Spider Solitaire with two suits: reading a game where colour matters
Two-suit Spider deals exactly like the one-suit game — ten columns, 44 of the 54 table cards face down — but half the cards are now hearts. You can still stack a heart on a spade of the next rank, but only runs of a single suit move together and only a full single-suit run from King to Ace counts. Every mixed stack you build is one you will later have to take apart.
We played 4,000 two-suit games with the same computer player we used for one suit: it sees only face-up cards and never undoes. It won just 5.7%, down from 52.8% with one suit. It finished less than one full run per game on average (0.8 of 8). And the early empty column, so powerful at one suit, got rarer: our player managed it before the first stock deal in only 21 of 4,000 games, and won 15 of those.
Two big caveats before you despair. First, this player is a weak one at two suits; a careful person with Undo will do far better. Second, the deals themselves are not the problem: published research finds 98.5% of four-suit deals winnable when every card is visible.
Two-suit Spider, by the numbers
| What we looked at | Result | Games | How we got it |
|---|---|---|---|
| Cards face down after the deal | 44 of 54 | every deal | From the rules |
| Games with a column emptied before the first stock deal | 0.5% (21 games) | 4,000 | Look-ahead player, no peeking, no Undo |
| Wins in those games | 15 of 21 | 21 | Same player |
| Wins overall | 5.7% (95% range 5.0–6.5%) | 4,000 | Same player (a floor, not the game's winnability) |
| Full runs completed per game, of 8 | 0.8 on average | 4,000 | Same player |
| Games with at least 4 runs completed | 6.3% | 4,000 | Same player |
| Moves in a winning game | about 155 | 228 wins | Same player |
How two-suit Spider works here
Set-up
104 cards, half spades and half hearts. Ten columns (four of six cards, six of five), bottom card face up, 50 cards in the stock.
Winning
Complete eight runs from King down to Ace, each in a single suit. Finished runs leave the table automatically.
Moves
- A card may go on any card one rank higher, whatever its suit.
- Only a run in one suit, in sequence, can be moved as a group.
- Any card or suited run can go into an empty column.
- The stock deals one card to each column; there is no second pass. You'll be warned if you try to deal with a column empty.
Reading a two-suit game
Look for same-suit pairs in the bottom row
Ten face-up cards is all you get at the start. A 7♥ that can go on the 8♥ is worth far more than one that can only go on the 8♠. Count how many of your possible first moves keep the suit together.
Mixed stacks are a loan, not a gift
Putting the 6♠ on the 7♥ uncovers a card now, but it locks those two together until you can break them apart. That's fine when the uncovered card is valuable. Just remember you're borrowing.
The empty column is rarer — so guard it
Our player emptied a column before the first deal in only 0.5% of games, down from 1.3% with one suit. When you do get one, use it to sort mixed stacks back into single suits rather than to store a stray card.
Aim for your first full run
Our player averaged less than one completed run per game. Your first King-to-Ace run in one suit clears 13 cards from the table and is a real turning point; steer towards it.
Undo is your look-ahead
The player in our study had no Undo and could not see hidden cards. That is exactly where you have the edge. Try a line, see what turns over, and back out if it's bad.
Why the jump from one suit is so big
With one suit, every sequence is a movable run. With two, many sequences are "false" runs that look tidy but can't be moved. Our player's wins fell from 52.8% to 5.7%, and its completed runs from about 5 per game to under 1. That's a measure of how much harder the sorting gets — but it's measured by a player without memory or Undo, so read it as the floor under your own results, not a prediction of them.
A first-minute checklist for two-suit Spider
The opening of a two-suit game gives you ten face-up cards and 44 hidden ones. This routine is built around the one thing that makes two suits hard: runs only move if they are in one suit.
- List the same-suit moves first. These build runs you can move later: a heart on a heart, a spade on a spade.
- Then list the moves that uncover a card. Some will mix suits. Take them if the uncovered card is likely to help, and remember you'll need to separate that stack later.
- Look at the short columns. Clearing one before the first deal is rare — it happened in 21 of our 4,000 games — but it's the strongest start you can have.
- Keep one column as a sorting space once you have it. Use it to rebuild mixed stacks into single-suit runs rather than as a dumping ground.
- Tidy before every deal. The stock puts a card on every column, which often lands on a run you just built.
- Use Undo to look. When two moves both uncover a card, try one, see what turns over, and back out if the other would have been better.
Our player didn't have that last option, which is a big part of why it won only 5.7% of games. You do.
Play two-suit Spider on klondike.nz
Free in any modern browser on desktop, tablet or phone, with nothing to install and no account or email. Big, clear cards with zoom, Hint, unlimited Undo, Restart and Auto Play; drag or tap to move; wins, time and moves tracked for the current game; sound off, fullscreen and table colour options. For an easier game, try one-suit Spider.
Where these numbers come from
The klondike.nz team ran this study in September 2026.
- Rules from our code. Two suits (we used spades and hearts), one card per column from the stock, no redeal, same-suit runs move together. We assumed you can't deal with an empty column, based on the game's warning; we didn't test it in a browser.
- A program version. The same engine as one-suit Spider, with two suits dealt.
- Numbered deals. Seeds 100,001 to 104,000, kept separate from the deals used to tune the player.
- The player. Sees face-up cards only, never undoes.
- Running it. 4,000 games.
- Comparison. No published figure exists for two suits; the four-suit research figure (98.5% with every card visible) shows the deals are not the barrier.
- Error bars. 5.0–6.5% is the 95% range around 5.7%.
- Limits. The player is weak at two suits. Don't read its win rate as how often the game can be won.
Under the bonnet
Python 3.11, standard library only, in spider.py. A shared runner uses multiprocessing to spread the games over 19 worker processes; the two-suit run took about 14 minutes. Each turn, the player searches 250 positions ahead over the visible cards with a best-first search, rewarding same-suit links and empty columns and penalising hidden cards. Every percentage comes with a Wilson score 95% interval. Claude, an AI assistant made by Anthropic, helped us write and review the code and draft this page; the numbers themselves come only from running the code, and we checked them against the published research.
Questions players ask
Is two-suit Spider twice as hard as one suit?
For our player it was much more than that: 5.7% wins against 52.8%. A human with Undo does better, but the jump is real.
Should I avoid building on the other suit?
Not always — sometimes it's the only way to uncover a card. Just know you'll have to undo that stack later.
How long is a winning two-suit game?
About 155 moves for our player, counting each stock deal.
Can most two-suit deals be won?
We didn't measure that directly. For four suits, published research says 98.5% of deals can be won with every card visible, which suggests the deal is rarely the problem.
Where can I find more of your findings?
Our blog has more from our research, and the games page lists everything you can play.
By Admin, klondike.nz. Study date: 18 September 2026.