Date: 2026-02-23
Our current Dynamic Grid strategy uses a fixed 2.59% spacing between buy and sell orders. But not all cryptos move the same way — Solana swings 5% in a day while Ethereum might only move 1.5%.
What if we adapted the grid spacing to match each crypto's actual volatility?
That's what Adaptive Grid v2.1 does. It uses the Average True Range (ATR) — a standard measure of how much an asset moves — to set the grid spacing for each crypto individually.
Fixed Grid (v1.0) Adaptive Grid (v2.1)
───────────────── ────────────────────
Same spacing for Wider spacing for
every crypto volatile assets (SOL, DOGE)
(2.59% always) Tighter spacing for
stable assets (ETH, BNB)
We ran both approaches — fixed and adaptive — through the same 3 test periods with 13 cryptocurrencies. Same data, same blind forward test, only the spacing changed.
Across all assets and test periods (39 total observations):
Adaptive vs Fixed — Per-Asset Results
─────────────────────────────────────────────────────
Return win rate: 10/39 (26%)
Mean return change: -2.5pp
Mean Calmar change: +0.04
Range: -17.4pp to +14.5pp
Calmar CDaR Portfolio: Fixed vs Adaptive
─────────────────────────────────────────────────────
W1 Fixed: 77.2% Adaptive: 70.9% Δ -6.2pp
W2 Fixed: 45.8% Adaptive: 39.8% Δ -6.0pp
W3 Fixed: 10.8% Adaptive: 8.3% Δ -2.5pp
In this test, the adaptive approach did not consistently outperform fixed spacing on raw returns. However, it may still offer benefits for risk management and drawdown reduction that require further tuning.
The key insight: one spacing doesn't fit all. High-volatility assets like SOL and DOGE need more room to breathe, while stable assets like ETH can profit from tighter grids that capture smaller price movements.
This is Phase 1 validation — proving the concept works with a one-time ATR calculation. Phase 2 would implement continuous ATR monitoring that adjusts the grid as market conditions change in real time.
Study date: 2026-02-23 | 13 cryptocurrencies | 3 test periods | Blind forward test