Solitaire Turn 3

New Game
Restart
Turn 1
Turn 3
Undo
Hint
Auto Play
Fullscreen Button

Solitaire Turn 3: the same game, a much steeper hill

Turn 3 is classic Klondike with one change: each time you click the stock, three cards flip over and only the top one is yours to play. The layout, the columns and the goal are identical to the Turn 1 game. The difference in results is not small.

We had a computer player work through 10,000 shuffled decks the way a relaxed, first-time player would — take the obvious move, never look ahead, never undo. At Turn 1 it won 42.3%. At Turn 3, on the very same kind of deal, it won just 10.6%. That's about one game in ten, and it's the fairest answer we have to "how often will I win Turn 3 if I just play?"

The good news: that figure is a floor, not a ceiling. Below you'll find what it looks like when you add a plan and the Undo button to the mix.

Turn 3, measured deal by deal

MeasurementTurn 3 resultDealsHow it was measured
Win rate with no look-ahead and no undo10.6% (95% range 10.0–11.2%)10,000Rule-following bot on visible cards only
Same bot at Turn 1, for comparison42.3%10,000Same bot, one-card draw
Passes through the stock in a winning gameabout 3.8 (vs. 2.3 at Turn 1)1,063 winsSame bot
Moves in a winning gameabout 111, including stock flips1,063 winsSame bot
Deals proven impossible even with every card visible9.5% (vs. 2.1% at Turn 1)1,000Solver that sees face-down cards
Deals proven winnable with every card visibleat least 66.2%1,000Same solver

Two things stand out. The casual win rate falls to a quarter of its Turn 1 value, and the share of deals that are simply dead — unwinnable no matter what — more than quadruples. Turn 3 isn't only harder to play well; more deals are lost before you start.

Why flipping three cards changes so much

At Turn 1 you see every stock card, one at a time, on every pass. At Turn 3 you only see every third card. The other two are stuck underneath until you play the top one — and playing a card from the waste shifts which cards come up on the next pass.

That's why winning games needed more trips through the stock (3.8 on average against 2.3 at Turn 1). Each pass, the cards line up a little differently depending on what you took off the waste last time. A card that was hidden on your first pass may pop up on the third.

The Turn 3 rules on this site

  • Seven columns dealt from one to seven cards, with only the bottom card of each face up.
  • A 24-card stock. Each click turns three cards onto the waste; only the top waste card is playable.
  • Unlimited passes. When the stock is empty, you can turn the waste over and go again as often as you like.
  • Build down in alternating colors in the columns, and move any face-up run as a group.
  • Only a King (or a run starting with a King) can fill an empty column.
  • Foundations climb by suit, Ace first and King last. One card at a time may be taken back down from a foundation if you need it.

Beating the one-in-ten baseline

Count the stock before you touch it

On your first pass, pay attention to which cards show up on top. You'll see eight of the 24. Knowing that a black 7 is coming up third tells you whether a column move now is worth making.

Change the rhythm on purpose

Every time you play a card off the waste, the three-card groups on the next pass shift by one. If a card you need keeps landing second or third in a group, playing one waste card — even a so-so one — can bring it to the top next time around.

Don't fear the buried Aces

At Turn 3 our bot won 10.1% of deals with two or more Aces hidden in the columns and 12.3% of deals with none hidden. That's a real difference, but a modest one. The stock, not the opening layout, decides most Turn 3 games.

Lean on Undo harder than you would at Turn 1

Because waste plays change the next pass, a single choice can decide the whole game. When two moves look equally good, try one, run through the stock, and undo back if it goes nowhere. The bot can't do that, which is a big part of why it wins only 10.6%.

Know when a deal is dead

Our solver proved 9.5% of Turn 3 deals impossible even when it could see every card. If you've tried two or three honest lines and nothing moves, it may be one of those. Start a new deal without guilt.

