What is gap-up and gap-down in trading?

During trading hours, price moves one transaction at a time. To travel from one level to another it has to trade through every price in between, because every tick on the chart is a real trade: a buyer and a seller who met at that number and agreed. That continuity is not a drawing convention. It is a consequence of how an order book works, and it is why an intraday line has no holes in it. A gap is the single exception, the one case where the line jumps and leaves a hole behind.

The hole is a band of prices at which nothing happened. It is worth being precise here, because the loose definition and the strict one differ and the difference matters later. The number quoted in the press and on your screen is the simple distance from the prior close to the new open. The band that was actually skipped runs from the prior session's high to the new session's low on a gap-up, and the other way round on a gap-down. Inside that band not one share changed hands: no buyer met a seller at any of those prices, so there is literally nothing to plot. A gap-up is that empty band sitting above the prior session; a gap-down is the same thing below it.

The width of the band is a rough gauge of how far the balance between buyers and sellers moved while the exchange was shut. It is the fossil record of a repricing that happened with no trading to carry price there. The clearest way to see this is to hold everything constant and change one thing: not the news, not the price action before or after, only whether the doors were open while the news was being priced in.

Read the right-hand panel again, because the argument lives in what is missing from it. Eight price levels are drawn between the prior high and the new open. On the left, every one of them carries a print, because a market that is open has no choice: to get price from 1,272 to 1,300 it must find a willing counterparty at 1,276, then at 1,280, then at every level after. On the right, the same eight levels are drawn and every one is bare. Nothing about the information differed. The exchange was simply not open to express it, so when it reopened, it reopened past them. That is the whole definition, and everything else on this page follows from it.

Indian markets trade only during set hours, but the world does not keep those hours. Company results are declared after the close, precisely so the market has time to digest them. Global indices move through the night, currencies and commodities shift, corporate actions are announced through exchange filings, and macro data such as rate decisions and inflation prints land before the open. None of it can be priced in through trading while the exchange is shut. So it queues. Buy and sell interest builds up as intent, waiting for the market to reopen. This is the fact the whole subject rests on: the information does not wait, only the trading does.

On the NSE that reopening is not a free-for-all. It runs through a pre-open call auction. Orders are collected from 9:00 to 9:08, then matched from 9:08 to 9:15. The system computes one equilibrium price: the price at which the maximum quantity can be executed. When more than one price would execute that same maximum volume, the tie is broken first by the least unmatched quantity, and only then by the price closest to the prior close. That equilibrium becomes the day's opening price, for the whole market, at once.

Now read that rule in the order it is actually written, because the answer is hiding in the sequence. The auction's first objective is not to stay near yesterday. Proximity to the prior close is the third criterion, a tie-breaker invoked only when two candidate prices are otherwise identical, which is to say almost never. The primary objective is volume: find the single price at which the most shares can change hands. If overnight orders have piled up on the buy side, the price that clears the most volume moves, and the market opens there, because that is what the rule instructs it to do. The chart below draws that decision as the auction actually makes it.

The shape of that chart is the reason a gap is not a malfunction. A market that opened at yesterday's close every morning would be a market that refused to trade: at 1,268 the buy book is enormous and the sell book is nearly empty, so barely a quarter of the available quantity would find a counterparty and the rest of the book would sit there unfilled. The auction's job is to clear, and clearing means finding the price where supply and demand actually meet. Overnight news moved that price. The gap is simply the distance the meeting point travelled while nobody was allowed to trade.

That last row deserves its coral flag, because it is the one case where the gap is not information at all. When the book is thin, the maximum-volume price can be dragged a long way by a quantity that would not move a liquid name at all, and the resulting gap reflects the state of the order book rather than the state of the business. Index gaps on Nifty and Bank Nifty almost never work this way, because a broad index dilutes any single participant across hundreds of constituents; they trace overwhelmingly to global cues and macro data. A single company gaps most violently around its own results, where the surprise is concentrated in one name and nothing dilutes it.

The classic classification comes from the technical-analysis literature of Edwards and Magee, and its one genuinely useful move is to sort gaps by where in a trend they appear rather than by how they look. This is the right instinct, because on their own they all look the same. A gap is a jump. The jump carries no information about itself; the context around it does.

There are four. A common or area gap forms inside a quiet, sideways range and usually means very little: it is noise, typically on light volume, and it tends to fill quickly. A breakaway gap launches price out of a range or a completed pattern, often on the heaviest volume the chart has seen in weeks, and marks the start of a new move. A continuation gap, also called a runaway or measuring gap, appears partway through an already-running trend as it accelerates, with participation holding up. An exhaustion gap comes near the end of an extended move on a final, climactic surge, and can precede a reversal rather than more of the same.

