Rule Collider

Backtests · GOLD · h4_momentum_atr · long

Does h4_momentum_atr above 0.3392 work on GOLD? 2023 to 2026 backtest

GOLD long when Close vs H4 EMA20 (H4 ATR) is above the threshold, tested at five thresholds taken from the feature’s own distribution on GOLD over this window: its 10th, 25th, 50th, 75th and 90th percentiles. Window 2023-09-30 to 2026-09-30, the last 30 percent held out of sample; stop 2.5 ATR, target 3 R, held overnight, max spread 30 pips, one-minute confirmation 2 pips, no hour filter. Fills and spread as measured on a demo fleet (how a result is judged). The direction and the exit are the ones the research catalogue carried for this pair; the thresholds are not.

Five thresholds, one rule shape

PercentileConditionIn sample R / tradesOut of sample R / tradestTrades to confirmLast quarter R / tradesVerdict
p10h4_momentum_atr > -1.556-152.7 R / 3179-116.2 R / 1179-2.11,114-55.0 R / 420Active, net negative
p25h4_momentum_atr > -0.6324-101.1 R / 2709-75.0 R / 970-1.51,819-38.7 R / 338No edge shown
p50h4_momentum_atr > 0.3392-50.8 R / 1964-45.3 R / 635-1.12,127-31.8 R / 203No edge shown
p75h4_momentum_atr > 1.466+17.4 R / 1037+8.7 R / 3000.313,659+1.7 R / 90Net positive, not distinguishable from chance
p90h4_momentum_atr > 2.444-9.6 R / 481+36.9 R / 1231.8146+8.7 R / 35Net positive, not distinguishable from chance

R is the risk on one trade. The out-of-sample split carries the verdict: below 100 trades it is “no edge shown” whatever the t; a positive mean with t below 2 is “net positive, not yet distinguishable from chance”; the last column is how many trades that mean would need to reach t = 2.

Each threshold in full

h4_momentum_atr > -1.556 (p10)

Honesty panelActive, net negative

Traded 1179 times out of sample, net negative beyond chance (-116.2 R, -0.099 R per trade, t = -2.1); still active in the last quarter (-55.0 R over 420 trades).

In sample
2023-09-30 to 2025-11-05
-152.7 R
3179 trades · 33% wins
-0.048 R per trade · t = -1.6, within chance
about 4,936 trades to confirm this mean (-0.048 R per trade) at t = -2 · shown: 3179
spread 0.104 R per trade
Out of sample
2025-11-05 to 2026-09-30
-116.2 R
1179 trades · 30% wins
-0.099 R per trade · t = -2.1, beyond chance
about 1,114 trades to confirm this mean (-0.099 R per trade) at t = -2 · shown: 1179
spread 0.064 R per trade
Last quarter
last 91 days of the window
-55.0 R
420 trades · 29% wins
-0.131 R per trade · t = -1.6, within chance
about 630 trades to confirm this mean (-0.131 R per trade) at t = -2 · shown: 420
spread 0.085 R per trade
Costs. Spread 0.093 R per trade at 1.25 times the bar’s minimum (more than the trades made before spread); 733 fills in measured hot hours (01:00 at 3.9 times). Fill offset 0.08 stop-distances on every entry, measured on 732 demo fills. Note: the symbol's quoting precision changed on 2025-08-04; the spread field is converted with the point of each era. At 1.5 times the bar’s minimum the out-of-sample figure would read -131.2 R (-0.111 R per trade over 1179 trades).

Entry: stop order 2 pips past the signal close, cancelled after 4 bars; 4,576 signals placed, 4,358 filled, 218 expired; trigger state, an entry on every bar the conditions hold while flat.

Exit: stop 2.5 ATR (mean 86.7 pips, 2.50 ATR, 29 raised to the 10-pip floor); target 3 R; exits 65% at the stop, 22% at the target, 13% at the day end.

Filters: signals on every weekday; no entries 22:00 to 01:00 broker time (2,484 bars); max spread 30 pips (0 bars over it); M1 confirmation 2 pips (883 signals rejected).

