DIARY
Does time matter?
Timeframes, holding time, sessions, weekdays and months.
RESULTConclusion
- The same method lost more on shorter timeframes. MACD made −834.7% on M5, −41.7% on H1, −8.2% on H4 and −5.2% on D1.
- The reason is simple. On a 1-hour hold the spread is about 27% of the typical move. On a 1-week hold it falls to about 1.6%.
- Longer timeframes only shrank the losses; they did not turn positive.
- Of 308 weekday and month tests, 16 were consistent in both halves with t>2, excluding Sunday bars. With this many tests, chance alone produces some. Only Monday in stock indices stood out; it is tested in What survived.
TIMEFRAMESTimeframes
Before costsAfter costs
M5 is left out of the chart because its losses are far larger; it is in the table below.
| Condition | Trades | Win rate | Before costs | After costs | 1st half | 2nd half | Random avg |
|---|---|---|---|---|---|---|---|
| MA20/50 M5 | 143,170 | 30.6% | −18.8% | −251.1% | −153.9% | −97.2% | −235.3% |
| MACD M5 | 493,130 | 28.1% | −20.8% | −834.7% | −525.6% | −309.1% | −816.3% |
| Donchian20 M5 | 147,423 | 30.5% | −63.0% | −305.7% | −189.4% | −116.3% | −243.4% |
| MA20/50 H1 | 11,537 | 34.7% | −11.0% | −29.4% | −10.1% | −19.3% | −14.7% |
| MACD H1 | 39,932 | 34.0% | +19.7% | −41.7% | −34.2% | −7.5% | −63.5% |
| Donchian20 H1 | 11,794 | 35.5% | −17.4% | −35.0% | −18.7% | −16.3% | −16.3% |
| MA20/50 H4 | 2,990 | 36.4% | +1.1% | −3.6% | −4.2% | +0.6% | +2.0% |
| MACD H4 | 10,509 | 34.9% | +8.2% | −8.2% | −7.9% | −0.3% | −18.1% |
| Donchian20 H4 | 3,039 | 36.1% | −6.1% | −10.5% | −11.5% | +1.0% | −0.1% |
| MA20/50 D1 | 554 | 37.0% | +0.5% | −0.4% | −14.5% | +14.1% | +11.5% |
| MACD D1 | 2,090 | 34.4% | −2.2% | −5.2% | +3.8% | −9.0% | −2.3% |
| Donchian20 D1 | 617 | 35.3% | −8.9% | −9.9% | −9.4% | −0.5% | +11.1% |
HOLDING TIMEHolding time and the spread
Median of 17 symbols.
| Holding time | Typical move | Spread | Spread as share of the move |
|---|---|---|---|
| 1 hour | 7.7 bp | 2.8 bp | 26.7% |
| 4 hours | 16.6 bp | 2.8 bp | 11.9% |
| 1 day | 55.7 bp | 2.8 bp | 3.8% |
| 1 week | 135.2 bp | 2.8 bp | 1.6% |
| 4 weeks | 283.2 bp | 2.8 bp | 0.8% |
SESSIONSMarket sessions
MACD used only during each session. No session made much difference.
| Condition | Trades | Win rate | Before costs | After costs | 1st half | 2nd half | Random avg |
|---|---|---|---|---|---|---|---|
| Tokyo | 10,204 | 42.3% | −1.1% | −16.5% | −12.8% | −3.7% | −16.4% |
| London | 15,509 | 43.7% | +6.6% | −16.6% | −13.5% | −3.1% | −24.6% |
| New York | 12,060 | 42.0% | +4.3% | −15.6% | −11.0% | −4.7% | −19.0% |
HOUR OF DAYWhich hour to enter
Hours that were positive in the same direction in both halves, across 7 symbols and 24 hours (UTC). Most fall at 20–23 UTC, when the spread widens around 17:00 New York. Bars use Bid and the actual spread is wider, so these numbers can't be trusted.
| Symbol | Hour | Side | Avg (bp) | 1st half | 2nd half |
|---|---|---|---|---|---|
| USDJPY | 2:00 | long | 0.09 | 0.16 | 0.02 |
| EURUSD | 20:00 | short | 0.17 | 0.20 | 0.15 |
| GBPUSD | 20:00 | short | 1.42 | 1.85 | 0.96 |
| XAUUSD | 20:00 | long | 0.97 | 0.71 | 1.25 |
| XAUUSD | 21:00 | long | 3.05 | 1.26 | 5.38 |
| XAUUSD | 23:00 | long | 0.68 | 0.05 | 1.35 |
This diary records tests on historical data. It does not recommend any method or symbol and does not indicate future profit or loss. Please make investment decisions at your own risk.