Register
Tutorial

Pause or Stop Automation in QuantPower

This tutorial covers every way to slow down or stop automation in QuantPower: pausing one algo's new entries, stopping it with or without squaring off, pausing or disarming signals and webhooks, acting on several algos at once, and the account-wide controls for an emergency: the kill switch, Exit all and Smart Exit. It also covers what each choice does to your open positions and to your tokens, because the right choice depends on both. Reading it takes fifteen minutes; doing any one of the steps takes a few seconds.

QuantPower Research · Updated 15 Nov 2026 · 17 min read

QuantPower algos are fully automated once switched on: they enter, watch and exit by their rules with no further tap. That is the point of an algo, and it is also why you need to know, before you need it, how to stop one. Every stop control is on the algo's own card or on the M2M tab, and none of them is hidden behind a menu.

Before you start

  • Know where your algos are. Open Algos. It has four tabs: Option bots (your algos), Signal algos, Webhooks and M2M (Smart Exit, the daily loss limit and the kill switch).
  • Set up the second factor. The kill switch, Exit all and Smart Exit ask for it, and you do not want to be setting it up during a fast market.
  • Decide your rules in advance. "If the market gaps more than 2%, I pause everything until 10 AM" is a rule. "I will decide on the day" is not.

The choices in one table

ChoiceNew entriesOpen positionsWho watches themWhere
PauseStoppedKept, with their targets, stops and exit timeThe algoThe algo's card
Stop algo and square off its positionsStoppedClosedNobody: they are closedThe algo's card
Stop algo, keep positionsStoppedKeptYou: the algo no longer watches themThe algo's card
Switch off (algo with no positions)StoppedNoneNobodyThe algo's card
Kill switchStopped in the chosen segmentsKept; open orders cancelledYouAlgos → M2M
Exit allAlgos stopped firstClosedNobody: they are closedPortfolio → Positions
Smart ExitAlgos in its segments stopped when it firesClosed when it firesNobody: they are closedAlgos → M2M

Read the table across a row before you act, because two choices that look alike can leave you in very different places. The most important column is the fourth. After "Stop algo, keep positions" or the kill switch, the positions are yours: no stop-loss and no exit time watch them unless you add one.

A short history of the off switch

Every automated trading system has needed a way to stop. Exchanges added market-wide circuit breakers after the crash of October 1987, and kill switches became a regulatory expectation for automated trading after the "flash crash" of May 2010 and the Knight Capital loss of August 2012, when a faulty system sent millions of orders in under an hour.1 In India, SEBI's 2012 guidelines on algorithmic trading asked brokers to put risk controls, including the ability to stop algo orders, in front of every algo.2

For a retail trader, the lesson is smaller in scale but the same in kind: an algo that keeps trading after its rules no longer fit the market can lose quickly, and the time to learn the off switch is before it is needed.

Step 1: Pause one algo

Open Algos → Option bots and find the algo's card. Tap Pause.

A paused algo:

  • Opens no new entries.
  • Keeps watching the legs it already holds: their targets, their stop-losses and the exit time still apply.
  • Shows "Paused" on its card, with a Resume new entries button.

Pause is the gentlest choice. Use it when you want an algo to finish what it has started but not to start anything new: before a news event, after a gap you do not like, or when you want to watch the market for a while before the algo acts again. It costs nothing in tokens, changes none of the algo's settings, and can be undone with one tap, which makes it the right first move whenever you are unsure what to do.

Step 2: Resume

Tap Resume new entries on the paused card. The algo goes back to entering by its rules, from the next entry its rules allow. Read the card once more before you resume, so you know what it will do next. If its entry time for today has already passed, it waits for its next entry day; resuming does not make it enter late.

Step 3: Stop one algo

Two stop choices are on the card, and they do different things:

Stop algo and square off its positions. The algo closes its open legs with market orders, then stops. Sold legs are bought back before bought legs are sold, so a hedge is never removed while the short option is still open. Use this when you want the strategy finished for the day.

Stop algo, keep positions. The algo stops at once and leaves its positions in your account. It no longer watches them: there is no stop-loss and no exit time on them from that moment. Use this when you want to take over the positions by hand, for example to close one leg and keep the other. Add your own stop straight away; the tutorial on managing open positions shows how.

Step 4: Restart today, or switch on again later

A stopped algo can be run again:

  • Restart today runs it again today, in the same place it ran before. An algo that ran in Forward test restarts in Forward test; one that ran live restarts at the same broker. It never moves from test money to real money, or from one broker to another, on its own.
  • Switch on again on another day works as it did the first time, with all the switch-on checks.

