Replay your trades thousands of times in different orders and mixes to see the range of outcomes a single backtest hides: drawdown, probability of ruin, losing streaks and a return band - plus a stress test of your portfolio against real market crises.
Free · no AI creditsOne snapshot, multiple scenarios
Compare a spot purchase with current holdings using one price and history snapshot. This preview never places an order.
Free: 2 legs, 1 saved result, 1 basic portfolio preview/day. Pro: 4 legs, 20 saved results, 3 comparable structures, 20 instruments and 20 new snapshots/day across both tools. All paid plans share these caps.
Private snapshots expire after 30 days. Download before expiry. Sliders and cached views use no new job or AI credit.
Live quote refresh and AI narrative are disabled pending rights, specifications and complete budget mapping. Manual mode is ready.
Manual simulation. Quotes, contract rules and fees are user declarations, not verified executable prices. No automatic venue or currency substitution.
Plain decimal with a dot, no separators, up to 12 decimal places. No missing value is assumed zero.
ISO timestamp with timezone, for example 2026-10-11T12:00:00Z.
Compare declared portfolio exposure with requested crypto, forex and equity history. Sources, native currencies, fallback and missing periods stay visible.
Loads required daily history from existing market sources. Crypto: Binance, Coinbase, Bybit, OKX or KuCoin; forex/equities: Yahoo then configured Twelve Data; Vietnam: existing broker sources. USD and USDT are distinct. Up to 64 weighted requests per job. A pinned venue stays on that venue; reference mode may use fallback. Current prices remain your declarations.
Equity closes use split adjustments where specified; Vietnamese sources retain provider closes with unverified corporate-action adjustments. Dividends, leverage, contract multipliers and automatic FX conversion are excluded.
A backtest or journal shows one path: the order your trades happened to arrive in. The stress test resamples those same closed trades thousands of times and reports the spread of outcomes - the percentiles, not a single number.
Percent-of-equity results compound; money results (journal, broker) are added to the starting capital you enter. Equity that reaches zero stays at zero. Every probability carries a ± margin from the number of paths.
Below 30 closed trades nothing is estimated. The simulation is seeded, so the same inputs always give the same answer.
It re-runs your list of closed trades thousands of times in random orders (and, with the bootstrap, random mixes) to show the range of drawdowns and returns the same edge could have produced - instead of the single path a backtest reports.
The share of simulated paths whose equity fell below your ruin threshold (for example 50% below the starting capital) at any point. It is shown with a ± margin because it is itself an estimate from a finite number of paths.
At least 30 closed trades; below that the tool shows no percentages. Results become noticeably more stable above 100 trades.
Shuffle isolates sequence risk (same trades, new order). Bootstrap also varies how many wins and losses you get, so it gives a return range. Block bootstrap keeps runs of consecutive trades together, which matters when your losses tend to cluster.
Your current holdings are replayed through real past episodes such as the March 2020 COVID crash, the 2022 LUNA and FTX collapses and the October 2025 liquidation cascade, using the prices each asset actually traded at. Assets that did not exist then are reported as not measured, never estimated.
No. It resamples history. It shows how fragile a track record is to luck and ordering; it cannot know about conditions that are not in your data.