaire Laboratory
5. Baker's Game — Wikipedia
6. FreeCell — Wikipedia
7. Eight Off — Wikipedia
8. Interview with Paul Alfille — Dennis Cronin, freecell.net, May 4, 2000
9. One Down, 31,999 to Go: Surrendering to a Solitary Obsession — The New York Times, 2002 (Internet Archive copy)
10. Seahaven Towers — Wikipedia
11. Longwood Wizards Guild — Longwood, 2016 (Internet Archive copy)
12. Brain Jam — official site
13. BR903/brainjam — Brian Raiter, GitHub
14. Windows 3.1 Shareware: brainj15 — Internet Archive
15. Peter Liepa — VGMPF Wiki
16. Baker's Game — SolSuite
17. The Winnability of Klondike Solitaire and Many Other Patience Games — Charlie Blake and Ian P. Gent, arXiv 2019 / JAIR 2026
18. Baker's Game — Solitaire Win Rates — Mark Masten
19. Play Baker's Game Solitaire — Goodsol (Internet Archive copy)
20. Freecell Solver — Features — Shlomi Fish
21. Free Online Competitive Freecell — freecell.net
22. Baker's Dozen (card game) — Wikipedia

Keep reading: Baker's Game Strategy Guide · FreeCell: History, Records and Competitions · A History of Solitaire (Patience) · Solitaire Competitions and Communities