Volume Profile POC

Overview

The Volume Profile POC strategy exploits the principle that price gravitates back toward the area of highest traded volume — the Point of Control (POC). When price drifts too far from the POC, the strategy enters a counter-trend position expecting reversion.

Discovered 2026-03-27 during a multi-strategy sweep on 10 coins. The original run showed near-zero correlation with the other strategies, which made it look like an independent diversifier. The important caveat: 10 coins is below our n≥30 minimum-sample standard, so this is not enough evidence to confirm an independent edge. The numbers below are reproducible but should be read as a small-sample lead, not a verified result.

How It Works

1. Volume Profile — computed over a rolling 168-bar (1-week) window using only completed bars (no look-ahead bias)
2. POC identification — the price level with the highest traded volume in the window
3. Deviation trigger — entry fires when current price deviates >3% from POC
4. LONG — price is below POC (sell pressure pushed it too far down; expect recovery)
5. SHORT — price is above POC (buy pressure pushed it too far up; expect reversion)
6. Exit — TP 5% / SL 2% / max-hold timeout

Why It Works (Thesis)

High-volume price levels represent market consensus — prices where the most participants are comfortable transacting. When price departs significantly from this consensus zone, it is often driven by short-term momentum or thin-liquidity moves rather than a genuine change in fair value. The 3% deviation filter keeps the strategy out of small, noisy moves. The 1-week volume window is long enough to identify stable equilibrium zones but short enough to adapt to changing market regimes.

The near-zero correlation with trend and breakout strategies would make it valuable for portfolio construction if it held up — it appears to profit during choppy, range-bound markets when trend strategies underperform. But on a 10-coin sample this remains a hypothesis, not a confirmed diversification benefit.

Results (2026-03-27 backtest)

Metric	Value
Coins tested	10
Coins profitable	10/10
Win rate	53.0%
Profit factor	1.14
Sharpe ratio	2.76
PF range	1.02–1.37

OOS 6/6 windows PASS — all 6 out-of-sample rolling windows produced PF > 1.00 (range: 1.03–1.15). This was encouraging, but all of it rests on the same 10-coin universe, which is below our n≥30 minimum-sample standard — so it does not meet the bar for a confirmed edge.

Portfolio correlation ≈ 0 with the other strategies in the original run. If it holds on a larger universe, this would offer genuine diversification — but on 10 coins it is a lead, not a confirmed property.

Default Parameters

Parameter	Value	Notes
Volume window	168 bars	1 week of 1H candles
Deviation threshold	3.0%	Minimum drift from POC to enter
Stop loss	2%	Tight SL appropriate for high-WR mean reversion
Take profit	5%	2.5:1 reward-to-risk ratio

Caveats

• Insufficient sample (n=10 coins), below our n≥30 standard. The original run only ever had a small-sample lead, never a confirmed edge — the re-measurement below is what settled it.
• Thin edge (PF 1.14) — small even in the original run. Position sizing should reflect this.
• Not yet validated on the full coin universe.
• Works best in range-bound markets; may underperform during strong sustained trends.
• Not live-tracked on OKX. Backtest only.

Update (2026-08-08): killed

The numbers above are the original 2026-03-27 run (10 coins, 1H). A current measurement on the simulator’s universe reversed the sign, and this preset is now published in the failures section of the grid:

	Original (2026-03-27)	Current (2026-08-08)
Profit factor	1.14	0.82
Win rate	53.0%	29.8%
Total return	—	-1,650.3%
Sharpe	—	-3.25
Trades	—	12,749

Source: src/config/simulator-presets.ts (volume-profile, measuredAt: 2026-08-08). We are leaving the original numbers in place rather than overwriting them — the gap between a small-sample lead and a full-universe measurement is the point.