Five Fronts▶Play

Split twenty troops across five fronts. There is no best split — and the rival is watching yours.

How to play

You have twenty troops and five fronts, and you split them however you like. The rival splits their own at the same moment, and neither side sees the other until both are committed. Whoever has more troops on a front takes it, equal numbers means nobody takes it, and taking three of the five wins the round. Tap a front to add a troop and use the minus button to take one back; deploy once all twenty are placed. There is no best split, because whatever you choose something beats it, so the rival simply watches what you have been doing and aims at it. Repeating one split wins about a third of its rounds at first and under three percent once you have been read, while varying every round holds near half. The rival also gains a troop every few rounds, so the run always ends. Your score is the number of rounds you win.

Tips & strategy

The first thing to understand is why an even spread feels safe and is not. Four on every front means you win a front only by out-guessing nobody: any opponent who gives up one front entirely and puts five on the other four beats you three to two, and it costs them nothing. Spreading evenly is not caution, it is the single most predictable thing you can do.

The second thing is that the counter to that also has a counter. Five-five-five-five-zero loses to six-six-four-four-zero, which loses to something else again. This is not a flaw in the game, it is the game: there is no allocation that beats every other allocation, which is exactly why the rival cannot be solved and neither can you.

What follows is the only real advice. Do not decide on a shape you like and keep using it. The rival averages your recent rounds and aims at the average, so a habit is a target — and it does not need to be a conscious habit. Favouring the left two fronts, or always leaving exactly one empty, or always making your biggest pile the same size, are all patterns the average will pick up even if the exact numbers change.

The practical method is to let something outside your taste choose. Pick a different front to abandon each round, or deliberately place a pile where you put your smallest last time, or simply scatter without thinking and only check the total. Measured over thousands of runs, a player who varies genuinely lasts about four times as long as a player using the best single fixed split — and the gap is almost entirely in the late rounds, after the rival has had enough of your history to aim with.