π Cumulative Trajectory β Project-Level View
Cumulative numbers aggregate gate versions of different rule compositions. N (count) and per-metric stability are gate-agnostic and statistically valid to sum. Excess vs control is weighted by n_combined per gate; market-regime drift may bias the sum. Cumulative WIN RATE is intentionally NOT shown β denominator depends on gate composition (loose vs tight gates pass different ticker counts; pooling produces a misleading scalar). See references/COMBINED_SIGNAL_LAYER_DESIGN.md for the full methodology.
1οΈβ£ Cumulative N (count of recommendations)
2,606 total Β· 86 formal Β· 2,520 shadow
Per-gate breakdown (9 gates)
| Gate | Title | Status | Formal | Shadow | Combined |
|---|---|---|---|---|---|
| v5_pretrack | Pretracking 5-rule v5 (placeholder gate) | retired | 8 | 0 | 8 |
| v4_pretrack | Pretracking 6-rule v2 (FCF-EBITDA+ROIC-FCF) | retired | 12 | 0 | 12 |
| v3_pretrack | Pretracking 6-rule v1 | retired | 15 | 0 | 15 |
| v12 | User Curated Lean Gate | active | 35 | 227 | 262 |
| v10 | New Metrics Gate | retired | 6 | 482 | 488 |
| v9 | Watson-ROIC Reintroduction | retired | 0 | 1,811 | 1,811 |
| v8 | Weighted FCF + P/S-P/C Rebalance | retired | 4 | 0 | 4 |
| v7 | FCF-EBITDA Convergence Swap | retired | 6 | 0 | 6 |
| v6 | Close Call Decision Combo | retired | 0 | 0 | 0 |
2οΈβ£ Cumulative Excess vs Control (weighted by N)
-12.00% across 2,606 recommendations
Excess vs control group, weighted by n_combined per gate. Spans bullβflat market periods (MarβApr 2026).
Per-gate breakdown
| Gate | N | Excess (pp) |
|---|---|---|
| v5_pretrack | 8 | -14.55% |
| v4_pretrack | 12 | -19.11% |
| v3_pretrack | 15 | -20.00% |
| v12 | 262 | -13.08% |
| v10 | 488 | -14.71% |
| v9 | 1,811 | -10.94% |
| v8 | 4 | -26.06% |
| v7 | 6 | -15.56% |
| v6 | 0 | β |
3οΈβ£ Per-Metric Direction Stability (gate-agnostic)
71 metrics scored by Signal Intelligence Layer Β· top 10 shown
Per-metric direction stability is gate-agnostic by design (SYSTEM_STANDARDS Β§3). Aggregation across gate versions is statistically valid.
| Metric | Stars | Confidence | MPS |
|---|---|---|---|
| eps-roa pe score | β β β β β | 87 | B |
| fcf efficency | β β β β β | 84 | B |
| quality value score | β β β β β | 82 | C |
| rsi-net ebitda signal | β β β β β | 76 | B |
| roic/ebitda spread | β β β β β | 75 | A |
| sm watson 3 | β β β β β | 71 | C |
| rev growth efficiency | β β β β β | 66 | C |
| sm2: growth-momentum premium | β β β β β | 65 | C |
| fcf quality score | β β β β β | 65 | D |
| quality value score (test version) | β β β β β | 62 | A |
π Worked Example: v9 β v10 transition β Composition shifted
Both gates' returns shown side-by-side AND combined β illustrates apples-to-apples math vs composition-shift risk.