Monthly R, h4_momentum_atr > -1.556 on GOLDOne bar per month: the R earned that month; the held-out months in the full colour.2023-10: +25.5 R over 117 trades2023-102023-11: -1.8 R over 105 trades2023-12: +1.8 R over 116 trades2024-01: -34.7 R over 148 trades2024-02: -25.6 R over 129 trades2024-03: +20.1 R over 102 trades2024-04: -25.5 R over 142 trades2024-042024-05: -27.8 R over 136 trades2024-06: -37.4 R over 137 trades2024-07: +18.0 R over 125 trades2024-08: -17.0 R over 159 trades2024-09: +11.9 R over 134 trades2024-10: +5.2 R over 149 trades2024-102024-11: -10.6 R over 105 trades2024-12: -31.5 R over 123 trades2025-01: +10.0 R over 126 trades2025-02: +5.8 R over 129 trades2025-03: +18.2 R over 79 trades2025-04: +1.2 R over 121 trades2025-042025-05: -15.6 R over 115 trades2025-06: -21.5 R over 129 trades2025-07: -25.9 R over 134 trades2025-08: -31.1 R over 140 trades2025-09: +30.5 R over 140 trades2025-10: +12.6 R over 132 trades2025-102025-11: +3.9 R over 113 trades2025-12: -0.6 R over 132 trades2026-01: +21.0 R over 102 trades2026-02: +1.0 R over 80 trades2026-03: -32.5 R over 86 trades2026-04: -25.1 R over 117 trades2026-042026-05: -8.6 R over 65 trades2026-06: -27.8 R over 71 trades2026-07: -32.9 R over 146 trades2026-08: +17.7 R over 140 trades2026-09: -39.8 R over 134 trades+31 R-40 R

h4_momentum_atr > -0.6324 (p25)

Honesty panelNo edge shown

No edge shown out of sample: -75.0 R over 970 trades is -0.077 R per trade, within what chance produces (t = -1.5); about 1,819 trades would be needed to confirm this mean at t = -2, shown: 970.

In sample
2023-09-30 to 2025-11-05
-101.1 R
2709 trades · 33% wins
-0.037 R per trade · t = -1.1, within chance
about 8,241 trades to confirm this mean (-0.037 R per trade) at t = -2 · shown: 2709
spread 0.103 R per trade
Out of sample
2025-11-05 to 2026-09-30
-75.0 R
970 trades · 31% wins
-0.077 R per trade · t = -1.5, within chance
about 1,819 trades to confirm this mean (-0.077 R per trade) at t = -2 · shown: 970
spread 0.063 R per trade
Last quarter
last 91 days of the window
-38.7 R
338 trades · 30% wins
-0.114 R per trade · t = -1.3, within chance
about 826 trades to confirm this mean (-0.114 R per trade) at t = -2 · shown: 338
spread 0.082 R per trade
Costs. Spread 0.092 R per trade at 1.25 times the bar’s minimum (more than the trades made before spread); 609 fills in measured hot hours (01:00 at 3.9 times). Fill offset 0.08 stop-distances on every entry, measured on 732 demo fills. Note: the symbol's quoting precision changed on 2025-08-04; the spread field is converted with the point of each era. At 1.5 times the bar’s minimum the out-of-sample figure would read -87.1 R (-0.090 R per trade over 970 trades).

Entry: stop order 2 pips past the signal close, cancelled after 4 bars; 3,868 signals placed, 3,679 filled, 189 expired; trigger state, an entry on every bar the conditions hold while flat.

Exit: stop 2.5 ATR (mean 83.8 pips, 2.50 ATR, 23 raised to the 10-pip floor); target 3 R; exits 65% at the stop, 23% at the target, 13% at the day end.

Filters: signals on every weekday; no entries 22:00 to 01:00 broker time (4,522 bars); max spread 30 pips (0 bars over it); M1 confirmation 2 pips (730 signals rejected).

Monthly R, h4_momentum_atr > -0.6324 on GOLDOne bar per month: the R earned that month; the held-out months in the full colour.2023-10: +28.4 R over 104 trades2023-102023-11: +5.6 R over 91 trades2023-12: +0.0 R over 100 trades2024-01: -27.0 R over 113 trades2024-02: -33.7 R over 105 trades2024-03: +14.7 R over 93 trades2024-04: -17.9 R over 125 trades2024-042024-05: -28.2 R over 117 trades2024-06: -31.5 R over 106 trades2024-07: +20.0 R over 110 trades2024-08: -20.6 R over 138 trades2024-09: +13.6 R over 119 trades2024-10: +5.2 R over 129 trades2024-102024-11: -7.3 R over 79 trades2024-12: -22.3 R over 90 trades2025-01: +4.4 R over 118 trades2025-02: +3.6 R over 114 trades2025-03: +20.3 R over 93 trades2025-04: +2.1 R over 97 trades2025-042025-05: -6.2 R over 84 trades2025-06: -25.4 R over 116 trades2025-07: -14.8 R over 108 trades2025-08: -24.7 R over 116 trades2025-09: +27.6 R over 133 trades2025-10: +16.3 R over 108 trades2025-102025-11: +8.2 R over 95 trades2025-12: +4.2 R over 122 trades2026-01: +18.9 R over 106 trades2026-02: +2.7 R over 75 trades2026-03: -17.4 R over 53 trades2026-04: -27.1 R over 90 trades2026-042026-05: -9.6 R over 55 trades2026-06: -19.4 R over 39 trades2026-07: -31.5 R over 111 trades2026-08: +19.5 R over 135 trades2026-09: -26.6 R over 92 trades+28 R-34 R

