Home / Risk Management / Circuit Breakers
Hard controlIntermediate

Circuit Breakers

A circuit breaker is an automatic halt that stops trading when a defined threshold is breached, existing both at the market level (set by the exchange) and at the system level (set by the trader).

Quick answer: A circuit breaker is an automatic halt that stops trading when a defined threshold is breached, existing both at the market level (set by the exchange) and at the system level (set by the trader).

In simple words

A circuit breaker is a tripwire that stops trading when things go too far, too fast. At the market level, exchanges like the NSE halt a stock or the whole market when prices move beyond set percentages, to cool panic. At the system level, your own algo should have breakers too: stop trading after a daily loss limit, or after a burst of errors or repeated orders, so a bad day or a bug cannot run unchecked.

Purpose

Circuit breakers exist to interrupt runaway feedback loops, whether a market panic or a malfunctioning algorithm, giving humans and systems a moment to stop before damage compounds.

Professional explanation

Market-level circuit breakers

Exchange circuit breakers halt trading when prices move beyond pre-set bounds. India's NSE operates an index-level circuit breaker on the market as a whole, triggered by moves in a benchmark index at defined percentage thresholds, which halts equity and derivative trading for a period that lengthens with the size of the move and the time of day. Individual securities also have daily price bands (circuit limits) beyond which they cannot trade in a session. These mechanisms are designed to arrest disorderly falls or rises, provide a pause for information to disseminate, and prevent cascading forced liquidations. They are set and enforced by the exchange and regulator, not by participants.

Why market halts matter to an algo

A halt is an event your system must handle gracefully. During a halt no orders execute, quotes may freeze or vanish, and on resumption prices can gap significantly, so an algorithm that assumes continuous trading can misbehave, for example by queuing orders that all fire at once on reopen or by mis-marking positions on a stale price. Systems must detect halts, suspend order generation, avoid acting on frozen or stale data, and treat the reopen as a discontinuity. Ignoring exchange halts is a common source of erroneous fills and risk mis-estimation at exactly the worst moments.

System-level daily-loss breakers

Independent of the exchange, a robust system enforces its own breakers, the most important being a daily-loss limit: once realised plus unrealised loss for the day reaches a threshold, the system stops opening new positions and often flattens existing ones. This caps the damage a bad regime or a subtly broken strategy can do in a single session, and it enforces the ruin-threshold thinking mechanically rather than relying on a human noticing. The threshold is a rules-of-thumb policy choice tied to the account's risk tolerance and the strategy's expected daily distribution, not a universal number.

Error and behaviour breakers

Beyond loss, breakers should trip on anomalous behaviour that signals a malfunction: an order rate far above normal, repeated rejects, the same order sent many times, a position or exposure exceeding limits, a stale or gapping data feed, or a mismatch between the system's expected state and the broker's reported state. These are early-warning breakers that catch bugs before they become losses, since a runaway loop can place hundreds of orders in seconds. Detecting the symptom (rate, rejects, state mismatch) is often faster and safer than waiting for the loss to accumulate.

Actions on trip: halt, cancel, flatten

A breaker must define not just the trigger but the response. Common actions, in increasing severity, are: block new entries while managing existing positions, cancel all resting orders, flatten (close) all positions to go to cash, and lock the system out until a human resets it. The right action depends on the trigger: a daily-loss breach might flatten and lock out, while a data-feed anomaly might cancel orders and pause pending recovery. The response must itself be robust, since flattening in a stressed market incurs slippage, and a poorly designed breaker can worsen the situation it was meant to contain.

Circuit breakers versus kill switches

The terms overlap, but a useful distinction is that circuit breakers are threshold-based, often graduated and sometimes automatically self-resetting controls that pause specific activity, while a kill switch is the ultimate, decisive halt of everything, usually requiring manual re-arming. A daily-loss circuit breaker that stops new entries is a softer control than a kill switch that cancels every order and flattens every position. In practice a system layers them: multiple graduated circuit breakers for specific conditions, with a single kill switch as the last resort for catastrophic or unclassified failures.

