Contents
Strategy pattern familiesGrid mechanics in detailAveraging ladders and safety ordersStatistical reversion approachesRange breakout entriesConfiguration starting pointsHow these strategies failStrategy pattern families
Automated trading strategies organize into four broad families: trend-following, statistical reversion, relative-value arbitrage, and liquidity provision. This catalog concentrates on the first two because they dominate deployed retail bots. Within those families, four patterns account for the overwhelming majority of running systems: interval grids, averaging ladders, oscillator reversion, and range breakouts.
| Pattern | Market hypothesis | Typical win rate | Dominant risk |
|---|---|---|---|
| Interval grid | Price oscillates within a range | sixty to seventy-five percent | One-sided accumulation in trends |
| Averaging ladder | Asset returns to average entry | Varies with depth | Permanently impaired asset |
| Oscillator reversion | Statistical snap-back to mean | fifty-five to sixty-five percent | Regime shift to trending |
| Range breakout | Momentum persists through levels | thirty-five to forty-five percent | Failed breakouts and fake-outs |
Grid mechanics in detail
A grid system places alternating buy and sell limit orders at fixed price intervals around a reference price. Each time the market moves through a level, one side fills and the bot immediately posts a counter-order at the next level. Profit accumulates from the interval spread, not from directional movement.
The critical engineering decisions are grid spacing (percentage or volatility-based), level count (which determines maximum inventory), and the inventory cap (which determines worst-case exposure).
Configuration starting point
grid: spacing_mode: atr # fixed, percent, or atr spacing_value: 0.5 # percent, or ATR multiplier levels_above: 6 levels_below: 6 order_size: 0.5 percent # of portfolio per level inventory_cap: 8 # max one-sided levels stop_loss: 12 percent # emergency exit
Averaging ladders and safety orders
Averaging bots open a base position and add at predetermined price intervals below entry. Each addition is typically larger than the last (the safety order pattern), which pulls the average entry down aggressively. The profit target is set relative to the averaged price, not the original entry.
Statistical reversion approaches
Reversion bots use a statistical indicator such as RSI, z-score of price against a moving average, or Bollinger band distance to identify when price has stretched away from its recent mean. They enter against the stretch and exit at the mean.
The setup works when markets are range-bound with persistent volatility. It fails when a regime shift turns stretched into a new normal, which is why regime filters (trend strength, volatility bands) are essential companions.
Range breakout entries
Breakout systems enter when price crosses a defined level (prior high or low, trendline, volatility band) with confirmation. The confirmation is usually volume or a close beyond the level, not just a touch. Exits use trailing stops or fixed targets rather than reversal signals.
Configuration starting points
Each pattern has a canonical configuration shape. Templates make backtests reproducible and parameters auditable.
shared_base:
risk:
max_portfolio_risk: 2 percent
max_drawdown_stop: 15 percent
correlation_cap: 0.7
execution:
order_type: limit
slippage_budget: 0.1 percent
timeout: thirty seconds
How these strategies fail
- Grid without inventory cap. In a sustained trend, the bot accumulates a position it cannot exit profitably. The cap is the difference between a grid and a slow-motion liquidation.
- Averaging without max drawdown. Safety orders exhaust capital before recovery. The stop preserves remaining capital, it does not catch the bottom.
- Reversion in a new regime. An oscillator can stay oversold for weeks when a market re-rates. Regime filters are not optional.
- Breakout on thin pairs. Low liquidity produces fake breakouts that immediately reverse. Volume confirmation filters most of these.