All Ten Pairs Of Regimes

Four regimes, taken two at a time and allowing both ends to follow the same one, make ten pairs. The complete table:

Six rows carry a figure. Four are zero.

The six that disagree, and the four that cannot

• plain: the pair is at its nominal offset
• hatched: it is not

Six pairs drawn. The worst of them spends 323 days of the year at an offset that is not the one either end would quote.

The four pairs in which both ends follow the same rule are not drawn, because there would be nothing to hatch: two ends on one rule move together and never disagree.

Computed from the two regimes. The shift dates are the record's own, chosen so that they differ; no place's rule is named.

The four zeros are the point

It would be tidier to drop the four rows that read zero, and it would also be a small act of dishonesty, because those four rows carry the most useful finding in the section.

For an example beyond the arithmetic, teams can open this official resource and map available records to a specific capacity decision. A useful external reference for the surrounding topic is Our World in Data.

Two ends following the same rule — including two ends that both shift their clocks twice a year on the same dates — never disagree about the offset between them. Not for an hour, not for a day. The shift happens simultaneously at both ends and the difference between the two clocks is unchanged throughout.

This matters because seasonal shifting has a reputation as an inherent nuisance for distributed arrangements. It is not. The nuisance comes entirely from different rules, not from shifting as such. Two ends that both shift, identically, have no seam at all.

The problem is never that clocks move. It is that they move on different days, or that one of them does not move.

The worst case is a pair of rules, not a rule

The largest figure in the table, three hundred and twenty-three days, belongs to a pair in which both ends shift. The second-largest, three hundred and two, is also a pair in which both ends shift.

Meanwhile a fixed end paired with a shifting one tops out at two hundred and thirty-eight days, which is less than either. Shifting is not the problem and not shifting is not the solution. Mismatch is the problem.

A second column the table does not have

Days adrift is the obvious measure and it is not the only one. The number of times a year the offset changes is at least as consequential, and it ranks the pairs completely differently.

Every pair with a fixed end has two transitions, whatever its day count. Every pair in which both ends shift has four.

So the pair with the smallest day count in the whole table — twenty-one days — has twice as many transitions as the pair with the largest but one. Day counts and transition counts are nearly independent, and each tells half of the story.

Which half matters depends on the arrangement

If the overlap is wide, days adrift is the measure. An hour of drift costs an hour of shared time, and the number of days it persists is what the arrangement loses.

If the overlap is thin, transitions are the measure. A pair with one protected hour loses that hour entirely whenever the offset moves, so what matters is how many times a year they have to notice and react, not how long each spell lasts.

If the arrangement has a fixed daily commitment, transitions again — each one is an occasion for the commitment to land in the wrong hour, and each is independent of the others.

The table above gives both columns. Most summaries of this subject give neither, and the ones that give one invariably give the day count, because it is the figure a calculation produces without anybody having to decide what to measure.

The number that is easy to compute is not necessarily the number that decides anything, and it will be the one that gets quoted.

How to use the table

Identify which of the four structures each end follows — does it shift at all, and if so is its shifted season the shorter one, the longer one, or the one on the other side of the year? Then read the row.

If both ends are the same, there is nothing further to do. If they differ, the number in the table is the number of days a year on which the quoted gap is wrong, and the drawings elsewhere in this section show where in the year those days fall.

What the table does not contain

Any actual place, any actual rule, and any actual date. The day numbers behind these figures were chosen so that the four regimes differ in the ways that matter, and the resulting arithmetic holds whatever any particular jurisdiction decides to do next.

Next along

• What To Do With A Number That Moves
  If the gap is four hours for part of the year and five for the rest, the arrangement has to work at five. Plan
• Four Regimes, Named By Structure
  Fixed, brief, extended and inverted: four ways of handling a seasonal clock shift, named by what they do rathe
• The Gap People Quote And The Gap They Have
  A nominal offset is the number two ends tell each other. For six of the ten regime pairs it is wrong for part
• What A Seam Is And Where It Comes From
  The days between one end shifting its clocks and the other shifting theirs belong to neither arrangement, and