Market-level vs System-level circuit breakers

AspectMarket-level (exchange)System-level (your algo)
Set byExchange / regulatorThe trader
TriggerIndex or stock price moveDaily loss, error rate, state mismatch
ScopeWhole market or a securityYour own strategy or account
PurposeCool panic, orderly marketCap your loss, catch malfunctions
Your controlMust handle, cannot setYou design and own it

Practical example

Illustrative example (Indian market)

You set a system daily-loss breaker at 3 percent of a ₹5,00,000 account, or ₹15,000. On a volatile Bank Nifty expiry your strategy takes three quick losing trades totalling ₹14,200 by mid-morning; the breaker is not yet tripped but is close, so no new entry is allowed that would risk pushing past it. A fourth managed position then loses another ₹1,500, taking the day to ₹15,700; the breaker trips, cancels all resting orders, flattens the open position, and locks the system until you review it manually. Separately, had the NSE index-level circuit breaker halted the market during that session, your system would have detected the halt, stopped generating orders, and treated the reopen price as a discontinuity rather than trading on stale quotes.

NSE's index-based market-wide circuit breakers are triggered at defined percentage moves in the benchmark and coordinated across exchanges, halting all equity and derivative trading, while individual stocks have daily price bands; an algo must subscribe to and respect these states, because attempting to trade a security at its circuit limit simply results in unexecuted orders piling up.

Advantages

  • Interrupts runaway feedback loops before damage compounds
  • Daily-loss breakers cap single-session damage mechanically
  • Error and rate breakers catch software malfunctions early
  • Graduated responses allow proportionate reactions to different triggers

Limitations

  • Flattening into a stressed market incurs slippage and can realise the worst price
  • A poorly tuned breaker trips on normal volatility and takes you out of good trades
  • Market halts create gaps and stale data that a naive system mishandles
  • Breakers add complexity and their own failure modes that must be tested
  • A breaker that only blocks new entries still leaves existing losing positions open

Why it matters in practice

  • Circuit breakers convert vague intentions to stop into enforced, automatic limits
  • They are the mechanism that makes daily-loss and error policies real rather than aspirational

Common mistakes

  • Assuming the exchange circuit breaker protects your account; it protects the market, not you
  • Setting a daily-loss limit but relying on manually watching instead of automating the halt
  • Not handling market halts, so queued orders all fire on reopen at a gapped price
  • Tuning the breaker so tight it trips on ordinary volatility, or so loose it never protects
  • Defining a trigger but not the action, so the breaker alerts but does not stop trading
  • Building a breaker that flattens without accounting for the slippage of doing so in stress

Professional usage

Serious trading operations run layered circuit breakers as core infrastructure: exchange-halt handling, per-strategy and account-level daily-loss limits, and behavioural breakers on order rate, reject rate and state reconciliation, each with a defined and tested response. Thresholds are calibrated against the strategy's normal daily distribution so they trip on genuine anomalies rather than noise, and every breaker is exercised in testing, including the flatten-in-stress path. Breakers sit in the risk engine, independent of strategy code, so a bug in the strategy cannot disable its own safety limit.

Key takeaways

  • Circuit breakers exist at the market level (exchange-set) and the system level (you set)
  • NSE halts the market on large index moves and bands individual stocks; your algo must handle halts
  • Build daily-loss and behavioural (error, rate, state-mismatch) breakers into the risk engine
  • Define the action, not just the trigger, and test the flatten-in-stress path

