Longest Run Without a New Closed-Balance High in 9 EA Backtests: 0 Months for One, 65 Months for Another

We took the published MT5 backtests of 9 of our EAs, added up the closed trades month by month, and asked one question: once the closed balance reached a month-end high, how many months did it take to close a month above it again? This article uses backtests only. It contains no forward-account figures. Figures are as of 2026-10-24.

- The longest run of month-ends below a previous high ranges from 0 months (GOLD KING) to 65 months (GOLD NEURON). 65 months is 5 years and 5 months: the month-end high of 2015-07 was first beaten in 2021-01.

- 4 of the 9 tests had a run of 12 months or more: GOLD NEURON 65, ETHEREUM TREND 23, MEGAMAX DONCHIAN 16, DONCHIAN ENGINE 12. The other 5 stayed at 7 months or less.

- GOLD KING closed every one of its 89 months at a new high. The same test has a 42.8% equity drawdown, so this is a statement about closed trades only.

- 6 of the 9 tests ended below their last high. The longest of those open runs is 7 months (ETHEREUM TREND, high of 2025-11).

- A long wait and a deep fall are not the same thing: DONCHIAN ENGINE waited 12 months with a whole-test balance drawdown of 7.31%, GOLD VIPER 6 months with 19.6%.

- Each closed trade is assigned to the month in which it closed. Commission and swap are included. Open positions are not valued, so a month-end balance can be a new high while the account is carrying a floating loss.

- The balance at a month-end is the 10,000 opening deposit (1,000 for GOLD TRIPWIRE) plus all closed results up to that month. A month counts as below the high when its month-end balance is not above the highest earlier month-end balance. Months without a closed trade count too, because the calendar keeps running. The opening deposit is the first high.

- We only see month-ends. A balance high set in the middle of a month, and a fall inside a month, are not visible in this count. The run lengths are our own arithmetic on the monthly split of each report.

- The drawdown in the table is the balance drawdown from the MT5 report for the whole test, measured on closed-trade balance. It is not the equity drawdown, which includes open trades and is larger; we mention the equity figure for GOLD KING only in the text.

- 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 | Longest run of month-ends below a previous high (months) | High it waited for, and the month it was beaten | Calendar months from first to last trade | Months that closed at a new high (share) | Run still open at test end (months) | Total return (MT5 report) | Balance drawdown (MT5 report, whole test)

GOLD NEURON | M30, 2004.01.01-2026.09.01, 100% real ticks | 65 | 2015-07 to 2021-01 | 270 | 49 (18.1%) | 3 | 4160.4% | 22.84%

ETHEREUM TREND | H4, 2019.01.01-2026.06.28, 100% real ticks | 23 | opening deposit to 2021-01 | 89 | 24 (27.0%) | 7 | 323.8% | 6.7%

MEGAMAX DONCHIAN (USDJPY) | H1, 2017.01.01-2026.06.01, 100% | 16 | 2018-07 to 2019-12 | 113 | 43 (38.1%) | 2 | 2155.4% | 15.19%

DONCHIAN ENGINE (free) | H1, 2017.01.01-2026.06.01, 100% real ticks | 12 | 2024-06 to 2025-07 | 113 | 42 (37.2%) | 2 | 165.2% | 7.31%

BITCOIN COMET | H4, 2018.01.01-2025.12.31, 98% | 7 | 2022-07 to 2023-03 | 94 | 40 (42.6%) | 1 | 12164.8% | 14.06%

ATLAS PORTFOLIO | H4, 2019.01.01-2026.06.28, 100% | 6 | 2019-02 to 2019-09 | 90 | 42 (46.7%) | 2 | 4941% | 16.29%

GOLD VIPER | M30, 2019.01.01-2026.09.01, 8% real ticks | 6 | 2024-11 to 2025-06 | 92 | 55 (59.8%) | 0 | 19565.4% | 19.6%

GOLD TRIPWIRE | M15, 2020.01.01-2026.10.14, 11% real ticks | 5 | 2024-09 to 2025-03 | 82 | 59 (72.0%) | 0 | 760.4% | 6.28%

GOLD KING | M1, 2019.01.01-2026.06.01, 5% real ticks | 0 | none | 89 | 89 (100.0%) | 0 | 6177.7% | 7.45%

Sorted by the longest run, largest first. Total return and balance drawdown come from the same MT5 Strategy Tester report as the monthly split. The tests cover different periods and data quality, so read each row on its own and do not rank the returns against each other.

The GOLD NEURON test is the longest in the table, 270 calendar months. Its closed balance set a month-end high in 2015-07 and did not close a month above it until 2021-01. Inside that run there were winning years, but they were not enough to get back above the old high before the next losing year started. Only 49 of its 270 months, 18.1%, closed at a new high.

The balance drawdown of the same report for the whole test is 22.84%. The run of 65 months is a separate measurement: it counts how long the balance stayed below a high, not how far it fell.

ETHEREUM TREND opened its first trade in 2019-02 and its closed balance stayed below the 10,000 opening deposit until 2021-01. Its balance drawdown for the whole test is 6.7%, the second lowest in the table, so the run was long rather than deep. It is also the test with the longest run still open at the end: 7 months below the 2025-11 high when the test stops in 2026-06.

GOLD KING is a grid-type EA: it adds positions while the price moves against it and closes the basket when the price comes back. The month-end count only sees closed baskets, so all 89 months can close at a new high while an open basket carries a loss. The equity drawdown of the same test, which includes open trades, is 42.8% against a balance drawdown of 7.45%. For this EA the zero says little about how long the account can be under water on equity.

- If you start an EA and the balance has not made a new high for some months, compare that wait with the longest run in its own backtest. For 4 of these 9 tests it was 12 months or more.

- Read the length of the wait next to the depth of the fall. The table shows both, and they do not line up: 12 months with 7.31% for one EA, 6 months with 19.6% for another.

- For an EA that holds positions open for a long time, a short wait on closed balance does not mean a short wait on equity.

- A backtest value is a measurement of the past on one history, not a limit. A future run can be longer. Nothing here is a promise of future results.