VALIDACIÓN DE SEÑALES — RESULTADOS SIMULADOS
Todas las cifras de esta página proceden de operaciones hipotéticas evaluadas sobre precios reales posteriores a cada señal (evaluación histórica). Ninguna métrica es una ganancia obtenida ni una probabilidad de éxito.
¿QUÉ ESTÁ REALMENTE DEMOSTRADO?
Conclusiones deterministas con reglas transparentes: n < 30 → INSUFFICIENT DATA; IC 95% del R medio entero < 0 → NEGATIVE SAMPLE; incluye 0 → NO EVIDENCE; entero > 0 → como mucho PRELIMINARY POSITIVE ASSOCIATION. Nada de lo que aparece aqui es una probabilidad ni una prediccion.
| Pregunta | Etiqueta | n | Metrica | Intervalo 95% | Comparacion | Razon |
|---|---|---|---|---|---|---|
| ¿Los setups (unicos, todos los modos) tienen R medio > 0? | NEGATIVE SAMPLE | 268 | mean R: -0.15 | [-0.29, -0.01] | R medio vs 0 | el IC 95% de la media esta entero por debajo de 0 |
| ¿Se sostiene en TIEMPO REAL (decisiones en tiempo real)? | NO EVIDENCE OF EDGE | 234 | mean R: -0.09 | [-0.25, +0.06] | R medio en tiempo real vs 0 | el IC 95% de la media incluye 0 |
| Win rate (realized_r > 0) | DESCRIPTIVE | 268 | win rate: 34.3% | [28.9%, 40.2%] | Wilson 95% | descriptivo; no es una probabilidad de ganar la próxima operación simulada |
| ¿El Opportunity Ranking ordena mejores oportunidades? (#1 vs resto) | NO EVIDENCE OF RANKING EDGE | 268 | diff mean R: +0.24 | [-0.11, +0.61] | #1 (n=233) vs #2+ (n=35) | el IC 95% de la diferencia incluye 0 |
| Ranking dentro del mismo instante T | NO EVIDENCE OF RANKING EDGE | 30 | mean diff R (mejor - resto): -0.14 | [-0.52, +0.24] | pares del mismo T | IC de la diferencia mismo-T incluye 0 |
| ¿LONG vs SHORT difieren? | NO EVIDENCE OF ASSOCIATION | 268 | diff mean R (LONG-SHORT): -0.04 | [-0.39, +0.29] | LONG n=60 vs SHORT n=208 | el IC 95% de la diferencia incluye 0 |
| ¿Immediate vs Conditional difieren? | PRELIMINARY NEGATIVE ASSOCIATION | 268 | diff mean R (IMM-COND): -0.42 | [-0.76, -0.04] | IMMEDIATE n=54 vs CONDITIONAL n=214 | el IC 95% de la diferencia excluye 0 por abajo |
| ¿Mayor Activity en T se asocia con mejores outcomes del setup? | PRELIMINARY POSITIVE ASSOCIATION | 234 | diff mean R (HIGH/ELEV - NORMAL/QUIET): +0.37 | [+0.07, +0.67] | setups con Activity registrada en T | el IC 95% de la diferencia excluye 0 por arriba (asociacion, no causalidad) |
| ¿Activity selecciona activos que despues se mueven mas? (seleccion, no direccion) | PRELIMINARY POSITIVE ASSOCIATION | 38211 | Spearman score vs |mov 24h|: +0.08 | [+0.07, +0.09] | todas las evaluaciones con Activity (WAIT incluido) | mayor Activity en T, mayor movimiento absoluto posterior (asociacion) |
| ¿El setup selecciona algo distinto a apostar en la misma direccion sobre otros activos del mismo T? (24h) | NO EVIDENCE OF EXCESS MOVE | 475 | mean excess dir. move %: +0.23 | [-0.02, +0.48] | setup vs WAIT de otros activos en el mismo T | el IC 95% de la media incluye 0 |
| Validacion temporal (walk-forward) | VALIDATION PENDING | 268 | operaciones simuladas únicas con R: 268 | — | necesarios ~120 (30 por bloque) | mecanismo preparado; se activa solo con bloques temporales suficientes |