Frequently asked questions

What is a circuit breaker in trading?
A circuit breaker is an automatic halt that stops trading when a threshold is breached. It exists at the market level, where exchanges pause trading on large price moves, and at the system level, where a trader's algo halts on a daily-loss or error limit.
What are NSE circuit breakers?
NSE operates index-based market-wide circuit breakers that halt all equity and derivative trading when a benchmark index moves beyond set percentage thresholds, with halt durations that vary by the size of the move and the time of day. Individual stocks also have daily price bands beyond which they cannot trade.
Do exchange circuit breakers protect my account?
No. Market-level breakers are designed to keep the overall market orderly and cool panic, not to limit your personal loss. Protecting your account requires your own system-level breakers, such as a daily-loss limit.
What is a daily-loss limit?
A daily-loss limit is a system-level breaker that halts new trades, and often flattens positions, once the day's realised plus unrealised loss reaches a set threshold. It caps how much a bad session or a broken strategy can lose in one day.
What should trigger a system circuit breaker?
Common triggers are a daily-loss threshold, an abnormal order rate, repeated order rejects, the same order sent many times, exposure exceeding limits, and a mismatch between the system's expected state and the broker's reported state. These catch both losses and malfunctions.
What actions should a circuit breaker take?
In increasing severity: block new entries, cancel resting orders, flatten all positions, and lock the system until a human resets it. The right action depends on the trigger, and the response must account for slippage when flattening in a stressed market.
What is the difference between a circuit breaker and a kill switch?
Circuit breakers are threshold-based, often graduated and sometimes self-resetting controls that pause specific activity, while a kill switch is the ultimate, decisive halt of everything that usually requires manual re-arming. Systems layer several breakers with one kill switch as the last resort.
How should an algo handle a market halt?
It should detect the halt, stop generating orders, avoid acting on frozen or stale quotes, and treat the reopening price as a discontinuity. Failing to do so can cause queued orders to fire at once on reopen or positions to be mis-marked on stale data.
How do I set a daily-loss limit level?
It is a policy choice tied to your risk tolerance and the strategy's normal daily loss distribution, not a universal figure. It should be large enough to survive ordinary bad days but small enough to prevent a catastrophic session, and calibrated against historical daily losses.
Can a circuit breaker make things worse?
Yes, if poorly designed. Flattening into a stressed market realises slippage, and a breaker tuned too tight exits good trades on normal volatility. This is why breaker thresholds and their flatten paths must be calibrated and tested.
Where should circuit breakers live in the system?
In the risk engine, independent of the strategy code, so that a bug in the strategy cannot disable its own safety limit. The breaker must be able to act on the order flow and positions regardless of what the strategy logic is doing.
Are stock circuit limits the same as market circuit breakers?
They are related but distinct. Stock circuit limits are daily price bands on individual securities beyond which they cannot trade, while the market-wide circuit breaker halts the entire market on large index moves. Both are exchange mechanisms an algo must respect.

Voice search & related questions

Natural-language questions people ask about Circuit Breakers.

What is a circuit breaker in the stock market?
It is an automatic pause on trading when prices move too far too fast. Exchanges use it to cool panic, and you should build your own to stop your algo after a bad loss.
Does the exchange circuit breaker save my money?
No. It keeps the whole market orderly, not your account safe. For that you need your own daily-loss limit that halts your trading.
How do I stop my algo after a bad day?
Set a daily-loss limit in your risk engine. When the day's loss hits it, the system cancels orders, closes positions and locks itself until you review.
What is the difference between a circuit breaker and a kill switch?
A circuit breaker pauses a specific thing when a threshold trips, often gently. A kill switch is the big red button that stops everything and needs a human to turn it back on.
What happens to my algo when the market halts?
It should notice the halt, stop sending orders and ignore frozen prices. If it does not, queued orders can all fire at a gapped price when the market reopens.
When should my system stop trading itself?
On a daily-loss limit, on a burst of errors or repeated orders, or when its view of positions does not match the broker. Any of those signals trouble.

Sources & references

    Last reviewed 11 July 2026. Educational content only — not investment advice. Markets and rules change; verify current conventions with SEBI, NSE/BSE and your broker.

    ← Previous
    Stop-Loss Concepts
    Educational content only — not investment advice. Examples use illustrative numbers and simplified models. Algorithmic trading and derivatives involve substantial risk. See our Risk Disclosure and SEBI Disclaimer.