The volume lane underneath is doing real work, and it is worth saying why. Position in the trend is only visible once you know where the trend ended, which is to say afterwards. Volume is the one piece of evidence available while the gap is still forming. A gap out of a range on no participation is a shrug; the same gap on the heaviest volume in months is a crowd arriving. That is genuinely useful, and it is also genuinely insufficient, because the heaviest volume of all often belongs to the exhaustion gap, where the crowd is not arriving but finishing.

Set the third and fourth rows side by side and the problem is plain. A runaway gap and an exhaustion gap are the same picture: a jump partway up a trend, on strong volume, in a market that has been rising. Every feature that would separate them lies in the future. Traders resolve this discomfort by adding qualifiers, and the qualifiers are usually just the reversal described in advance. The honest reading is that the taxonomy is a vocabulary for describing structure after the fact, and a poor instrument for deciding anything at nine fifteen in the morning.

The single most repeated claim about gaps is that they always fill, meaning price eventually returns to the prior close and covers the empty band. It is repeated because it is memorable, and it survives because it is unfalsifiable in casual use: given enough time, price returns to most levels, and if it has not yet, the claim has simply not come true yet. As a mechanism it is empty. Nothing in the order book owes the chart a visit to a level it already rejected.

What actually drives the two outcomes is the question the gap itself cannot answer: did anything change? If the overnight move was an overreaction, or a thin book dragged the auction somewhere the business does not justify, then the new price has no support underneath it and buyers who wanted the old level are still there, so price works back and the band closes. If the move was a genuine repricing on real information, then the old level is a price nobody wants any more, and there is nothing to pull price back to it. Same picture on the chart, opposite mechanism underneath.

A fill is a possibility, not an entitlement. Trading it as a rule is not a strategy. It is a bet that nothing important happened, placed without checking.

Here is the consequence that turns a chart curiosity into the single most important reason an overnight position carries extra risk. A stop-loss is a trigger, not a price. That sentence sounds like a technicality and it is the whole thing. A stop is a condition held at the exchange: when the last traded price crosses your level, the order is released. What it is released as, and what the market happens to be doing at that instant, decides your actual exit. You never bought a price. You bought a condition and a release.

A stop-loss market order becomes a plain market order the moment it triggers, and a market order fills at the first price available. Intraday, with price trading continuously, the first available price is usually within a tick or two of your level, which is why the illusion holds up for months at a time: the stop appears to be a price, because continuous trading keeps supplying one. A gap is exactly the case where that supply stops. If a stock you hold gaps clean past your stop while the market was closed, the first available price is the gapped open, and that is where you exit.

Nothing malfunctioned in that chart. The trigger fired correctly, the order was released correctly, and it filled at the best price on offer, exactly as specified. The stop enforced an attempt to exit. What it could never do was hold a price, because there was no trading at your price to hold: your stop level spent the whole night sitting inside a band where no buyer and no seller existed. A stop is not a floor under your position. It is an instruction to leave, and the market decides the fare.

A stop-loss limit order does not rescue you either; it fails the other way, and the symmetry is worth understanding because people reach for it as the fix. A stop-limit caps how bad your fill can be. But if the market gaps clean through your limit and never trades back to it, the order does not fill at all, and you are left still holding the position with no protection and a loss that keeps running. Capping the price can cost you the exit entirely. Between the two, the market order gives you a certain exit at an uncertain price; the limit order gives you a certain price and an uncertain exit. Across a gap there is no third option that gives you both, because both were being manufactured by the trading that is no longer happening.

This is why overnight and gap risk is described as uncontrollable by design rather than by misfortune. You cannot place, move, cancel or trigger an order while the market is shut. A position carried through the close is exposed to whatever the auction decides the next morning, with no opportunity to act in between and no instrument that changes that. It is not a matter of a better stop or a faster platform. It is a property of the closure itself, and it applies to every participant equally.

Not every gap on a chart carries information, and confusing the two is a genuine trap rather than a pedantic one. When a stock goes ex-dividend, ex-bonus or ex-split, its opening price is mechanically reduced by a set amount to reflect the corporate action. The exchange adjusts positions so that value is preserved, not destroyed. Nothing was repriced, nothing was decided, and no opinion changed. The arithmetic simply moved.

Take a dividend. A stock trading near ₹335 that declares a ₹15 dividend opens on the ex-date near ₹320. On the chart that is a ₹15 gap-down, and it looks exactly like a piece of bad news. The holder is no poorer: the ₹15 is leaving the company as cash, so the share is worth ₹15 less and the shareholder receives the ₹15 separately. A bonus or split works the same way in reverse arithmetic: the share count changes and the price is scaled to match, with the total value of the holding untouched. NSE Clearing states the principle directly, that adjustments are set so the value of a participant's position on the cum and ex dates remains the same as far as possible. All figures here are illustrative.