h4_momentum_atr > 0.3392 (p50)

Honesty panelNo edge shown

No edge shown out of sample: -45.3 R over 635 trades is -0.071 R per trade, within what chance produces (t = -1.1); about 2,127 trades would be needed to confirm this mean at t = -2, shown: 635.

In sample
2023-09-30 to 2025-11-05
-50.8 R
1964 trades · 33% wins
-0.026 R per trade · t = -0.7, within chance
about 17,246 trades to confirm this mean (-0.026 R per trade) at t = -2 · shown: 1964
spread 0.102 R per trade
Out of sample
2025-11-05 to 2026-09-30
-45.3 R
635 trades · 32% wins
-0.071 R per trade · t = -1.1, within chance
about 2,127 trades to confirm this mean (-0.071 R per trade) at t = -2 · shown: 635
spread 0.062 R per trade
Last quarter
last 91 days of the window
-31.8 R
203 trades · 29% wins
-0.157 R per trade · t = -1.4, within chance
about 434 trades to confirm this mean (-0.157 R per trade) at t = -2 · shown: 203
spread 0.083 R per trade
Costs. Spread 0.092 R per trade at 1.25 times the bar’s minimum (more than the trades made before spread); 448 fills in measured hot hours (01:00 at 3.9 times). Fill offset 0.08 stop-distances on every entry, measured on 732 demo fills. Note: the symbol's quoting precision changed on 2025-08-04; the spread field is converted with the point of each era. At 1.5 times the bar’s minimum the out-of-sample figure would read -53.1 R (-0.084 R per trade over 635 trades).

Entry: stop order 2 pips past the signal close, cancelled after 4 bars; 2,745 signals placed, 2,599 filled, 146 expired; trigger state, an entry on every bar the conditions hold while flat.

Exit: stop 2.5 ATR (mean 81.0 pips, 2.50 ATR, 16 raised to the 10-pip floor); target 3 R; exits 64% at the stop, 23% at the target, 13% at the day end.

Filters: signals on every weekday; no entries 22:00 to 01:00 broker time (7,795 bars); max spread 30 pips (3 bars over it); M1 confirmation 2 pips (508 signals rejected).

Monthly R, h4_momentum_atr > 0.3392 on GOLDOne bar per month: the R earned that month; the held-out months in the full colour.2023-10: +27.1 R over 100 trades2023-102023-11: -2.1 R over 81 trades2023-12: -11.6 R over 77 trades2024-01: -9.9 R over 60 trades2024-02: -17.1 R over 54 trades2024-03: +8.8 R over 83 trades2024-04: -17.0 R over 98 trades2024-042024-05: -29.5 R over 78 trades2024-06: -22.0 R over 64 trades2024-07: +8.1 R over 85 trades2024-08: -33.1 R over 103 trades2024-09: +7.1 R over 95 trades2024-10: +4.4 R over 91 trades2024-102024-11: +7.1 R over 44 trades2024-12: -14.8 R over 51 trades2025-01: +3.6 R over 97 trades2025-02: +16.1 R over 87 trades2025-03: +4.2 R over 71 trades2025-04: +13.0 R over 67 trades2025-042025-05: -1.0 R over 56 trades2025-06: -11.0 R over 60 trades2025-07: -6.3 R over 67 trades2025-08: -13.3 R over 87 trades2025-09: +21.1 R over 122 trades2025-10: +21.5 R over 82 trades2025-102025-11: +2.0 R over 76 trades2025-12: -8.2 R over 91 trades2026-01: +16.9 R over 94 trades2026-02: -0.7 R over 53 trades2026-03: -9.7 R over 25 trades2026-04: -12.6 R over 50 trades2026-042026-05: +0.7 R over 30 trades2026-06: -6.4 R over 17 trades2026-07: -25.5 R over 60 trades2026-08: +11.4 R over 110 trades2026-09: -17.7 R over 33 trades+27 R-33 R

h4_momentum_atr > 1.466 (p75)

