Strategy robustness before capital deployment
Sign in

Published · Updated ·All research

Case Study: Validating a Freqtrade RSI Strategy with Kiploks (Real Numbers)

A practical case study of validating a Freqtrade RSI strategy: backtest checks, walk-forward evidence, cost drag, and final go/no-go decision.

This case study walks through a realistic workflow for validating a simple RSI strategy built in Freqtrade. The goal is not to show a perfect equity curve, but to show a disciplined decision process.

Setup

  • Pair universe: liquid USDT pairs only
  • Timeframe: 1h
  • Core logic: RSI threshold entries with symmetric exits
  • Validation stack: backtest, walk-forward, cost stress, deploy checklist

Step 1: Baseline backtest

The initial run looked attractive on headline return, but that alone is never enough. We inspected:

  • trade count adequacy,
  • drawdown depth and recovery shape,
  • period concentration of returns.

Step 2: Overfitting pressure test

Parameter sweeps can create false confidence. We checked whether performance survives small RSI threshold shifts.

Result:

  • one narrow parameter zone performed best,
  • neighboring zones degraded but stayed acceptable,
  • edge looked sensitive, but not purely singular.

Step 3: Walk-forward validation

We split the data into rolling IS/OOS windows and evaluated WFE consistency.

  • OOS retention remained positive in most windows,
  • failure windows appeared during high-volatility trend reversals,
  • WFE stayed near deployable range, but with caution flags.

Step 4: Cost realism

After adding realistic fees and stressed slippage assumptions:

  • net edge compressed materially,
  • a portion of trades moved near break-even,
  • turnover-heavy periods contributed most drag.

This changed position sizing recommendations.

Step 5: Final verdict framework

The strategy was not rejected, but also not approved for full-size deployment. Decision:

  • deploy at reduced size,
  • enforce tighter kill-switch thresholds,
  • re-evaluate monthly with fresh OOS windows.

Key lessons

  1. A strong backtest can still be fragile after costs.
  2. Walk-forward context is essential for deployment confidence.
  3. Partial deployment with monitoring often dominates all-or-nothing decisions.

Numbers change; the workflow stays

If you rerun this case study next quarter, the exact metrics will move.

The durable lesson is the sequence: export, stress costs, walk-forward, then size with governance.

Related articles