Informes
Recientes
Informe Diario
2026-06-20
Informe de Trading — 2026-06-20 (modo fallback)
Groq no disponible — informe generado directamente desde los modelos cuantitativos.
337 simbolos analizados. Distribucion: {"ALCISTA": 86, "BAJISTA": 11, "NEUTRAL": 240}
Corto plazo (top 5)
| Ticker | Precio | Pred | Dir | Conf | Vol |
|---|---|---|---|---|---|
| UBSG.SW | $41.06 | $39.10 (-4.78%) | ALCISTA ⚠ | 79.1% | 25% |
| LLOY.L | $106.05 | $101.80 (-4.01%) | ALCISTA ⚠ | 79.1% | 26% |
| ASML.AS | $1676.00 | $1614.72 (-3.66%) | ALCISTA ⚠ | 79.1% | 40% |
| ENI.MI | $21.14 | $20.00 (-5.40%) | ALCISTA ⚠ | 48.0% | 33% |
| BGEO.L | $11550.00 | $11158.50 (-3.39%) | ALCISTA ⚠ | 63.6% | 32% |
Medio plazo (top 5)
| Ticker | Precio | Pred | Dir | Conf | Vol |
|---|---|---|---|---|---|
| BESI.AS | $317.80 | $305.10 (-4.00%) | ALCISTA ⚠ | 55.3% | 51% |
| AAL.L | $3995.00 | $3805.00 (-4.76%) | ALCISTA ⚠ | 48.0% | 41% |
| GLE.PA | $78.05 | $75.22 (-3.63%) | ALCISTA ⚠ | 55.3% | 35% |
| EDV.L | $4180.00 | $3961.22 (-5.23%) | ALCISTA ⚠ | 41.0% | 64% |
| G24.DE | $76.50 | $79.89 (+4.43%) | BAJISTA ⚠ | 46.6% | 29% |
Largo plazo (top 5)
| Ticker | Precio | Pred | Dir | Conf | Vol |
|---|---|---|---|---|---|
| STAN.L | $2058.00 | $1997.05 (-2.96%) | ALCISTA ⚠ | 63.6% | 36% |
| AD.AS | $34.52 | $35.74 (+3.55%) | BAJISTA ⚠ | 54.0% | 23% |
| VOD.L | $108.70 | $104.26 (-4.08%) | ALCISTA ⚠ | 48.0% | 26% |
| AENA.MC | $27.04 | $26.13 (-3.38%) | ALCISTA ⚠ | 55.3% | 22% |
| DG.PA | $130.65 | $126.35 (-3.29%) | ALCISTA ⚠ | 55.3% | 19% |
Criptomonedas (top 5)
Sin candidatos.
Cartera diversificada
| Ticker | Peso | Precio | Dir | Vol | Horizonte |
|---|---|---|---|---|---|
| UBSG.SW | 14.2% | $41.06 | ALCISTA | 25% | Corto |
| LLOY.L | 13.4% | $106.05 | ALCISTA | 26% | Corto |
| ASML.AS | 8.8% | $1676.00 | ALCISTA | 40% | Corto |
| BESI.AS | 6.8% | $317.80 | ALCISTA | 51% | Medio |
| AAL.L | 8.5% | $3995.00 | ALCISTA | 41% | Medio |
| GLE.PA | 10.0% | $78.05 | ALCISTA | 35% | Medio |
| STAN.L | 9.6% | $2058.00 | ALCISTA | 36% | Largo |
| AD.AS | 15.3% | $34.52 | BAJISTA | 23% | Largo |
| VOD.L | 13.4% | $108.70 | ALCISTA | 26% | Largo |
Performance por modelo
| Modelo | Accuracy | MAE | Sesgo | N evals | Simbolos |
|---|---|---|---|---|---|
| ensemble | 56% | 2.88% | 0.66% | 26169 | 973 |
| gbm | 51% | 1.58% | -0.25% | 26112 | 965 |
| merton | 50% | 1.50% | -0.33% | 26169 | 973 |
| ml | 50% | 9.70% | -0.78% | 26169 | 973 |
| kalman_signal | 47% | — | — | 26169 | 973 |
| hmm_signal | 45% | — | — | 26169 | 973 |
| kalman | 44% | 2.09% | -0.56% | 26168 | 973 |
Metricas globales (ensemble, 7d)
MAE: 2.88% | RMSE: 22.37% | Hit rate: 56% | N: 26169
Disclaimer: contenido informativo, no es asesoramiento financiero.
12. Metricas Profesionales
Risk-adjusted metrics computadas sobre las ultimas 2000 evaluaciones del ensemble en los ultimos 30 dias. Proxy de retornos: +1%/-1% por hit/miss (bet de tamano fijo). Anualizacion: 252.
| Metrica | Valor |
|---|---|
| Sharpe (anualizado) | -0.02 |
| Sortino (anualizado) | -0.02 |
| Calmar | -0.02 |
| Max Drawdown | -60.56% |
| Profit Factor | 1.00 |
| Win Rate (= Hit Rate global) | 49.95% |
| Expectancy por trade | -0.001% |
| N evaluaciones | 2000 |
Hit rate por bucket de confianza
| Bucket conf. | N | Hit rate |
|---|---|---|
| 00-30 | 1549 | 51.3% |
| 30-50 | 383 | 45.7% |
| 50-60 | 43 | 53.5% |
| 60-70 | 17 | 23.5% |
| 70-80 | 7 | 14.3% |
| 80-100 | 1 | 100.0% |
13. Calibracion (post-isotonic)
Isotonic (PAV) calibrator entrenado sobre 560 muestras historicas el 2026-06-20T00:30:00Z. Las NUEVAS predicciones del sistema ya entregan confidence calibrada (raw_confidence se conserva para auditoria). Tabla comparativa: mismas 5000 evaluaciones recientes, mostradas por bucket antes (raw) y despues (calibrado) del mapeo.
