ADX Trend Short

Overview

The ADX Trend Short strategy combines two components of the Directional Movement System: ADX (Average Directional Index) measures raw trend strength regardless of direction, while DMI− (Minus Directional Indicator) confirms the bearish direction. A SHORT entry fires only when both agree — strong trend and downside direction confirmed by a fresh cross.

Discovered 2026-05-04 via a 4,116-combination parameter sweep across 50 coins. The 12H timeframe with SL 15% / TP 5% produced an OOS/IS profit-factor ratio of 1.01 in that run. At the time we read the lack of decay as a sign of a durable edge. A fresh out-of-sample re-test on 2026-06-28 (product engine, regime-robustness + 5-axis adversarial kill) overturned that: like the other crypto SHORT presets, the result is bear-beta — both windows were falling markets, so a short book wins on direction rather than on the ADX/DMI filter. The backtest is reproducible; we no longer present this as a confirmed live directional edge.

How It Works

1. ADX filter — ADX(14) must be above 25, confirming the market is in a trending (not ranging) regime
2. DMI cross — DMI− (selling pressure) crosses above DMI+ (buying pressure) on the current bar
3. Entry — SHORT on the next bar’s open after the cross is confirmed
4. Exit — TP 5% / SL 15% / 4-bar (2-day on 12H) timeout

Why It Works (Thesis)

ADX above 25 filters out choppy, low-volatility markets where crossover signals are noise. By requiring a concurrent DMI− cross, the strategy enters after directional commitment has been established — not on anticipation. The result: a high-WR setup (65.5%) where most entries are aligned with genuine trend momentum rather than mean-reverting noise.

The 12H timeframe eliminates intraday noise that plagues 4H and 6H signals, producing cleaner trend confirmation with more decisive follow-through. The wide SL (15%) accommodates the natural volatility at this timeframe while the tight TP (5%) locks in profits quickly — the market rarely gives back gains once a 12H trend is established.

Results (2-year backtest, IS/OOS split, measured 2026-05-04)

Metric	IS (May24–May25)	OOS (May25–May26)	Combined
Total trades	239	248	487
Win rate	~65%	~66%	65.5%
Profit factor	1.68	1.70	1.68
Coins profitable	37/50	36/50	—

OOS/IS ratio: 1.01 — OOS profit factor (1.70) exceeded IS (1.68) in the original run. This looked like a no-decay edge, but the 2026-06-28 re-test shows both the IS and OOS windows were bear markets, so the stability reflects persistent bear-beta rather than an edge that survives a regime change.

Multi-window (4 rolling 6-month periods, 6H scan): W1=2.02 · W2=1.69 · W3=1.28 · W4=1.27 — profitable in all 4 original windows. These numbers are reproducible but, like the headline metrics, are attributed to bear-beta after the fresh OOS re-test.

Caveats

• Failed fresh out-of-sample regime-robustness (2026-06-28). The earlier “zero decay” framing did not survive — the SHORT result is bear-beta, not a regime-independent edge. Status downgraded from verified to testing for this reason.
• 12H signals fire roughly once per coin every 2–3 days — lower frequency than shorter timeframes.
• Wide SL (15%) means individual losses can be substantial; manage position size accordingly.
• 487 total trades across 50 coins over 2 years — thin per-coin sample; diversification across coins is important.
• Not live-tracked on OKX. Backtest only — “verified” would mean the backtest is reproducible, not that the live directional edge is confirmed.