When to Stop Trading 24/7: Set a Boundary

In a 24/7 or near-continuous market, the trader cannot rely on a useful market close to define when the intended trading session should end. The stopping point has to come from a session the trader defines directly: a start anchor that activates it, a terminal boundary that ends new-entry eligibility, and a reset rule that states when the next session may begin. Attempt-count, loss-limit, and time-based shutdown triggers then run inside that self-defined session the same way they would run inside a regular session’s hours. Once the terminal boundary or an inside trigger fires, new-entry eligibility stays off until the stated reset condition occurs — the market remaining open does not re-arm the entry gate.

That distinction — a session container the trader defines, versus the shutdown triggers that operate inside it — is the entire problem this article solves. Session shutdown already defines what happens once a shutdown trigger fires or the terminal boundary is reached: lock the entry gate, reconcile open exposure, log the evidence. Matching an overtrading driver to the right shutdown trigger already defines which trigger — attempt count, loss limit, time boundary — fits which recurring behavioral pattern. Both assume the container already exists, because in a regular-session market the exchange’s hours supply it for free. A continuous market does not supply one. This article covers the step that has to happen first: building the session container itself, so the existing trigger framework has something to run inside.

Why “the market is open” is not a stopping variable

Three instrument categories are commonly described as “24/7” or “always on,” and they do not share one stopping-relevant structure.

Instrument type	Actual availability	What this means for a session container
CME cryptocurrency futures and options	Trade 24 hours a day, seven days a week since CME’s 24/7 launch on May 29, 2026, interrupted only by brief scheduled maintenance windows; Cryptocurrency Spot-Quoted futures remain on a separate five-day schedule outside this program¹	A maintenance interruption is a scheduled venue pause, not a trader’s session close — it does not reconcile exposure or reset a self-defined container
Many other futures contracts	Trade nearly around the clock with a scheduled daily maintenance period; exact open, close, and maintenance timing is set contract by contract rather than uniformly across an exchange²	Continuous trading does not imply constant liquidity — volume and spread can vary materially by contract and time of day³
Conventional spot forex and similar 24-hour-weekday instruments	Trades continuously across the global business week through a decentralized OTC market structure rather than a single exchange session; sales desks in the UK, US, Singapore, and Hong Kong together accounted for roughly three-quarters of global FX trading in the BIS’s 2025 survey⁴	Within the global business week, there is no single centralized exchange close that defines the trader’s intended intraday session boundary

The shared fact across all three rows is that “the venue will still accept an order” is a statement about access, not about whether continuing to trade is still the trader’s planned activity. A venue-defined regular session can supply an external boundary for trading activity: the exchange’s own close gives the trader’s intended session an edge to end at. None of these three structures reliably supplies a close that matches the trader’s intended discretionary session. The trader has to build the container that would otherwise be supplied for free. A separate, earlier BIS observation from the 2013 survey also found that intraday liquidity in conventional spot FX was typically deepest around the London open and the London/New York overlap — a historical pattern from that survey, not a current 2026 measurement.⁵

The trigger framework still works — it just has nothing to attach to yet

The driver-to-trigger mapping — an attempt-count trigger for frequency deviation, a loss-limit trigger for recovery pressure, a time-boundary trigger for boredom-driven drift — does not require a market close to function. An attempt-count trigger fires when the trader’s predefined attempt ceiling is reached, regardless of whether the venue remains open. A loss limit fires when realized loss reaches the threshold, regardless of session structure. What those triggers assume, implicitly, is that a session container already exists for the count or the clock to run inside — a defined window with a start, so “fourth attempt this session” and “2:30 p.m. this session” are meaningful phrases.

In a regular-session market, the exchange’s own hours supply that container for free. A trader does not have to decide when the trading day starts and ends; the market decides it, and the trigger framework operates inside a container someone else already built. A continuous market does not build that container. Without one, a time-based shutdown trigger has no reference point — 2:30 p.m. relative to what session? — and an attempt count risks becoming a rolling, never-resetting tally across a day that never technically ends. The design work this article owns is building the container first, so the existing trigger framework has something to attach to.

Build the session container before attaching a trigger

A usable session container needs exactly three elements defined before trading begins, the last of which is what lets it recur on a fixed schedule rather than a convenient one.

Choose the start anchor. State exactly what activates the session — a specific clock time, or an explicit activation event such as the start of a predefined monitoring window. The anchor has to be a fact that is true or false at a given moment, not “whenever I start looking at the market.”

Choose the terminal boundary. State exactly when new discretionary entries stop, even though the venue itself keeps quoting a price. This can be a clock time (for example, 12:00 UTC) or a fixed duration measured from the start anchor — for example, four hours from a stated monitoring-start event, not from the first filled entry, so time spent evaluating setups before an entry does not quietly extend the container.

Write the re-arm/reset rule. State exactly when the next session becomes eligible to begin — a fixed calendar-day reset in a stated time zone, or a new stated start-anchor event for the next planned session. The reset point is precommitted and externally identifiable from the trader’s written rule; it does not move because the trader wants one more trade, and it does not depend on whether the venue happens to still be open.

Then attach shutdown triggers inside the container. Once the container exists, the driver-to-trigger framework above — attempt count, loss threshold, or an earlier time cutoff — can end activity before the terminal boundary is reached. An attempt ceiling is a trigger operating inside the container; it is not itself the container, and reaching it does not define when the next session may begin — the separately stated reset rule does.

