concept
Historical Intelligence
Historical Intelligence surfaces pattern likelihoods from your own playbook history — per instrument, once you have enough labeled trades for Chartnaut to measure outcomes responsibly.
Think of it as memory for a strategy: "When I take this setup on ES, how often does +1R hit before -1R?"
Where to find it
Playbooks → your playbook → Manage Playbook → Historical Intelligence tab.
Readouts are scoped to one playbook and one instrument at a time — not global market truth.
The ~100 trades threshold
Chartnaut needs roughly 100 trades per instrument in the book before likelihood views are meaningful.
Below that, the UI may show progress counters or early hints — treat them as directional, not verdicts.
What it shows
Depending on data volume and version:
Hit rates at configured R-levels
Tag-conditioned outcomes ("A+ only" vs all trades)
Comparisons across quarters from Session Map collection
Exact charts follow the live Historical Intelligence tab in Manage Playbook.
Improve the signal
Collect deliberately via Trade Book Session Map and Collecting Data sessions — see Collect playbook data
Tag consistently — see Tags and R-levels
Choose training sources — Logged, Backtest, Live — see Training sources
Retrain when you add a large new batch or fix mis-tagged history
Garbage labels produce garbage likelihoods. Intelligence reflects what you logged, not what you wish you traded.
Agents and intelligence
Agents can discuss playbook stats using the same data you see. They do not replace disciplined logging.
Humans and agents share tools — you decide what gets tagged and which trades enter training.
Important distinctions
Historical Intelligence = playbook execution memory (~100 trades/instrument).
Research studies on Definitions = hypothesis testing on setup events — different data pipeline.
In-terminal pins show study results on chart — not the same as Historical Intelligence readouts.
Next
Training sources — Logged vs Backtest vs Live
Collect playbook data — fill quarters on Session Map