| Hipotesis pre-registradas (out-of-sample en tiempo real) | EVALUATED | 229 | pendientes: 0 | — | solo datos en tiempo real posteriores al registro | 0 de 5 hipotesis sin muestra suficiente todavia |
Rendimiento simulado
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | Total R | Desv. | Payoff medio + | Perdida media | Profit factor | Max DD (R) | R por setup emitido | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Setups unicos (principal) | 268 | -0.15 | [-0.29, -0.01] | -1.00 | 34.3% | [28.9%, 40.2%] | -39.28 | 1.21 | +1.49 | -1.00 | 0.78 | 41.51 | -0.08 | |
| Todas las evaluaciones (n no independiente) | 272 | -0.14 | [-0.28, -0.00] | -1.00 | 34.6% | [29.2%, 40.4%] | -38.24 | 1.21 | +1.49 | -1.00 | 0.79 | 41.51 | -0.08 | |
| Unicos · reconstrucción histórica | 34 | -0.51 | [-0.81, -0.15] | -1.00 | 20.6% | [10.3%, 36.8%] | -17.29 | 0.99 | +1.39 | -1.00 | 0.36 | 19.99 | -0.24 | |
| Unicos · en tiempo real | 234 | -0.09 | [-0.25, +0.06] | -1.00 | 36.3% | [30.4%, 42.7%] | -22.00 | 1.23 | +1.49 | -1.00 | 0.85 | 41.51 | -0.05 |
Percentiles R (unicos): p5 -1.00 · p10 -1.00 · p25 -1.00 · p50 -1.00 · p75 +1.29 · p90 +1.65 · p95 +1.87. Win = realized_r > 0 (definicion canonica). AMBIGUOUS no tiene R y se excluye (se cuenta aparte). "R por setup emitido" cuenta NEVER_TRIGGERED / INVALIDATED / ENTRY_SKIPPED como 0 R. Max DD = mayor caida de la curva acumulada en R de esta muestra simulada (descriptivo).
R acumulado (setups unicos, orden de entrada)
Distribucion de R realizado (bins 0.5R)
LONG vs SHORT
LONG − SHORT: diff mean R -0.04 [-0.39, +0.29] · n 60 vs 208 · NO EVIDENCE OF ASSOCIATION el IC 95% de la diferencia incluye 0
LONG / SHORT
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| LONG | 60 | -0.18 | [-0.49, +0.11] | -1.00 | 33.3% | [22.7%, 45.9%] | |
| SHORT | 208 | -0.14 | [-0.31, +0.03] | -1.00 | 34.6% | [28.5%, 41.3%] |
Immediate vs Conditional
IMMEDIATE − CONDITIONAL: diff mean R -0.42 [-0.76, -0.04] · n 54 vs 214 · PRELIMINARY NEGATIVE ASSOCIATION el IC 95% de la diferencia excluye 0 por abajo
Immediate / Conditional
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| CONDITIONAL | 214 | -0.06 | [-0.22, +0.09] | -1.00 | 38.8% | [32.5%, 45.5%] | |
| IMMEDIATE | 54 | -0.49 | [-0.77, -0.16] | -1.00 | 16.7% | [9.0%, 28.7%] |
Alignment
Alineacion (TF a favor)
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| 2/4 | 67 | -0.08 | [-0.37, +0.22] | -1.00 | 35.8% | [25.4%, 47.8%] | |
| 3/4 | 86 | -0.10 | [-0.35, +0.14] | -1.00 | 38.4% | [28.8%, 48.9%] | |
| 4/4 | 115 | -0.22 | [-0.43, +0.00] | -1.00 | 30.4% | [22.8%, 39.4%] |
Cautelas del motor
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| 0 | 27 | -0.68 | [-0.93, -0.37] | -1.00 | 14.8% | [5.9%, 32.5%] | INSUFFICIENT SAMPLE |
| 1 | 76 | -0.20 | [-0.47, +0.07] | -1.00 | 31.6% | [22.2%, 42.7%] | |
| 2 | 106 | +0.11 | [-0.11, +0.35] | -1.00 | 45.3% | [36.1%, 54.8%] | |
| 3 | 56 | -0.34 | [-0.63, -0.05] | -1.00 | 25.0% | [15.5%, 37.7%] | |
| 4 | 3 | +0.56 | IC N/D | +1.15 | 66.7% | [20.8%, 93.9%] | INSUFFICIENT SAMPLE |
Technical Rank — ¿el ranking ordena mejores oportunidades?