If the broker the algo used has since been disconnected, a restart is refused and says so. Connect the same broker again and the restart works.

Step 5: Act on several algos at once

Select several algo cards, then choose the action for all of them together: pause, stop and square off, or stop and keep positions. The same choice then applies to every algo you selected, each in its own place. This is the quickest way to pause every strategy before an event without touching each card.

Step 6: Pause signals and webhooks

Signals and webhooks can also start trades, so pausing automation means checking them too.

Signal algos. A signal in one-touch mode only notifies you; the order waits for your touch. A signal in auto mode places the order itself once armed. To stop it, disarm it or switch it off on its card in Algos → Signal algos. Any change to an armed signal's condition or order disarms it too, so you must arm it again on purpose.

Webhooks. A webhook is your own private URL that a charting tool or scanner calls. Each webhook has a pause switch in Algos → Webhooks. A paused webhook ignores its hits; the hit log shows that they arrived and were ignored. Pause it when the tool that calls it is sending alerts you do not want acted on.

A signal or webhook can also start or stop an algo. If you have stopped an algo by hand, check that no signal or webhook is set to start it again.

Step 7: The emergency controls

When something is clearly wrong and you want everything to stop now, use the account-wide controls in Algos → M2M or on the Positions screen.

Kill switch. Choose the segments and confirm with your second factor. QuantPower stops trading in those segments until each segment's next session, stops every running algo in them and cancels their open orders. It places no order and does not close positions. Use it when you want nothing new to happen while you look.

Exit all. On the Positions screen. It stops the running algos first, then squares off every position, or every position in one segment. It asks for your tap and your second factor. Use it when you want to be flat.

Smart Exit. Set in advance, not in the moment: it closes the day automatically when the day's MTM reaches a target, a stop, a trailing step or a locked profit. The MTM stop loss tutorial sets it up.

What each choice does to tokens

SituationTokens
Pause and resume inside the paid 28 daysNothing more
Stop and restart inside the paid 28 daysNothing more. A restart is never charged
Stop early in the 28 daysNo refund for the days left
An algo still on at the end of the 28 daysCharged the full price for the next 28 days
Switch off before the cycle endsNot charged for the next cycle

A live algo costs 200 tokens per 28 days and a Forward test algo 100, charged once per cycle. If you plan to stop running an algo, switch it off before the end of the cycle so it is not charged again.

When an algo stops by itself

Not every stop is yours. An algo also stops, or holds back, on its own when something goes wrong, and it tells you why on its card. Knowing these cases helps you read the card calmly:

What happensWhat the algo does
An entry order is refused (margin, a risk rule, the broker)Closes the legs already in, stops, and shows the reason
An exit order is refusedTries again, up to 8 times, with a growing gap. The card shows what is still owed
Prices or Greeks it needs are missingWaits and says why. It never picks a strike, target or stop on a guess
The broker is not answeringOrders are refused, not assumed. Nothing is treated as "no position"
The broker was disconnectedIt cannot place orders there. A restart is refused until the broker is connected again
A fixed contract expiredStops, and names the contract
Tokens run out at the end of the 28 daysStops, and tells you. Open positions are kept and no longer watched

In each case, read the card first, fix the cause, and then decide whether to restart today or wait for the next session.

Changing an algo safely

Pausing is often the first step towards changing an algo. The safe sequence:

  1. Pause or stop. An algo can be edited only when it is off. If you only paused it, switch it off.
  2. Square off if you are changing the legs. While an algo that is off still holds positions, its legs, product and quantity cannot be changed. Its targets, stop-losses and times can.
  3. Edit. Change one thing at a time, so you know which change made the difference.
  4. Backtest the changed version before it trades. A rule backtest costs 2 tokens, and the first 10 each week are free.
  5. Run it in Forward test if the change is large: a new entry rule, a new strike rule, more legs.
  6. Switch it on. All the switch-on checks run again.

An edit and a switch-on at the same moment can never mix: if they clash, one of them is told to try again. If the algo was edited after your screen loaded, the switch-on asks you to check it again.

Pausing in Forward test

Algos running in Forward test pause, stop and restart in exactly the same way as live ones. Two differences:

  • Nothing reaches a broker. Stopping and squaring off a Forward test algo closes its test positions in the Forward test account only.
  • The kill switch, Exit all and Smart Exit cover your real brokers. The Forward test account has its own separate book and never mixes with real money. "Reset" on the Forward test account clears it and stops the Forward test algos; live algos are not touched.

Practise the pause and stop controls in Forward test first. Knowing exactly what "Stop algo, keep positions" leaves behind is easier to learn with test money.