Tabla comparativa por bucket
| Bucket | N raw | Hit rate raw | N calibrado | Hit rate calibrado |
|---|---|---|---|---|
| 00-30 | 3531 | 55.4% | 0 | N/A |
| 30-50 | 1193 | 55.4% | 0 | N/A |
| 50-60 | 172 | 56.4% | 0 | N/A |
| 60-70 | 83 | 38.6% | 0 | N/A |
| 70-80 | 19 | 57.9% | 0 | N/A |
| 80-100 | 2 | 50.0% | 5000 | 55.2% |
Curva isotonic (knots)
| raw_conf | calibrated |
|---|---|
| 29.0 | 99.8% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 60.0 | 100.0% |
| 62.0 | 100.0% |
| 62.0 | 100.0% |
| 67.0 | 100.0% |
| 67.0 | 100.0% |
| 67.0 | 100.0% |
| 70.0 | 100.0% |
| 70.0 | 100.0% |
| 70.0 | 100.0% |
| 70.0 | 100.0% |
| 70.0 | 100.0% |
| 70.0 | 100.0% |
| 70.0 | 100.0% |
| 70.0 | 100.0% |
| 70.0 | 100.0% |
| 70.0 | 100.0% |
| 70.0 | 100.0% |
| 70.0 | 100.0% |
Criterio de exito (bucket 80+ >= 70% hit_rate): NO (hit rate calibrado = 55.2%, n=5000). Razon: el raw signal esta anti-calibrado (hit rate DECRECE con confidence en los datos historicos). PAV es monotona no decreciente, por lo que no puede restaurar magnitudes altas cuando el input no las soporta. Para elevar el bucket 80+ al 70% habria que corregir el generador de confianza (no el calibrador) o flippear el signo de la direccion en el master_agent.
Notas: trained on 560 (raw_conf, hit) pairs
14. Sizing y Riesgo
Pipeline profesional: Kelly fraccional (0.25) -> Vol-target (15% anual) -> Cap de correlacion (max 20%). Muestreadas las ultimas 8 predicciones con sizing calculado.
Resumen
- Kelly medio: 7.58% (cap 20%)
- Risk score medio: 71.1/100 (100 = mas seguro)
- Predicciones con Kelly = 0: 4/8
Distribucion de Kelly size
| Rango Kelly | N | % |
|---|---|---|
| 0% | 4 | 50.0% |
| 0-5% | 0 | 0.0% |
| 5-10% | 1 | 12.5% |
| 10-15% | 1 | 12.5% |
| 15-20% | 2 | 25.0% |
Top-5 riesgo mas elevado (risk_score bajo)
| Symbol | Asset | Direction | Conf | Kelly | Risk |
|---|---|---|---|---|---|
| USO | Oil (WTI) | ALCISTA | 55 | 6.71% | 56.1 |
| SPY | S&P 500 | ALCISTA | 57 | 20.00% | 57.7 |
| QQQ | NASDAQ 100 | ALCISTA | 57 | 13.91% | 60.9 |
| EURUSD=X | EUR/USD | ALCISTA | 70 | 20.00% | 61.0 |
| BTC-USD | Bitcoin | NEUTRAL | 50 | 0.00% | 74.6 |
Sin alertas de riesgo extremo (todos los risk_score >= 40).
Caveats: vol anualizada estimada desde ATR% (rough ~16x daily). Cap de correlacion actualmente usa matriz vacia (penalty=1) cuando el agente analiza un unico simbolo; el cap duro del 20% sigue aplicandose. Kelly full nunca se usa — siempre 1/4 Kelly.
15. Top Picks del dia
Seleccion diaria por horizonte. Criterio: maximo score combinado = master_score * conf_cal/100 * sign(direction) * (1 - risk/100). CORTO usa GBM/Kalman, MEDIO stacking ensemble, LARGO fundamental+macro. Ver /top-picks para histograma y equity.
| Horizonte | Asset | Ticker | Dir | Conf cal | Master | Kelly | Precio | Target | Rationale |
|---|---|---|---|---|---|---|---|---|---|
| CORTO (1-7d) | ABN.AS | ABN.AS | ALCISTA | 100.0% | 50.0 | 0.00% | 37.5800 | 38.3316 | GBM/Kalman ensemble: conf_raw=52.8% (cal=100.0%), dir=ALCISTA, vol_annual=28.2%, regime=BULL |
| MEDIO (1-3m) | A2A.MI | A2A.MI | ALCISTA | 100.0% | 50.0 | 0.00% | 2.2500 | 2.2545 | Stacking ensemble (ridge): pred_return=0.20%, conf_raw=54.6% (cal=100.0%), dir=ALCISTA |
| LARGO (6-12m) | S&P 500 | SPY | NEUTRAL | 99.8% | 57.0 | 20.00% | 746.7400 | 746.7400 | Fundamental=50/100 (neutral) + Macro=45/100 (RISK-OFF); blend=47.5 -> NEUTRAL, cal_conf=99.8% |
Caveats: target_price es indicativo (derivado del modelo y horizonte), NO es un precio objetivo formal. El rationale combina senales cuantitativas; no es analisis fundamental humano. El equity por horizonte en /top-picks necesita varios dias de acumulacion antes de ser estadisticamente significativo.