MACD signal cross fails on 35 percent of tickers, MACD zero cross on 3 percent

MACD zero cross returns an average of 6.61 percent per ticker. MACD signal cross, built on the same MACD indicator with a different trigger rule, returns 1.18 percent. Zero cross fires when the MACD line crosses above or below zero. Signal cross fires when the MACD line crosses its own signal line, a smoother and more frequent trigger in theory. Same underlying calculation, two different rules for when to act on it.

The return gap is not the most striking difference between the two studies. MACD signal cross failed to produce a usable backtest on 77 of the 220 tickers in its universe, a 35.0 percent failure rate. MACD zero cross failed on only 6, 2.7 percent, the same low failure rate most studies on this site show. One rule completed 65 percent of its universe. The other completed 96 percent of the same universe, tested over the same ten years.

What is nearly identical between them

Win rate and drawdown, the two numbers usually blamed for a weak return, are almost the same across both studies. MACD zero cross wins 36.35 percent of trades with an average maximum drawdown of negative 7.86 percent. MACD signal cross wins 36.98 percent with a drawdown of negative 7.75 percent, both figures within a point of each other. Whatever separates a 6.61 percent return from a 1.18 percent one here, it is not that one rule trades noticeably worse than the other on a per-trade basis. Tenachine's guide on reading a win rate makes the case that win rate alone rarely explains a return gap, and this pair is a clean example: two nearly identical win rates sitting next to two very different returns.

The number that does move together with return

Trade frequency tracks the gap better than win rate does. MACD zero cross generated 9,052 trades across its 214 completed tickers, about 42.3 per ticker. MACD signal cross generated 2,363 trades across 143 tickers, about 16.5 per ticker, both because it completed fewer tickers and because the signal-line trigger fired less often per ticker that did complete. A rule with more chances to compound a real edge, on a wider slice of the universe, produced the bigger number here. Tenachine's guide on survivorship bias already flagged MACD signal cross's 35.0 percent failure rate as one of the worst coverage gaps on the site. This guide adds the other half: the same indicator, given a rule that keeps more of the universe in play, produced a markedly different headline number.

Tenachine does not publish why any individual run failed, so it would be a guess to say the zero-line trigger is mechanically more reliable than the signal-line one, rather than say the two studies simply drew different tickers into their failure pool. What can be said without guessing is narrower: two rules built on the same indicator, tested on the same 220 stocks over the same decade, produced return numbers nearly six times apart, and the more complete one is also the one with the bigger number. Past performance does not predict future results, and neither rule's decade is a forecast for the next one.