Session notices
One line when monitoring starts (how many stocks, until when), one when it ends (how many Strong / Normal / Weak, how many excluded), and a warning if the market feed drops. Each carries a link back to this dashboard.
Every F&O stock, checked on each 5-minute candle. When one does something worth a look, it shows up below — tap it to see exactly what happened and why.
Each NSE sector index against yesterday's close. Tap a row for its F&O stocks.
| Sector | Change | Trend | Advancing | Level |
|---|
Stocks trading past their daily, weekly, monthly, 3-month, 6-month, 52-week or all-time high or low. Only the longest period cleared is listed.
| Time | Stock | Breakout | Prior level | Price |
|---|
Every signal since Monday. Tap a row for the candles that produced it.
| Day | Time | Signal | Tier | Stock | Strategy | At | By |
|---|
Choose exactly which signals reach your phone. Detection never stops — this only decides what gets sent. The dashboard always shows everything.
Type / in the chat and Telegram lists these with descriptions — nothing to memorise. Buttons appear under most replies, so you rarely type at all.
Every signal message has the same five parts.
Real messages, rendered by the same code that sends them.
One line when monitoring starts (how many stocks, until when), one when it ends (how many Strong / Normal / Weak, how many excluded), and a warning if the market feed drops. Each carries a link back to this dashboard.
Every Friday after the session the bot sends the week's signals as a CSV and a PDF, whether or not you downloaded them here. Ask any time with /week.
The stock universe is rebuilt from the exchange's own file each morning. If NSE adds or drops a stock, you get a note naming what changed.
It never places an order, never sends Weak signals unless you ask, never messages during a replay, and never sends anything when two strategies conflict within 15 minutes — that stays on the dashboard, marked as a conflict.
Share the first one — it works on every network, including Jio. The direct Railway link is a backup for networks that aren't blocking it.
Mainpivotstrategy.pages.dev
— recommended, opens everywhereDirectpivotstrategy.up.railway.app
— backup; may not open on some ISPsSend /start once from the chat you want to control it from — that chat becomes the owner and no other chat can change your settings. Settings live on the server, so a change here shows up in the bot and the other way round, instantly. The bot reads nothing except the commands you send it.
Four patterns, checked on every 5-minute candle of every F&O stock. Here's each one in plain words — what sets it off, what doesn't, and the exact numbers behind it.
Used by two of the four patterns. Worth understanding first.
Take yesterday's highest price, lowest price and closing price, and average them. That number is the pivot — a rough "fair price" for today that a lot of traders watch. Above it, buyers are winning the day; below it, sellers are.
From the same three numbers we also work out levels above (R1, R2, R3 — where rising prices often stall) and levels below (S1, S2, S3 — where falling prices often find a floor). You'll see all seven in every signal's detail view.
"It couldn't stay down."
Up signal: the stock is above the pivot, slips below it, then takes it straight back. Three closes — A,B,A. That reclaim is the signal: the dip got bought.
Down signal: the exact mirror — B,A,B. It was below, popped above, then got pushed back below. Sellers won the argument.
It won't fire if: the stock never crosses the pivot at all · only one crossing happens (two closes is not a pattern) · the sequence doesn't complete before your cut-off time · it already signalled that way today (one each way, per stock, per day).
Every 5-minute close is stamped A if it finished at or above the pivot and B if it finished below. Wicks are ignored entirely — only the close counts. Repeats collapse, so A,A,B,B,A is read as A,B,A: what matters is how many times price actually crossed the line, not how long it loitered.
Counting starts wherever the stock happens to be — there is no anchor and no preference for either side. A stock that opens above the pivot is treated exactly like one that opens below it.
There is no noise zone. A close five paise past the pivot counts the same as one five rupees past it. We tested a volatility-scaled dead zone across 211 stocks over 12 months and it moved the accuracy by less than one percentage point — so it was removed rather than kept for the look of it. The same test retired the CPR-width, moving-average and "wait for a decisive break" filters.
