Escape Route fairness and verification
Escape Route is built so that its fairness can be checked, not just claimed. This page explains the mathematics in plain terms and how the numbers were verified.
The model in plain terms
At every junction the player picks one of two routes. The Safe street has three roads, one blocked, so it holds with probability 2/3 and pays x1.5. The Risky alley has two roads, one blocked, so it holds with probability 1/2 and pays x2.0.
Each step pays exactly 1 divided by its survival probability. That means every junction is a fair bet on its own: the expected value of taking a step equals the value of stopping before it, before the house edge is applied.
The house edge is applied once, as a single 0.96 factor at cash-out. It is never hidden inside the steps, which is why the return is 96.00% no matter which routes the player picks or where they stop.
Why strategy cannot change the RTP
Because every step is individually fair and the edge is a single constant factor, no sequence of choices can raise or lower the long run return. An all-Safe player and an all-Risky player face the same 96.00% return. What changes is the shape of the ride: Safe gives frequent small climbs, Risky gives rare deep climbs.
The extreme cases make this visible. All-Safe to the top pays 124x and happens about once in 130 rounds. All-Risky to the top pays 3,932x and happens about once in 4,096 rounds. Same return, very different variance.
How it was verified
| Check | Value | Note |
|---|---|---|
| Method | Monte Carlo simulation | The full game logic is simulated, not approximated |
| Sample size | 2M+ rounds per strategy | Over two million simulated rounds for each tested strategy |
| Strategies tested | 6 | Including all-Safe, all-Risky, random choice and mixed stopping rules |
| Observed deviation | Under 0.5 percentage points | Every strategy lands within 0.5pp of the 96.00% target |
| Result | Strategy independent RTP | Player choices move variance only, never the return |
What this page is not
This is our own verification of our own prototype, not a certificate. The demo runs on an uncertified random number generator and play money only. Production for licensed markets requires a certified RNG, RGS integration and jurisdiction specific approval, and the mathematics would be re-verified by an independent test laboratory as part of that process.
The full math specification and simulation results are available to publishers and operators on request.
Questions about the model: sabine@humbleswede.com