The fifth row is the one that costs real money, and it is the reason this section is not trivia. Your stop cannot tell the difference between the two columns. It sees a price below its level and it does what it was told. An ex-date adjustment can therefore trigger a stop and close a position you had every intention of keeping, on a fall that never happened to anybody, and you will find the loss on your contract note all the same. The defence is not clever: it is knowing the ex-dates of what you hold, and understanding that a chart which has not been adjusted for corporate actions is showing you a jump that never reflected any change in what the asset was worth.

Put the last two sections together and the picture is uncomfortable but clean. A stop does not cap a loss by decree; it caps it by finding you a counterparty at your level, and counterparties only exist while the market is trading. So the protection you thought you owned is not a property of the order. It is a property of continuous trading, and it is switched off every evening at half past three along with everything else.

That has one blunt consequence for how a position is carried. You cannot flatten the coral line: no order type bends it, no platform is fast enough to beat it, and no amount of attention helps, because the exposure exists precisely during the hours you are not allowed to act. Every lever you might reach for is on the wrong side of the close. Every lever except one: the scale of the vertical axis. Position size survives the close for the same reason nothing else does, that it was settled before the doors shut. A 6 percent gap against a position sized so that 6 percent is survivable is a bad morning. The same gap against a position sized on the assumption that the stop would hold is something else entirely, and the difference was decided the previous afternoon.

This is also why leverage and overnight risk are such a poor combination, and why the Indian retail context makes the point sharply. According to SEBI, about 93% of individual traders in equity derivatives made net losses over FY22 to FY24, aggregate net losses exceeding ₹1.8 lakh crore (SEBI, September 2024). Leverage does not create gaps and it does not widen them. What it does is multiply the vertical axis of that chart, so that a gap which would have been a bad morning on an unleveraged position becomes an account event on a leveraged one, with the stop that was supposed to prevent it sitting uselessly inside the untraded band the whole time.

A gap is one of the most information-dense events on a chart precisely because it is a repricing compressed into a single point. That density is why it attracts slogans, and it is also why the slogans fail: a single print does not contain the reason it happened. Reading a gap well is not pattern recognition. It is a short sequence of questions asked in the right order, none of which the jump itself answers.

What a gap is not is a signal you can act on mechanically. The volatile open is where the initial reaction and its reversal whip against each other, and it is where a hasty entry is punished hardest. The discipline that matters is upstream of all of it: understanding the mechanism, sizing so that a single adverse gap cannot do outsized damage, and knowing that a position carried through the close is exposed to a move no order type can catch. That upstream judgement, reading context and deciding in advance rather than reacting to a jump, is exactly what the method we teach is built around.

The gap is not the risk. The closed door is the risk. The gap is only the receipt it hands you in the morning.

Take that away and the rest is manageable. Continue to the companion strategy guide for the fill-versus-continuation decision and the risk of trading gaps deliberately, or to circuit limits for the case where a gap does not merely hurt you but locks you in.

• NSE pre-open call auction. The equilibrium price is the price at which the maximum quantity is executable, with ties broken by the least unmatched quantity and then by proximity to the prior close; order entry runs 9:00 to 9:08 and matching 9:08 to 9:15. This ordering is the mechanism that sets a gapped open, and the reason yesterday's close has no first claim on it. nseindia.com
• NSE price bands and circuit filters. Daily price bands are set in categories such as 2, 5, 10 and 20 percent of the prior close, based on liquidity and volatility, beyond which a stock cannot trade that day. This is why a strong enough gap can lock a stock at its band and leave a holder with no exit at all. nseindia.com
• NSE Clearing, corporate-action adjustments. Adjustments for dividends, bonuses and splits are set so the value of a participant's position on the cum and ex dates remains the same as far as possible, which establishes that an ex-date gap is a book-keeping step rather than a repricing. nseclearing.in
• Indian retail derivatives outcomes. The Securities and Exchange Board of India study of individual traders in the equity derivatives segment reports that about 93% of individual traders in equity derivatives made net losses over FY22 to FY24, aggregate net losses exceeding ₹1.8 lakh crore (SEBI, September 2024). It is cited here only as context for why leverage and overnight exposure combine badly. sebi.gov.in
• Gap classification, Edwards and Magee. The common (area), breakaway, continuation (runaway) and exhaustion taxonomy comes from the classical technical-analysis literature, which is also the source of the caveat repeated throughout this page: the labels are confirmed by subsequent price action rather than by the gap itself.