A checklist before a known event

Budget day, an election result, a central bank decision or a large company's results can move the market further and faster than a strategy was tested for. A checklist the evening before:

  1. List the algos that would enter during the event window.
  2. Decide for each: pause it for the day, stop and square off before the event, or leave it running with a tighter account limit.
  3. Check every signal in auto mode and every webhook that could start a trade during the event.
  4. Set or lower Smart Exit's stop-loss for the day, and check the daily loss limit.
  5. Make sure the second factor works, so the kill switch is one step away.
  6. After the event, resume algos one at a time, reading each card first.

Pauses that belong in the rules

If you pause an algo by hand on the same kind of day again and again, the pause belongs in the algo's own rules, where it happens every time without you:

  • Days. An algo runs on every trading day or only on the weekdays you pick. If you never want a strategy on expiry day, leave that weekday out. Market holidays are skipped automatically.
  • Entry and exit times. An entry time of 9:45 instead of 9:20 keeps the algo out of the opening minutes on every day, not only the ones you remember. The exit time is never later than the market's own square-off time.
  • Wait and trade, or opening range breakout. Instead of entering at a fixed time, an algo can wait for the premium or the underlying to move by a set amount, or for a break of the first N minutes' range. On a quiet day it may not enter at all.
  • Re-entries. Setting re-entries to zero means a strategy stopped out once stays out for the day.

Manual pauses are for the unusual day. Rules are for the usual one.

Account limits that pause for you

Three account settings in Algos → M2M stop new trading without your tap, once you have set them:

  • The daily loss limit refuses new orders once today's loss reaches it, from algos, signals and your own taps. Square-offs still go through.
  • The maximum orders a day, if you set one, refuses orders beyond that count, which guards against a strategy that loops.
  • Smart Exit squares off and stops the algos in its segments when the day's MTM reaches your target, stop, trailing step or locked profit.

These are the controls that work when you are not looking. The pause and stop buttons are for when you are.

A worked example: away from the screen

A trader has a meeting from 1 PM to 3 PM on a day when two algos are running: a Nifty iron condor entered at 10:00 and a crude oil strategy due to enter at 1:30 PM.

  1. 12:50. She pauses the crude oil algo: she does not want a new commodity position she cannot watch. The iron condor keeps running with its own leg stops, its M2M stop and its 3:10 PM exit.
  2. 12:52. On the M2M tab she arms Smart Exit for NSE F&O with a stop-loss of ₹9,000 and a lock that keeps ₹3,000 once the day reaches ₹6,000. The daily loss limit is already ₹11,000.
  3. 2:20. During the meeting the market falls; the iron condor's call side is untouched, the put side moves against it. The algo's own put-leg stop exits that leg. The day's MTM never reaches Smart Exit's line.
  4. 3:05. Back at her desk, she reads the card: one leg stopped at 2:21, the rest due to exit at 3:10. She resumes the crude oil algo; its entry time has passed, so it waits for tomorrow.

Nothing in the meeting needed her. That was the purpose of the pause and the limits.

On a phone

Every control in this tutorial is on the phone too: the algo cards in Algos, the pause and stop buttons on each card, the M2M tab, and Exit all on the Positions screen. On a small screen, select several cards to pause them together rather than opening each one. The kill switch and Exit all ask for the same second factor on a phone as on a laptop.

When to pause: a short list

Each trader's list differs. These are common entries:

SituationA common choice
A large gap at the open, outside the range your strategy was tested onPause new entries until the first fifteen minutes are over
A known event: a budget, an election result, a central bank decisionPause, or stop and square off, before it
A strategy has lost more than its normal worst dayStop and square off; review before restarting
Two orders you did not expectKill switch; read the activity log
A broker session endedAlgos at that broker are refused entries already; sign in, then check each card
You will be away from the screen with positions openSmart Exit armed; daily loss limit set

Reading the card before you act

Before you pause or stop an algo, read its card. It shows:

  • The state: waiting for entry, running, paused, stopped, and why.
  • Today's legs with their fills and the M2M.
  • A line for anything the algo is waiting on: a broker sign-in, a missing price, an entry condition not yet met.
  • The activity log: every entry, exit, re-entry, change and refusal, one plain line each. Copy log copies the day.

Two minutes with the card often changes the choice. An algo that looks stuck may be waiting for a price; an algo that looks wrong may be doing exactly what its rules say on a day the rules do not suit.

A worked example: a gap-up morning

A trader runs three Nifty algos: a 9:20 short straddle, an opening range breakout on futures and an iron condor entered at 10:00. At 9:08, pre-open shows Nifty opening 1.8% higher on overnight news.

