Rule Collider

Backtests · GOLD · h4_rsi · long

Does h4_rsi above 53.526 work on GOLD? 2023 to 2026 backtest

GOLD long when H4 RSI (14) 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_rsi > 36.766-177.9 R / 3219-135.5 R / 1218-2.4863-55.3 R / 423Active, net negative
p25h4_rsi > 44.574-124.7 R / 2786-93.1 R / 978-1.81,177-46.6 R / 337No edge shown
p50h4_rsi > 53.526+9.8 R / 1947-25.8 R / 612-0.66,153-20.5 R / 178No edge shown
p75h4_rsi > 63.476+15.0 R / 1054+21.7 R / 2780.81,956+10.8 R / 78Net positive, not distinguishable from chance
p90h4_rsi > 71.118-5.9 R / 472+16.9 R / 1130.9511+2.9 R / 22Net 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_rsi > 36.766 (p10)

Honesty panelActive, net negative

Traded 1218 times out of sample, net negative beyond chance (-135.5 R, -0.111 R per trade, t = -2.4); still active in the last quarter (-55.3 R over 423 trades).

In sample
2023-09-30 to 2025-11-05
-177.9 R
3219 trades · 32% wins
-0.055 R per trade · t = -1.9, within chance
about 3,708 trades to confirm this mean (-0.055 R per trade) at t = -2 · shown: 3219
spread 0.102 R per trade
Out of sample
2025-11-05 to 2026-09-30
-135.5 R
1218 trades · 30% wins
-0.111 R per trade · t = -2.4, beyond chance
about 863 trades to confirm this mean (-0.111 R per trade) at t = -2 · shown: 1218
spread 0.062 R per trade
Last quarter
last 91 days of the window
-55.3 R
423 trades · 29% wins
-0.131 R per trade · t = -1.6, within chance
about 631 trades to confirm this mean (-0.131 R per trade) at t = -2 · shown: 423
spread 0.084 R per trade
Costs. Spread 0.091 R per trade at 1.25 times the bar’s minimum (more than the trades made before spread); 720 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 -150.6 R (-0.124 R per trade over 1218 trades).

Entry: stop order 2 pips past the signal close, cancelled after 4 bars; 4,663 signals placed, 4,437 filled, 226 expired; trigger state, an entry on every bar the conditions hold while flat.

Exit: stop 2.5 ATR (mean 87.9 pips, 2.50 ATR, 30 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 (2,317 bars); max spread 30 pips (0 bars over it); M1 confirmation 2 pips (858 signals rejected).

Monthly R, h4_rsi > 36.766 on GOLDOne bar per month: the R earned that month; the held-out months in the full colour.2023-10: +35.2 R over 107 trades2023-102023-11: -5.0 R over 111 trades2023-12: -0.8 R over 112 trades2024-01: -28.8 R over 157 trades2024-02: -30.8 R over 132 trades2024-03: +20.1 R over 102 trades2024-04: -26.0 R over 142 trades2024-042024-05: -28.6 R over 141 trades2024-06: -39.2 R over 140 trades2024-07: +7.5 R over 125 trades2024-08: -15.2 R over 162 trades2024-09: +12.2 R over 133 trades2024-10: +1.9 R over 147 trades2024-102024-11: -9.7 R over 101 trades2024-12: -38.9 R over 124 trades2025-01: +9.8 R over 126 trades2025-02: +9.6 R over 127 trades2025-03: +18.6 R over 84 trades2025-04: -2.2 R over 122 trades2025-042025-05: -19.9 R over 122 trades2025-06: -24.6 R over 128 trades2025-07: -23.9 R over 141 trades2025-08: -30.8 R over 147 trades2025-09: +31.0 R over 140 trades2025-10: +6.3 R over 140 trades2025-102025-11: +6.1 R over 119 trades2025-12: -0.8 R over 141 trades2026-01: +16.9 R over 111 trades2026-02: +3.0 R over 87 trades2026-03: -35.4 R over 90 trades2026-04: -25.5 R over 115 trades2026-042026-05: -9.6 R over 70 trades2026-06: -40.8 R over 68 trades2026-07: -37.0 R over 147 trades2026-08: +19.3 R over 143 trades2026-09: -37.6 R over 133 trades+35 R-41 R

h4_rsi > 44.574 (p25)

Honesty panelNo edge shown

No edge shown out of sample: -93.1 R over 978 trades is -0.095 R per trade, within what chance produces (t = -1.8); about 1,177 trades would be needed to confirm this mean at t = -2, shown: 978.

