Golf Solitaire: quick games, big swings, and a rule that changes the odds
Golf is solitaire at its most direct. Seven columns of cards, one pile in the middle, and a single rule: play any card that's one higher or one lower than the card on the pile. Clear the columns and you win. There are only 35 column cards and 16 stock cards, so games are short.
So how often does a casual player clear the board? We built a simple "no peeking" player that picks the best-looking card it can see and only draws from the stock when stuck. It won 22.4% of 10,000 games — a bit better than one in five. But when we searched every deal completely, with the order of the stock known, 92.7% could be cleared. Golf is mostly a game of choices, not of luck.
Golf in numbers
| Measurement | Result | Deals | Player |
|---|---|---|---|
| Wins for a player who can't see the stock and plays greedily | 22.4% (95% range 21.6–23.2%) | 10,000 | Greedy bot |
| Deals that can be cleared with perfect play | 92.7% (range 92.2–93.2%) | 10,000 | Exhaustive search, stock order known |
| Deals that can't be cleared at all | 7.3% | 10,000 | Same search |
| Deals where the first pile card can't be played on | 27.4% | 10,000 | Reading the deal |
| Stock cards still unused in a winning line | about 1.1 on average | 9,272 winnable deals | Exhaustive search, first solution found |
That last row is worth a second look. In winnable deals, the first winning line the search found usually had about one stock card left over. Golf wins are tight: you generally need nearly the whole stock.
The King-and-Ace rule on this site
Many Golf versions stop at the ends of the deck: you can't play a King on an Ace or an Ace on a King. This site's Golf allows it — the ranks wrap around, so after a King, an Ace (or a Queen) will do, and after an Ace, a King (or a 2).
We tested both versions, and the difference is huge:
| Version | Winnable with perfect play | Greedy bot wins | First card has no play |
|---|---|---|---|
| With King–Ace wrap (this site) | 92.7% | 22.4% | 27.4% |
| Without the wrap | 44.0% | 3.2% | 31.2% |
Without the wrap, fewer than half of deals can be cleared at all, and a casual player wins about one in thirty. With it, almost every deal is winnable. That's why Golf here feels generous.
How Golf is set up
- Seven columns of five cards, all face up.
- One card starts on the foundation pile.
- A 16-card stock. Turn one card onto the pile when you have no play. There's no second pass.
- Playing: move the bottom card of any column onto the pile if it's one rank above or below the pile card. Suits don't matter.
- Wrap-around: Kings and Aces count as next to each other.
- Win by clearing all 35 cards from the columns. Leftover stock cards don't matter.
Casual play vs. careful play
A greedy player looks at the pile card, sees which column cards fit, and picks one. Our bot was slightly smarter than that — it chose the card that led to the longest chain it could see — and it still won only 22.4%. The full search, which knew the stock order, won 92.7%. You can't know the stock order either, but you can do much of what the search does: look for long chains, and keep your options open.
Five ways to beat the greedy baseline
Play the longest chain, not the first card
Before you play, trace the run you can make: 6, 7, 8, 7, 6… A chain of five beats a chain of two, even if the short one is more obvious.
Choose which way to go
From a 9 you can go up to 10 or down to 8. Pick the direction that leaves more cards playable afterward.
Ration the 16 stock cards
You have only 16 stock cards and winning lines use almost all of them. Only turn a stock card when you truly have no play.
Work the deep columns
All columns start with five cards, but as the game goes on, the columns you've ignored become the problem. Spread your plays so no column is left full at the end.
Remember the wrap
Players used to other Golf games forget that a King can go on an Ace here. It's often the play that saves the game.
Where Golf games are lost
- Taking a short chain when a longer one was available.
- Drawing from the stock while a legal play is still on the table.
- Emptying the easy columns and leaving one full column of awkward ranks.
- Forgetting the King–Ace wrap.
Play Golf Solitaire here
Free, in your browser, on a computer, tablet or phone — no download, account or email address. Large, clear cards with zoom; Hint, unlimited Undo, Restart and Auto Play; and drag or tap to move. Wins, time and moves are tracked for the current game, and you can turn sound off, go fullscreen or change the table color. No gambling and nothing to buy.
Like quick games? Pyramid is another fast one, built on pairs that add to 13. If you'd like something longer, try classic Klondike or Diplomat.
The research behind these numbers, step by step
Golf was tested in September 2026 by the play-solitaire-free-online.us team at Solitairen OOD, using the rules as coded on this site.
- Rules from our game code. Seven face-up columns of five, one card on the foundation, a 16-card stock with no redeal, plays one rank up or down, and — in the only Golf version on this site — Kings and Aces wrap around. Clearing the columns wins the game.
- Game rebuilt in code. We programmed Golf with and without the wrap so we could compare the two.
- Numbered deals. Seeds 1 to 10,000 with the wrap and 1 to 3,000 without, all our own numbering.
- Two players. A greedy bot that doesn't know the stock order, plays the column card with the longest visible chain and draws only when stuck (a floor). And an exhaustive search that knows the stock order and tries every possibility, so every deal ends as definitely winnable or definitely not (a ceiling no person can reach, because nobody knows the stock order).
- Running it. Both players on every deal, with and without the wrap.
- Published comparison. Blake & Gent (2026) report 45.109% of Golf deals winnable without the wrap, with the stock known. Our no-wrap result was 44.0% (range 42.2–45.8%), which includes their figure. That agreement is what gives us confidence in the with-wrap number, for which we know of no published figure.
- Margin of error. The "95% range" shows how much our sample could be off. For the site's version, the winnable share is between 92.2% and 93.2%.
- Limits. The search knows the stock order, which you don't, so 92.7% is a ceiling. The greedy bot is simple, so 22.4% is a floor. Your results should fall in between.
The code behind the test
Golf is coded in golf.py, in Python 3.11 with the standard library only. The exhaustive search is a depth-first search over the column heights, the position in the stock and the rank on the foundation, and it remembers dead ends so it never explores the same hopeless position twice. That's what lets it settle every single deal. The greedy bot is a few lines on top of the same game model. runner.py spread the deals over 19 worker processes on a 20-core machine: about 12 minutes for the 10,000 wrap deals and about 7 minutes for the 3,000 no-wrap deals. Ranges are Wilson score intervals.
An AI assistant, Claude by Anthropic, helped us write and review the Golf code and draft this page. The figures are not AI-generated: they came from running the code, and our no-wrap result was checked against the published study.
By Admin, play-solitaire-free-online.us (Solitairen OOD). Research dated September 18, 2026.
Questions new players ask
Do Kings and Aces connect in this version?
Yes, on this site. Kings and Aces wrap around, and that single rule more than doubles the share of winnable deals, from 44.0% to 92.7%.
Do suits matter in Golf?
No. Only the rank counts: one up or one down.
Why did I lose a deal that looked easy?
Usually a short chain was taken where a longer one was possible, or a stock card was turned too soon. Only 7.3% of deals here can't be won at all.
What does "no play on the first card" mean?
In 27.4% of deals, no column card fits the starting pile card, so your first move is a draw from the stock. That's normal and not a bad sign by itself.
Is Golf good for a short break?
It's one of the shortest games here: 35 column cards and a 16-card stock, with no second pass.