| Rank en T | n operaciones simuladas | Mean R | IC 95% | Win rate | MFE medio (R) | MAE medio (R) | Mov. direccional 24h (todos los setups) | |
|---|---|---|---|---|---|---|---|---|
| #1 | 233 | -0.11 | [-0.27, +0.04] | 35.2% | +0.96 | -0.79 | -0.21% (n=410) | |
| #2-3 | 34 | -0.34 | [-0.68, +0.03] | 29.4% | +0.83 | -0.84 | -0.35% (n=64) | |
| #4+ | 1 | -1.00 | IC N/D | 0.0% | +0.93 | -1.00 | +0.89% (n=1) | INSUFFICIENT SAMPLE |
#1 vs #2+: diff mean R +0.24 [-0.11, +0.61] · n 233 vs 35 · NO EVIDENCE OF RANKING EDGE el IC 95% de la diferencia incluye 0
Top 3 vs #4+: diff mean R +0.86 IC N/D · n 267 vs 1 · INSUFFICIENT DATA n=267 vs n=1; se necesitan >= 30 por grupo
#1 vs #2: diff mean R +0.18 [-0.24, +0.57] · n 233 vs 29 · INSUFFICIENT DATA n=233 vs n=29; se necesitan >= 30 por grupo
#2 vs #3: diff mean R +0.30 [-0.59, +0.98] · n 29 vs 5 · INSUFFICIENT DATA n=29 vs n=5; se necesitan >= 30 por grupo
#3 vs #4: diff mean R +0.41 IC N/D · n 5 vs 1 · INSUFFICIENT DATA n=5 vs n=1; se necesitan >= 30 por grupo
Mismo instante T (mejor rank − resto): n pares 30 · media -0.14 [-0.52, +0.24] · NO EVIDENCE OF RANKING EDGE IC de la diferencia mismo-T incluye 0
El rank #1 suele ser el unico setup del instante: rank y categoria (SETUP > CONDICIONAL) estan confundidos. La comparacion mismo-T es la mas limpia pero requiere instantes con >= 2 setups. El rank es el calculado EN T con los snapshots de ese instante (sin Activity).
Activity — seleccion de activos vs calidad del setup
Cobertura: 39576 de 47181 evaluaciones con Activity registrada en T · 234 setups con Activity y R final · excluidas por Activity posterior a T: 0
A. Calidad del setup segun Activity en T
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| ELEVATED | 67 | -0.20 | [-0.50, +0.12] | -1.00 | 29.9% | [20.2%, 41.7%] | |
| HIGH | 80 | +0.25 | [-0.02, +0.53] | -1.00 | 48.8% | [38.1%, 59.5%] | |
| NORMAL | 44 | +0.18 | [-0.19, +0.53] | -0.26 | 50.0% | [35.8%, 64.2%] | |
| QUIET | 43 | -0.84 | [-0.96, -0.68] | -1.00 | 9.3% | [3.7%, 21.6%] |
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| #1-5 | 89 | +0.47 | [+0.20, +0.74] | +1.23 | 56.2% | [45.8%, 66.0%] | |
| #11+ | 64 | -0.63 | [-0.82, -0.42] | -1.00 | 17.2% | [9.9%, 28.2%] | |
| #6-10 | 81 | -0.29 | [-0.53, -0.04] | -1.00 | 29.6% | [20.8%, 40.3%] |
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| Q1 (bajo) | 78 | -0.49 | [-0.70, -0.28] | -1.00 | 23.1% | [15.1%, 33.6%] | |
| Q2 | 78 | -0.07 | [-0.35, +0.24] | -1.00 | 35.9% | [26.1%, 47.0%] | |
| Q3 (alto) | 78 | +0.28 | [-0.02, +0.56] | -0.09 | 50.0% | [39.2%, 60.8%] |
HIGH/ELEVATED vs NORMAL/QUIET: diff mean R +0.37 [+0.07, +0.67] · n 147 vs 87 · PRELIMINARY POSITIVE ASSOCIATION el IC 95% de la diferencia excluye 0 por arriba (asociacion, no causalidad)
SETUP LONG: HIGH/ELEV vs NORMAL/QUIET: diff mean R +1.40 IC N/D · n 3 vs 6 · INSUFFICIENT DATA n=3 vs n=6; se necesitan >= 30 por grupo
SETUP SHORT: HIGH/ELEV vs NORMAL/QUIET: diff mean R +0.22 [-0.50, +0.88] · n 25 vs 15 · INSUFFICIENT DATA n=25 vs n=15; se necesitan >= 30 por grupo
CONDITIONAL LONG: HIGH/ELEV vs NORMAL/QUIET: diff mean R +1.81 [+1.30, +2.25] · n 17 vs 9 · INSUFFICIENT DATA n=17 vs n=9; se necesitan >= 30 por grupo
CONDITIONAL SHORT: HIGH/ELEV vs NORMAL/QUIET: diff mean R +0.09 [-0.29, +0.47] · n 102 vs 57 · NO EVIDENCE OF ASSOCIATION el IC 95% de la diferencia incluye 0
B. Seleccion: ¿los activos con mas Activity se mueven mas despues? (todas las evaluaciones)
| Activity level en T | n | |mov 24h| medio | IC 95% | Rango 24h medio | |
|---|---|---|---|---|---|
| HIGH | 6360 | 2.82% | [+2.75, +2.89] | 6.53% | |
| ELEVATED | 13657 | 3.22% | [+3.17, +3.27] | 6.71% | |
| NORMAL | 6514 | 2.82% | [+2.75, +2.88] | 6.24% | |
| QUIET | 11680 | 2.53% | [+2.49, +2.58] | 5.53% |
Spearman(Activity score, |mov 24h|) = 0.08 [+0.07, +0.09] · n=38211
Activity solo se registro en snapshots v2 (en tiempo real Top Active desde 2026-10-04). reconstrucción histórica y v1 = N/D: no se reconstruye Activity a posteriori. Activity no interviene en el ranking tecnico.
