← All articles

Time Perception

Why Deadlines Feel Like They Arrive in Dog Years

October 5, 2026

On 19 November 1955, The Economist published C. Northcote Parkinson's essay and opened with a postcard. An elderly lady of leisure, he wrote, can spend a whole day sending one card to a niece in Bognor Regis: an hour finding the card, an hour hunting spectacles, half an hour on the address, more than an hour composing, twenty minutes deciding about an umbrella. A busy person, he said, would have finished the same errand in about three minutes. The sentence people still quote is the setup, not the bureaucracy formula that follows: work expands so as to fill the time available for its completion.

That is the feeling this site is named for, minus the civil-service charts. A deadline six months out looks like a roomy afternoon. The same deadline, two weeks out, feels like it has been aging in dog years. The clock did not speed up. The work got elastic, and the forecast you made on a calm Tuesday did not include the umbrella debate.

The 7-to-1 rule is a party trick

This site's converter treats one human year as seven dog years. That ratio is a playful compression, the same folk rule people use at dinner when a Labrador turns four and someone announces the dog is twenty-eight. It is not a medical formula, and veterinary sources have been saying so for years.

The American Kennel Club's guide to dog-to-human age, updated in September 2026, quotes the American Veterinary Medical Association directly: contrary to popular belief, dogs do not age at a rate of seven human years for each year in dog years. The AKC reports the AVMA's rough guideline for a medium-sized dog as about fifteen human years for the first dog year, about nine for the second, and about five human years for each dog year after that. Size matters. The same AKC page says small dogs are often treated as senior around seven, while larger breeds are often treated as senior at five or six, because veterinarians start seeing more age-related problems earlier in big dogs.

A methylation study of 104 Labrador Retrievers, summarized on that AKC page, produced a different curve: human-equivalent age roughly equals 16 times the natural log of the dog's age, plus 31. The AKC is explicit about the limit. The formula came from one breed. A Great Dane and a dachshund do not share a lifespan, so one equation cannot be a universal clock.

Use the converter the way you would use the postcard story. Turn a date into dog years, full moons, or Mondays when a calendar number feels abstract. Then put the joke down. The planning problem is not that your project is secretly a senior Labrador. The planning problem is that early forecasts ignore the hours that will get eaten.

What the thesis students actually predicted

Roger Buehler, Dale Griffin, and Michael Ross published five studies of this bias in the Journal of Personality and Social Psychology in 1994. They called it the planning fallacy, borrowing the name from Daniel Kahneman and Amos Tversky's 1979 description: a confident belief that your own project will go as planned, even when you know similar projects ran late.

In the first study, undergraduates working on an honors thesis gave a best-guess completion time. On average they predicted 33.9 days. They actually took 55.5 days. Fewer than a third, 29.7 percent, finished by the date they had called their most accurate prediction. Asked to assume everything went as well as it possibly could, they predicted about 21.4 days, and only 10.8 percent finished that early. Asked to assume everything went as poorly as it possibly could, they predicted 48.6 days, and fewer than half, 48.7 percent, finished even by that pessimistic date. Four more students had not finished when the researchers stopped recording, two semesters after the survey.

The predictions were not noise. Best guesses correlated with actual completion times at r = .77. Students who said they would take longer than their classmates usually did take longer. The bias sat on top of real information. People were informative and still systematically early.

A classroom survey in the same paper found students reporting that they had finished about 68 percent of previous projects later than they expected. The authors' point is the uncomfortable part: people can remember a string of late projects and still treat the current forecast as the realistic one. They know the past and still plan as if this time the spectacles will be on the desk.

Keep the sample in view. These were undergraduates, the thesis analysis rested on a few dozen completers, and a psychology lab is not a construction site or a freelance invoice. The Channel Tunnel figures Buehler and colleagues mention in the introduction, a June 1993 opening that slipped and a budget that swelled, are there to show the pattern is not only a student problem. They are not a license to treat 55.5 days as your personal multiplier.

Why the last fortnight feels heavier than the first month

Parkinson's postcard is elastic demand. Buehler's students are optimistic supply. Put them together and the dog-years feeling has a mechanism.

Early in a project the available time is large, so the work spreads. You research a little, rename the file, wait for a reply that is not on the critical path. None of that feels like failure because the due date is still a rumor. Near the end, the same tasks have to fit a smaller box, and every interruption is visible. The week did not shrink on the calendar. The slack you spent in month one is gone, and the tasks you described as "almost done" still have the umbrella decision left in them.

Buehler, Griffin, and Ross also watched what people thought about while predicting. Think-aloud notes were full of future scenarios, the smooth path from here to done, and thin on the last three times a similar task slipped. Past misses got explained as external, temporary, and specific: the roommate, the broken printer, the one weird professor. Those explanations let the next forecast start from zero.

That is why a deadline can feel both far away and suddenly on top of you. Far away, you are narrating a best-case film. On top of you, you are living the cuts the film left out.

A forecast that survives contact with last time

Study 4 in the 1994 paper is the useful twist. Asking students to recall how similar assignments had gone was not enough. They could say, out loud, that they usually finished about a day before the deadline, then predict the new computer assignment as if that history belonged to someone else. The optimistic bias shrank only when the instructions made them connect those past episodes to the plan in front of them.

Even then, the authors are careful. Connecting the past removed the average optimistic bias in that experiment. It did not make each guess more accurate in absolute terms. Students became as likely to guess long as short. Unbiased is not the same as precise. A personal deadline system that promises precision is selling a mood.

Here is a small practice that follows that result, not a guarantee. Pick one live deadline. Write three lines before you touch the task list:

  1. The last similar piece of work: what you predicted, and the date it actually left your hands.
  2. One cause you are tempted to call a one-off. Say whether that cause has shown up twice.
  3. The date you will treat as your own finish, set earlier than the external due date only if line 1 gives you a reason. If your last three similar jobs finished the night before, do not invent a cushion you have never used. Record the pattern instead.

Then run the external date through the deadline converter once, read the dog-year number, and close the tab. The number is a nudge. The three lines are the estimate.

If you want the calendar version of protecting the early hours, the companion piece is how a scheduled block beats the heroic all-nighter. This page is about the perception. That one is about where the hours go.

What this will not fix

A clearer forecast does not move a due date set by a court, a client contract, or a professor. It does not diagnose time-blindness, ADHD, or a sleep disorder. If lateness is tangled up with distress that is spilling into the rest of life, that is a clinical question, not a converter question. The studies above are about prediction error in the samples they measured. They do not say you are broken, and they do not say a three-line ritual will make the next project land on the day you name.

Parkinson was satirizing staff growth in public offices, not handing you a law of nature. The postcard is an illustration inside an essay. The thesis numbers are means from one university sample in the early 1990s. The AVMA aging guideline is a rough veterinary communication tool, and the Labrador methylation formula is one breed's clock. The 7-to-1 switch on this website is a joke with a serious neighbor: your early plan is probably a film, and the last week is the cutting room.

Which of your open deadlines still has a forecast you have not compared with the last time you did that kind of work?

Sources

← All articles · Try the deadline converter