The calibration dashboard answers a simple question: when Tradient Score says a trade is an 80, do those paper trials actually win more often (and make more money) than the 50s? Closed paper trials are bucketed by score and charted against realized win rate and mean P&L.
Where to find it
Open Portfolio. The calibration card appears once you have paper activity — especially helpful after enabling auto paper so trials accumulate without babysitting every fill.
How to read it
- Closed trials / total — how much data the chart has. Thin samples are noisy; treat early buckets as directional.
- Score buckets — predicted score ranges at entry. A well-behaved score rises roughly monotonically across buckets (higher score → better realized outcomes).
- Win rate & mean P/L — what actually happened after exit rules or manual closes.
How to get better data
Manually paper high-conviction results from Focus, or leave auto paper on for a few weeks. Prefer closing trials with the same rules you would use live so the buckets reflect a real process, not random hold-to-expiry noise.