Dow futures night session: YM hours in Taiwan time and backtests
YM trades 23 hours a day, from 17:00 to 16:00 Central Time. That is 07:00 to 06:00 in Taiwan in winter and one hour earlier in summer. Define sessions in exchange time, build daily bars on the exchange trading date, and test overnight fills with higher slippage.
By Soren Huang · Published · updated · Editorial policy
Contents
- When do Dow futures trade in US and Taiwan time?
- Is the Taiwan "night session" the same as the US overnight session?
- How does daylight saving time move the Taiwan hours?
- What happens at the daily break, the weekly open and the quarterly roll?
- Which session should a YM backtest use?
- How much slippage should you assume overnight?
- What should you check before trusting a YM backtest?
- Frequently asked questions
Dow futures (YM) trade on CME Globex from 17:00 to 16:00 the next day, US Central Time, Sunday to Friday, with a one-hour daily break. In Taiwan that is 07:00 to 06:00 the next day in winter and 06:00 to 05:00 in summer. In a backtest, define sessions in exchange time and assume higher slippage overnight.
When do Dow futures trade in US and Taiwan time?
YM trades almost around the clock: 23 hours a day, five days a week, with a one-hour break each afternoon in Chicago.
E-mini Dow futures (YM) is a futures contract on the Dow Jones Industrial Average. It is worth $5 times the index, it is listed on the Chicago Board of Trade (CBOT), and it trades on CME Globex. CME Globex is the electronic trading platform of CME Group. The minimum price move is one index point, worth $5 per contract. The Micro E-mini Dow (MYM) follows the same index at $0.50 per point.
Central Time (CT) is the time zone of Chicago, and CME publishes its trading hours in Central Time. Taiwan uses UTC+8 all year. This article uses the 24-hour clock for every time.
| Part of the trading day | Central Time | Taiwan time, US winter | Taiwan time, US summer |
|---|---|---|---|
| Globex session opens | 17:00 | 07:00 | 06:00 |
| US stock market opens (New York 09:30) | 08:30 | 22:30 | 21:30 |
| US stock market closes (New York 16:00) | 15:00 | 05:00 (next day) | 04:00 (next day) |
| Globex session closes | 16:00 | 06:00 (next day) | 05:00 (next day) |
| Daily break | 16:00–17:00 | 06:00–07:00 | 05:00–06:00 |
The week opens on Sunday at 17:00 CT, which is Monday morning in Taiwan. It closes on Friday at 16:00 CT, which is early Saturday morning in Taiwan. Futures keep trading for one more hour after the US stock market closes.
Is the Taiwan "night session" the same as the US overnight session?
No. The two terms describe almost opposite parts of the day.
When a Taiwan trader looks up the Dow night session (道瓊夜盤), they usually mean the Taiwan evening. During those hours the US stock market is open. In US terms, that is the regular session, not the night.
The US overnight session is the Globex trading outside US stock market hours. It runs from 17:00 to 08:30 CT. In Taiwan, that is from early morning to late evening, which is the Taiwan daytime. Many guides, data vendors and platform settings use the US meaning. So when a source says "overnight liquidity is thin", it is describing Taiwan daytime hours.
How does daylight saving time move the Taiwan hours?
From the second Sunday in March to the first Sunday in November, every YM time in Taiwan moves one hour earlier, because the US changes its clocks and Taiwan does not.
The conversion is simple:
Taiwan time = Central Time + 14 hours in US winter, and + 13 hours in US summer.
Here "Central Time" is the clock time in Chicago, and the added hours carry you into the next Taiwan day when the total passes midnight.
The exchange schedule itself does not change. The session still opens at 17:00 in Chicago all year. Only your view of it from Taiwan moves. This matters in a backtest in two ways:
- Fixed Taiwan-time rules drift. Suppose a rule allows entries only from 22:30 to 05:00 Taiwan time. In winter it covers the US cash session exactly. In summer it misses the first cash-session hour (21:30–22:30) and adds the hour after the stock market close (04:00–05:00). About two-thirds of each year in your history then follows a different rule.
