Singapore to London Heathrow in January is an 8-hour clock difference that the body clock closes by running later, about 5.3 days at the published averages. Singapore to San Francisco, 16 hours behind, is 8 hours ahead the other way round the dial, and that 8-hour advance comes to 8 days at the same averages. Both figures are our arithmetic, on time zone offsets for 15 January 2026 and field rates of about 1.5 hours a day delaying and 1 hour a day advancing. East costs more, at least in part, because the human body clock runs, on average, a little longer than 24 hours, so it slides later more readily than it moves earlier.
Body clocks rehearse a small delay every day
In 157 people studied in a laboratory, Duffy and colleagues (2011) measured an average intrinsic period of 24.15 hours, or 24 hours 9 minutes, close to the 24.18 hours Czeisler's group reported in 1999. With no daylight to correct it, a clock of that period would drift about nine minutes later each day, by our arithmetic.
Flying west asks the clock for a phase delay, running later; flying east asks for a phase advance, running earlier. Eastman and Burgess, in a 2009 review, explain the easier westward direction "at least in part" by that slightly long period. Duffy found periods below 24 hours in 35% of the women and 14% of the men.
Early flight estimates get top billing in the day-per-zone rule
The same review cites "early estimates from real jet travelers", reviewed by Aschoff's group in 1975: about 92 minutes a day of delay after westward flights and 57 minutes a day of advance after eastward ones. The review works these through as 1.5 and 1 hour a day. On our reading, the familiar "a day per time zone" restates the eastward average, with westward coming to about two-thirds of a day for each hour. The unit that matters is the hour of clock difference; Singapore to Delhi is 2.5 hours.
These are averages, and light moves them. After a 12-hour laboratory shift with timed bright light, the review reports 1.6 hours a day advancing and 2.4 delaying over the first four days, and says the flight averages are "altered by the actual pattern of light and dark".
Daylight saving recasts the hour count on routes out of Singapore
Our offsets come from the IANA time zone database (release 2026c) at noon on 15 January and 15 July 2026; we counted both ways round the dial and kept the shorter.
| Route | Clock difference, Jan / Jul (hours) | Shorter way | Days at the published averages, Jan / Jul |
|---|---|---|---|
| Singapore to London | 8 / 7 behind | Delay | 5.3 / 4.7 |
| Kuala Lumpur to Jeddah | 5 / 5 behind | Delay | 3.3 / 3.3 |
| Singapore to Sydney | 3 / 2 ahead | Advance | 3.0 / 2.0 |
| Singapore to San Francisco | 16 / 15 behind | Advance of 8 / 9 hours | 8.0 / 9.0 |
| Hong Kong to New York | 13 / 12 behind | Delay | 8.7 / 8.0 |
Singapore, Malaysia and Hong Kong keep one offset all year. Many destinations keep summer time (daylight saving time), so each European, Australian and North American route in the table moves an hour between January and July. In 2026 London falls back on 25 October and New York on 1 November. The body follows the clock difference, whatever path the aircraft takes.
According to the IANA release notes, British Columbia moved to permanent UTC−7 on 9 March 2026 and Alberta to permanent UTC−6 on 18 June. The Northwest Territories did the same as Alberta on 21 August, and Manitoba is due to move to permanent UTC−5 on 31 October. Toronto still changes its clocks. Vancouver is therefore 15 hours behind Hong Kong all year from November 2026, where it used to be 16 each winter.
The long way round steals the scene on the largest eastward shifts
Advancing at 1 hour a day, or delaying the rest of the 24 at 1.5 hours a day, the two take equal time at an advance of 9.6 hours on our arithmetic, and beyond it delaying is shorter. The other rate pairs in the same review put that point between 9.2 and 10.3 hours, again by our arithmetic, since only the ratio of the rates matters.
Observation agrees. Eastman and Burgess write that antidromic re-entrainment, shifting the long way round, "is common following eastward travel, especially when 8 or more time zones are crossed". Gundel and Wegmann (1989) estimated the critical advance at +7 to +10 hours. In a 2001 study by a Tokyo team, seven of eight men delayed after an 11-hour eastward flight to New York. The samples are small. Hong Kong to New York would need an advance of 11 or 12 hours; at the published averages delaying takes 8.0 to 8.7 days against 11 to 12 advancing.
Others said first that 12 hours east and 12 west are one clock change: a University of Maryland model published in 2016 gave the same recovery for 12 time zones either way, as Live Science reported, and Wikivoyage puts the maximum jet lag at 12 hours in either direction. Our jet lag calculator missed it: until 14 October 2026 it multiplied time zones crossed, up to 14, by a day or two-thirds of a day, so 12 hours east came out at about half as many days again as 12 hours west, with the same eastward light advice at 2 zones as at 14. The revision of that date works in hours of clock difference reduced to a 24-hour day, shows both ways round and flags advances of 8 to 12 hours.
Morning light switches roles at the body's temperature low
In 21 people given 6.7 hours of bright light, Khalsa and colleagues (2003) found that light centred before the core body temperature minimum, the nightly low point, delayed the clock, and light centred after it advanced the clock.
Eastman and Burgess's worked example puts that low point at 04:00, three hours before waking. On our arithmetic for that textbook case, an 8-hour advance such as Singapore to San Francisco in January places the unshifted low at noon on the destination clock, so morning light there falls on the side Khalsa found delays the clock. Individual low points vary; this describes a mechanism and is no timetable.
Sleep timing often missed its cue by the end of a 15-day window
A large real-world dataset comes from Willoughby and colleagues (2025) at the National University of Singapore, using Oura data: 64,847 trips of at least 1,000 km by 57,240 ring users. Sleep duration came back to within about 12 minutes of baseline after about two days. Sleep timing had often not returned to baseline by day 15, the end of the window. Disruption was more severe after eastward travel and across more time zones.
The study read sleep timing from a ring rather than measuring the body clock, and the authors note that the earlier circadian estimates came from very different conditions. Its largest group pooled 7 to 9 time zones, so it says nothing about the long-way-round band. Its trips all began in North America and Europe.
View author profile → · Editorial policy