Where Turn 3 games slip away

  • Clearing the waste too eagerly. Playing every waste card the moment it fits can scramble a stock order that was about to give you what you needed.
  • Moving a King into an empty column with nothing to follow. You may need that space for a different King that's holding cards down.
  • Racing cards to the foundations. At Turn 3 you need every place to park a card; a foundation card can't catch anything.
  • Quitting after one pass. Winning games went through the stock almost four times on average.

Features of Turn 3 on this site

It's free and plays in any modern browser on a desktop, laptop, tablet or phone — nothing to download, no account, no email. Cards are big and clear, and there's a zoom control. You have Hint, unlimited Undo, Restart for the same deal, and Auto Play to finish a game that's won. Drag a card or tap it and tap its target. The game keeps track of wins, time and moves for the game in front of you, and you can turn off the sound, go fullscreen or change the table color.

If Turn 3 is wearing you down, go back to Turn 1 for a warm-up, or try FreeCell, where every card is visible from the start. For a completely different puzzle, Pyramid is all about pairs that add to 13.

The research behind these numbers, step by step

This study was run in September 2026 by the play-solitaire-free-online.us team at Solitairen OOD, using the game exactly as it plays on this site.

  1. Rules from our own code. From the game's JavaScript: three cards per draw, unlimited redeals, alternating-color builds, whole face-up runs can move, Kings only into empty columns, one card allowed back from a foundation.
  2. The game, rebuilt. We wrote Turn 3 Klondike as a program and ran a rules check over 1,000 Turn 3 games (plus 1,000 at Turn 1) to confirm every bot move was legal.
  3. Numbered deals. Seeds 1 to 10,000 produce the shuffles, so the test can be repeated card for card. These are our own numbers, not Windows or site deal numbers.
  4. Two players. A bot that plays only the top waste card it can see and never undoes (a floor), and a solver that knows every card (a ceiling that no person can reach).
  5. Running it. 10,000 deals for the bot; 1,000 for the solver, each search capped at 150,000 positions. Deals not finished by then are counted as undecided (24.3%), not as wins or losses.
  6. Published comparison. Blake & Gent (2026) put Turn 3 winnability at 81.945% when every card is known. Our solver's range — 66.2% proven, up to 90.5% if every undecided deal were winnable — includes it.
  7. Margin of error. The 95% range is where the true value almost certainly lies. The bot's 10.6% is really somewhere from 10.0% to 11.2%.
  8. Limits. The bot is simple and a patient human will beat it. The solver's range is wide because of the undecided deals. Neither predicts your next game.

The code behind the test

The code is Python 3.11 with nothing but the standard library. Turn 3 shares klondike.py with Turn 1; the bot is the same checklist player, restricted to the top card of each three-card group. The trickier part is the solver: at Turn 3 you can't reach any stock card you like, so it models exactly which cards can come to the top of the waste given what's already been played. It's a depth-first search that keeps a record of visited positions. runner.py spreads the deals over 19 worker processes on a 20-core machine; the bot's 10,000 games took about 8 seconds, the solver's 1,000 deals about 17 minutes. Error bars use the Wilson score interval.

An AI assistant (Claude, by Anthropic) helped us write and check the code and draft this text. It didn't produce the results — those came from running the simulations — and we compared them with the published research before publishing.

By Admin, play-solitaire-free-online.us (Solitairen OOD). Research dated September 18, 2026.

Questions new players ask

Is Turn 3 the "real" Solitaire?

Both are real; they're just different settings of the same game. Turn 1 is kinder, Turn 3 is the bigger challenge. Pick whichever you enjoy.

How often should I expect to win Turn 3?

Playing on instinct without undoing, roughly one game in ten, based on our bot's 10.6%. With careful use of Undo and some stock-counting you can do noticeably better.

Can a Turn 3 deal be unwinnable from the start?

Yes. Our solver proved 9.5% of deals can't be won even with every card visible, against 2.1% at Turn 1.

Why does the same deal play differently the second time through?

Because every card you take off the waste shifts the three-card groups on the next pass. That's the main puzzle of Turn 3.

What should I play when I need a break from Turn 3?

Golf Solitaire is quick and relaxed, and one-suit Spider wins more often than you'd think.