← Stokey

Strategy

A few real decisions, made well.

These assume you already know the rules — see How to Play first if you don't.

Bank or push? Read the chip, not the row

Pushing grows your row for free on your own turn, but every turn it stays unbanked is a turn an opponent could torch it. The board gives you a live read on exactly that risk. The same row is a different decision depending on what the chip says:

Most Kings already spent, one seat acts before you.

→ Push. A third King is worth 132; the odds of losing 62 are slim.


Kings still live, three seats act before you.

→ Bank the 62. It's close to a coin flip, and you lose the whole row, not the difference.


Identical cards, identical 62 on the table. The only thing that changed is the chip — that's the number the decision actually turns on.

There's also a safety net worth knowing: a still-lit row banks automatically when the shared deck runs dry. There's no benefit to banking early "to be safe" right near the very end — the risk that matters is an opponent's next turn, not the deck running out.

Torching is denial, not a payout

A torch pays you nothing. The burned row and your triggering card just become a bigger pot in the Bonfire, up for grabs by whoever draws the key next — you included, but not guaranteed. So the only question worth asking is how much it denies:

Worth it

Their row Q Q Q K ×3

Denies 147


Usually isn't

Their row 3 3 ×2

Denies 12


You gain 0 either way, and either way you give up growing your own row. The number on the left is the whole argument.

Track spent ranks

Every pair and every capture pulls cards out of play for good. A rank that's exhausted is dead — it can never pair you, never torch anyone, and never capture the Bonfire again. At a 4-player table the pool holds 16 copies of each rank (four decks comingled), so ranks die slowly, and knowing which ones have is a real edge:

A live rank. Your pair of 7s can still grow — and it's still a target.

A dead rank. Your pair of Kings can never be extended — and can never be torched. Bank it whenever you like; there's no rush and no risk.


The same count tells you whether a Bonfire whose key is a dying rank is likely to sit there untouched for a long time.

Watch the pot, but don't plan around it

One torch can swing the Bonfire sharply — and move the key while it does it, because the triggering card lands on top:

Before

You've been hoping to draw a 6.


Someone torches a ×2 Queens row with the Q♣

The pot tripled — and the rank you needed changed.


A fat Bonfire is worth watching, but capturing it takes one specific rank and you can't aim your flips. Let the pot pull your attention, not your plan.

Two lines: split the risk, not just the cards

Every placement is safe, so the choice is never about danger — it's about how much one opponent decision can cost you:

Everything on one line

Line 2 K cold

One torch on the 7s takes all 75


Spread across both

Line 2 K K ×2

A torch on the 7s costs 36 — the 52 survives


Stacking one rank deep is worth more per card, but it hands a single opponent one very valuable decision. Two lit lines also mean two ranks would have to come up against you.

Play your standing

The same move is right or wrong depending on the scoreboard:

Rival158
You96
Bot84
Bot71

Behind by 62. Torching the leader's 147 row closes more ground than any row you could realistically build this turn — and denial is one of the few ways to close a gap at all.


Flip the scores and the answer flips with them: from the front, protecting what you've banked beats spending turns on denial that no longer changes the standings.

Read the table, not just your own row.

Playing against the computer? Difficulty 1–10 changes how cautiously the bots bank and how eager they are to torch — higher difficulty plays closer to the ideas above, with less randomness in the calls.