2/9 combinations achieve PF > 1.0 (22%). Variant A leads on portfolio PF. PF range: 0.6612–1.0816.
Best Portfolio PF
1.0816
Variant A, hold=2
Best Portfolio Exp
0.0171%
Variant A, hold=2
PF > 1.0
2/9
22% of combinations
PF > 1.2 (pass)
0/9
full threshold
Stable Instruments
0
≥70% approval rate
Candidate Instruments
0
≥50% approval rate
RESEARCH DECISION
FAIL
No combination achieves PF > 1.2 with ≥3 approved instruments. Best portfolio PF: 1.0816 (Variant A, hold=2). 0 instrument(s) stable across combinations. Expansion filter does not unlock momentum edge in FX D1. Breakout momentum hypothesis is not supported by this data. Kill this strategy line.
STABLE INSTRUMENTS (≥70%)
STABLE INSTRUMENTSapproved in ≥70% of 9 parameter combinations0 instruments meet ≥70% threshold
No instruments approved in ≥70% of parameter combinations.
PORTFOLIO CANDIDATE SET
PORTFOLIO CANDIDATE SET
No instruments approved in ≥50% of combinations.
1. Summary Table — All 9 Combinations
Variant
Hold
Trades
Avg/Inst
Port. PF
Port. Exp%
Avg MaxDD
Approved
Approved Instruments
A — ATR Expansion (range > 1.5×ATR14)
1
436
18.2
0.6978
-0.0563%
-1.97%
0
—
A — ATR Expansion (range > 1.5×ATR14)
2
423
17.6
1.0816
0.0171%
-2.13%
0
—
A — ATR Expansion (range > 1.5×ATR14)
3
419
17.5
1.0090
0.0023%
-2.47%
0
—
B — Range Breakout (range > max prior 10)
1
460
19.2
0.6612
-0.0658%
-2.09%
0
—
B — Range Breakout (range > max prior 10)
2
445
18.5
0.9745
-0.0057%
-2.10%
0
—
B — Range Breakout (range > max prior 10)
3
438
18.2
0.9136
-0.0239%
-2.82%
0
—
C — Percentile Exp. (range > 80th pct prior 50)
1
775
32.3
0.7820
-0.0405%
-2.96%
0
—
C — Percentile Exp. (range > 80th pct prior 50)
2
762
31.8
0.8460
-0.0378%
-4.05%
0
—
C — Percentile Exp. (range > 80th pct prior 50)
3
757
31.5
0.8923
-0.0306%
-4.44%
0
—
2. Parameter Region Analysis
Variant
Combos
Avg Port. PF
Avg Port. Exp%
PF > 1.0 (pass rate)
PF > 1.2
A — ATR Expansion (range > 1.5×ATR14)
3
0.9295
-0.0123%
2/3 (67%)
0/3
B — Range Breakout (range > max prior 10)
3
0.8498
-0.0318%
0/3 (0%)
0/3
C — Percentile Exp. (range > 80th pct prior 50)
3
0.8401
-0.0363%
0/3 (0%)
0/3
▶ Edge is concentrated in the A — ATR Expansion (range > 1.5×ATR14) variant (highest avg portfolio PF). Performance in other variants is substantially lower.
3a. Heatmap — Portfolio Profit Factor
Rows = hold_bars | Columns = Variant |
red = low PF
→
green = high PF
hold \ Variant
A
B
C
1
0.6978
0.6612
0.7820
2
1.0816
0.9745
0.8460
3
1.0090
0.9136
0.8923
3b. Heatmap — Portfolio Expectancy (%)
Rows = hold_bars | Columns = Variant |
red = negative
→
green = positive
hold \ Variant
A
B
C
1
-0.0563%
-0.0658%
-0.0405%
2
0.0171%
-0.0057%
-0.0378%
3
0.0023%
-0.0239%
-0.0306%
3c. Heatmap Interpretation
Hold axis: Portfolio PF is non-monotonic with respect to hold_bars; peaks at hold=2. Variant axis: Variant A leads (0.9295 mean PF) vs Variant C (0.8401). Performance gap of 0.0894 indicates filter type materially affects signal quality. Concentration: Large value range (0.4204) indicates high concentration — performance is not evenly distributed across the grid.
4. Trade Frequency vs Edge
By Hold Bars
hold_bars
Avg Trades
Avg PF
Avg Exp%
1
557
0.7137
-0.0542%
2
543
0.9674
-0.0088%
3
538
0.9383
-0.0174%
By Variant
Variant
Avg Trades
Avg PF
Avg Exp%
A
426
0.9295
-0.0123%
B
448
0.8498
-0.0318%
C
765
0.8401
-0.0363%
▶ Hold bars: more trades (shorter holds) correlates with lower PF — trade frequency dilutes edge. / Variant: higher-frequency variants correlate with lower PF — stricter filters improve signal quality.