Assets
Activo
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| AAVE | 8 | -1.00 | [-1.00, -1.00] | -1.00 | 0.0% | [0.0%, 32.4%] | INSUFFICIENT SAMPLE |
| ADA | 3 | +1.44 | IC N/D | +1.46 | 100.0% | [43.9%, 100.0%] | INSUFFICIENT SAMPLE |
| ASTER | 1 | -1.00 | IC N/D | -1.00 | 0.0% | [0.0%, 79.3%] | INSUFFICIENT SAMPLE |
| AVAX | 2 | +1.35 | IC N/D | +1.35 | 100.0% | [34.2%, 100.0%] | INSUFFICIENT SAMPLE |
| BCH | 18 | +1.27 | [+0.74, +1.74] | +1.54 | 83.3% | [60.8%, 94.2%] | INSUFFICIENT SAMPLE |
| BNB | 13 | -0.83 | [-1.00, -0.50] | -1.00 | 7.7% | [1.4%, 33.3%] | INSUFFICIENT SAMPLE |
| BTC | 6 | +1.08 | [-0.00, +2.17] | +1.99 | 66.7% | [30.0%, 90.3%] | INSUFFICIENT SAMPLE |
| DOGE | 18 | -0.12 | [-0.59, +0.37] | -1.00 | 44.4% | [24.6%, 66.3%] | INSUFFICIENT SAMPLE |
| DOT | 8 | -1.00 | [-1.00, -1.00] | -1.00 | 0.0% | [0.0%, 32.4%] | INSUFFICIENT SAMPLE |
| ENA | 4 | -1.00 | IC N/D | -1.00 | 0.0% | [0.0%, 49.0%] | INSUFFICIENT SAMPLE |
| ETH | 14 | -1.00 | [-1.00, -1.00] | -1.00 | 0.0% | [0.0%, 21.5%] | INSUFFICIENT SAMPLE |
| GRAM | 13 | +1.10 | [+0.38, +1.67] | +1.43 | 76.9% | [49.7%, 91.8%] | INSUFFICIENT SAMPLE |
| HBAR | 20 | -1.00 | [-1.00, -1.00] | -1.00 | 0.0% | [0.0%, 16.1%] | INSUFFICIENT SAMPLE |
| ICP | 1 | -1.00 | IC N/D | -1.00 | 0.0% | [0.0%, 79.3%] | INSUFFICIENT SAMPLE |
| LINK | 17 | -0.02 | [-0.52, +0.48] | -1.00 | 47.1% | [26.2%, 69.0%] | INSUFFICIENT SAMPLE |
| LTC | 6 | -0.67 | [-1.00, -0.01] | -1.00 | 16.7% | [3.0%, 56.4%] | INSUFFICIENT SAMPLE |
| NEAR | 1 | -1.00 | IC N/D | -1.00 | 0.0% | [0.0%, 79.3%] | INSUFFICIENT SAMPLE |
| PEPE | 10 | +1.35 | [+1.16, +1.50] | +1.48 | 100.0% | [72.2%, 100.0%] | INSUFFICIENT SAMPLE |
| QNT | 2 | -1.00 | IC N/D | -1.00 | 0.0% | [0.0%, 65.8%] | INSUFFICIENT SAMPLE |
| SHIB | 9 | -0.44 | [-1.00, +0.40] | -1.00 | 22.2% | [6.3%, 54.7%] | INSUFFICIENT SAMPLE |
| SKY | 1 | -1.00 | IC N/D | -1.00 | 0.0% | [0.0%, 79.3%] | INSUFFICIENT SAMPLE |
| SOL | 20 | -0.65 | [-1.00, -0.24] | -1.00 | 15.0% | [5.2%, 36.0%] | INSUFFICIENT SAMPLE |
| SUI | 22 | -0.64 | [-1.00, -0.23] | -1.00 | 13.6% | [4.7%, 33.3%] | INSUFFICIENT SAMPLE |
| TAO | 1 | -1.00 | IC N/D | -1.00 | 0.0% | [0.0%, 79.3%] | INSUFFICIENT SAMPLE |
| TRX | 7 | -0.65 | [-1.00, +0.04] | -1.00 | 14.3% | [2.6%, 51.3%] | INSUFFICIENT SAMPLE |
| UNI | 6 | -0.59 | [-1.00, +0.23] | -1.00 | 16.7% | [3.0%, 56.4%] | INSUFFICIENT SAMPLE |
| WLD | 2 | -1.00 | IC N/D | -1.00 | 0.0% | [0.0%, 65.8%] | INSUFFICIENT SAMPLE |
| WLFI | 15 | +1.63 | [+1.57, +1.68] | +1.65 | 100.0% | [79.6%, 100.0%] | INSUFFICIENT SAMPLE |
| XLM | 2 | -1.00 | IC N/D | -1.00 | 0.0% | [0.0%, 65.8%] | INSUFFICIENT SAMPLE |
| XRP | 18 | -0.36 | [-0.80, +0.13] | -1.00 | 27.8% | [12.5%, 50.9%] | INSUFFICIENT SAMPLE |
Distribución de resultados simulados (señales únicas)
| Outcome canonico | n |