Honesty panelNet positive, not distinguishable from chance

Net positive out of sample, not yet distinguishable from chance: +8.7 R over 300 trades is +0.029 R per trade, t = 0.3; about 13,659 trades would be needed to confirm this mean at t = 2, shown: 300; last quarter +1.7 R over 90 trades.

In sample
2023-09-30 to 2025-11-05
+17.4 R
1037 trades · 34% wins
+0.017 R per trade · t = 0.3, within chance
about 41,043 trades to confirm this mean (+0.017 R per trade) at t = 2 · shown: 1037
spread 0.091 R per trade
Out of sample
2025-11-05 to 2026-09-30
+8.7 R
300 trades · 34% wins
+0.029 R per trade · t = 0.3, within chance
about 13,659 trades to confirm this mean (+0.029 R per trade) at t = 2 · shown: 300
spread 0.051 R per trade
Last quarter
last 91 days of the window
+1.7 R
90 trades · 36% wins
+0.019 R per trade · t = 0.1, within chance
about 32,245 trades to confirm this mean (+0.019 R per trade) at t = 2 · shown: 90
spread 0.066 R per trade
Costs. Spread 0.082 R per trade at 1.25 times the bar’s minimum (81% of the gross); 219 fills in measured hot hours (01:00 at 3.9 times). Fill offset 0.08 stop-distances on every entry, measured on 732 demo fills. Note: the symbol's quoting precision changed on 2025-08-04; the spread field is converted with the point of each era. At 1.5 times the bar’s minimum the out-of-sample figure would read +5.6 R (+0.019 R per trade over 300 trades).

Entry: stop order 2 pips past the signal close, cancelled after 4 bars; 1,406 signals placed, 1,337 filled, 69 expired; trigger state, an entry on every bar the conditions hold while flat.

Exit: stop 2.5 ATR (mean 83.1 pips, 2.50 ATR, 3 raised to the 10-pip floor); target 3 R; exits 63% at the stop, 23% at the target, 14% at the day end.

Filters: signals on every weekday; no entries 22:00 to 01:00 broker time (11,254 bars); max spread 30 pips (7 bars over it); M1 confirmation 2 pips (221 signals rejected).

Monthly R, h4_momentum_atr > 1.466 on GOLDOne bar per month: the R earned that month; the held-out months in the full colour.2023-10: +12.9 R over 65 trades2023-102023-11: -5.1 R over 48 trades2023-12: -6.4 R over 46 trades2024-01: -9.8 R over 12 trades2024-02: -9.6 R over 19 trades2024-03: +10.1 R over 64 trades2024-04: -3.4 R over 44 trades2024-042024-05: -2.3 R over 29 trades2024-06: -9.1 R over 29 trades2024-07: -3.0 R over 48 trades2024-08: -16.6 R over 45 trades2024-09: -2.6 R over 64 trades2024-10: +3.5 R over 48 trades2024-102024-11: +10.4 R over 31 trades2024-12: -6.3 R over 17 trades2025-01: -0.7 R over 52 trades2025-02: +6.4 R over 43 trades2025-03: +19.7 R over 26 trades2025-04: +16.2 R over 45 trades2025-042025-05: +5.2 R over 29 trades2025-06: -7.8 R over 20 trades2025-07: +1.3 R over 31 trades2025-08: -19.1 R over 47 trades2025-09: +18.3 R over 77 trades2025-10: +15.2 R over 58 trades2025-102025-11: -5.0 R over 39 trades2025-12: +5.2 R over 48 trades2026-01: +29.2 R over 54 trades2026-02: -8.5 R over 19 trades2026-03: -2.6 R over 5 trades2026-04: -3.7 R over 21 trades2026-042026-05: -1.5 R over 17 trades2026-06: -6.1 R over 7 trades2026-07: -3.5 R over 18 trades2026-08: +8.3 R over 68 trades2026-09: -3.2 R over 4 trades+29 R-19 R

h4_momentum_atr > 2.444 (p90)

Honesty panelNet positive, not distinguishable from chance

Net positive out of sample, not yet distinguishable from chance: +36.9 R over 123 trades is +0.300 R per trade, t = 1.8; about 146 trades would be needed to confirm this mean at t = 2, shown: 123; last quarter +8.7 R over 35 trades.

