A retail investor sees a stock close up 8 percent on heavy volume, opens the broker breakdown, and concludes that “big money is buying.” Then the stock drifts down for three weeks. What went wrong was not the price. It was the reading. The NEPSE floorsheet, the public record of every executed trade on the Nepal Stock Exchange, is the most underused free dataset available to ordinary investors in Nepal. It is also one of the most misread.
The floorsheet tells you exactly what happened in a trading session: which broker bought, which broker sold, how many shares, and at what rate. That is genuinely useful. The problem is that most retail traders treat it like a crystal ball, when it is closer to a receipt. This guide does two things. First, it explains what every column in a NEPSE floorsheet actually means, in plain terms. Then it covers the harder part, which is the analytical edge: separating real accumulation from shares being shuffled in a circle, reading large single transactions correctly, and understanding why broker numbers mislead more often than they reveal.
What a floorsheet is, and where to find it
After each trading session, NEPSE publishes a complete log of all trades that were matched and executed that day. This is the floorsheet. It is raw and granular. Every single matched order becomes one row.
You can pull it straight from the source on the official Nepal Stock Exchange site at nepalstock.com (NEPSE). Most retail investors, though, read it through mirror sites that make it searchable and sortable: ShareSansar at sharesansar.com/floorsheet and Merolagani both republish the same NEPSE data with filters by company, by broker, and by date. The underlying numbers are identical because they all come from NEPSE. The mirrors just make it easier to slice.
One thing worth saying upfront. The floorsheet is end-of-session data on the official site, not a live tape, although third-party platforms show a “live” running version through the day. For learning to read demand, the full daily floorsheet is what matters.
The columns, one by one
A NEPSE floorsheet row carries a small number of fields. Learn these and you can read any of them.
Contract number. A unique ID for each matched trade. NEPSE assigns it sequentially. It is mostly an identifier for reference and ordering. Beginners can largely ignore it, but it does tell you the sequence in which trades executed, which occasionally matters.
Stock symbol. The traded scrip, for example NABIL, NTC, or UPPER. On the full floorsheet this is how you filter to one company.
Buyer broker. A number, not a name. This is the broker firm that executed the buy side of the trade. NEPSE identifies brokers by code, so you will see something like “Buyer 58” rather than the firm’s name. More on why this matters below.
Seller broker. Same idea, on the sell side. The broker code of the firm that sold.
Quantity. The number of shares (technically “kitta”) in that single matched trade.
Rate. The price per share at which the trade executed, in NPR.
Amount. Quantity multiplied by rate. The rupee value of that one trade. If a row shows 500 shares at NPR 1,200, the amount is NPR 6 lakh.
Most published floorsheets also carry the business date and, on the live versions, a trade time so you can see the order in which trades hit during the session.
That is the whole structure. Six or seven fields. The data is simple. The interpretation is where almost everyone slips.
Reading volume and price together (the basic literacy)
Start with the obvious and get it right before going further. If you filter the floorsheet to one company, you are looking at every trade in that scrip for the day. Sum the quantity column and you have the day’s total traded volume. Sum the amount column and you have the day’s turnover in rupees.
Now do something most people skip. Look at the rate column across the session and ask a simple question: were the larger trades happening at higher rates or lower rates? Volume on its own says nothing about direction. A stock can trade 2 lakh shares in a day where most of that volume executed as the price was falling, which means sellers were hitting bids, not buyers chasing. The floorsheet lets you see this, because each row pairs a quantity with the rate it cleared at.
This is the first real edge, and it costs nothing. Heavy volume with most large trades clustering at the higher end of the day’s range looks like buyers paying up. Heavy volume with the big trades printing at the low end looks like distribution. Neither is a guarantee. But it is a far better read than “high volume equals bullish,” which is what most of the market assumes.
If you are still building the basics of how trades, accounts, and settlement fit together, our guide on how to open a Demat and Meroshare account in Nepal covers the plumbing behind every one of these rows.
The hard part: real demand versus shares going in circles
Here is the trap. You see a stock with very high volume, you see the same two or three buyer broker codes appearing again and again, and you conclude that strong hands are accumulating. Sometimes that is exactly right. Often it is not.
The thing the floorsheet cannot tell you directly is whether a buyer and a seller are actually different people. A broker code is a firm, not a person. Behind Buyer 58 might be hundreds of separate retail clients. Behind Seller 58 on the same day, also hundreds. So when you see broker 58 buying heavily and broker 45 selling heavily, you genuinely do not know if that is real demand meeting real supply, or a small coordinated group moving shares between accounts to manufacture the appearance of activity.
