Flexa Intel logo

Trading Strategy

Resample a real trade history to see the full range of outcomes — and the chance of ruin

Overview

A simulation that takes your historical per-trade profit and loss and replays it thousands of times in different random orders. One backtest shows a single path; this shows the whole distribution of paths the same edge could have produced — including the bad ones that a lucky sequence hid.

Answer the question: "My backtest returned 40% — but what is the realistic range of outcomes, how deep a drawdown should I expect, and what are the odds this strategy blows up the account?"

Data — What data do you need

Field Power BI field / example Description
Trade DateTradeDateClose date of each trade. Orders the history and sets the period the forecast runs from.
Profit / LossSum of PnLResult of a single closed trade in account currency — negative for losers (e.g. -55.74, 152.55).
⚠️
One row per trade, not per day. The simulation resamples individual trade results. If your table already aggregates P&L to a daily or monthly total, it has fewer, larger observations to draw from and the outcome range will look artificially narrow.

Use Case — Position Sizing for a Systematic Strategy

Scenario: A trader has 190 closed trades from a systematic strategy with a 58% win rate. The backtest looks strong, but it is one ordering of those trades. Before committing more capital the trader wants to know how the same edge behaves across thousands of re-orderings, and whether 2% risk per trade survives a bad run.

Configuration:

  • ▸Problem Type: Trading Strategy
  • ▸Date: TradeDate
  • ▸Values: Sum of PnL — one row per closed trade
  • ▸Initial balance: $10,000
  • ▸Risk per trade: 2%
  • ▸Forecast horizon: 12 months

Sample output — outcome range across 10,000 resampled paths:

PercentileEnding BalanceReturnMax DrawdownLongest Losing Run
P5 (bad run)$8,400−16%−31%9 trades
P25$11,900+19%−22%7 trades
P50 (median)$14,600+46%−17%6 trades
P75$17,800+78%−13%5 trades
P95 (good run)$22,300+123%−9%4 trades
P(ruin)1.8% of paths———
📉
Reading the result: The median path is the number to plan around — not the backtest. The gap between the 5th and 95th percentile is how much of the backtest was sequence luck. Maximum drawdown is what the account has to survive before the edge pays out, so it decides position size; P(ruin) is the share of paths that never got there.
Trading Strategy simulation output showing Most Possible, Worst and Best scenario details

The visual reports three scenarios side by side — Most Possible, Worst and Best — so the downside is never one click away.