Klondike Turn 3: a practical playbook

Klondike Turn 3 is the same patience as ordinary Klondike with one change: when you turn the stock, three cards come over at once and only the top one of the three can be played. That single rule turns a relaxed game into a puzzle about timing. This page is a hands-on playbook for managing the stock: how the three-card turn really works, which cards you can reach on each pass, when to play from the waste and when to hold back, and a worked example. For our measured odds and the comparison with the published research, see Solitaire Turn 3, which plays exactly the same game.

How the three-card turn actually works

After the deal, 24 cards sit in the stock. Each click turns three of them onto the waste, but only the uppermost is available. Twenty-four divides neatly into eight groups of three, so on a first pass through an untouched stock you see eight playable cards: the 3rd, 6th, 9th and so on down to the 24th card of the stock.

The other sixteen are not lost. They are simply out of reach until something changes the grouping. And the one thing that changes it is playing a card from the waste.

Why playing a waste card shifts everything

Suppose the top card of a group of three is playable and you use it. The card beneath it now becomes the top of the waste and is available straight away. More importantly, the stock no longer splits into the same triples on the next pass: every later group is offset by one card, so cards that were buried in the middle of a triple can surface. Each card you take from the waste reshuffles which cards you will see next time round, without any actual shuffling.

On this site you can go through the stock as many times as you like, so the question is never "will I see this card again?" but "how many passes will it take to line it up?"

Reading the stock, pass by pass

Pass one: survey

Treat the first run through the stock as reconnaissance. Note which of the eight reachable cards are useful, and roughly where the important ones (Aces, 2s, the cards your tableau is crying out for) sit in the cycle. Play the obvious ones, but don't feel obliged to use every playable card on the first go.

Pass two: take what the tableau needs

By now the tableau has usually changed. Cards that had nowhere to go on the first pass may fit, and every waste card you use slides the grouping along.

Later passes: work the offsets

If a card you need keeps landing in the middle of a triple, look for a waste card you can play just before it comes round. Using one card earlier in the pass shifts that later triple by one, and the card you want may become the top card.

In our own simulation, games that our computer player won at Turn 3 went through the stock 3.8 times on average, compared with 2.3 times at Turn 1. Several passes are normal; there is nothing wrong with going round again.

Play it now or hold it back?

At Turn 1 you can almost always take a playable card. At Turn 3, taking a card has a cost as well as a benefit, because it changes which cards appear later. A few rules of thumb:

• Play low cards to the foundations as soon as you can. An Ace or a 2 sitting in the waste is doing nothing useful.
• Play a card that frees a face-down card in the tableau. Revealing hidden cards is still the top priority.
• Think twice before playing a card that only tidies, such as moving a card from the waste onto the tableau when it neither frees anything nor leads anywhere. It shifts the triples and may push a card you need out of reach.
• Remember the card underneath. When you play a waste card, the next card down becomes available. Two useful cards in a row can be worth taking together.

A worked example

Here is an illustrative situation (not a real deal from our study) to show the thinking.

1. The waste shows 4♥ on top, with J♣ and 2♠ underneath it. Your tableau has a black 5 showing, and the A♠ is already on a foundation.
2. Playing the 4♥ onto the black 5 is legal. It also exposes the J♣, which you have no home for — but beneath that sits the 2♠, which you want.
3. You can't reach the 2♠ this turn, because the J♣ is in the way. But by playing the 4♥ now, you have shifted the stock by one card for the next pass.
4. Next time round the J♣ and 2♠ will fall into different triples. Count the cards: if the 2♠ lands in the top position of a group, it is yours.
5. If it doesn't, one more waste card played earlier in that pass will move it again.

That counting habit — "which cards will be on top next time?" — is the heart of good Turn 3 play.

Turn 3 mistakes that cost games

• Clicking through the stock without looking. Every pass tells you where the cards are. Rushing through wastes the information.
• Using every playable waste card by habit. Sometimes holding back keeps a better alignment for the next pass.
• Forgetting the tableau. The stock is only half the game. Reveal face-down cards in the columns before you start turning.
• Emptying a column too early. As in all Klondike on this site, only a King can go into it.

How much harder is it?

A great deal. In our tests the same simple computer player that won 42.3% of deals at Turn 1 won just 10.6% at Turn 3 over 10,000 deals (95% range 10.0% to 11.2%). The published St Andrews study puts the share of Turn 3 deals that could be won with every card visible at 81.945%, so most deals are still winnable in principle; the difficulty is reaching the right cards at the right time. The full comparison is on the Solitaire Turn 3 page.

Where those figures come from, step by step

1. Rules from our code. Three cards per turn, only the top one playable, unlimited passes through the stock, Kings only into empty columns.
2. The program. A Klondike game written in Python, with a rule-based computer player that sees only face-up cards, takes the top waste card only, and never takes a move back.
3. The deals. 10,000 deals, each shuffled from a numbered seed so the test can be repeated.
4. Counting passes. For each won game we counted how many times the player went through the stock.
5. The published figure. Blake and Gent's peer-reviewed study (Journal of Artificial Intelligence Research, 2026) used a solver that sees every card, so its 81.945% describes the deals, not what a player achieves.
6. What we can't say. Our player is simple; a thoughtful human with Undo will do better than 10.6%.

The technical side

The simulation is Python 3.11 with no outside libraries, in klondike.py, run across the cores of one machine with Python's multiprocessing. The 95% range is a Wilson score interval. Claude, Anthropic's AI assistant, helped us write and review the code and draft this playbook; none of the numbers came from the AI — they come from running the program and were checked against the St Andrews study.

Turn 3 questions

Is Klondike Turn 3 different from Solitaire Turn 3?

No. It is the same game and the same rules; the two pages simply look at it from different angles.

Can I switch to one card at a time?

Yes: Klondike Solitaire turns one card at a time.

Is there a limit on passes through the stock?

No, go round as often as you need.

What happens to the two cards under the top one?

They stay in the waste. Play the top card and the next one down becomes available straight away.

What is good practice for Turn 3?

FreeCell trains the same planning habits with every card visible.

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