The break-even rule
A contract bought at a price of p cents wins (100 − p)¢ and loses p¢. It breaks even when your win rate equals the price. Fees push the bar a little higher.
| Price paid | Win | Loss | Break-even win rate |
|---|---|---|---|
| 50¢ | +50¢ | −50¢ | 50% |
| 70¢ | +30¢ | −70¢ | 70% |
| 90¢ | +10¢ | −90¢ | 90% |
| 97¢ | +3¢ | −97¢ | 97% |
Two traders, same win rate
- Trader A buys at 90¢ and wins 9 of 10. Nine wins earn 9 × 10¢ = 90¢. One loss costs 90¢. Net: zero, and negative after fees.
- Trader B buys at 70¢ and wins 9 of 10. Nine wins earn 9 × 30¢ = $2.70. One loss costs 70¢. Net: +$2.00.
Same 90% win rate, very different results. What separates them is the price each paid relative to how likely the contract really was to win.
Edge is the number that matters
Edge is the difference between the true chance of a contract winning and the price you pay for it. If a contract really wins 93% of the time and you pay 90¢, you have about 3¢ of edge per contract before fees. If it wins 88% of the time, you lose about 2¢ per contract, however often you win.
That's why late-market contracts at 95–99¢ are rarely the "sure thing" they look like. They do usually win, but the price already reflects that, and one reversal takes back a long run of small wins.
How SpectralBot uses this
SpectralBot doesn't signal just because it's confident. A signal fires only when its calibrated confidence clears a fixed threshold and the market price leaves a minimum edge. A side the model likes, but that's already priced in, gets no signal. How SpectralBot decides YES or NO explains the rest.