This second pattern has a name. Circular trading, sometimes called wash trading, is when shares are bought and sold repeatedly within a closed group to create artificial volume and a rising price, with no real change in beneficial ownership. SEBON and market commentators have flagged it as a recurring problem in NEPSE, enabled by people opening multiple Demat and TMS (trading) accounts under family members and associates and shuffling shares between them (Securities Board of Nepal / SEBON, reform commentary). The point for you is mechanical: circular trading produces exactly the floorsheet signature that naive readers misinterpret as accumulation. High volume. Rising price. A handful of recurring broker codes.
So how do you tell the difference? You cannot do it with certainty from public data, and anyone who claims they can is selling something. But you can stack the odds.
Genuine accumulation tends to leave a messier, broader footprint. Many different buyer broker codes, varied trade sizes, buying that absorbs supply at rising prices over several sessions, and crucially, the price holding when the volume fades. Manufactured volume tends to be narrow: the same small set of codes appearing on both the buy and sell side across days, big jumps in price on volume that has no obvious news behind it, and a price that sags the moment the activity stops. If a stock spikes hard on volume and then cannot hold its level once the floorsheet goes quiet, that is your tell. Real buyers do not evaporate overnight.
My view, stated plainly: if you are a beginner, treat any single-day broker concentration as a question, never an answer. The floorsheet flags something worth investigating. It does not confirm anything.
Large single transactions, and what they really mean
A single row showing, say, 50,000 shares at NPR 900 (an amount near NPR 4.5 crore) jumps off the page. Big block. Must be institutional conviction. Maybe.
Large single trades on the floorsheet usually mean one of a few things, and they are not equally bullish. It could be a genuine institutional or large investor taking a position. It could be a negotiated transfer that was arranged off-market in substance and simply matched on the system, in which case it tells you nothing about live demand. It could be one large holder exiting to one large buyer, which is a transfer of ownership at an agreed price, not a signal of broad market appetite. The floorsheet shows the trade. It does not show the intent.
The useful habit is to ask who needed it more. A large block printing at the top of the day’s range, where the buyer clearly paid up to get filled, leans bullish. A large block at the low end, where a seller dumped size into whatever bid existed, leans bearish. A block that prints right at the previous close with no struggle in either direction is most likely a pre-arranged transfer and is close to information-free for your purposes. Same row, three very different readings, all from looking at the rate relative to the rest of the session.
Why broker numbers mislead
This deserves its own warning because it is the single most common mistake. NEPSE has dozens of licensed brokers, in the order of 90 firms (NEPSE / SEBON). Retail traders memorise a few broker codes and build entire theories on them: “broker 34 is the smart money,” “when broker 49 buys, follow it.”
This is mostly noise, for a simple reason already stated. A broker code aggregates every client of that firm. A large, popular brokerage will show up constantly on both sides of the market simply because it has the most clients, not because it has a market view. ShareSansar has reported that some brokers carry far more online clients than others, with one broker reported as having the highest number (ShareSansar). A high-client broker dominating the floorsheet is a function of size, not signal.
There is also no public mapping from a broker code to who is actually trading through it on a given day. The firm is not the investor. Building conviction on “which broker” without knowing “which client” is reading tea leaves. The broker columns are useful for one honest thing: spotting unusual concentration that is worth a second look. They are not a leaderboard of smart money.
A worked example
Suppose you filtered the floorsheet for a mid-cap on a day it closed up 7 percent. Total volume looks heavy, around 1.5 lakh shares, turnover roughly NPR 1.8 crore. You feel bullish. Now read properly.
You sort by amount and look at the ten largest trades. Eight of them printed in the bottom third of the day’s price range, early in the session. The price closed high mainly because volume thinned out in the afternoon and a few small trades nudged it up on no size. Buyer codes are spread across many firms, but the big early prints came from two codes that also appear as large sellers the same day. That combination, big size clearing low, recurring codes on both sides, a close that floated up on thin afternoon trade, is not the picture of strong accumulation. It is closer to the opposite. The headline (“up 7 percent on heavy volume”) and the floorsheet tell two different stories. The floorsheet is the one to trust.
The verdict: a receipt, not a forecast
The floorsheet is the best free window into NEPSE that most retail investors never properly open. Learn the columns, and you can see, with real precision, what actually happened in a session: who paid up, who dumped, where the size cleared, and whether the close was earned or floated up on air. That is a genuine edge, and almost nobody uses it well.
But here is the position I will defend. The floorsheet shows you what happened, not what will happen. Over-reading it is its own trap. Every pattern you find is a hypothesis about behavior that already occurred. The moment you start treating recurring broker codes as a buy signal, or a single big block as institutional endorsement, you have stopped analyzing and started storytelling. Combine the floorsheet with the things it cannot see, the company’s fundamentals, the sector, the NRB monetary policy backdrop that moves the whole market, and for banks the core ratios like NPL, CD and CAR that actually drive value.
Read the floorsheet to understand the day. Do not ask it to predict tomorrow. Used that way, it is one of the sharpest tools you have. Used the other way, it will lead you confidently in the wrong direction.
This is analysis, not financial advice.