|---|---|
| ENTRY_SKIPPED_OUTSIDE_ZONE | 48 |
| INVALIDATED_BEFORE_TRIGGER | 151 |
| NEVER_TRIGGERED | 8 |
| PENDING | 30 |
| SL_HIT | 176 |
| TP2_TARGET | 92 |
AMBIGUOUS = TP y SL en la misma vela 15m sin gap que fije el orden: nunca se asume el resultado favorable.
MFE / MAE
MFE vs MAE por operación simulada (R) · ● TP2 ● SL ● otros
| Medida | n | Media | IC 95% | Mediana |
|---|---|---|---|---|
| MFE hasta salida | 268 | +0.94 | [+0.86, +1.02] | +0.87 |
| MAE hasta salida | 268 | -0.80 | [-0.84, -0.76] | -1.00 |
Del outcome canonico v1. Limitacion: en la vela de salida por TP2 no se incluye su minimo y en la de SL no se incluye su maximo (orden intrabar desconocido).
Reconstrucción histórica vs tiempo real
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| reconstrucción histórica | 34 | -0.51 | [-0.81, -0.15] | -1.00 | 20.6% | [10.3%, 36.8%] | |
| en tiempo real | 234 | -0.09 | [-0.25, +0.06] | -1.00 | 36.3% | [30.4%, 42.7%] |
reconstrucción histórica = reconstruido con velas cerradas <= T obtenidas despues (sin Activity); en tiempo real = decidido en tiempo real. Rangos: reconstrucción histórica: → · en tiempo real: → . Nunca se mezclan para validar: la validacion out-of-sample usa en tiempo real.
Baseline / Control
B1. Movimiento absoluto posterior: setups vs WAIT
| Modo | Horizonte | n setup / WAIT | |mov| setups | |mov| WAIT | IC diff | |
|---|---|---|---|---|---|---|
| reconstrucción histórica | 4h | 73 / 7502 | 1.00% | 1.07% | [-0.32, +0.23] | NO EVIDENCE OF ASSOCIATION |
| reconstrucción histórica | 24h | 73 / 7502 | 2.77% | 3.00% | [-0.93, +0.52] | NO EVIDENCE OF ASSOCIATION |
| en tiempo real | 4h | 402 / 37871 | 1.01% | 1.19% | [-0.28, -0.08] | PRELIMINARY NEGATIVE ASSOCIATION |
| en tiempo real | 24h | 402 / 37871 | 2.25% | 2.88% | [-0.87, -0.38] | PRELIMINARY NEGATIVE ASSOCIATION |
B2. Movimiento en la direccion del setup desde T (sin entrada ni stop)
| Horizonte | n | Media | IC 95% | Mediana | |
|---|---|---|---|---|---|
| 1h | 475 | +0.08% | [-0.00, +0.16] | -0.00% | NO EVIDENCE OF EDGE |
| 4h | 475 | +0.08% | [-0.05, +0.21] | -0.01% | NO EVIDENCE OF EDGE |
| 12h | 475 | -0.11% | [-0.34, +0.12] | -0.26% | NO EVIDENCE OF EDGE |
| 24h | 475 | -0.22% | [-0.54, +0.09] | -0.30% | NO EVIDENCE OF EDGE |
B3. Exceso frente a la misma apuesta sobre los otros activos (WAIT) del mismo T
| Horizonte | n | Exceso medio | IC 95% | |
|---|---|---|---|---|
| 4h | 475 | +0.15% | [+0.05, +0.24] | PRELIMINARY POSITIVE ASSOCIATION |
| 24h | 475 | +0.23% | [-0.02, +0.48] | NO EVIDENCE OF EXCESS MOVE |
Comparaciones descriptivas en % de precio desde el cierre en T (no son operaciones simuladas: no hay entrada, stop ni coste). B3 resta la misma apuesta direccional sobre los demas activos WAIT del mismo T (controla el movimiento general del mercado).
