How to choose a Pomodoro interval

Updated 20 September 2026

There is no correct Pomodoro interval. There is the one Cirillo used, there are a few that other people have popularized, and there is the one that fits the work you are doing today. This guide goes through the common options, gives you a rule for picking, and tells you what to watch for when the one you picked stops fitting.

The five common patterns

25/5. The original. Wikipedia (retrieved 2026-09-20) gives 25 minutes as the typical timer setting, with a short break of 5 to 10 minutes and a long break of about 20 to 30 minutes after 4 rounds. It is the only pattern on this list that a randomized comparison has tested directly, for better and worse. Smits, Wenzel and de Bruin (2025) tested exactly this, 4 cycles of 25/5 over 2 hours, and found no productivity advantage over self-chosen breaks. It is short enough that you can commit to one when you are avoiding something, and short enough that a bad block only costs you 25 minutes. The 25/5 timer runs it with a 15-minute long break.

30/10. A small step up. Five extra minutes of work rarely changes what you get done, but doubling the break does. Ten minutes is long enough to leave the room, which 5 often is not. If you like the rhythm of 25/5 but find yourself still sitting at your desk when the break ends, 30/10 is the smallest change that fixes it.

50/10. Two pomodoros with the break between them removed. This suits work where the first 10 or 15 minutes are spent loading the problem back into your head, so a 25-minute block gives you only 10 minutes of real progress. Code, proofs, dense reading, and anything with a lot of open files. The cost is that a wasted block wastes 50 minutes, and 10 minutes may not be enough recovery after that long. The 50/10 timer pairs it with a 25-minute long break.

52/17. This pattern comes from DeskTime, a time-tracking company, which reported it in 2014 from an analysis of its own users' activity data: about 52 minutes of work, then 17 minutes away. It is an observation about one company's customers, not a study. It has survived because the numbers are sensible. 52 minutes is close to 50/10 on the work side, and 17 minutes is a real break, long enough to eat or walk around the block. The 52/17 timer has the numbers preset.

90/20. The longest common block. The 90-minute figure comes from the idea that attention runs in roughly 90-minute cycles through the day, which traces back to Nathaniel Kleitman's work on ultradian rhythms. Applied to focused work, that is a hypothesis, not an established fact, and nobody has shown that a 90-minute block beats a 50-minute one. What 90/20 does offer is room for work that genuinely needs an uninterrupted stretch: a long draft, a model you cannot leave half built, a problem you have to hold entirely in your head. The 90/20 timer gives a 20-minute break after each block and a 30-minute one after every 2.

A decision rule

Two questions decide most of it.

How long does the task take to get going? If you can be productive within 2 minutes of sitting down, because the task is a list of small items like email, flashcards, grading, or data entry, use 25/5 or 30/10. The bell will rarely interrupt anything that matters, and the frequent breaks keep a dull task from grinding you down. If the task has a setup cost, say 10 minutes of reading your own notes before you can write the next line, use 50/10 or 52/17. If the setup cost is closer to 20 minutes and the work falls apart when you step away, use 90/20.

How long before you drift? Be honest about this one. If you notice yourself reaching for your phone at minute 15 of a 25-minute block, the answer is not a longer block. Stay at 25/5, or drop to the 15-minute blocks on the ADHD timer, or even the 5-minute timer when starting is the problem. If instead the bell keeps landing mid-sentence and you resent it most blocks, you have outgrown the interval. Go up one step.

The rule of thumb that comes out of this: the bell should ring within about 5 minutes of when you would have wanted a break anyway. Much earlier than that and it is an interruption. Much later and you have already stopped working without noticing.

Ogut (2025) makes a related point in the limitations of that scoping review: rigid timing may not suit every learner, and may interrupt flow on complex spatial reasoning tasks. That is the case for longer blocks on complex work. Smits et al. (2025) make the case for caution in the other direction. In their comparison, fatigue rose faster over 2 hours in the 25/5 group than with self-chosen breaks. Short breaks, taken often, did not prevent tiredness. If you are tired by the third block, lengthen the break before you lengthen the work.

A worked example: a proposal on 52/17

Say it is 1 pm and you need a first draft of a 6-page project proposal by the end of the day. You have your notes, an outline, and a headache from the morning. Here is an afternoon on the 52/17 timer.

Block 1, 1:00 to 1:52. Task: write the background section from the outline. The first 12 minutes go to rereading your notes. By 1:20 you are writing. When the bell goes at 1:52 you are two paragraphs into the second section and mildly annoyed to stop.

Break 1, 17 minutes. You leave the desk, walk to get coffee, and eat something. You do not check email. You are back at 2:09 and the headache is mostly gone.

Block 2, 2:09 to 3:01. Task: finish the second section and draft the third. Because you stopped mid-flow, you know exactly where to pick up and there is no reread. The section finishes by 2:35. The third section is harder, and at 3:01 it is rough but complete.

Break 2, 17 minutes. Ten minutes outside, 7 minutes answering the two messages that actually needed you.

Block 3, 3:18 to 4:10. Task: the budget section and a first pass at the summary. Numbers go faster than prose. The summary is a paragraph by 4:10.

Break 3. You look at what you have: 4 of 6 sections drafted, one to tidy, one to write. One more block gets a full draft. Three blocks of 52 have produced roughly 2 hours and 36 minutes of writing in just over 3 hours of clock time, with 3 real breaks. On 25/5 you would have heard 6 bells in the same span and probably reread your notes 3 times instead of once.

When to change

Give an interval at least 4 full cycles before judging it. One bad afternoon says nothing. After that, these are the signals.

Go shorter if you are abandoning blocks. A block you stop at minute 30 of 50 is not a 50-minute block, it is a 30-minute block with guilt attached. Drop to an interval you actually finish.

Go longer if you resent the bell in most blocks and can honestly say you were still working when it rang.

Lengthen the break, not the block, if you feel worse at 3 pm than at 1 pm. The Smits et al. (2025) fatigue result is a reason to take the break seriously. Five minutes is enough to stand up. It is not enough to recover from a block that took real effort.

Change by task, not by mood. Keep 25/5 for admin days and 50/10 or 52/17 for building days, and set the timer to match the task you are about to do. Do not change the interval in the middle of a session because a block went badly. Finish the session, then decide.

And keep in mind that the whole question of whether timed breaks help is open. The guide on whether the Pomodoro technique works sets out both sides. If you have tried 3 intervals and none of them beats just working until you want to stop, that is a legitimate result, and you are allowed to act on it.