  1. 9:10. The trader selects the straddle and the iron condor cards and taps Pause on both. Neither will enter until resumed.
  2. 9:15. The market opens. The opening range breakout algo is left running: its rules are built for gaps, and it has a stop on its own.
  3. 9:30. The breakout algo enters long on the futures. Its stop and exit time are set.
  4. 10:15. The market has settled. The trader resumes the iron condor; its entry time of 10:00 has passed, so it waits for its next entry day rather than entering late. The straddle stays paused for the day.
  5. 11:40. The futures trade reaches its target and the algo exits.

Pausing two algos for the morning cost nothing in tokens and kept them from entering on a day their tests did not cover.

A worked example: a strategy that misbehaves

At 11:05 a trader sees two new orders from a Bank Nifty algo that was not due to trade again.

  1. 11:05. She sets the kill switch for NSE F&O. Trading there stops until the next session, the algo stops, and its open orders are cancelled. Her positions stay.
  2. 11:06. She opens the algo's card and reads the activity log. The algo re-entered twice after its stop was hit, because its re-entry setting allowed up to three re-entries a day.
  3. 11:10. The behaviour is the rule doing what it says, not a fault. She decides she does not want re-entries on this strategy.
  4. 11:12. Her open positions now have no algo watching them. She exits them from the Positions screen. The kill switch does not block a square-off.
  5. After the close. She edits the algo (possible only when it is off), sets re-entries to zero and runs a backtest of the changed version before switching it on again tomorrow.

Common mistakes

  • Stopping and keeping positions, then forgetting them. After "Stop algo, keep positions", nothing watches the positions. Add a stop or close them.
  • Using the kill switch to close positions. It stops new trading but places no order. Use Exit all to close.
  • Pausing the algo but not the signal or webhook that starts it. A signal or webhook can start an algo again. Check them too.
  • Editing a running algo. An algo can be edited only when it is off. While an algo that is off still holds positions, its legs, product and quantity cannot be changed until you square off.
  • Expecting a resumed algo to catch up. If its entry time has passed, it waits for its next entry day.

Common problems

ProblemWhat to check
The algo still entered after I paused itCheck the time on the activity log: the entry may have gone out just before the pause
Restart today is refusedThe broker it used is not connected or not signed in. Connect and sign in again
Edit is greyed outThe algo is on. Switch it off first
Delete does nothingDelete works only when the algo is off, and asks twice
A webhook hit started a stopped algoPause the webhook, or remove the algo from it
New orders are refused everywhereThe kill switch or the daily loss limit is on

Where automation controls are going

Automation is easy to start and, on many platforms, hard to stop cleanly. The direction in India, since SEBI's retail algo framework came into force for every broker from 1 April 2026, is clearer control at every level: each strategy registered with the exchange, brokers responsible for every algo order, and account-level limits that act across strategies.3 For a trader, that means the off switches matter as much as the strategies.

Next steps

Start with 500 free tokens

Connect your broker, build a strategy and test it on paper before a single live order goes out.

Frequently asked questions

Does pausing an algo close its positions?

No. A paused algo opens no new entries but keeps watching its open legs: their targets, stop-losses and exit time still apply.

What is the difference between the kill switch and Exit all?

The kill switch stops new trading in the chosen segments until the next session, stops the algos there and cancels their open orders, but places no order. Exit all stops the algos and then squares off every position.

Do I pay tokens again if I restart an algo the same day?

No. A restart is never charged, and switching off and on again inside the paid 28 days costs nothing more.

What happens to a paused algo at its exit time?

It still exits. Pausing stops new entries only; the legs a paused algo holds keep their targets, stop-losses and exit time, so they are closed at the exit time as usual.

Can a webhook or signal start an algo I stopped?

Yes, if it is linked to that algo. A webhook entry hit or a signal set to start the algo switches it on again in the place it ran before. Pause the webhook, disarm the signal or remove the algo from it if you want it to stay off.

Can I pause every algo at once?

Yes. Select the algo cards and pause them together. For an emergency, the kill switch stops every algo in the chosen segments in one step.

References

  1. U.S. SEC and CFTC, "Findings regarding the market events of May 6, 2010"; SEC, order in the matter of Knight Capital Americas LLC, October 2013 (sec.gov)
  2. SEBI, "Broad guidelines on algorithmic trading", circular of 30 March 2012 (sebi.gov.in)
  3. SEBI, circular on safer participation of retail investors in algorithmic trading, 4 February 2025, in force for every broker from 1 April 2026 (sebi.gov.in)
Trading involves substantial risk. Stopping or pausing an algo does not guarantee an exit at a specific price. This page is for information and is not investment advice.