—
Loading…
Trading mode
Profit over time
Running total after each settled trade, fees included.
Is the bot's confidence right?
For trades it rated at each confidence level: what it predicted vs. how often they actually won.
By city
Every morning the bot compares the last 90 days of each model's forecasts for your settlement stations with the
official NWS highs. It learns each model's bias ("runs 1.4° cool") and how far off it usually is, then
shifts and sizes its forecasts to match. The backtest replays those days using only information available at the time.
Backtest: does the tuned confidence match reality?
Every bucket on every past day, all cities: the chance the tuned forecast gave it vs. how often it actually happened.