Spider Solitaire with 4 suits: nearly every deal can be won, almost nobody wins on autopilot
Four-suit Spider is the full game: two decks, all four suits, and runs that only move when they're in one suit. It has a reputation as the hardest solitaire most people ever play, and our tests explain why in a single pair of numbers.
Published research found that about 98.5% of four-suit deals can be won by a solver that knows where every card is. Our casual computer player — one that looks at the visible cards, plans a short way ahead and never undoes — won 4 games out of 2,000. That's 0.2%.
Almost every game is winnable, and almost none are won by playing on instinct. Everything in between is skill, patience and the Undo button.
Four suits in figures
| What was measured | Result | Games | Source |
|---|---|---|---|
| Wins by a player who sees only face-up cards and never undoes | 0.2% (4 games; 95% range 0.08–0.51%) | 2,000 | Our look-ahead bot |
| Suits completed per game, out of 8 | 0.05 on average | 2,000 | Same bot |
| Games where a column was cleared before the first stock deal | 6 games (0.3%), of which 2 were won | 2,000 | Same bot |
| Deals winnable when every card is known | 98.487% (±1.513%) | published sample | Blake & Gent, 2026 |
| Reported win rate for real players | 11% or less | not stated | solitairex.io, method not given |
Put those rows side by side and the message for a new player is clear: the four-suit deal in front of you is very probably winnable. Whether you win it depends on how you play, not on luck.
The difficulty ladder
We ran the same bot at every Spider level, which gives a clean view of how fast the game gets harder as suits are added.
| Suits | Bot win rate | Suits completed per game (of 8) | Games reaching 4+ suits |
|---|---|---|---|
| 1 (play) | 52.8% | 4.96 | 56.9% |
| 2 (play) | 5.7% | about 0.8 | 6.3% |
| 4 (this page) | 0.2% | 0.05 | 0.2% |
Each step up the ladder cuts the casual win rate by about ten times or more. At four suits, the bot typically doesn't complete a single suit.
Four-suit Spider rules on this site
- 104 cards from two full decks.
- Ten columns: 6, 6, 6, 6, then 5 in each of the other six. Only the bottom card is face up.
- A 50-card stock, dealt one card onto each column at a time, five times, with no redeal.
- Any card can go on a card one rank higher, whatever the suit.
- Only runs of a single suit can be moved together.
- Empty columns take any card or single-suit run.
- A full King-to-Ace run of one suit is removed automatically. Remove all eight to win.
The game warns you if you deal from the stock while a column is empty.
How real players close the gap
Think of Undo as part of the game
The difference between 0.2% and 98.5% is almost entirely about exploring. Our bot had to live with every move it made; you don't. Back up to test which column to open, as often as you need to. There's no penalty for it here.
Turn cards over first, tidy later
Hidden cards are the enemy. When you have a choice, make the move that turns a face-down card over, even if it's a messy mixed-suit move.
Protect every empty column
Empty columns are how you untangle mixed runs and re-sort them by suit. Don't fill one with a single card just because it's there.
Clear the table before each deal
A stock deal buries every column under a new card. Make every move worth making first, and fill any empty column, before you deal.
Measure progress by suits, not wins
Our bot averaged 0.05 completed suits per game. If you complete even one or two, you're already doing far better than play-by-instinct. Winning comes later.
What sinks most four-suit games
- Building long mixed-suit piles that can never move again.
- Dealing from the stock too early out of frustration.
- Spending an empty column on a single loose card.
- Not using Undo — the one thing our bot couldn't do.
What the four-suit game offers here
Free in any current browser, on computer, tablet or phone; no download, no account, no email. Big, readable cards and a zoom control help with ten crowded columns. You have Hint, unlimited Undo, Restart and Auto Play, and can drag cards or tap to move them. Wins, time and moves are shown for the game in progress. Mute the sound, switch to fullscreen, or change the table color any time. No gambling, nothing to buy.
If four suits is too much today, two suits is a fairer fight. For something else that's hard but fully visible, try FreeCell or Baker's Solitaire.
The research behind these numbers, step by step
This test was run in September 2026 by the play-solitaire-free-online.us team at Solitairen OOD, using four-suit Spider as it plays on this site.
- Rules from the game code. 104 cards in ten columns (6,6,6,6,5,5,5,5,5,5); a 50-card stock dealt one per column, no redeal; any card on one higher; only single-suit runs move; any card to an empty column; full suits removed automatically. Assumed: dealing isn't allowed while a column is empty. The game warns about this, and we didn't check in a browser whether it blocks the deal.
- Game rebuilt in code. We programmed Spider with all four suits, two of each.
- Numbered deals. Seeds 100001 to 102000. The bot was tuned on seeds 1 to 30, which were kept out of the results.
- The player. A bot that sees only face-up cards, searches up to 250 positions ahead per turn and never undoes. At four suits it is a very weak player: its result shows how hard the game is without Undo and deep planning, not how winnable the deals are.
- Running it. 2,000 complete games. We built no solver of our own for Spider, so for the ceiling we rely on the published figure.
- Published comparison. Blake & Gent (2026) estimate that 98.487% (±1.513%) of four-suit Spider deals are winnable when every card is known. That's our ceiling; the bot's 0.2% is our floor.
- Margin of error. With only 4 wins, the bot's true rate could be anywhere from 0.08% to 0.51%. The range is wide in relative terms because wins were so rare.
- Limits. Four wins and six early-empty-column games are tiny samples. Treat them as a picture of difficulty, not as precise rates.
The code behind the test
Four-suit Spider uses spider.py, written in Python 3.11 with the standard library alone. The bot is a best-first look-ahead: each turn it explores up to 250 positions over the cards it can see, treating anything unturned as unknown, and chooses moves by a score that heavily rewards same-suit links and uncovered cards and penalizes hidden ones. runner.py spread the 2,000 games over 19 worker processes on a 20-core machine, and the run took about three and a half minutes. Percentage ranges are Wilson score intervals, which behave sensibly even when the count is as small as four.
An AI assistant (Anthropic's Claude) helped us write and review the Spider code and draft this page. The results are from running the code, not from the AI, and we compared our framing with the published study before publishing.
By Admin, play-solitaire-free-online.us (Solitairen OOD). Research dated September 18, 2026.
Questions new players ask
Is four-suit Spider even winnable?
Very much so. Published research puts it at about 98.5% of deals when every card is known. The hard part is finding the path without seeing the hidden cards.
Why is my win rate so low?
You're in good company. A simple player without Undo won 0.2% in our test, and solitairex.io lists four-suit Spider at 11% or less for its players.
What should I aim for as a beginner?
Complete one suit per game. Then two. Our bot averaged 0.05.
How many moves did the bot's four wins need?
Our bot's four wins took about 209 moves on average, counting stock deals. That's from only four games, so take it as a rough guide.
Should I learn one or two suits first?
Yes. One-suit Spider teaches the rhythm of the game, and two suits teaches you to care about suits.