In sample
2023-09-30 to 2025-11-05
-124.7 R
2786 trades · 33% wins
-0.045 R per trade · t = -1.4, within chance
about 5,681 trades to confirm this mean (-0.045 R per trade) at t = -2 · shown: 2786
spread 0.100 R per trade
Out of sample
2025-11-05 to 2026-09-30
-93.1 R
978 trades · 30% wins
-0.095 R per trade · t = -1.8, within chance
about 1,177 trades to confirm this mean (-0.095 R per trade) at t = -2 · shown: 978
spread 0.062 R per trade
Last quarter
last 91 days of the window
-46.6 R
337 trades · 29% wins
-0.138 R per trade · t = -1.6, within chance
about 558 trades to confirm this mean (-0.138 R per trade) at t = -2 · shown: 337
spread 0.084 R per trade
Costs. Spread 0.090 R per trade at 1.25 times the bar’s minimum (more than the trades made before spread); 608 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 -105.2 R (-0.108 R per trade over 978 trades).

Entry: stop order 2 pips past the signal close, cancelled after 4 bars; 3,963 signals placed, 3,764 filled, 199 expired; trigger state, an entry on every bar the conditions hold while flat.

Exit: stop 2.5 ATR (mean 85.1 pips, 2.50 ATR, 22 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 (4,043 bars); max spread 30 pips (0 bars over it); M1 confirmation 2 pips (742 signals rejected).

Monthly R, h4_rsi > 44.574 on GOLDOne bar per month: the R earned that month; the held-out months in the full colour.2023-10: +33.0 R over 102 trades2023-102023-11: +12.5 R over 92 trades2023-12: -6.3 R over 99 trades2024-01: -24.7 R over 117 trades2024-02: -39.0 R over 106 trades2024-03: +19.8 R over 99 trades2024-04: -20.9 R over 131 trades2024-042024-05: -30.6 R over 118 trades2024-06: -39.5 R over 108 trades2024-07: +17.4 R over 114 trades2024-08: -17.0 R over 147 trades2024-09: +8.7 R over 130 trades2024-10: +2.1 R over 137 trades2024-102024-11: +3.2 R over 75 trades2024-12: -24.3 R over 92 trades2025-01: +8.6 R over 126 trades2025-02: +12.0 R over 114 trades2025-03: +14.2 R over 89 trades2025-04: +1.4 R over 102 trades2025-042025-05: -11.4 R over 91 trades2025-06: -25.6 R over 114 trades2025-07: -17.4 R over 116 trades2025-08: -37.9 R over 111 trades2025-09: +29.5 R over 138 trades2025-10: +13.4 R over 113 trades2025-102025-11: +8.3 R over 100 trades2025-12: +1.9 R over 125 trades2026-01: +18.0 R over 100 trades2026-02: -1.2 R over 82 trades2026-03: -30.3 R over 58 trades2026-04: -26.1 R over 99 trades2026-042026-05: -10.2 R over 56 trades2026-06: -12.6 R over 26 trades2026-07: -38.8 R over 113 trades2026-08: +17.6 R over 141 trades2026-09: -25.4 R over 83 trades+33 R-40 R

h4_rsi > 53.526 (p50)

Honesty panelNo edge shown

No edge shown out of sample: -25.8 R over 612 trades is -0.042 R per trade, within what chance produces (t = -0.6); about 6,153 trades would be needed to confirm this mean at t = -2, shown: 612.

In sample
2023-09-30 to 2025-11-05
+9.8 R
1947 trades · 34% wins
+0.005 R per trade · t = 0.1, within chance
about 461,992 trades to confirm this mean (+0.005 R per trade) at t = 2 · shown: 1947
spread 0.099 R per trade
Out of sample
2025-11-05 to 2026-09-30
-25.8 R
612 trades · 32% wins
-0.042 R per trade · t = -0.6, within chance
about 6,153 trades to confirm this mean (-0.042 R per trade) at t = -2 · shown: 612
spread 0.057 R per trade
Last quarter
last 91 days of the window
-20.5 R
178 trades · 29% wins
-0.115 R per trade · t = -0.9, within chance
about 811 trades to confirm this mean (-0.115 R per trade) at t = -2 · shown: 178
spread 0.079 R per trade
Costs. Spread 0.089 R per trade at 1.25 times the bar’s minimum (more than the trades made before spread); 406 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 -32.8 R (-0.054 R per trade over 612 trades).