- Bar timestamps may be in different time zones. Data can be stored in exchange time, in UTC or in your local time. If two files use different zones, a time filter or a join between them will be off by one hour for part of each year.
What happens at the daily break, the weekly open and the quarterly roll?
Each of these boundaries creates a price gap or a change in contract, and a backtest has to handle each one on purpose.
| Event | When (exchange time) | What to do in the backtest |
|---|---|---|
| Daily break | 16:00–17:00 CT | No bars exist here. Treat the 17:00 bar as a new start, and decide whether the strategy may hold a position through the break. |
| Trading date boundary | 17:00 CT | Build daily bars from 17:00 to 16:00 CT, not from Taiwan midnight to midnight. |
| Weekly open | Sunday 17:00 CT | Expect weekend gaps. Check how stops and entries fill on the first bar. |
| Quarterly roll | March, June, September, December | Use a continuous contract that is adjusted at each roll, so roll gaps do not appear as price moves. |
| US holidays | US exchange calendar | Use the CME calendar, not the Taiwan calendar. |
The trading date is the date CME assigns to a session, and a new trading date starts at 17:00 CT. So the session that opens on Monday evening in Chicago counts as Tuesday. In Taiwan in winter, that Tuesday session runs from Tuesday 07:00 to Wednesday 06:00. If you cut daily bars at Taiwan midnight instead, each US cash session is split across two daily bars, and daily highs, lows and closes no longer match the exchange.
A roll is the switch from an expiring contract to the next one. YM lists quarterly contracts. Trading in an expiring contract ends at 09:30 New York time on the third Friday of the contract month. Volume moves to the next quarter in the days before expiry. The two contracts trade at different prices, mainly because of interest rates and expected dividends. An unadjusted continuous series shows that difference as a sudden jump. A strategy can "trade" that jump in a backtest even though nobody could profit from it. A back-adjusted series removes the jumps by shifting older prices, so use it for signals.
YM follows the US exchange calendar, so it trades on most Taiwan holidays, including Lunar New Year, and closes or shortens its hours on some US holidays. A data filter built on the Taiwan trading calendar will drop valid YM days. For the details on holiday sessions, see market holidays in backtests: gaps, reopenings and calendar changes.
Which session should a YM backtest use?
Test the session you will actually trade, and build the bars from that session. If you are unsure, run a full-session version and a cash-hours version and compare them.
A session template is the setting that tells the backtest which hours belong to a trading day. US regular trading hours (RTH) are the hours when the US stock market is open, 08:30–15:00 CT. Some platform templates end the regular session a little later, so check the end time in your template. The choice of session changes more than the trade times:
| Choice | What the bars contain | Use it when |
|---|---|---|
| Full Globex session | All 23 hours | You hold positions overnight, or you trade the US overnight from Taiwan during the day. |
| US cash hours only | 08:30–15:00 CT; overnight moves appear as a gap at the open | You trade only in the Taiwan evening and close positions before the US close. |
| Full-session bars, entries in a window | All 23 hours, but entries only inside a time window you set | You want indicators to see overnight prices, but you only enter when you are at the screen. |
Indicators react to the session choice. A 20-bar moving average on full-session bars includes many quiet overnight bars. On cash-hours bars it covers only active trading, and the overnight move arrives as one large gap. The same rules can therefore give very different trades. For more on full-session and day-only bars, see night sessions in futures backtests: day-only vs. full-session bars. The same CME hours and session choices apply to the S&P and Nasdaq contracts, covered in ES and NQ futures backtesting: five things to set first.
Make the exit rules match how you will trade. If you trade from Taiwan in the evening, plan for positions that are still open at your bedtime. Either test a rule that closes them, or test holding them through the US overnight with the higher slippage described below.
How much slippage should you assume overnight?
Use a larger slippage for fills in the US overnight than for fills in US cash hours. Then check whether the strategy still works when you double it.
Slippage is the difference between the price the backtest assumes and the price you actually get. In the US overnight, fewer orders rest in the order book. A market order or a stop order then moves the price more before it is filled. The first minutes after the daily break and after the Sunday open can be especially thin. Many major US economic reports come out at 08:30 New York time, before the stock market opens. That is 21:30 Taiwan time in winter and 20:30 in summer, and prices can jump in seconds.
