Manual Strategies

Six rules run with identical code on 13,736 gold and 16,926 EURUSD 5-minute bars. Every EURUSD scalping rule lost after a 1-pip spread, including two fades that captured +0.44 and +0.35 ATR before costs; on gold the same rules sat near break-even. The EMA 50/200 cross made +$501 on gold and lost 137.5 pips on EURUSD.

Four gold scalping rules — EMA 9/21 cross, RSI 7 fade, 1-hour breakout and Bollinger fade — run through Asia, London, the London–New York overlap and the New York afternoon on 13,736 5-minute bars. The overlap moved gold the most and averaged -$0.41 per trade with 2 of 8 tests positive; the quiet New York afternoon averaged +$0.85 with 7 of 8 positive.

Swing highs and lows coded as a rule and measured across 8,092 touches on 16,926 EURUSD 5-minute bars. The levels held 56.3% of the time — and 2,778 levels invented at random prices held 59.8%, in all five runs. The bounce rate turned out to be a function of zone width: 49.5% at 1 pip, 82.8% at 10, 94.3% at 20. Both the bounce and the break lost money, and first touches held better than third ones.

Engulfing bars, pin bars and stars coded as rules and run over 16,926 EURUSD 5-minute bars. All three were followed by a continuation of the prior move more often than a reversal, and 60% of engulfing bars formed with no move left to reverse. The market printed exactly 822 bullish and 822 bearish engulfings. Doji describes 20% of all bars, every one of them with a body of exactly zero.

MACD 12/26/9 on 16,926 EURUSD 5-minute bars produced 1,482 crossings, one an hour, split exactly 741 up and 741 down. The median signal was cancelled by the opposite signal after 45 minutes, and only 0.5% lasted four hours. Every holding period lost money before spread; traded always-in, it won 24.6% and lost 913 pips.

104 divergences on 16,926 EURUSD 5-minute bars. Entered at the pivot, they won 66.3% and made 544 pips. Entered at the bar you could actually see them, the same signals won 50.0% and made 133. The 25-minute wait for confirmation costs three quarters of the edge.

Traded the way Wilder wrote it, Parabolic SAR won 28.5% of 1,182 trades and lost 1,094 pips. The entry was not the problem. Held four hours instead of to the next dot, the same signals made +0.65 pips each.

I optimised 129 parameter sets on half of 16,926 EURUSD bars and ran the winners on the half they had never seen. The grid champion kept 0.8% of its result, and the ranking between halves correlated at 0.063.

Four common free arrow logics, 16,926 EURUSD 5-minute bars, no discretion. Win rates of 43.8% to 46.7%, and every one of them negative after a single pip of spread. One exception, and it is the slowest arrow on the longest hold.