"It punched through the ceiling."
Which way we watch depends on where it opens. Opens above the pivot → we only watch it break upward through R1, then R2, then R3. Opens below → we only watch it break downward through S1, S2, S3. A stock is never watched both ways at once.
Every level is its own signal. Clears R1 at 9:45 and R2 at 10:20? That's two signals, and the second one is treated as stronger — it's already proven itself once.
If it opens already above R1, we skip R1 and watch R2 next. No credit for ground it never had to fight for.
It won't fire if: the price only pokes through and falls back — the candle must close past the level · it opened above the pivot but collapses through supports (wrong side — we're not watching that) · it gapped above every resistance at the open (nothing left to break) · that same level already fired today.
"A sharp run, a breather, then off again."
Three parts. First a sharp run — one to four candles in the same direction, covering at least three times the stock's normal 5-minute move, and mostly in one push rather than a messy back-and-forth.
Then a breather — three to ten small candles, each less than half the size of the run's candles, drifting gently and giving back no more than half of the run. A narrowing, triangle-shaped pause counts too.
Then the signal: a candle that closes out of that pause, in the direction of the original run. The pause is over; the move is resuming.
It won't fire if: the pause drifts on past ten candles (that's not resting, that's stalling) · it gives back more than half the run · it breaks the wrong way · it breaks out before three calm candles have formed (too twitchy to trust).
"Three candles coiled, the fourth let go."
Three candles in a row: green, red, green — where the middle red one is small enough to sit inside the two green ones. Buyers pushed, sellers pushed back but couldn't get anywhere, buyers pushed again. Tension.
The signal is the very next candle: it has to close above the highest point of all three. That's the release.
The mirror image — red, green, red then a close below all three — is the down signal. The colours decide the direction; we never read a green-red-green group as a down signal.
It won't fire if: a candle opens and closes at the same price (no colour, so the sequence is broken) · the middle candle's body pokes outside the outer two (not a real sandwich) · the fourth candle only wicks past and closes back inside · the fourth candle does nothing (the setup just expires, and a later trio can form a fresh one).
Every signal is scored, so you know which deserve attention first.
It becomes Strong when: two different patterns fire on the same stock the same way at the same time · it broke a second or third level rather than the first · the flag's pause was tight (held two-thirds of the run) · the deciding close finished well clear of the pivot · the wider market was moving the same way.
It drops to Weak when: the wider market was pushing the other way · or there was an opposite signal on the same stock within 15 minutes.
That last one matters. If one pattern says up and another says down within a few minutes, we don't send you both and let you guess — we mark them both as conflicted and stay quiet. Fewer, clearer alerts.
By default Strong and Normal reach your phone. Weak stays on the board. You can change that any time from the Telegram bot's Filters menu.
A sandwich is often just the tail end of a flag. When both fire on the same candle in the same direction, we merge them into one signal rather than two — it's one thing happening, not two independent confirmations. Only a pivot pattern agreeing with a chart pattern counts as genuine agreement.
Only F&O stocks — the ones with futures and options — about 210 of them. The list is rebuilt from the exchange's own file every morning, so if NSE adds or drops a stock, we follow automatically and you get a note about it.
We watch 5-minute candles from 9:15 until whatever end time you set (11:30 by default; change it any time, for today or for good, and it takes effect immediately).
Nifty and Sensex don't produce signals — they're only there to tell us the market's mood, which feeds into the grading. Sector indexes are the same: context, not signals.
If a stock's data has a gap — a missing 5-minute candle — we stop watching it for the day and list it under “Skipped”. A pattern read from incomplete data is worse than no pattern at all.
This tool tells you where to look, not what to buy. It watches 210 stocks so you don't have to, and shows you exactly what happened so you can judge it yourself in ten seconds instead of ten minutes.
We tested these patterns over a full year of real data. They're good at finding activity worth your eyes. They are not a money printer, and anyone who tells you a 5-minute pattern is one is selling something. Always pull up the chart before you act.