Entry: stop order 2 pips past the signal close, cancelled after 4 bars; 2,686 signals placed, 2,559 filled, 127 expired; trigger state, an entry on every bar the conditions hold while flat.

Exit: stop 2.5 ATR (mean 83.3 pips, 2.50 ATR, 14 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 (7,453 bars); max spread 30 pips (3 bars over it); M1 confirmation 2 pips (471 signals rejected).

Monthly R, h4_rsi > 53.526 on GOLDOne bar per month: the R earned that month; the held-out months in the full colour.2023-10: +36.4 R over 101 trades2023-102023-11: +7.0 R over 70 trades2023-12: -8.1 R over 74 trades2024-01: -15.5 R over 46 trades2024-02: -10.5 R over 54 trades2024-03: +11.7 R over 75 trades2024-04: -11.7 R over 93 trades2024-042024-05: -18.1 R over 81 trades2024-06: -24.5 R over 62 trades2024-07: +13.2 R over 82 trades2024-08: -21.1 R over 116 trades2024-09: +12.1 R over 97 trades2024-10: -13.4 R over 104 trades2024-102024-11: +1.4 R over 46 trades2024-12: -11.2 R over 40 trades2025-01: +2.2 R over 106 trades2025-02: +9.1 R over 99 trades2025-03: +6.2 R over 67 trades2025-04: +17.5 R over 65 trades2025-042025-05: +3.1 R over 52 trades2025-06: -5.3 R over 60 trades2025-07: -5.1 R over 60 trades2025-08: -19.0 R over 85 trades2025-09: +27.8 R over 126 trades2025-10: +25.3 R over 86 trades2025-102025-11: +2.0 R over 69 trades2025-12: -0.3 R over 101 trades2026-01: +21.6 R over 89 trades2026-02: -6.4 R over 55 trades2026-03: -3.5 R over 15 trades2026-04: -19.9 R over 63 trades2026-042026-05: +1.9 R over 26 trades2026-06: -0.7 R over 16 trades2026-07: -21.0 R over 51 trades2026-08: +12.0 R over 110 trades2026-09: -11.5 R over 17 trades+36 R-25 R

h4_rsi > 63.476 (p75)

Honesty panelNet positive, not distinguishable from chance

Net positive out of sample, not yet distinguishable from chance: +21.7 R over 278 trades is +0.078 R per trade, t = 0.8; about 1,956 trades would be needed to confirm this mean at t = 2, shown: 278; last quarter +10.8 R over 78 trades.

In sample
2023-09-30 to 2025-11-05
+15.0 R
1054 trades · 34% wins
+0.014 R per trade · t = 0.3, within chance
about 56,605 trades to confirm this mean (+0.014 R per trade) at t = 2 · shown: 1054
spread 0.090 R per trade
Out of sample
2025-11-05 to 2026-09-30
+21.7 R
278 trades · 35% wins
+0.078 R per trade · t = 0.8, within chance
about 1,956 trades to confirm this mean (+0.078 R per trade) at t = 2 · shown: 278
spread 0.052 R per trade
Last quarter
last 91 days of the window
+10.8 R
78 trades · 36% wins
+0.138 R per trade · t = 0.7, within chance
about 665 trades to confirm this mean (+0.138 R per trade) at t = 2 · shown: 78
spread 0.070 R per trade
Costs. Spread 0.082 R per trade at 1.25 times the bar’s minimum (75% of the gross); 214 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 +18.8 R (+0.068 R per trade over 278 trades).

Entry: stop order 2 pips past the signal close, cancelled after 4 bars; 1,398 signals placed, 1,332 filled, 66 expired; trigger state, an entry on every bar the conditions hold while flat.

Exit: stop 2.5 ATR (mean 84.2 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 (11,243 bars); max spread 30 pips (5 bars over it); M1 confirmation 2 pips (233 signals rejected).

