Ten Across deals a single pack into ten columns of five in a strikingly symmetric shape. The two outer columns are entirely face up; moving inwards, each column hides one more card than the last, until the two middle columns hide four beneath a single open card. Two more cards go to the two cells, one each, and the pack is exhausted.
The symmetry is not decoration — it is the whole strategic map. The edges of the board are open and workable from the first move; the centre is where the hidden cards live. Play tends to run from the outside in, and a game is usually decided by how early you can crack columns four and five open.
You build downwards in suit: a card may land on the card one rank above it of the same suit, and only of the same suit. When a column contains a complete run from king down to ace, all of one suit, that run leaves the board and counts as one collected set. Gather all four sets and the game is won. There are no foundation piles to feed one card at a time; the run is assembled in the tableau itself and removed whole.
The Scorpion move is the whole character of this line, and it is worth stating plainly. In almost every other solitaire you may only lift a tidy sequence. Here you may lift any face-up card, and every card resting on it comes along, in whatever jumbled order it happens to be. A king with a stray four and a nine piled on top travels as a unit of three. Nothing is asked of the tail beyond every card in it being face up.
The strategy that follows from this is unusual. Because a disordered tail travels, burying a card is far less permanent than it looks; what actually costs you the game is a card trapped under a face-down one, because you cannot carry what you cannot see. So the opening priority is always the same: turn hidden cards over. Prefer any move that exposes a face-down card, even a move that looks untidy, and be suspicious of a move that merely rearranges cards you can already see.
Ten Across is one of the games where implementations genuinely disagree, so it is worth saying plainly which one you are playing. Several other sites give Ten Across four ordinary foundation piles that you feed one card at a time, starting with the aces. This version does not. Here a foundation will only take a complete king-to-ace run of a single suit, assembled in the tableau and removed whole — the Spider rule, not the Klondike one.
We follow the open-source PySolFC collection, whose Ten Across is defined with Spider_SS_Foundation: a foundation that accepts exactly thirteen cards at once and nothing less. That single choice is what makes the game a long exercise in assembly rather than a race to unload aces, and it is why Ten Across sits beside Spider in this catalogue even though its board is Yukon’s. If you learned the card-by-card version elsewhere, both readings exist; this one is the cited one.
Before touching a card, count the face-down cards in each column. Those numbers are the real difficulty of the deal, and they tell you where the work is. Look for cards of the same suit sitting one rank apart in different columns, because those are the joins that will eventually become a collected set. Look also for aces buried deep: an ace is the last card of a set and a set cannot leave the board without it, so an ace under four face-down cards is a problem you will have to solve eventually and may as well start on early.
The first is treating the Scorpion move as a free pass. It is powerful, but every card you pile onto a column is a card sitting above something else, and depth eventually becomes its own obstacle. The second is chasing a tidy-looking sequence at the cost of leaving a face-down card buried; sequences can always be rebuilt, whereas a hidden card that never gets uncovered ends the game. The third is emptying a column with no plan for it. A space is the most valuable thing on the board, and filling it back in with the first card that fits usually wastes it.
This is a thinking game rather than a fast one. Most deals reward slow, deliberate play, and a single careless move early can quietly decide the outcome twenty moves later. Use undo freely while you are learning the shape of the board — the point is to understand why a line of play failed, not to protect a statistic.