Heikin Ashi Trend

Status: KILLED (2026-08-14 live re-measure). At the preset defaults this configuration ruins the account (-100% under fixed nominal sizing). Its defining trait is volume: 8,944 trades — the highest trade count of the six strategies in this measurement batch — and every one of them paid fees.

Overview

Heikin Ashi candles average and smooth raw OHLC data so that trends render as unbroken runs of same-colored candles. The classic reading — which this engine implements literally — is that a smoothed candle with no wick against the trend direction signals strong, one-sided pressure. The preset enters after a run of such candles, in either direction.

How It Works

1. HA candles — computed sequentially with no look-ahead: HA close = (open + high + low + close) / 4; HA open = (previous HA open + previous HA close) / 2; HA high/low = the extremes of the raw high/low and the HA body
2. LONG — 3 consecutive HA bullish candles (HA close > HA open) and the last one has no lower wick (HA low not below the body bottom, with a tolerance of 0.1% of the body size)
3. SHORT — mirror image: 3 consecutive HA bearish candles and no upper wick on the last one
4. Entry — next bar’s open after the signal candle
5. Exit — TP 8% / SL 7% (preset defaults)

Results (live engine re-measure, 2026-08-14)

Scope: top-10 coins, 1H candles, leverage 5x, fee 0.05% per side, window 2023-12-10 → 2026-08-14, fixed nominal sizing.

Metric	Value
Profit factor	0.91
Total return	-100.0% (account ruin)
Win rate	46.85%
Max drawdown	100.0% (account equity)
Sharpe	-2.22
Trades	8,944

Why It Loses

• Smoothing means lateness. HA candles lag raw price by construction, and the 3-candle confirmation adds more delay — by the time the signal fires, the move it is chasing is already mature, and the entry often lands near local exhaustion.
• The signal is too easy to satisfy on 1H crypto. Wick-free smoothed candles occur constantly in normal volatility, which is how the preset racked up 8,944 trades. High frequency is not an edge; it is a fee bill — at 0.05% per side and 5x leverage, every marginal signal pays a toll the thin per-trade edge cannot cover.
• Near-breakeven gross, dead net. PF 0.91 with a mildly favorable exit asymmetry (TP 8% vs SL 7%) and a 46.85% win rate means the raw signal is roughly a coin flip — and a coin flip minus thousands of fee payments compounds to 0. The Sharpe of -2.22 reflects that steady, high-frequency bleed.

Caveats

• -100% is this repo’s ruin notation: under fixed nominal sizing the account equity reached 0 inside the window.
• The verdict covers the combined both-direction book at the preset defaults on the stated scope. No claim is made about other timeframes, coin sets, or parameter values.
• Killed status: reproducible in the simulator, presented for education only.