A working notebook
Build a signal. Test the assumption.
This guide covers Signal Studio in the Beta interface: current matches, saved strategies and hypothetical portfolios. Start simple; open the controls you need.
1. Define what you want to observe
The default pool is all US stocks Ducky currently covers, not every US listing. Choose your watchlist or a custom verified pool when you want a narrower scope. Market cap is an optional filter.
Use daily, weekly or monthly RSI, supported drawdown windows, or reported insider purchases. ALL requires every condition; ANY requires at least one. Missing evidence remains visible.
Which high is the price below?
Year to date compares with the highest close since January 1. Past year uses a rolling calendar year. These are drawdowns from a high, not returns from the first day of the period.
2. Inspect the current matches
Start with stocks that matched during the selected history window: a small price chart, the latest reconstructed signal, and each stock’s return beside SPY and QQQ over the same dates. The current match count stays separate. History is reconstructed, not a record of notifications sent.
3. Run a hypothetical portfolio
Select stocks from the list and choose Simulate selected stocks to create a separate portfolio. Drag its history start or holding period from 1 to 60 trading days; release to update. Each position counts from its own entry, and the previous curve stays visible while the update runs.
Equal allocation gives each resolved stock its own starting cash sleeve. Custom weights can leave some capital in cash. A stock enters at the next regular open after its trigger; a repeated trigger while it is held does not restart the holding clock.
What the simulation measures
Entry day counts as holding day one; exit uses the selected trading day’s close. Prices are split-adjusted; cash distributions are excluded. The default cost is 10 basis points per side. The same daily portfolio path underlies return, maximum drawdown, volatility and Sharpe. Sharpe uses a zero risk-free rate and needs at least 20 daily return observations.
4. Read the timeline and constituents
The main chart fills the saved simulation window. Use the small date strip beneath it to extend the next run up to 12 months; it does not invent returns for uncalculated dates. The fixed chart marker is the first reconstructed matching signal within the applied window, not a past notification. Metrics, applied conditions and historical trades follow below.
Incomplete condition or price history can make portfolio headline metrics unavailable. A blank metric is not a zero return. Today’s covered pool is replayed through history; it does not reconstruct the membership of a historical index.
5. Save, return, and decide what to monitor
Edits save asynchronously; confirm they are saved before leaving. Reopening a portfolio restores its last dated report without recalculating. After you start that portfolio’s simulation, a settled chart gesture updates it with the new settings.
Monitoring is a separate confirmed action with its own channels and status. Manage existing monitors in My playbooks. Deleting a saved strategy does not stop a separate monitor. A historical simulation does not send past alerts.
Open Signal Studio Beta →Take the same definition to your AI
Export the structured playbook, ask your AI to propose a change, then validate it before import. The connection guide explains the exact schema and which tools can save or simulate it.
Connect your AI →Checked against the product · October 19, 2026