In sample
2023-09-30 to 2025-11-05
-9.6 R
481 trades · 35% wins
-0.020 R per trade · t = -0.3, within chance
about 27,367 trades to confirm this mean (-0.020 R per trade) at t = -2 · shown: 481
spread 0.078 R per trade
Out of sample
2025-11-05 to 2026-09-30
+36.9 R
123 trades · 41% wins
+0.300 R per trade · t = 1.8, within chance
about 146 trades to confirm this mean (+0.300 R per trade) at t = 2 · shown: 123
spread 0.045 R per trade
Last quarter
last 91 days of the window
+8.7 R
35 trades · 40% wins
+0.248 R per trade · t = 0.8, within chance
about 207 trades to confirm this mean (+0.248 R per trade) at t = 2 · shown: 35
spread 0.061 R per trade
Costs. Spread 0.071 R per trade at 1.25 times the bar’s minimum (61% of the gross); 89 fills in measured hot hours (01:00 at 3.9 times). Fill offset 0.08 stop-distances on every entry, measured on 732 demo fills. Note: the symbol's quoting precision changed on 2025-08-04; the spread field is converted with the point of each era. At 1.5 times the bar’s minimum the out-of-sample figure would read +35.8 R (+0.291 R per trade over 123 trades).

Entry: stop order 2 pips past the signal close, cancelled after 4 bars; 631 signals placed, 604 filled, 27 expired; trigger state, an entry on every bar the conditions hold while flat.

Exit: stop 2.5 ATR (mean 87.0 pips, 2.50 ATR); target 3 R; exits 62% at the stop, 23% at the target, 15% at the day end.

Filters: signals on every weekday; no entries 22:00 to 01:00 broker time (13,368 bars); max spread 30 pips (7 bars over it); M1 confirmation 2 pips (106 signals rejected).

Monthly R, h4_momentum_atr > 2.444 on GOLDOne bar per month: the R earned that month; the held-out months in the full colour.2023-10: +1.7 R over 44 trades2023-102023-11: -3.8 R over 20 trades2023-12: +3.5 R over 20 trades2024-01: -2.2 R over 2 trades2024-02: -5.5 R over 5 trades2024-03: +12.3 R over 50 trades2024-04: -11.0 R over 16 trades2024-042024-05: -5.6 R over 19 trades2024-06: -3.8 R over 8 trades2024-07: -2.4 R over 17 trades2024-08: -13.5 R over 18 trades2024-09: -5.1 R over 24 trades2024-10: +1.6 R over 17 trades2024-102024-11: +7.2 R over 17 trades2024-12: -0.5 R over 7 trades2025-01: +3.2 R over 17 trades2025-02: -0.8 R over 16 trades2025-03: +11.0 R over 20 trades2025-04: +13.9 R over 27 trades2025-042025-05: -4.0 R over 14 trades2025-06: -4.7 R over 8 trades2025-07: -3.9 R over 15 trades2025-08: +0.6 R over 7 trades2025-09: +10.1 R over 47 trades2025-10: -7.7 R over 26 trades2025-102025-11: +4.8 R over 17 trades2025-12: +11.4 R over 21 trades2026-01: +21.4 R over 29 trades2026-02: +1.2 R over 6 trades2026-03: -2.2 R over 2 trades2026-04: -4.6 R over 6 trades2026-042026-05: -2.7 R over 6 trades2026-06: -1.1 R over 1 trades2026-07: -0.5 R over 4 trades2026-08: +9.2 R over 31 trades+21 R-14 R

What h4_momentum_atr measures

Close vs H4 EMA20 (H4 ATR) (atr), on H4 bars. M5 close minus the H4 EMA20, in multiples of the H4 ATR(14).

Test your own variant

Rule Collider runs one rule on one symbol over years of broker M5 data with the fills and the spread a demo fleet measured, and shows the out-of-sample figure first. Early access goes to the waitlist in signup order.

Join the waitlist

Historical simulation on broker M5 data with a measured fill offset, a measured spread factor, swap and a one-minute confirmation model. Past results do not predict future results. This is a research tool, not investment advice. Hours are broker time, the chart time in MT5. Built 2026-11-04 from runs ending 2026-09-30; refreshed monthly.