Documentation

Backtest methodology

The numbers we publish on the homepage all come from a single, repeatable backtest with the parameters below. There is no curve-fitting per pair, no manual signal-selection, and no separate “optimisation” backtest you don't see.

Test parameters

InstrumentBTC/USDT (Binance perpetual)
Timeframe4 hour candles
Period2021-01-01 → 2025-12-31 (5 yrs)
Strategy modeLong + Short, signal-driven
Position size100% of equity per trade (compounding)
Stop lossATR-based, 1.5×
Take profitATR-based, 3.0× (R:R = 1:2)
Commission0.05% per side (taker fee modelled)
Slippage2 ticks per fill
RepaintNone — every signal evaluated on candle close
Min confluence score60 / 100 (default)

Headline results

Profit Factor2.05
Net Return+320% (compounded)
Win Rate65% on closed trades
Avg Winner+14.25% per trade
Avg Loser−3.61% per trade
Risk / Reward1 : 2 (stop : target)

What we did NOT do

  • No look-ahead bias. Every signal uses barstate.isconfirmed in Pine Script — signals fire on candle close only and never re-evaluate.
  • No per-pair tuning. Default parameters are the same on every market the indicator runs on. We did not search for the best parameters per coin.
  • No selective trade reporting.Every signal that fired in the test window is included — no cherry-picking winners.
  • No surviorship-bias data.BTC/USDT has been a continuously listed instrument over the test window. We don't test on coins that don't exist anymore.

Live ↔ backtest alignment

The Pine Script that produces backtest signals on TradingView is byte-for-byte the same script that fires live alerts to the SniperEdge AI webhook. There is no separate “live” vs “backtest” mode. If a signal fired in the backtest, the same conditions today fire the same signal live. This is uncommon — most indicators have a hidden optimisation step that doesn't exist in production.

Caveats — read these

  • Past performance ≠ future results.Markets change regimes. A strategy that worked 2021–2025 is not guaranteed to work in 2026.
  • Compounding amplifies drawdowns.The +320% number assumes 100%-of-equity sizing on every trade. Most real traders size smaller for sleep-at-night risk; your real returns will scale roughly linearly with sizing.
  • Funding fees on perpetuals are not modelled. On BTC perpetuals these are typically <0.01%/8h and don't change the headline numbers materially, but they do compound on long held positions.
  • Live execution introduces latency.Webhook → exchange round-trip is typically 200–800ms; on fast moves this can mean the fill is a few ticks worse than the modelled slippage above.

Want to run it yourself?

Subscribers get the indicator with TradingView's built-in Strategy Tester — you can re-run any of the numbers above on your own chart with your own settings.

Risk disclosure: Trading is risky. Backtested performance is hypothetical and may not reflect actual trading conditions. This page is technical information, not financial advice. Trade only with capital you can afford to lose.