Build

Reversion’s algorithm builder lets you define rule-based trading strategies, backtest them against historical data, and optimize parameters — all without writing code.

How It Works

Define Algorithm (entry/exit rules, indicators, position sizing)
        ↓
Backtest (simulate against historical candles)
        ↓
Analyze (metrics: Sharpe, win rate, drawdown, indicator usefulness)
        ↓
Optimize (Bayesian search over parameter space)
        ↓
Deploy (run live with the same config)

Algorithms are versioned — each update creates a new immutable version, so you can backtest and compare iterations without losing previous configs.

What’s in This Section

Guide

End-to-end walkthrough: pick a market, choose indicators, define entry/exit rules, run a backtest, interpret results, and optimize. Start here if you’re new.

Algorithms

Reference for algorithm configuration: types (LONG/SHORT/BOTH), entry and exit conditions, indicator groups, position sizing, order types, stop loss and take profit, timeout modes, and versioning.

Backtesting

How the backtesting engine works: running backtests, the simulation pipeline, understanding swap metrics and algo metrics, equity curves, capital scaling, and fee modeling.

Optimization

Parameter optimization: Bayesian search with Optuna, rolling cross-validation for robustness, custom scoring functions, and the suggest-score-iterate workflow.