The same start-anchor, terminal-boundary, and reset-rule structure applies when what removes a natural closing bell is a fixed work schedule rather than continuous market hours. See trading around a full-time job for deriving that container from the job’s fixed hours instead of the market’s own structure.

These elements only work if they match real behavior. A four-hour container is only useful if the trader can actually apply the same standard at hour one and hour four; extending it because the venue is still open reintroduces the exact problem the container was built to remove. Once the container is defined, everything session shutdown already describes applies without modification: the entry gate locks at the terminal boundary, open exposure is reconciled against its own management plan, and the evidence is logged before the trader steps away. The market staying open past that point does not change any of those three steps — it only means nothing outside the trader’s own written rule is going to enforce them.

Table: session container vs. the false cue it replaces

Instrument context	False stopping cue	Container element that replaces it
CME 24/7 crypto futures and options	Waiting for the scheduled maintenance window as if it were a session close	A terminal boundary and reset rule set independently of the maintenance schedule
Near-continuous futures	Assuming a quiet, thin-liquidity stretch means the session is effectively over	A terminal boundary aligned with the contract’s intended liquidity window, with separate attempt-count or loss triggers operating inside it, not a judgment about how the chart looks right now³
Decentralized spot forex	Treating “still technically open” as permission to keep evaluating setups past the planned window	A terminal boundary tied to the trader’s planned monitoring window, defined independently of venue availability

This is a new-entry boundary, not a hold-authorization decision

Setting when new activity should stop is a different question from what happens to a position that is already open when the terminal boundary arrives. A session container determines whether new discretionary risk may be initiated; it does not automatically answer whether already-authorized exposure should be held, reduced, exited, or managed under an existing contingency. Overnight trading behavior owns that second question: whether carrying an existing position past a session boundary was authorized by the active risk state, and what exit contingency applies if the invalidation point is reached with reduced monitoring. A continuous market does not remove that question — if anything it sharpens it, because reduced monitoring can extend well past the length of a single overnight window in a market that never technically closes. The container this article covers decides when the entry gate locks; what happens to exposure still open at that moment is governed by the position-management plan already in place, not by this framework.

Frequently asked questions

If a market never closes, do I even need a session container?

Yes. The market not closing removes the container the market would otherwise supply for free; it does not remove the reason one exists. A specific, precommitted start anchor, terminal boundary, and reset rule are easier to hold to under pressure than a general intention to stop when it “feels right” — the same planning principle session shutdown relies on.⁶ Continuous evaluation of new setups without a defined container is a frequency-deviation risk regardless of whether the venue happens to still be open.

Does a maintenance window count as my session close?

Not by itself. A maintenance window is a scheduled pause in trading, not a reconciliation of exposure or a reset of a trader’s own container.¹ Treat it as a scheduling fact about the venue, and set the actual terminal boundary and reset rule independently of when that window happens to occur.

How is this different from just picking a loss limit?

A loss limit is one trigger type within the driver-to-trigger framework above, and it still works in a continuous market once a session container exists for it to operate inside. The problem this article addresses comes first: without a defined container, there is no stated point at which “today’s loss limit” resets, which can let a single threshold silently govern activity across many days instead of one.

Where Costante fits

Costante can keep predefined session guardrails — such as a session cutoff, a re-entry limit, and an after-loss risk state — visible during the trading session, alongside the trader’s written plan and later review evidence. Where the trader has defined a session container and its shutdown triggers, that guardrail visibility and logging can help keep the definition in view in the moment and reviewable afterward.

Costante does not connect to a broker or exchange, does not execute or block an order, does not automatically know an instrument’s trading hours or maintenance schedule, and does not determine what session container or terminal boundary is appropriate for a given market or trading style. The trader remains responsible for defining the container, confirming the instrument’s actual trading hours and liquidity structure with their broker or venue, and every decision made before and after the boundary is reached.

Costante provides educational workflow tools, not financial advice. Trading involves risk.

CME Group. 24/7 Crypto Futures and Options Trading. CME Group. CME Group Announces Launch of 24/7 Cryptocurrency Futures and Options Trading (May 29, 2026 launch date). Maintenance windows for the 24/7 program run Monday–Friday from 4:00–4:02 p.m. CT and Saturday from 2:00–4:00 a.m. CT; Cryptocurrency Spot-Quoted futures are excluded and remain on a five-day schedule. ↩ ↩²
CME Group. CME Group Holiday and Trading Hours. ↩
CME Group. CME Liquidity Tool (current and historical bid-ask spreads, book depth, and cost-to-trade statistics across CME Group products and Chicago, London, and Singapore time zones). ↩ ↩²
Bank for International Settlements. 2025 Triennial Central Bank Survey of Foreign Exchange and Over-the-counter (OTC) Derivatives Markets (April 2025 survey; preliminary results released September 2025; complete turnover data released December 2025; final results, including FX settlement data, released June 2026; sales desks in the UK, US, Singapore, and Hong Kong SAR accounted for roughly 75% of FX trading). ↩
Bank for International Settlements. The anatomy of the global FX market through the lens of the 2013 Triennial Survey (historical liquidity concentration around the London open and London/New York overlap). ↩
Gollwitzer, P. M., & Sheeran, P. (2006). Implementation Intentions and Goal Achievement: A Meta-analysis of Effects and Processes. Advances in Experimental Social Psychology, 38, 69–119. This meta-analysis examined if-then planning across goal-pursuit domains generally, not trading specifically; it supports the general planning mechanism, not a trading-performance outcome. ↩