Trader Brain

Read the market the way one trader does.

Replay any trading day. See the call Intraday Hunter's logic makes, what happened next, and the verified track record.

Run a read See the track record
Bank Nifty one-minute chart for 2026-08-04 up to 09:45 with the prior close, round number and his levels drawn
Bank Nifty, 2026-08-04 at 09:45. His read: lean short. By the close the index moved 0.109%.

The numbers before the pitch

57.3%his trades right at the close, 110 days
48 to 66%95% interval. It includes 50.
75.7%on trend days at +60 min, 37 days, p = 0.0026
149of his trades studied, 118 evening plans, 17 explicit rules

His actual trades land 57.3% at open to close; the 95% interval includes 50%. Direction is not a proven edge. The one slice that holds up: trend days. Every read you run here is logged and scored against real bars, so the record keeps growing with use: 10 reads logged so far.

What a read looks like

Bank Nifty · 2026-08-04 · 09:45 IST grounded
Biaslean short confidence 0.496 (from his history, not the model)
Who is trappedYesterday's buyers. After a 2-day up-run they carried longs into a close at 58247.95 and above the 58000 round number, where buyers park. Today's small gap-down has broken the prior close and is holding 404 pts below it, and price is also under 58000. Those longs are underwater and under pressure. Anyone who bought the current 5-bar bounce off 57651 is a fresh weak hand: if 58000 rejects them, they become the next layer of exits. I see no big trapped seller crowd above who could force a short-covering rally.
Stop logicThe line is the prior close at 58247.95. Price is below it now (broke_down_hold), and the short stays valid only while that holds. A reclaim of 58247.95 means the market has disproved the trapped-buyer thesis, so I cut and don't argue. An earlier warning sign is a strong reclaim and hold above the 58000 round number and my 58100 resistance; if that happens I trim or exit rather than wait for the full prior-close hit. Never add to the short if it moves against me.
ReasoningThis is the flat-to-gap-down branch. Yesterday's buyers, sitting near the 58247.95 close after a two-day up-run, are now exposed, and price has broken below the prior close and is holding there. That is the condition I need for trapped buyers to start exiting. 58000 is the round number where buyers had parked, and we are trading under it, so those parked buyers are the fuel if price fails to reclaim it. This bounce from 57651 looks like retail asking 'market upar aa raha hai, hum bhi try karte hain'. If 57920–58000 rejects it, those late buyers get added to the trapped pile, similar to 6zppfK_n4yg where a slight gap-down left buyers above under pressure. The SL is the prior close; if Bank Nifty reclaims it, I cut, because we can't chase the market saying only we are right.
What happened h30: flat -0.021% (wrong)  h60: down -0.171% (right)  close: up 0.109% (wrong)  

Levels on the chart come only from numbers in the data: prior close, day high and low, the opening range, VWAP on the future, round numbers and the levels he named the evening before. The model narrates inside those numbers; it cannot invent one.

How it works

  1. Replay. Pick an instrument, a past trading day and a time. The engine rebuilds the market exactly as it looked then, from one-minute bars, with nothing from later in the day.
  2. Read. His explicit rulebook fires on the state, three of his most similar past trades are retrieved, and a language model writes the read in his framing. Hard rules are enforced after the model, not just suggested to it.
  3. Score. The read is resolved against what the index did 30 minutes, 60 minutes and a session later. Wins and losses are counted the same way for everyone.
  4. Log. Every read is stored with its full state and shown on the track record, split into reads he has evidence for (mornings) and reads that extrapolate his logic (afternoons).

What this is not

It is not a signal service and does not tell you what to buy. The best-documented profile here is statistically indistinguishable from a coin flip on direction over the full sample. What it does offer is a disciplined way to see a market through one experienced trader's eyes, with every claim tied to a number you can check. Sizing is shown as an illustration of what a read implies for a given amount of risk, with every assumption visible.