Two coins. One clock. No oracle.
A battle is not a bet on whether a coin goes up. It is a bet on whether yours outruns theirs. Both can crash and you still win, as long as yours crashes less.
Someone opens a battle
They pick the two coins, set the clock, and put up seed money so the odds start at even. Anyone can open one.
You pick a side
Stake, and you get shares in that corner. Shares are what pay out — one share of the winning side redeems for an equal slice of the whole pot.
The odds move as people back a side
Pile into one corner and its shares get expensive; the other gets cheap. A cheap share pays more. That is the entire reward for backing early or backing the underdog — priced by the crowd, not by a formula.
You can cash out any time
Changed your mind, or the odds swung your way and you want to lock it in? Sell your shares back before the clock runs out. You are never trapped in a position.
The clock hits zero
The contract reads each coin’s own pool and asks one question: which climbed further, in percent, over the window? Not which is worth more — which moved more. A coin at a fraction of a cent beats a five-dollar coin if it outran it.
Winners redeem
Winning shares split the whole pot — both sides’ money — plus every fee collected while the odds were being traded. Losing shares are worth nothing.
Why there is no oracle
Most prediction markets bet on something outside the chain — an election, a match, a price on an exchange. A blockchain cannot see any of that on its own, so it needs an oracle: a feed, a reporter, a dispute window, and somebody to trust.
A battle asks a question the chain can already answer. Each coin’s own liquidity pool keeps a running record of its price, on the chain, as a matter of arithmetic. Settlement reads that record at the start and again at the end and subtracts. There is nothing to report and nothing to dispute, and once the clock runs out anyone can settle it — including you.
Why a whale cannot just buy the win
The number that settles a battle is a time-weighted average across the whole window, not the price at the final second. Pumping a coin in the last minute moves the average by almost nothing — to shift it you would have to hold the price up for hours, which is indistinguishable from the coin genuinely winning.
The same is true of the odds: backing a side late costs more and pays less, because the crowd has already priced it. There is no edge in waiting.
What can go wrong
If both coins move identically, if a corner has nobody in it, or if a coin’s pool stops answering, the battle is called off and every stake is refundable. No admin can seize a pot, and no failure path leaves money stuck in the contract.