Radar / Ideas / RuleGate — pre-trade discipline…

RuleGate — pre-trade discipline copilot for intraday index traders

daily ideaambitiousJEV confidence 0.752026-09-24

OutcomeA retail intraday trader who actually follows their own rulebook — enforced pre-trade discipline that cuts the unforced errors (revenge trades, overtrading, ignored stops) behind most blown accounts.

The problem

SEBI's FY22-FY24 study found 93% of India's 1+ crore individual F&O traders lost money (Rs.1.8 lakh crore aggregate, ~Rs.2 lakh average), while 97% of FPI profits came from algos. The edge retail lacks is not information but discipline: revenge trades, overtrading, ignored stop-losses.

The idea

A pre-trade gate between the trader and the broker. The trader writes a rulebook once; before every order, Jev answers typed questions — does this trade match the plan? (noul), is this revenge trading? (noul), risk budget remaining? (score) — with hard vetoes enforced in code. A post-market journal classifies every mistake (fomo / revenge / overtrading / plan-followed). Paper-trading first; live mode needs explicit per-day arming. Decision-support tooling, not financial advice.

Why now

NSE's 2026 retail algo circular lets individuals automate up to 10 orders/sec with just a static IP and no registration, and Zerodha made personal Kite Connect APIs free. The architecture pattern is proven in the open by buberlo/jev-trader: 'Jev should not own the trading system — it should own selected judgments inside it.'

What it combines

JEV (typed pre-trade judgments with hard vetoes) + broker execution APIs (Zerodha Kite Connect) + a journaling loop — the collision turns a static rulebook into a living gate that learns which rules actually save money.

Distribution

B2B2C: sell to brokers (Zerodha, Upstox) as an embedded discipline layer — their UI keeps showing the trader's rule-based decisions (plan-match, revenge-trade flag, risk budget) at order time. Brokers get retention and a responsible-trading story; RuleGate gets distribution to crores of traders with zero CAC. The direct-to-trader paper-trading app is the wedge MVP that proves it works.

Build it with

Jev (TypeSafe AI)Decision layer: noul/score/choice questions on trade state, ~33ms per judgment.
Zerodha Kite ConnectOrder placement and positions; personal tier free, data APIs Rs.500/month.
jev-trader patternCopy the separation: deterministic state engine in code, judgment battery in Jev, calibration log of state-decision-outcome triples.

Evidence

Get the week's best AI launches, plus 3 ideas worth building

One email every Saturday. Ranked by traction, not hype. Free.