Every 12-Month Window in 8 EA Backtests: 4 Never Lost; Worst −23.0% for MEGAMAX DONCHIAN; GOLD KING DD Still 91.3%
14 November 2026, 00:00
0
7
A yearly table only shows January to December. A buyer can start an EA in any month, so we looked at every run of 12 calendar months in a row, not only calendar years. For the backtests of 8 EAs we list on the Market, we slid a 12-month window one month at a time and checked the closed balance at the end against the start. Result: 4 of the 8 tests never had a losing 12-month window. The worst window in the other 4 ranged from −1.2% to −23.0%. But one of the four with no losing window, GOLD KING v6, also has a maximum relative equity drawdown of 91.3%. Everything here is backtest. No forward result is used. Figures are as of 2026-11-14.
How the windows are counted
- Source: the deal list of each published MT5 test report, grouped by the month a trade was closed (commission and swap included). Floating profit and loss of open positions is not included.
- A window is 12 consecutive calendar months. Its return is the net profit of those 12 months divided by the closed balance at the start of the first month.
- Only full months of the test period are used. Where a test ends in the middle of a month, that month is left out: ATLAS June 2026 (−3,491.29), ETHEREUM EA June 2026 (−17.50) and GOLD TRIPWIRE October 2026 (+108.67).
- The drawdown column is the maximum relative equity drawdown of the whole test, the figure we publish for each EA. It is not a drawdown of the window.
The table
Sorted by the share of losing windows, highest first.
EA | Backtest period | History quality | 12-month windows | Losing windows | Share losing | Worst window | Median window | Best window | Max relative equity drawdown (whole test)
ETHEREUM EA | 2019.01.01–2026.06.28 | 100% real ticks | 78 | 17 | 21.8% | −10.3% (2023-02 to 2024-01) | +15.1% | +131.3% | 17.1%
MEGAMAX DONCHIAN | 2017.01.01–2026.06.01 | 100% | 102 | 19 | 18.6% | −23.0% (2018-09 to 2019-08) | +23.7% | +525.2% | 58.6%
DONCHIAN ENGINE (free) | 2017.01.01–2026.06.01 | 100% real ticks | 102 | 8 | 7.8% | −1.9% (2024-05 to 2025-04) | +9.9% | +42.1% | 10.8%
BITCOIN COMET | 2018.01.01–2025.12.31 | 98% | 85 | 3 | 3.5% | −1.2% (2022-08 to 2023-07) | +45.8% | +681.6% | 31.2%
ATLAS | 2019.01.01–2026.06.28 | 100% | 78 | 0 | 0% | +10.6% (2020-01 to 2020-12) | +80.8% | +234.1% | 26.3%
GOLD TRIPWIRE | 2020.01.01–2026.10.14 | 11% real ticks | 70 | 0 | 0% | +14.7% (2023-04 to 2024-03) | +30.8% | +722.8% | 19.0%
GOLD VIPER | 2019.01.01–2026.09.01 | 8% real ticks | 81 | 0 | 0% | +16.9% (2021-12 to 2022-11) | +101.7% | +316.7% | 30.3%
GOLD KING v6 | 2019.01.01–2026.06.01 | 5% real ticks | 78 | 0 | 0% | +35.0% (2020-10 to 2021-09) | +61.1% | +195.4% | 91.3%
Two tests lost in about one window in five
ETHEREUM EA lost money in 17 of 78 windows (21.8%) and MEGAMAX DONCHIAN in 19 of 102 (18.6%). MEGAMAX DONCHIAN also has the deepest window in the table: −23.0% from September 2018 to August 2019. Its median window is +23.7% and its best +525.2%, so a buyer who started in the wrong month and one who started in the right month would have seen very different first years from the same test.
Small losses when they happened
DONCHIAN ENGINE had 8 losing windows out of 102, the worst −1.9%. BITCOIN COMET had 3 out of 85, the worst −1.2%. In both, a losing 12-month stretch did occur, but it ended close to flat in closed balance.
No losing window is not the same as a small drawdown
The four tests with no losing window show a maximum relative equity drawdown from 19.0% (GOLD TRIPWIRE) to 91.3% (GOLD KING v6). GOLD KING v6 has the highest worst window in the table, +35.0%, and the deepest drawdown. Two things explain the gap. A window compares month-end closed balances only, so a dip inside a month that was recovered before the month closed does not appear. And floating losses on open positions are not in the closed balance at all. The window column and the drawdown column answer different questions, so read them together.
Limits of this table
- GOLD NEURON is left out today: the deal list we rebuilt does not add up to the net profit in its report, so we do not use its monthly figures. TRINITY and AIFW are left out because the original report files are not available.
- Windows overlap. 78 windows are not 78 independent results. Neighbouring windows share 11 of their 12 months.
- The tests differ in symbol, timeframe, period, tick model and risk setting. GOLD KING v6, GOLD VIPER and GOLD TRIPWIRE ran on 5% to 11% real ticks. The row order is not a ranking.
- These are EAs we chose to keep listed, so this is a survivor set. All of it is historical simulation, not a live record.
For buyers
- Look at the worst 12 months, not only the worst calendar year. Your first year will start in whatever month you buy.
- If an EA never lost over any 12 months in its backtest, still check the maximum relative equity drawdown before you choose a lot size.
- Past windows do not tell you which window comes next.
The two EAs discussed most in this article: MEGAMAX DONCHIAN https://www.mql5.com/en/market/product/181838 · GOLD KING https://www.mql5.com/en/market/product/180879
Our published backtests state period, balance and settings, and our forward accounts are labelled as demo accounts wherever we show them. The full list: fxea365.com/ea/ranking