| Gate | Title | N | Excess vs Control (pp) |
|---|---|---|---|
| v9 | Watson-ROIC Reintroduction | 1,811 | -10.94% |
| v10 | New Metrics Gate | 488 | -14.71% |
| Combined (weighted by N) | 2,299 | -11.74% | |
C v12 #949 User Curated Lean Gate 58% win Β· +9.7% avg Β· n=237 β ACTIVE
π§± Composition (5 rules Β· stars + separation per rule)
RSI-Net EBITDA Signalβ β β β β 13.2pp Β· strongβ₯ 248.62 (higher wins)48% data
direction: higher wins
MPS grade: B
SIL confidence: 76/100
Control separation (sep_pp): -13.20pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
FCF Efficencyβ β β β β 13.1pp Β· strongβ€ 0.05 (lower wins)46% data
direction: lower wins
MPS grade: B
SIL confidence: 84/100
Control separation (sep_pp): -13.08pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
ROIC-EBITDA Quality Compositeβ β β β β 2.9pp Β· moderateβ₯ 153.57 (higher wins)44% data
direction: higher wins
MPS grade: C
SIL confidence: 51/100
Control separation (sep_pp): +2.92pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: higher
Avg. EPS Growthβ β β β β 5.3pp Β· strongβ₯ 37.03 (higher wins)70% data
direction: higher wins
MPS grade: B
SIL confidence: 62/100
Control separation (sep_pp): +5.26pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: higher
FCF Quality Scoreβ β β β β 13.4pp Β· strongβ€ 0.19 (lower wins)32% data
direction: lower wins
MPS grade: D
SIL confidence: 65/100
Control separation (sep_pp): -13.45pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
π Performance (Real Money + Shadow + Combined)
πͺ£ Profile (buckets + coverage)
C v10 #8473 New Metrics Gate 56% win Β· +7.4% avg Β· n=485 βΈ RETIRED
π§± Composition (5 rules Β· stars + separation per rule)
Quality Value Scoreβ β β β β 22.1pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: C
SIL confidence: 82/100
Control separation (sep_pp): -22.06pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
FCF Efficencyβ β β β β 13.1pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: B
SIL confidence: 84/100
Control separation (sep_pp): -13.08pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
FCF-EBITDA Convergence Scoreβ β β β β 6.4pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: D
SIL confidence: 55/100
Control separation (sep_pp): -6.42pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
EPS-ROA PE Scoreβ β β β β 12.9pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: B
SIL confidence: 87/100
Control separation (sep_pp): +12.95pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: higher
ROI Liquidity / FCFβ β β β β 6.5pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: F
SIL confidence: 32/100
Control separation (sep_pp): +6.55pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: higher
π Performance (Real Money + Shadow + Combined)
πͺ£ Profile (buckets + coverage)
D v9 #7600 Watson-ROIC Reintroduction 48% win Β· +11.1% avg Β· n=1810 βΈ RETIRED
π§± Composition (5 rules Β· stars + separation per rule)
Quality Value Scoreβ β β β β 22.1pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: C
SIL confidence: 82/100
Control separation (sep_pp): -22.06pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
FCF Efficencyβ β β β β 13.1pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: B
SIL confidence: 84/100
Control separation (sep_pp): -13.08pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
FCF-EBITDA Convergence Scoreβ β β β β 6.4pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: D
SIL confidence: 55/100
Control separation (sep_pp): -6.42pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
P/S v P/C Multipleβ β β β β 18.2pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: A
SIL confidence: 61/100
Control separation (sep_pp): -18.18pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
Watson-ROIC Confidence Indexβ β β β β 0.0pp Β· noisethreshold n/acoverage n/a
direction: unknown wins
MPS grade: C
SIL confidence: 12/100
Control separation (sep_pp): +0.00pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
π Performance (Real Money + Shadow + Combined)
πͺ£ Profile (buckets + coverage)
? v8 #5031 Weighted FCF + P/S-P/C Rebalance 0% win Β· -5.3% avg Β· n=3 βΈ RETIRED
π§± Composition (5 rules Β· stars + separation per rule)
Weighted FCF Efficiencyβ β β β β 3.8pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: D
SIL confidence: 51/100
Control separation (sep_pp): -3.82pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
SM Watson 5β β β β β 3.8pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: D
SIL confidence: 31/100
Control separation (sep_pp): -3.82pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
Quality Value Scoreβ β β β β 22.1pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: C
SIL confidence: 82/100
Control separation (sep_pp): -22.06pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
