π‘οΈ Weather Ensemble
Phase 3: Late Entry + Calibrated
11 Sources
31 Cities
Bias Corrected
Paper trading performance & forecast accuracy β auto-refreshes every 60s
Refreshing...
β Portfolio
Analysis
π Trading
π― Forecast Accuracy
π Calibration
π¬ Ensemble Health
$0
Total P&L
β
Win Rate
0
Open Positions
$10,000
Bankroll
βοΈ Active Config
Stop-Loss:
30%
Min Bet:
$500
Min Edge:
20%
Entry Window:
3-24h
Daily Cap:
-$2,000
Scan:
30 min
Min Sources:
2
Calibration:
Isotonic (883 brackets)
π Equity Curve
ποΈ P&L by City
π‘ Open Positions
π Performance by Segment
No resolved trades yet
π° Win Rate by Entry Price
No resolved trades yet
π Recent Trades
No resolved trades yet β positions will appear here when they close
π― Forecast Accuracy β MAE by City
Open-Meteo historical forecasts vs TWC actuals. 40 dates per city. Lower = better forecasting = bigger edge.
π City Calibration Detail
πΊοΈ City Tiers
π Empirical RMSE by City Γ Horizon
RMSE captures both bias and variance. Used in
prob_in_range()
for probability calculation. Higher RMSE = wider uncertainty = fewer trades.
π Bias by City
Systematic forecast error. Positive = forecast too low (actual hotter). Negative = forecast too high (actual colder). Corrected in real-time.
β
Active Sources
β
Avg Sources/Signal
β
Cached Forecasts
β
Degraded / Offline
π‘ Source Status
Loading ensemble status...
π Source Distribution
How many sources contribute per cached signal
No cached forecasts yet
πΊοΈ City Γ Source Coverage
Green = source contributing for this city
No cached forecasts yet