Pomodoro Arithmetic: What a 25/5 Day Actually Costs
The Pomodoro technique is usually described in terms of its rhythm: twenty-five minutes on, five off, a longer break every fourth session. What rarely gets described is what that adds up to — how long a run of sessions actually takes, and how much of that elapsed time is focus rather than break.
Both matter if you are trying to fit focused work into a day that has other things in it.
What a run actually costs
From the focus session planner, on the classic 25/5 cadence with a 15-minute long break every fourth session, starting at nine in the morning:
| Sessions | Focus | Breaks | Elapsed | Focus ratio | Finish |
|---|---|---|---|---|---|
| 4 | 100 min | 3 short = 15 min | 115 min | 87.0% | 10:55 |
| 6 | 150 min | 4 short + 1 long = 35 min | 185 min | 81.1% | 12:05 |
| 8 | 200 min | 6 short + 1 long = 45 min | 245 min | 81.6% | 13:05 |
The headline: to get 200 minutes of focused work — a little over three hours — you need to block out four hours and five minutes. That gap is the thing people leave out when they plan a morning.
The trailing break is not counted, and that is deliberate
Notice that four sessions produce three breaks, not four. The planner does not add a break after the final session, because the run is over — whatever you do next is not part of it.
This is a small detail with a real consequence for scheduling. If you plan four sessions to end at a hard deadline, the last thing that happens is work, not a break. If you actually want to arrive at the next thing having had a pause, you need to schedule that pause yourself, outside the run.
Why the focus ratio moves around
The ratio is not constant, and it is not monotonic either: 87.0% at four sessions, 81.1% at six, then back up to 81.6% at eight. That looks wrong at a glance.
The explanation is where the long breaks land. Four sessions attract no long break at all — the long break comes after the fourth, and there is no fifth session for it to precede — so the run is unusually efficient. Six sessions have absorbed one 15-minute long break across a shorter total, which drags the ratio down. By eight sessions that same single long break is spread over more focus time, so the ratio recovers slightly.
The practical reading: short runs look more efficient than long ones because they skip the recovery, not because they are better. A four-session morning at 87% is not a more productive cadence than an eight-session day; it is a cadence that has not yet paid for itself.
Twenty-five minutes is a starting point
Nothing about the number is sacred, and the planner takes every field as an input. Run four sessions at 50/10 with a 30-minute long break every third instead, and you get 200 minutes of focus in 250 minutes elapsed — an 80.0% ratio, finishing at 13:10.
Compare that with the eight-session 25/5 run: the same 200 minutes of focus, five minutes less elapsed time, and half as many transitions. For work with any warm-up cost, that is a meaningfully better arrangement, because each of those eight starts pays the reload cost again.
For work that is genuinely chunked — clearing a queue, reviewing a list, practising something in short reps — the shorter cadence wins instead, because there is no warm-up to lose and the frequent breaks keep the quality up.
What the breaks are for
The five minutes is not decoration and it is not slack to reclaim. It exists so that the next session starts from a recovered state rather than continuing a declining one.
Which means what you do with it matters. Switching from the task to email is not a break — it is a change of task at the same desk, with the same posture and the same screen. Standing up, walking, looking at something further than arm's length away: these are what the break is nominally for, and they are what people skip first.
The long break is the same argument at a larger scale. It is the one most likely to be deleted on a good day and the one whose absence shows up two sessions later.
Fitting a run to the time you have
The most useful direction to run the arithmetic is backwards. If you have a 90-minute window before a meeting, you do not have four sessions — you have three, taking 85 minutes, and five minutes of margin. Planning four means the meeting interrupts the fourth, which is the worst outcome available: you pay the start-up cost and get none of the block.
The general habit is to plan the number of sessions that fits inside the window with margin, rather than the number that fills it exactly. A run that finishes early is a run that finished. A run that gets cut off mid-session loses that whole session's work-in-progress along with the time.
Where this sits among the other planners
The session planner is the right tool for fragmented days and for work that does not need a long run-up. When you have a genuinely clear morning and the work has a warm-up cost, the longer blocks of the ultradian cycle planner generally serve better. And neither answers the prior question of how much of this kind of work your week can hold at all — that is what the deep work budget planner is for.
All three are planning aids rather than claims about attention, and none of them will make a badly-chosen task worth doing. They only tell you what a plan costs before you commit a morning to it.
Counting sessions instead of hours
One quiet benefit of running the arithmetic is that it gives you a unit for estimating work. "This will take about three hours" is a guess most people are bad at. "This is about six sessions" is a guess most people are noticeably better at, because a session is a concrete thing you have experienced many times.
After a few weeks of running sessions, you develop a calibrated sense of what fits in one. A first draft of a short piece: two sessions. Reviewing a colleague's work: one. Untangling a problem you do not yet understand: unknown, and that itself is useful information.
The estimate then converts back into elapsed time through the planner, including the breaks — which is where most estimates go wrong, because people estimate the work and then schedule as though the breaks did not exist.
When a session gets interrupted
Interruptions are the normal case rather than the exception, and it is worth having a policy rather than improvising each time.
The useful convention is that an interrupted session does not count. Not as punishment, but because a session broken in the middle usually produces less than half of what a whole one does — the value is concentrated in the uninterrupted stretch. Counting it as done inflates your sense of how much focused work you actually did and corrupts the estimates above.
The corollary is to make interruption less likely rather than to plan around it. Twenty-five minutes is short enough that "I'll come back to you in twenty minutes" is almost always an acceptable answer, which is one of the genuine advantages of a short cadence over a long one.