| INSTRUMENT | LAST | POS % | LOWER ◂ RANGE ▸ UPPER | VOL | STATUS |
|---|
⚑ SIGNAL JOURNAL — every fired signal today
A permanent log of each BUY / SELL / TRAP as it fired (persists across restarts). Click any row to open its TradingView chart. Delivery shows whether the Telegram alert was sent.
◎ TRADE PLAN — entry · stop-loss · targets · exit
For every BUY/SELL the bot lays out the actionable levels. 1R = the opening-range height. Stop is the opposite band; targets are 1R and 2R; exit by 15:15 IST or trail after T1. Live status updates each cycle. Click a row for the chart.
The ◆ AI VERDICT fuses each setup's 6-gate conviction structure with the LEARN model's realised win-rate & expectancy for that setup's feature bucket — no LLM, grounded in your own trades. Decision-support, not autotrading.
⚙ OPTIONS STRATEGY — buy the direction
Each ORB signal becomes an option buy: a bullish BUY → ATM Call (CE), a bearish SELL → ATM Put (PE), nearest expiry. Entry = the live premium; target & stop are projected from the underlying's 1R/2R move via delta (≈0.5 ATM), giving a clean 1:2. Risk / reward / capital are shown in ₹ for the full lot. Click the option link ↗ to open that exact contract's chart on Groww.
The list is a running book of today's signals — a row's TIME is when the signal fired, and ENTRY is the premium at that moment. Only FRESH (<10m) rows are realistically enterable; older rows are there to track how the day's calls played out.
Estimates use a 0.5 ATM delta and ignore theta/IV; treat targets as guides, not guarantees. Not investment advice.
✔ END-OF-DAY REVIEW — how the day played out
Two books side by side: the direct equity/futures trades, and the options buying strategy. Both update live through the session and settle into the final result at close.
Every trading day's options result — high-conviction only (★ qualified + ◆ TAKE), in ₹. Options are intraday, so each day settles on its own and nothing carries forward; the CUM ₹ column is just the running account total. Today's row updates live and settles at close.
◈ TRADER AI — reads the tape, decides on the spot
A professional trader brain reads every symbol's 1-minute moves — market structure, VWAP, momentum, volume and ORB location — and makes a live TAKE / WAIT / SKIP call with written reasoning. Every TAKE runs in a paper account (no real orders) so we can prove the edge before risking a rupee.
★ HIGH PROBABILITY — where two edges agree
Separate research (your core strategy is untouched). This fires only when the validated A+ intraday signal lands on a stock with heavy institutional OI activity (top-25% open-interest change on NSE). In out-of-sample testing this subset held at ~53% win, +0.37R — roughly 3–4× the baseline A+ edge. Fewer signals, better odds. Not 100% — no honest edge is; this is the highest-probability slice we could validate.
⚡ AUTO-TRADE — automated option execution on Groww
Buys the ATM option of each ★ high-conviction signal, sizes by your risk settings, and manages a stop + target (OCO) on every position. DRY logs orders without sending; LIVE places real orders. Daily-loss kill-switch and a manual KILL below.
↺ 365-DAY BACKTEST — the same setup, replayed on history
Runs the identical live strategy — ORB breakout → stop / 1R / 2R, plus the Scale & Trail profit-booking — over the last year on every underlying, day by day. P&L R is the disciplined hold-to-close outcome (stopped at −1R, +2R if the 2R target tags first); BOOKED R is the Scale & Trail result. Download every trade as CSV. Not investment advice.
◆ LEARN — why trades hit SL & an adaptive edge model
Every resolved trade is scored by its setup features — volume, conviction, time of day, the instrument's game-theory edge, and direction — and the model learns the win-rate & expectancy (R) of each. The reward is the realised R, so it up-weights what pays and flags what bleeds. This is an online contextual-bandit (reinforcement-style) that accumulates across every session — not a deep-RL agent (those overfit thin market data).
Reward = trade R · online contextual-bandit · accumulates in learning.json across days · not deep-RL, not investment advice.
∑ MATHEMATICS — ORB STRATEGY BACKTEST
Is the 0.2711% edge real, or curve-fit? Every BUY/SELL is replayed over history; the trade is risked against the band width (1R) and exited at the day's close. Expectancy > 0 and profit factor > 1 mean positive edge; a Sharpe above ~0.1/trade with shallow drawdown is what you want.
◉ PHYSICS — ECONOPHYSICS OF EACH MARKET
Markets behave like physical systems. The Hurst exponent reveals the regime (mean-revert / random / trend), kurtosis exposes fat tails (crash risk the bell curve hides), vol-clustering shows volatility begets volatility, and the OU half-life is how fast price springs back to its mean.
⚄ GAME THEORY — MICROSTRUCTURE & STOP-HUNTS
Every breakout is a game against other players. When a band breaks and holds (conviction), momentum traders win; when it's a trap (wick sweep, close back inside), larger players ran retail stops and reversed. A high trap rate means this instrument's ORB breakouts are mostly liquidity grabs — fade them.
◐ PSYCHOLOGY — CROWD BEHAVIOUR & SENTIMENT
Price is a poll of human emotion. The Fear/Greed gauge reads today's breadth + momentum + volume across your board. Per instrument: RSI shows over-extension, reversion score measures whether the crowd overreacts to spikes, and streaks reveal herding.
⇄ SWING DESK — momentum, minus the blow-offs
Separate research (your core strategies untouched). Every evening we rank the entire liquid NSE cash market (~1,000 stocks) by 1-month momentum, then skip the parabolic tops (already up >40%) so you capture the move, not the peak. In testing, skipping blow-offs cut disaster trades from 12% → 4%. And holding ~3 weeks (letting winners run) roughly doubled expectancy — momentum is a weeks-not-days effect. Buy strong-but-not-vertical, hold ~3 weeks, cash/spot — no options decay. Not 100%, and momentum still suffers in market crashes — this is safer, not risk-free. Size accordingly.
This is a tracked model portfolio: buy a name once when it appears in BUY TODAY, hold it through HOLDING, and exit only when it moves to SELL TODAY. No churn — a stock you already own never reappears as a new buy.
Ranked by 20-day return; filtered to ₹5cr+ daily turnover, price ≥ ₹20. Target/stop are ±guides. Momentum favours mid/small-caps — real gap & drawdown risk; size accordingly. Not investment advice.
⚡ BTST SCANNER — strong close, overnight gap
Separate research (your core strategies untouched). Every day near 15:10 we rank the watched F&O universe by how strongly each stock is closing — price far above its VWAP, high in the day's range, on a late-session volume surge. That buying pressure tends to spill into the next morning's gap up. Buy at 15:10, sell at next open. Validated walk-forward (40 stocks, ~3 months): the strongest-5% closers returned +0.68% net of cost out-of-sample (83% up) — the one edge that survived every test. This is a momentum bet: it will have losing streaks when the market chops or tops. Not risk-free. Size accordingly.
Ranked by close-above-VWAP %. Tiers & expected gaps are from the backtest, net of ~0.25% BTST cost (brokerage+STT+slippage). Scores are only final from ~15:00 — before that the close isn't formed. Not investment advice.