Anyone can show you a backtest that looks perfect. The hard question is whether it keeps working outside the picture frame — so every strategy here has to face four separate tests.
The oldest trap in trading is called overfitting: tune a strategy hard enough on history and it will fit that history perfectly — the way a student who memorized last year's answer sheet aces last year's exam. Show either one a new question and they fall apart.
A single beautiful equity curve proves almost nothing. What matters is how a strategy behaves on data it never saw and under conditions it wasn't tuned for. That's what the gates are for.
History is cut into several windows, and the strategy is judged window by window as the test marches forward through time. It has to keep delivering across those slices — at least half of them must come up profitable — so one lucky stretch of history can't carry the whole grade. Winning once could be luck; winning as the calendar marches is much harder to fake.
The most recent stretch of the test period is graded separately, and its result is held up against the earlier stretch. An edge that shone early and faded late — the classic signature of a strategy that merely memorized its past — fails here. The gate asks a simple question: was it still working at the end?
The strategy's parameters and conditions get nudged — a little more here, a little less there. A real edge bends without breaking; a coincidence wearing a costume shatters the moment anything shifts. Fragile, one-off fits die at this gate.
Finally, the strategy's trades are reshuffled a thousand times. If the smooth ride only exists in one lucky ordering of wins and losses — a deep drawdown hiding behind a fortunate sequence — the shuffle exposes it. An equity curve that survives randomization isn't down to luck.
On the report, each gate a strategy cleared shows up as a green badge under the verdict — hover one for a one-line reminder of what it tested. Four badges means it survived everything above.
Be clear about what that buys you: the gates strip away self-deception, not uncertainty. Markets change. A strategy that earned every badge has proven it wasn't a fluke in the past — never that tomorrow must cooperate.
Validation reduces the chance a result is a statistical accident — it does not predict the future, and nothing here is financial advice. Always do your own research.