Feature analysis (asociaciones descriptivas)
| Feature en T | campo | n | Spearman vs R | IC 95% (bootstrap) | |
|---|---|---|---|---|---|
| Alineacion | aligned_count | 268 | -0.06 | [-0.17, +0.06] | NO EVIDENCE OF ASSOCIATION |
| Cautelas | cautions | 268 | 0.08 | [-0.03, +0.19] | NO EVIDENCE OF ASSOCIATION |
| Rank tecnico en T | technical_rank | 268 | -0.07 | [-0.17, +0.03] | NO EVIDENCE OF ASSOCIATION |
| Setups simultaneos en T | slot_actionable | 268 | -0.15 | [-0.25, -0.04] | PRELIMINARY NEGATIVE ASSOCIATION |
| Riesgo (ATR15) | risk_atr | 268 | -0.16 | [-0.26, -0.05] | PRELIMINARY NEGATIVE ASSOCIATION |
| Distancia al gatillo (ATR15) | trigger_distance_atr | 268 | 0.05 | [-0.07, +0.17] | NO EVIDENCE OF ASSOCIATION |
| R:R TP3 | rr_tp3 | 191 | 0.14 | [-0.00, +0.26] | NO EVIDENCE OF ASSOCIATION |
| RSI 15m | rsi_15m | 268 | 0.00 | [-0.11, +0.12] | NO EVIDENCE OF ASSOCIATION |
| RSI 1h | rsi_1h | 268 | 0.02 | [-0.10, +0.13] | NO EVIDENCE OF ASSOCIATION |
| ATR% 15m | atr_pct_15m | 268 | 0.13 | [+0.00, +0.25] | PRELIMINARY POSITIVE ASSOCIATION |
| ATR% 1h | atr_pct_1h | 268 | 0.16 | [+0.04, +0.29] | PRELIMINARY POSITIVE ASSOCIATION |
| Vol. relativo 15m | rel_volume_15m | 268 | -0.17 | [-0.29, -0.04] | PRELIMINARY NEGATIVE ASSOCIATION |
| Activity rank | activity_rank_effective | 234 | -0.47 | [-0.56, -0.36] | PRELIMINARY NEGATIVE ASSOCIATION |
| Activity score | activity_score | 234 | 0.21 | [+0.10, +0.33] | PRELIMINARY POSITIVE ASSOCIATION |
| Hora UTC | hour_utc | 268 | 0.10 | [-0.03, +0.22] | NO EVIDENCE OF ASSOCIATION |
Solo features almacenadas en T. Correlacion de rangos con el R realizado: asociacion descriptiva, nunca causalidad; sin entrenamiento ni optimizacion de umbrales. Con ~15 features, alguna puede salir "significativa" por azar (comparaciones multiples).
Walk-forward / validacion temporal
READY operaciones simuladas únicas con R: 268 · necesarios ~120 · expanding window, 4 bloques cronologicos; seleccion en train, evaluacion en el bloque siguiente; minimo 30 operaciones simuladas por bloque
| Fold | n train | n test | Seleccion (train) | Mean R train | Mean R test | n test sel. |
|---|---|---|---|---|---|---|
| 1 | 67 | 67 | SHORT | +0.16 | +0.03 | 58 |
| 2 | 134 | 67 | SHORT | +0.08 | -0.35 | 52 |
| 3 | 201 | 67 | CONDITIONAL | -0.02 | -0.17 | 55 |
Estabilidad por periodo (semana ISO, por origen)
| Origen | Semana | n | Mean R | IC 95% | |
|---|---|---|---|---|---|
| reconstrucción histórica | 2026-W37 | 10 | -1.00 | [-1.00, -1.00] | INSUFFICIENT SAMPLE |
| reconstrucción histórica | 2026-W38 | 7 | -0.28 | [-1.00, +0.65] | INSUFFICIENT SAMPLE |
| reconstrucción histórica | 2026-W39 | 16 | -0.27 | [-0.75, +0.30] | INSUFFICIENT SAMPLE |
| reconstrucción histórica | 2026-W40 | 1 | -1.00 | IC N/D | INSUFFICIENT SAMPLE |
| en tiempo real | 2026-W40 | 4 | -1.00 | IC N/D | INSUFFICIENT SAMPLE |
| en tiempo real | 2026-W41 | 43 | -0.11 | [-0.48, +0.26] | |
| en tiempo real | 2026-W42 | 48 | +0.04 | [-0.29, +0.40] | |
| en tiempo real | 2026-W43 | 17 | +0.46 | [-0.13, +1.04] | INSUFFICIENT SAMPLE |
