Local automation runtime

The web app is the operator UI. A separate local worker process handles Shariah Tables sync, Upstox-backed scanning, and paper automation while Kite stays execution-only.

PAPERUpstox data readyACTIVEStopped off hours
Data providerUpstox Analytics Token

Ready through 15 Apr 2027, 11:59 pm

Broker sessionJuned Rahim Sayyed

Last auth 14 May 2026, 1:55 pm

Runtime postureStopped off hours

Upstox quote fetch failed with status 401.

Next stepsUpstox for data, Kite only for execution

Add UPSTOX_ANALYTICS_TOKEN for scans and backtests, then run npm run worker or npm run local. Kite login is only needed when live execution or broker sync is required.

Backtests

Compare strategy variants

Run the unchanged rotation strategy against the defensive EMA20 reference and the historical 10% max-loss reference over the same fixed windows before trusting any short-term improvement.

Back to history

New comparison

Runs 6M, 1Y, 3Y, 5Y, and MAX sequentially for the selected strategies.

Fixed-window validation

This creates 15 child backtests and runs them one by one. A plan is not accepted unless it beats Nifty across every fixed window and clears the long-window baseline check.

Strategies to compare

The current strategy is always included as the baseline, and the default validation round compares only the defensive EMA20 reference and the historical 10% max-loss reference against it. The selective fail-fast EMA experiment stays available for separate backtests until this shortlist is validated on Shariah Tables.

6M1Y3Y5YMAX / 20YSequential jobsNifty 50 benchmark

The comparison saves child runs individually so you can inspect every failed or passing window.

Comparison history

Each comparison groups child backtests into one pass/fail matrix.

No comparisons yet

Create a comparison to judge strategy variants across short and long windows together.