Losing Months in 8 EA Backtests: From 0 of 89 to 48.7%. Why the Zero Comes With a 42.8% Equity Drawdown

We split the published MT5 backtests of 8 of our EAs into calendar months and counted the months that closed with a net loss. This article uses backtests only. It contains no forward-account figures. Figures are as of 2026-10-17.

- GOLD KING closed all 89 of its 89 months in profit. The same test had a 42.8% equity drawdown.

- The 6 trend and breakout EAs with the most losing months lost money in 40.0% to 48.7% of their months.

- GOLD VIPER sits in between, with 24 losing months out of 92 (26.1%).

- 3 of the 8 tests had a run of 6 losing months in a row and still ended with a positive total return.

- Each closed trade is assigned to the month in which it was closed. Commission and swap are included.

- A losing month is a month whose closed trades add up to a net loss. Months with no closed trade are left out.

- Open positions are not counted until they close. A month can therefore close in profit while the account is carrying a large floating loss.

- TRINITY and AIFW are not in this article: the original report files behind their published backtests are not in our breakdown, so we could not split them by month.

EA | Test period and data | Months with trades | Losing months | Share | Longest run of losing months | Total return (MT5 report) | Equity drawdown (MT5 report, whole test)

GOLD KING | M1, 2019.01.01-2026.06.01, 5% real ticks | 89 | 0 | 0.0% | 0 | 6177.7% | 42.8%

GOLD VIPER | M30, 2019.01.01-2026.09.01, 8% real ticks | 92 | 24 | 26.1% | 3 | 19565.4% | 24.09%

ATLAS PORTFOLIO | H4, 2019.01.01-2026.06.28, 100% | 90 | 36 | 40.0% | 4 | 4941% | 19.74%

BITCOIN COMET | H4, 2018.01.01-2025.12.31, 98% | 94 | 41 | 43.6% | 5 | 12164.8% | 26.76%

MEGAMAX DONCHIAN (USDJPY) | H1, 2017.01.01-2026.06.01, 100% | 113 | 51 | 45.1% | 6 | 2155.4% | 16.27%

GOLD NEURON | M30, 2004.01.01-2026.09.01, 100% real ticks | 270 | 124 | 45.9% | 6 | 4160.4% | 28.5%

DONCHIAN ENGINE (free) | H1, 2017.01.01-2026.06.01, 100% real ticks | 113 | 52 | 46.0% | 6 | 165.2% | 10.34%

ETHEREUM TREND | H4, 2019.01.01-2026.06.28, 100% real ticks | 78 | 38 | 48.7% | 4 | 323.8% | 10.33%

Total return and drawdown are taken from the same MT5 Strategy Tester report as the monthly split. The drawdown column is the maximum equity drawdown in percent for the whole test, so it includes floating losses. The tests cover different periods, so read each row on its own and do not rank the returns against each other.

These two figures do not contradict each other. GOLD KING is a grid-type EA: it adds positions while the price moves against it and closes the basket when it recovers. The monthly count only sees trades when they close. While a basket is open, its loss shows up in equity, not in the closed result of the month.

The same report shows how long that can last. The longest single holding time in the test is 2,608 hours, about 108 days. Over that kind of stretch the calendar keeps turning, other trades keep closing in profit, and every month can still end green on paper.

So for a grid EA, "no losing month" is a statement about closed trades. The number to read next to it is the equity drawdown. Here it is 42.8%.

ATLAS PORTFOLIO, BITCOIN COMET, MEGAMAX DONCHIAN, GOLD NEURON, DONCHIAN ENGINE and ETHEREUM TREND lost money in 40.0% to 48.7% of their months. Their whole-test win rates are between 33.2% and 53.29%, and a month in which the winning trades are few or small easily ends red.

The longest losing runs are 4 to 6 months. MEGAMAX DONCHIAN, GOLD NEURON and DONCHIAN ENGINE each had 6 losing months in a row, and all 3 tests still ended with a positive total return (2155.4%, 4160.4% and 165.2%, with equity drawdowns of 16.27%, 28.5% and 10.34%).

GOLD VIPER trades far more often than the others (10,340 trades in this test, win rate 56.4%). One plausible reading is that more trades per month make each month less dependent on a single large winner; we have not tested that. Its share of losing months is 26.1%, with a longest run of 3. Its equity drawdown for the whole test is 24.09%.

- A few red months in a row are inside the backtest for most of these EAs. Before calling one broken, compare the run with the longest losing run in its own backtest: 3 to 6 months for 7 of the 8 here.

- A long record of green months is not the same as a low drawdown. For an EA that holds positions through adverse moves, check the equity drawdown and the longest holding time first.

- Monthly counts depend on the period. A test that covers 22 years, like GOLD NEURON, includes more market conditions than one that covers 7, and its share is not directly comparable.

- These are historical simulations. They describe the past test, not future months, and no figure here is a promise of future results.

The two EAs discussed most in this article: GOLD KING https://www.mql5.com/en/market/product/180879 · ETHEREUM TREND https://www.mql5.com/en/market/product/182358