Canfield

Canfield is the game that came out of a gambling house. Richard A. Canfield reportedly sold players a deck for $50 and paid out per card they got to the foundations — a bet that worked because the game looks far more winnable than it is.

The mechanism behind that is the reserve: thirteen cards face down that you can only reach one at a time and in the order they happen to be in, and which you are obliged to draw from whenever you empty a tableau column. Add a foundation base rank chosen at random by the deal, and you have a game where the difference between what is theoretically winnable and what a good player actually manages is enormous — about 71% against 35%. Almost all of that gap is information you do not have.

Thirteen cards go face down into a reserve with only the top one showing, four columns are dealt, and one card goes up to start the foundations — whatever rank that card happens to be becomes the base rank for all four. Columns build down in alternating colours and turn the corner, so a King can go onto an Ace; foundations build up in suit and wrap the same way. A gap in the tableau is filled from the reserve rather than by any card you like, which keeps the reserve moving. The stock deals three at a time and you may go through it as often as you like while moves remain.

The reserve is the game. Thirteen cards you can only take one at a time, in a fixed order, and gaps must be filled from it — so your real objective is working through the reserve rather than advancing the foundations. Open a column when you can, because that pulls a reserve card into play, and be wary of blocking your own tableau with a card the reserve will need to displace. The unlimited redeals mean patience genuinely pays: cycling the stock again is free, and a position that looks stuck often is not. The article puts around 71% of deals as theoretically winnable but expert play at roughly 35%, and that gap is almost entirely the hidden reserve.