The ~2.4% Floor
The house always wins. We measured the effective house edge on every Deriv binary contract type — and it sits at a flat ~2.4% floor. Across nine signal families and five indices — in-sample, out-of-sample, and against a permutation null — we searched for one strategy that beats the house. Zero survived. The edge is in the payout, not the market, so the only honest move is to show you the real odds and help you not blow up.
Try it yourself
Watch a simulated random walk with a 2.4% house edge. See how it inevitably drops to $0 over time.
Our Findings
The ~2.4% house edge floor
Universal across every binary contract; it's the payout, not the market.
View Paper →
Memoryless markets
Synthetic feeds are a random walk; charts show patterns (arcsine law) that carry zero information about the next tick.
View Paper →
Payout is the only lever
We pick the synthetics that pay the most back (~95.3%), lowering the break-even win-rate.
View Paper →
The martingale illusion
Deeper recovery raises the green-session rate but the mean stays negative and the tail explodes.
View Paper →
The circuit-breaker illusion
A session PnL stop doesn't help; only a hard depth cap bounds the loss.
View Paper →
Basket mean-reversion
The one real signal we found on real-underlying baskets, but the payout eats almost all of it. (Variance ratio)
View Paper →
Spike timing is Poisson clock
Step, Jump and Crash spikes arrive at random intervals; you can't time them.
View Paper →