Backtest it, then
try to break it.
A backtest that looks great proves almost nothing on its own. BuildYourPocket runs an idea through a full lifecycle — author, backtest, gauntlet, certify, forward-test, deploy — every stage draft-only and gated, designed to keep only the edges that survive honest scrutiny.
From idea to honestly-tested,
one gated stage at a time.
A promising backtest is a hypothesis to be attacked, not a promise to be trusted. Each stage has to be earned before the next unlocks — and nothing ever leaves the draft state to trade real money.
Author
Express a strategy as rules — entries, exits and sizing — using the shared condition engine. No code required.
Backtest + Gauntlet
Run it against history with realistic frictions, then subject the result to the statistical gauntlet.
Certify
Only an idea that clears the gauntlet is certified — and the certificate carries its caveats, not just its wins.
Forward-test
Watch it on unseen, out-of-sample conditions before you trust it — the honest test history can't give you.
Deploy — draft-only, gated
A deployed strategy still stops as a draft: it proposes and alerts, and every action is yours to place through your own broker. BuildYourPocket never auto-trades real money.
Statistics that keep
you honest.
The gauntlet applies well-known robustness techniques so a strategy has to earn your trust rather than flatter it. It is a tool for measuring whether an edge is real — not a promise that it will keep working. Its whole purpose is to guard against over-fitting: the trap of a rule that only ever worked in the past you happened to test on.
Illustrative gauntlet output — synthetic, no real strategy or returns.
Every idea, judged by
the gauntlet — in the app.
The Strategy Research surface as it appears in BuildYourPocket: each candidate carries its lifecycle stage, its gauntlet verdict and its over-fitting metrics. Demo strategies, hypothetical results.
Strategy Research
Author → backtest + gauntlet → certify → forward-test → deploy. Draft-only, gated.
Strategy validation
DemoDeflated Sharpe (DSR)
Overfit prob. (PBO)
The gauntlet judges every idea. Nothing auto-trades; results are hypothetical and shown with their limits.
Internal read-only research workspace · a visualization of existing reports · not investment advice · no orders · no execution · no forecasts.
The numbers that describe
an edge — and its cost.
A strategy is judged on more than one flattering figure. The gauntlet reports the measures that tell you about consistency, downside and durability — not just the headline return.
Sharpe & Sortino
Return earned per unit of risk — Sortino focusing on the downside that actually hurts.
Calmar & Max Drawdown
Return set against the deepest peak-to-trough fall — reward measured against pain.
Profit factor & expectancy
How much is won versus lost, and what an average trade is really worth over time.
Hypothetical is
labelled hypothetical.
Backtests and simulations are hypothetical by nature. They rely on assumptions, they can differ materially from live outcomes, and no test — however well it scores — makes future profit certain.
No profitability guarantee
Passing the gauntlet measures robustness, not future returns. Nothing here promises you will make money.
Past ≠ future
Past and modelled results do not guarantee future results. Live conditions can differ materially.
Draft-only, not advice
Nothing here is a recommendation, and BuildYourPocket never auto-trades real money.
Test your edge before
the market does.
Author an idea, run the gauntlet, and keep only what survives. Start free today.