Monthly R, h4_rsi > 63.476 on GOLDOne bar per month: the R earned that month; the held-out months in the full colour.2023-10: +22.7 R over 64 trades2023-102023-11: +0.1 R over 47 trades2023-12: -3.7 R over 45 trades2024-01: -5.8 R over 7 trades2024-02: -8.8 R over 12 trades2024-03: +9.3 R over 63 trades2024-04: -8.2 R over 58 trades2024-042024-05: -6.8 R over 34 trades2024-06: -14.6 R over 24 trades2024-07: +5.7 R over 52 trades2024-08: -6.5 R over 45 trades2024-09: +2.9 R over 58 trades2024-10: +7.8 R over 41 trades2024-102024-11: +4.2 R over 31 trades2024-12: -2.5 R over 13 trades2025-01: -7.9 R over 59 trades2025-02: +4.6 R over 51 trades2025-03: +16.1 R over 33 trades2025-04: +12.8 R over 54 trades2025-042025-05: -2.6 R over 22 trades2025-06: -10.7 R over 18 trades2025-07: -3.7 R over 20 trades2025-08: -24.8 R over 46 trades2025-09: +12.8 R over 93 trades2025-10: +22.6 R over 64 trades2025-102025-11: +1.5 R over 36 trades2025-12: +2.1 R over 47 trades2026-01: +27.1 R over 58 trades2026-02: -1.8 R over 12 trades2026-03: -3.3 R over 5 trades2026-04: -9.8 R over 24 trades2026-042026-05: -0.4 R over 9 trades2026-06: -4.5 R over 9 trades2026-07: -1.8 R over 5 trades2026-08: +12.5 R over 73 trades+27 R-25 R

h4_rsi > 71.118 (p90)

Honesty panelNet positive, not distinguishable from chance

Net positive out of sample, not yet distinguishable from chance: +16.9 R over 113 trades is +0.149 R per trade, t = 0.9; about 511 trades would be needed to confirm this mean at t = 2, shown: 113; last quarter +2.9 R over 22 trades.

In sample
2023-09-30 to 2025-11-05
-5.9 R
472 trades · 34% wins
-0.013 R per trade · t = -0.2, within chance
about 70,234 trades to confirm this mean (-0.013 R per trade) at t = -2 · shown: 472
spread 0.086 R per trade
Out of sample
2025-11-05 to 2026-09-30
+16.9 R
113 trades · 39% wins
+0.149 R per trade · t = 0.9, within chance
about 511 trades to confirm this mean (+0.149 R per trade) at t = 2 · shown: 113
spread 0.041 R per trade
Last quarter
last 91 days of the window
+2.9 R
22 trades · 41% wins
+0.133 R per trade · t = 0.4, within chance
about 605 trades to confirm this mean (+0.133 R per trade) at t = 2 · shown: 22
spread 0.056 R per trade
Costs. Spread 0.077 R per trade at 1.25 times the bar’s minimum (80% of the gross); 97 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 +16.0 R (+0.141 R per trade over 113 trades).

Entry: stop order 2 pips past the signal close, cancelled after 4 bars; 607 signals placed, 585 filled, 22 expired; trigger state, an entry on every bar the conditions hold while flat.

Exit: stop 2.5 ATR (mean 93.7 pips, 2.50 ATR); target 3 R; exits 62% at the stop, 22% at the target, 16% at the day end.

Filters: signals on every weekday; no entries 22:00 to 01:00 broker time (13,432 bars); max spread 30 pips (7 bars over it); M1 confirmation 2 pips (86 signals rejected).

Monthly R, h4_rsi > 71.118 on GOLDOne bar per month: the R earned that month; the held-out months in the full colour.2023-10: +4.0 R over 37 trades2023-102023-11: -2.6 R over 16 trades2023-12: -4.7 R over 14 trades2024-03: +7.1 R over 51 trades2024-04: -8.0 R over 26 trades2024-05: -4.0 R over 15 trades2024-07: -11.7 R over 25 trades2024-072024-08: -7.2 R over 15 trades2024-09: +7.0 R over 31 trades2024-10: -4.1 R over 21 trades2024-11: -0.3 R over 9 trades2024-12: -4.2 R over 5 trades2025-01: -10.1 R over 24 trades2025-012025-02: +4.9 R over 15 trades2025-03: +14.8 R over 24 trades2025-04: +4.6 R over 32 trades2025-05: -1.1 R over 5 trades2025-06: -0.9 R over 2 trades2025-07: -1.7 R over 6 trades2025-072025-08: -0.3 R over 2 trades2025-09: +7.4 R over 54 trades2025-10: +5.2 R over 43 trades2025-11: -1.7 R over 14 trades2025-12: +0.7 R over 30 trades2026-01: +21.6 R over 33 trades2026-012026-02: -3.3 R over 9 trades2026-03: -0.0 R over 2 trades2026-05: -3.3 R over 3 trades2026-08: +2.9 R over 22 trades+22 R-12 R

What h4_rsi measures

H4 RSI (14) (index), on H4 bars. RSI(14) of the last completed H4 bar.

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.