Pine Script Sessions and Time Zones: Exchange Time, IANA Zones and DST

Every session specification needs an owner clock

“09:30–16:00” is incomplete. Add timezone, days, whether the opening or closing timestamp is tested, overnight behavior and the target timeframe. When timezone is omitted, time() and time_close() use the exchange timezone represented by syminfo.timezone by default.

Save chart timezone with screenshots and calculation timezone with strategy settings.

Choose time, time_close and calendar functions by job

A timezone changes calendar interpretation, not the absolute meaning of a UNIX timestamp.

not na(time(timeframe.period, session, timezone)) is a common opening-membership test. If the bar must close inside as well, test time_close() separately. A one-hour bar opening at 15:30 and closing at 16:30 can open inside a 09:30–16:00 session but not be fully contained.

Decode time, days, overnight spans and bar alignment

A time-based session string follows the general form HHmm-HHmm:days. Day digits run from 1 for Sunday through 7 for Saturday; 23456 represents Monday through Friday. Multiple periods can be comma-separated. input.session() makes hours editable; a separate day input keeps the weekday assumption explicit.

Named exchange sessions and time-based session strings are different concepts.

When a session endpoint does not align with the chart timeframe, time functions can construct internal bars for the requested timeframe. Align endpoints to chart-bar boundaries when possible, or plot both opening and closing membership so the discrepancy is visible.

Use IANA identifiers for regional local time

UTC-4 is a fixed offset. America/New_York describes a region whose offset changes between UTC−4 and UTC−5 with daylight saving time. They coincide during part of the year but are not interchangeable.

Official Pine guidance recommends IANA when local regional time and DST must align.

1. 01Test the week before DST

   Record the same weekday and local-hour membership.

2. 02Test the transition week

   Highlight bars where fixed UTC and IANA disagree.

3. 03Test the week after

   Recheck session endpoints and the first eligible entry.

4. 04Switch exchange symbols

   Confirm whether exchange or explicit regional time was intended.

This is not a price or return comparison. It tests whether one local-session definition produces the same classification through the seasonal clock change.

Pair the same weekday, local time and chart timeframe in an ordinary week and the transition week. Store UNIX timestamp, exchange rendering, IANA rendering, fixed-UTC rendering and time/time_close results. Retain the fixed-offset version as a control that should drift by one hour so the test demonstrates that it can detect a DST defect.

Expose hours, days and the IANA zone as inputs

This educational strategy keeps hours, days and timezone separate and permits entries only when the bar opens in session. It closes an open position when the session ends. The moving-average cross is a mechanics test, not a recommended market or session.

//@version=6
strategy("IANA session filter audit", overlay = true,
         initial_capital = 100000, margin_long = 100, margin_short = 100)

string hoursInput = input.session("0930-1600", "Session hours")
string daysInput = input.string("23456", "Session days",
     options = ["23456", "1234567"])
string timezoneInput = input.string("America/New_York", "Session time zone",
     options = ["America/New_York", "Europe/London", "Asia/Tokyo", "Etc/UTC"])
string sessionString = hoursInput + ":" + daysInput

bool opensInSession = not na(time(timeframe.period, sessionString, timezoneInput))
bool sessionEnded = not opensInSession and opensInSession[1]

float fast = ta.ema(close, 20)
float slow = ta.ema(close, 50)
bool longSignal = ta.crossover(fast, slow)

if opensInSession and longSignal and strategy.position_size == 0
    strategy.entry("Long", strategy.long)

if sessionEnded and strategy.position_size != 0
    strategy.close_all(comment = "Session end")

bgcolor(opensInSession ? color.new(color.teal, 88) : na)
plot(fast, "Fast", color.aqua)
plot(slow, "Slow", color.orange)

For an audit build, format the same bar timestamp in exchange time, Etc/UTC and the selected IANA zone in a table or Pine Log. This reveals display/classification confusion. Keep production drawings minimal while preserving the selected timezone in Inputs and documentation.

sessionEnded becomes true only when the script evaluates the first outside-session bar. strategy.close_all() then creates a market order during that calculation, while a standard simulated fill occurs at the next available tick. Preserve session endpoint, outside-bar open/close, order creation and fill as four timestamps. A process-on-close variant is a separate version and still does not guarantee a live broker fill at the endpoint.

Test bar-open and bar-close session membership separately

The mini calculator works on a simplified 24-hour clock and returns opening-endpoint and closing-endpoint membership separately. Those memberships do not prove which boundary was crossed or how many crossings occurred. Enter values after converting them to one local timezone; the calculator does not perform timezone or DST conversion.

• Does the first bar open exactly at the session start?
• Does the final bar close beyond the session end?
• Do equivalent local bars remain eligible across DST?
• Are overnight entries assigned to the intended trading day?
• Is an expected boundary bar missing because the market was closed?

Attach a session-boundary probe sheet to timestamped output

1. 01Freeze the clock

   Save session string, days, IANA timezone and chart timeframe.

2. 02Visualize boundaries

   Inspect session start/end, opening membership and closing membership.

3. 03Test three DST periods

   Compare the first and last eligible bars before, during and after transition.

4. 04Export the trade list

   Attach entry/exit timestamp and timezone notes to the CSV/XLSX.

5. 05Analyze conditionally

   Use session, weekday and time-of-day analysis only when timestamp granularity exists.

Do not stop at a boundary screenshot; retain a session-boundary probe sheet. On an ordinary date and a DST-transition week, capture bars immediately before the start, exactly at the start, immediately before the end and exactly at the end, including IANA zone, session string, weekday, open/close timestamps and time()/time_close() results.

Bind the probe sheet and trade CSV with one version ID. Keep chart timezone as display metadata, exchange timezone as Pine’s default clock and explicit IANA as the strategy-selected clock. A version that fails any boundary probe is not eligible for session-performance comparison.

Correct classification does not guarantee an in-session exit fill. Pass bar membership and order lifecycle separately.

Reproduce three cases: an ordinary week, an overnight session and a DST-transition week. For each, freeze the last inside and first outside bars, then record endpoint membership, sessionEnded, close-order creation and actual fill. Change only timeframe next and verify that missing or misaligned boundary bars remain explainable. Keep runs with correct classification but delayed fills as exit-design failures.