Profit Factor by Weekday in 8 EA Backtests: Monday Led in 6 of 8. Why We Still Don't Recommend a Day Filter
18 October 2026, 00:00
0
52
We split the published MT5 backtests of 8 of our EAs by the weekday on which each trade closed, and compared the profit factor of Monday to Friday. This article uses backtests only. It contains no forward-account figures. Figures are as of 2026-10-17.
The answer first
- Monday had the highest profit factor of the five weekdays in 6 of the 8 backtests. In the other 2, the best day was Tuesday (BITCOIN COMET) and Wednesday (ETHEREUM TREND).
- In 6 of the 8, every day with closed trades ended with a net profit.
- The two exceptions are BITCOIN COMET on Thursday (126 trades, profit factor 0.96) and ETHEREUM TREND on Friday (14 trades, profit factor 0.25).
- We do not read any of this as a reason to switch an EA off on certain days. The reasons are below.
How we counted
- Each closed trade is assigned to the weekday of its close time in the Strategy Tester. A trade opened on Friday and closed on Monday counts as Monday.
- Commission and swap are included. Profit factor is gross profit divided by gross loss for the trades of that day.
- Some closes are stamped Saturday or Sunday. They are counted in the column for weekend closes and left out of the Monday-to-Friday comparison.
- 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 weekday.
The table
EA | Test period and data | Best weekday, Mon-Fri (profit factor, trades) | Worst weekday, Mon-Fri (profit factor, trades) | Days with a net loss | Weekend closes | Whole-test profit factor | Equity drawdown (MT5 report, whole test)
GOLD KING | M1, 2019.01.01-2026.06.01, 5% real ticks | Mon (9.61, 541) | Fri (3.81, 534) | none | 73 | 4.66 | 42.8%
GOLD VIPER | M30, 2019.01.01-2026.09.01, 8% real ticks | Mon (2.35, 1,709) | Fri (1.22, 1,903) | none | 235 | 1.64 | 24.09%
GOLD NEURON | M30, 2004.01.01-2026.09.01, 100% real ticks | Mon (1.8, 429) | Wed (1.09, 607) | none | 61 | 1.35 | 28.5%
MEGAMAX DONCHIAN (USDJPY) | H1, 2017.01.01-2026.06.01, 100% | Mon (1.78, 105) | Wed (1.3, 144) | none | 14 | 1.55 | 16.27%
DONCHIAN ENGINE (free) | H1, 2017.01.01-2026.06.01, 100% real ticks | Mon (1.8, 107) | Wed (1.24, 146) | none | 13 | 1.5 | 10.34%
ATLAS PORTFOLIO | H4, 2019.01.01-2026.06.28, 100% | Mon (1.89, 212) | Tue (1.08, 237) | none | 63 | 1.48 | 19.74%
BITCOIN COMET | H4, 2018.01.01-2025.12.31, 98% | Tue (2.16, 130) | Thu (0.96, 126) | Thu | 146 | 1.72 | 26.76%
ETHEREUM TREND | H4, 2019.01.01-2026.06.28, 100% real ticks | Wed (6.95, 25) | Fri (0.25, 14) | Fri | 33 | 2.83 | 10.33%
The whole-test profit factor and the drawdown come from the same test as the weekday split. The drawdown is the maximum equity drawdown in percent for the whole test; there is no separate drawdown per weekday. The tests cover different periods, so compare the days inside one row, not the rows with each other.
Why we compare profit factors and not dollars
Several of these tests compound: position size grows with the balance. A weekday whose losing trades happened to fall in the later, larger years looks much worse in dollars than one whose losses came early. The profit factor of a day is also weighted by size, so it is not immune to this, but it is less distorted than the dollar total. That is why the table shows profit factors and trade counts only.
Monday on top in 6 of 8: what it does and does not tell you
The pattern is consistent enough to note, and we have not established why. One mechanical part of it is the counting rule: trades held over the weekend can only close once the market reopens, so Monday collects most of them. Whether those trades are better on average is a separate question that this split cannot answer.
In some rows the gap between the best and the worst weekday is small. For GOLD NEURON the best and worst weekdays are 1.8 and 1.09; for MEGAMAX DONCHIAN 1.78 and 1.3. All five weekdays of these EAs were profitable in the test.
The two red days
- ETHEREUM TREND, Friday: 14 trades, which is 9.1% of its 154. With 14 trades, one or two results could change the sign of the day. A profit factor of 0.25 on 14 trades is a weak basis for any rule.
- BITCOIN COMET, Thursday: 126 trades with a profit factor of 0.96, close to break-even. Its Friday is 1.0 on 106 trades. The same EA's best day, Tuesday, is 2.16 on 130 trades. We have not tested whether a spread like this is larger than what chance alone would produce when 775 trades are split into 7 groups.
Why a weekday filter is not the lesson
- Any filter chosen after looking at a result will improve that same result. Removing the worst day from a backtest always makes that backtest look better, whether or not the day is worse in the future. To check a filter properly it has to be chosen on one period and tested on a different one. We have not done that here, so we do not recommend it.
- Skipping a day does not simply delete its trades. With compounding, a missed trade changes the size of every later one, and a position not opened on one day can change what the EA does next.
- The close-day rule assigns each trade to the day it closed, not the day it was opened. Blocking new entries on a day is a different rule from the one this table measures.
- These are historical simulations. They describe how the past test spread across weekdays, not how future weeks will go, and no figure here is a promise of future results.
The two EAs discussed most in this article: BITCOIN COMET https://www.mql5.com/en/market/product/181831 · ATLAS PORTFOLIO https://www.mql5.com/en/market/product/182751
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