If your data includes volume, measure liquidity yourself. Average the volume per bar for each hour of the trading day, in exchange time. The quiet hours are where your slippage setting matters most.
Hypothetical example
A YM strategy has 400 trades: 250 in US cash hours with an average gross result of $40 per trade, and 150 in the US overnight with an average gross result of $30. Commission and fees are $4 per round trip. One tick of slippage is $5 per side, so $10 per round trip.
| Cash-hours trades | Overnight, 1 tick per side | Overnight, 2 ticks per side | |
|---|---|---|---|
| Gross result per trade | $40 | $30 | $30 |
| Costs per round trip | −$14 | −$14 | −$24 |
| Net result per trade | $26 | $16 | $6 |
| Net result, all trades | $6,500 | $2,400 | $900 |
With one tick, the overnight trades look like a solid part of the result. With two ticks, they add little, and the decision to trade overnight depends on a cost estimate you cannot measure exactly.
For the full method of comparing costs with the average trade, see trading costs in backtests: commission, slippage and the average trade.
What should you check before trusting a YM backtest?
Run these checks in order. Each one removes an error that can make the result look better or worse than the trading would be.
- Time zone of the data. Confirm whether the bar timestamps are in Central Time, UTC or Taiwan time, and whether a timestamp marks the bar's open or its close.
- Session template. Set it in exchange time: 17:00–16:00 CT for the full session, or the cash hours you trade.
- Daily bars. Build them on the exchange trading date that starts at 17:00 CT.
- Time rules. Write entry windows in Central Time, then check a few trades in March and November to confirm they did not shift.
- Roll handling. Use a back-adjusted continuous contract, and look at the trades around each quarterly roll.
- Calendar. Use the US exchange calendar for holidays and shortened days.
- Costs by session. Set commission and slippage, with a higher slippage for overnight entries, and rerun with double slippage.
- Execution timing. Fill orders on the next bar after the signal, not on the close of the bar that created it.
In AlphaPrime (Valynox), backtesting includes trading sessions, fees and slippage, and next-bar execution in the research settings, so every candidate is evaluated under the same session and cost assumptions. The Performance Report shows equity, drawdown, the individual trades, periodic performance and long and short analysis. You can read each trade's entry time from the trades list and see whether the result depends on one part of the day.
What this means if you use AlphaPrime: set the YM session and the costs in exchange time before the first search, then use the Performance Report to see which hours the trades come from.
Frequently asked questions
What time does the Dow futures night session start in Taiwan?
If "night session" means the Taiwan evening when the US stock market is open, it starts at 22:30 Taiwan time in US winter and 21:30 in US summer. It ends at 05:00 or 04:00 the next morning. Futures keep trading for one more hour after that, until 06:00 or 05:00.
Why did my YM trade times shift by one hour in March and November?
The US switches to daylight saving time on the second Sunday in March and back on the first Sunday in November, and Taiwan does not change its clocks. The exchange hours stay the same in Chicago, so they move one hour in Taiwan time. Define session and time rules in Central Time to keep them stable all year.
Do Dow futures trade on Taiwan holidays?
Yes, on most of them. YM follows the US exchange calendar, so it trades during Taiwan holidays such as Lunar New Year. It closes or shortens its hours on some US holidays instead.
Should I backtest YM on the full session or on regular hours only?
Use the session you will actually trade. If you hold positions overnight or trade during the Taiwan daytime, use the full Globex session. If you trade only in the Taiwan evening and close before the US close, use cash hours or full-session bars with an entry window. Comparing both versions shows how much of the result comes from the overnight hours.
When do Dow futures roll to the next contract?
YM has quarterly contracts in March, June, September and December. Trading in the expiring contract ends at 09:30 New York time on the third Friday of the contract month, and volume moves to the next contract in the days before that. In a backtest, use a back-adjusted continuous contract so the price difference at the roll does not look like a market move.
What AlphaPrime does here
AlphaPrime's backtesting evaluates strategies with trading sessions, fees and slippage, and next-bar execution in every run. The Performance Report shows equity, drawdown, trades, periodic performance and long and short analysis for each result.