| en tiempo real | 2026-W44 | 66 | -0.13 | [-0.43, +0.17] | |
| en tiempo real | 2026-W45 | 56 | -0.25 | [-0.55, +0.05] |
Hipotesis pre-registradas (se evaluan SOLO con datos en tiempo real posteriores al registro)
| ID | Hipotesis | Registrada | n | Metrica | IC 95% | |
|---|---|---|---|---|---|---|
| H1 | Setups unicos activados (en tiempo real) tienen R medio > 0. | 229 | -0.07 | [-0.23, +0.09] | NO EVIDENCE OF EDGE | |
| H2 | Setups con rank tecnico #1 en T obtienen mayor R medio que los de rank >= 2. | 229 | +0.26 | [-0.16, +0.65] | NO EVIDENCE OF ASSOCIATION | |
| H3 | Setups con alineacion 4/4 obtienen mayor R medio que los de alineacion < 4. | 229 | -0.15 | [-0.46, +0.18] | NO EVIDENCE OF ASSOCIATION | |
| H4 | Setups en activos con Activity HIGH/ELEVATED en T obtienen mayor R medio que NORMAL/QUIET. | 229 | +0.35 | [+0.04, +0.66] | PRELIMINARY POSITIVE ASSOCIATION | |
| H5 | LONG y SHORT difieren en R medio (bilateral). | 229 | +0.27 | [-0.21, +0.77] | NO EVIDENCE OF ASSOCIATION |
Otros segmentos (contexto en T)
Hora UTC (bloques 4h)
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| 00-03h | 64 | -0.31 | [-0.57, -0.02] | -1.00 | 28.1% | [18.6%, 40.1%] | |
| 04-07h | 36 | -0.06 | [-0.46, +0.34] | -1.00 | 38.9% | [24.8%, 55.1%] | |
| 08-11h | 40 | -0.33 | [-0.67, +0.07] | -1.00 | 27.5% | [16.1%, 42.8%] | |
| 12-15h | 31 | -0.40 | [-0.76, -0.03] | -1.00 | 25.8% | [13.7%, 43.2%] | |
| 16-19h | 29 | +0.24 | [-0.24, +0.73] | -1.00 | 48.3% | [31.4%, 65.6%] | INSUFFICIENT SAMPLE |
| 20-23h | 68 | +0.02 | [-0.28, +0.31] | -1.00 | 39.7% | [28.9%, 51.6%] |
Dia de la semana (UTC)
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| Fri | 71 | -0.34 | [-0.58, -0.07] | -1.00 | 26.8% | [17.9%, 38.1%] | |
| Mon | 30 | -0.61 | [-0.90, -0.22] | -1.00 | 13.3% | [5.3%, 29.7%] | |
| Sat | 26 | -0.04 | [-0.50, +0.44] | -1.00 | 38.5% | [22.4%, 57.5%] | INSUFFICIENT SAMPLE |
| Sun | 23 | -0.58 | [-0.87, -0.23] | -1.00 | 21.7% | [9.7%, 41.9%] | INSUFFICIENT SAMPLE |
| Thu | 56 | +0.02 | [-0.30, +0.34] | -1.00 | 42.9% | [30.8%, 55.9%] | |
| Tue | 31 | -0.55 | [-0.92, -0.14] | -1.00 | 16.1% | [7.1%, 32.6%] | |
| Wed | 31 | +1.06 | [+0.66, +1.41] | +1.52 | 80.6% | [63.7%, 90.8%] |
Origen en tiempo real / reconstrucción histórica
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| reconstrucción histórica | 34 | -0.51 | [-0.81, -0.15] | -1.00 | 20.6% | [10.3%, 36.8%] | |
| en tiempo real | 234 | -0.09 | [-0.25, +0.06] | -1.00 | 36.3% | [30.4%, 42.7%] |
Volatilidad 15m (regimen en T)
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| HIGH | 14 | -0.42 | [-1.00, +0.18] | -1.00 | 21.4% | [7.6%, 47.6%] | INSUFFICIENT SAMPLE |
| LOW | 20 | -0.22 | [-0.75, +0.33] | -1.00 | 30.0% | [14.5%, 51.9%] | INSUFFICIENT SAMPLE |
| NORMAL | 234 | -0.12 | [-0.28, +0.03] | -1.00 | 35.5% | [29.6%, 41.8%] |
Volatilidad 1h (regimen en T)
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| HIGH | 65 | +0.31 | [-0.01, +0.64] | +0.83 | 50.8% | [38.9%, 62.5%] | |
| LOW | 8 | -0.70 | [-1.00, -0.09] | -1.00 | 12.5% | [2.2%, 47.1%] | INSUFFICIENT SAMPLE |
