Walkthroughs
End-to-end stories. One trading idea, walked all the way through the agent loop — propose, backtest, probe, modify, decide.
Run a research project
Track multiple theses through scan, build, and validate inside one kanban. Ideas / Tested / Built columns keep every experiment — including the rejections — visible in one place.
Scan a thesis for edge
Test a trading idea against forward returns in seconds. The agent runs signal_scan, returns four scalars — win rate, Sharpe, p-value, reliable y/n — and you decide whether to build it or skip it. No backtest required.
Building a strategy around system events
How to talk to the agent about Quawd's system events — news, earnings, dividends — and compose them with the boring building blocks: trend gates, volume confirmation, stops and targets. We build one strategy on AAPL that uses positive-sentiment news to time entries, filters with a 50-day trend, requires volume confirmation, and caps losses with tight risk. The whole strategy is one conversation.
Finding alpha in changing regimes with Quawd
How Quawd turns a one-line trading idea into a battle-tested strategy. We walk a Connors RSI(2) setup on SPY through Quawd's iteration loop — propose, backtest, probe, modify, decide — and watch it hold up across five years of changing market regimes.