P/S v P/C Multipleβ β β β β 18.2pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: A
SIL confidence: 61/100
Control separation (sep_pp): -18.18pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
PEG Adj. Hybrid Growthβ β β β β 5.5pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: C
SIL confidence: 42/100
Control separation (sep_pp): -5.54pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
π Performance (Real Money + Shadow + Combined)
πͺ£ Profile (buckets + coverage)
β v6 #β Close Call Decision Combo no priced data yet βΈ RETIRED
π§± Composition (5 rules Β· stars + separation per rule)
FCF Quality Scoreβ β β β β 13.4pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: D
SIL confidence: 65/100
Control separation (sep_pp): -13.45pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
SM Watson 5β β β β β 3.8pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: D
SIL confidence: 31/100
Control separation (sep_pp): -3.82pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
PFCF v EBITDAβ β β β β 0.4pp Β· noisethreshold n/acoverage n/a
direction: unknown wins
MPS grade: C
SIL confidence: 33/100
Control separation (sep_pp): -0.37pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
ROIC/Ebitda Spreadβ β β β β 9.2pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: A
SIL confidence: 75/100
Control separation (sep_pp): +9.23pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: higher
Quality Value Scoreβ β β β β 22.1pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: C
SIL confidence: 82/100
Control separation (sep_pp): -22.06pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
π Performance (Real Money + Shadow + Combined)
πͺ£ Profile (buckets + coverage)
A v7 #4920 FCF-EBITDA Convergence Swap 100% win Β· +6.5% avg Β· n=6 βΈ RETIRED
π§± Composition (5 rules Β· stars + separation per rule)
FCF Quality Scoreβ β β β β 13.4pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: D
SIL confidence: 65/100
Control separation (sep_pp): -13.45pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
SM Watson 5β β β β β 3.8pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: D
SIL confidence: 31/100
Control separation (sep_pp): -3.82pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
PFCF v EBITDAβ β β β β 0.4pp Β· noisethreshold n/acoverage n/a
direction: unknown wins
MPS grade: C
SIL confidence: 33/100
Control separation (sep_pp): -0.37pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
FCF-EBITDA Convergence Scoreβ β β β β 6.4pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: D
SIL confidence: 55/100
Control separation (sep_pp): -6.42pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
Quality Value Scoreβ β β β β 22.1pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: C
SIL confidence: 82/100
Control separation (sep_pp): -22.06pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
π Performance (Real Money + Shadow + Combined)
πͺ£ Profile (buckets + coverage)
A v5_pretrack #7814 Pretracking 5-rule v5 (placeholder gate) 75% win Β· +7.5% avg Β· n=8 βΈ RETIRED
π§± Composition (5 rules Β· stars + separation per rule)
FCF Qualityβ n/asep n/athreshold n/acoverage n/a
direction: unknown wins
SM Watson 5β β β β β 3.8pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: D
SIL confidence: 31/100
Control separation (sep_pp): -3.82pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: lower
PFCF vEBITDAβ n/asep n/athreshold n/acoverage n/a
direction: unknown wins
ROIC/Ebitda Spreadβ β β β β 9.2pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: A
SIL confidence: 75/100
Control separation (sep_pp): +9.23pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: higher
Mrkt Mispriceβ β β β β 6.6pp Β· strongthreshold n/acoverage n/a
direction: unknown wins
MPS grade: D
SIL confidence: 38/100
Control separation (sep_pp): +6.63pp (return spread between high-scoring and low-scoring tickers in the 500-stock control group β current snapshot, not the historical value at gate retirement)
Control direction: higher
π Performance (Real Money + Shadow + Combined)
πͺ£ Profile (buckets + coverage)
B v4_pretrack #4772 Pretracking 6-rule v2 (FCF-EBITDA+ROIC-FCF) 71% win Β· +5.1% avg Β· n=7 βΈ RETIRED
π§± Composition (6 rules Β· stars + separation per rule)
partial: FCF-EBITDAβ n/asep n/athreshold n/acoverage n/a
direction: unknown wins
ROIC-FCFβ n/asep n/athreshold n/acoverage n/a
direction: unknown wins
+4 unknownβ n/asep n/athreshold n/acoverage n/a
direction: unknown wins
π Performance (Real Money + Shadow + Combined)
πͺ£ Profile (buckets + coverage)
C v3_pretrack #1539 Pretracking 6-rule v1 57% win Β· +2.1% avg Β· n=7 βΈ RETIRED
π§± Composition (6 rules Β· stars + separation per rule)
unknownβ n/asep n/athreshold n/acoverage n/a
direction: unknown wins