| NORMAL | 195 | -0.28 | [-0.43, -0.12] | -1.00 | 29.7% | [23.8%, 36.5%] |
Filtro 30m
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| BEARISH | 140 | -0.07 | [-0.27, +0.13] | -1.00 | 37.9% | [30.2%, 46.1%] | |
| BULLISH | 41 | -0.22 | [-0.58, +0.15] | -1.00 | 31.7% | [19.6%, 47.0%] | |
| NEUTRAL | 87 | -0.23 | [-0.47, +0.03] | -1.00 | 29.9% | [21.3%, 40.2%] |
TP3 definido (nivel 1h/4h)
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| no | 77 | -0.03 | [-0.31, +0.23] | -1.00 | 39.0% | [28.8%, 50.1%] | |
| si | 191 | -0.19 | [-0.36, -0.02] | -1.00 | 32.5% | [26.2%, 39.4%] |
Contexto Pulse (solo registro)
| Grupo | n | Mean R | IC 95% | Mediana R | Win rate | IC win (Wilson) | |
|---|---|---|---|---|---|---|---|
| pulse N/D | 84 | +0.16 | [-0.11, +0.44] | -1.00 | 45.2% | [35.0%, 55.9%] | |
| pulse a favor | 3 | -1.00 | IC N/D | -1.00 | 0.0% | [0.0%, 56.1%] | INSUFFICIENT SAMPLE |
| pulse en contra | 2 | +1.13 | IC N/D | +1.13 | 100.0% | [34.2%, 100.0%] | INSUFFICIENT SAMPLE |
| pulse sin lado | 179 | -0.29 | [-0.46, -0.12] | -1.00 | 29.1% | [22.9%, 36.1%] |
Data quality
Activos cubiertos: 33 (AAVE, ADA, ASTER, AVAX, BCH, BNB, BTC, DOGE, DOT, ENA, ETH, GRAM, HBAR, ICP, LINK, LTC, NEAR, ONDO, PEPE, PUMP, QNT, SHIB, SKY, SOL, SUI, TAO, TRX, UNI, WLD, WLFI, XLM, XRP, ZEC)
Rango temporal: reconstrucción histórica: → · en tiempo real: →
Cobertura Activity: reconstrucción histórica: 0/7575 · en tiempo real: 39576/39606 · Activity posterior a T detectada: 0
Cobertura outcome: 45863 de 47181 evaluaciones con outcome final · pending 1318 (de ellos 0 bloqueados: —)
Ambiguous: 0 · DATA_GAP: 0 · setups 509 (unicos 505)
Campos faltantes (N/D: nunca almacenados en T o no aplicables)
| Campo | Faltan | De |
|---|---|---|
| activity_available | 7605 | 47181 |
| activity_rank_effective | 7607 | 47181 |
| price_at_T | 10 | 47181 |
| aligned_count | 0 | 509 |
| atr15 | 0 | 509 |
| rr_tp3 | 169 | 509 |
| trigger_distance_atr | 0 | 509 |
| rsi_15m | 10 | 47181 |
| pulse_global_direction | 17449 | 47181 |
Huecos entre instantes evaluados
| Modo | Instantes | Paso esperado | Slots ausentes | Mayores huecos |
|---|---|---|---|---|
| en tiempo real | 3265 | 15 min | 314 | 01.10.2026 14:45 CEST → 04.10.2026 20:00 CEST (77.2 h); 17.10.2026 11:00 CEST → 17.10.2026 11:30 CEST (0.5 h); 09.10.2026 12:00 CEST → 09.10.2026 12:30 CEST (0.5 h) |
| reconstrucción histórica | 505 | 60 min | 0 | — |
- MFE/MAE hasta la salida (outcome v1): en la vela de salida por TP2 no se incluye su minimo (orden intrabar desconocido) y en la de salida por SL no se incluye su maximo.
- Rank tecnico en T: calculado solo con los snapshots del mismo instante y modo.
- pending_stalled = pendientes cuyo par ya no se ingiere (salio del universo); no se resolveran sin nueva ingesta.
simulada-validation-v1 · simulada-validation-dataset-v1 · engine simulada-setup-v1 · outcome simulada-outcome-v1 · bootstrap B=2000 seed=20261005 (IC normal si n > 400) · min sample 30 · pagina generada . Reproducible: python -m src.paper_lab validate --export DIR