필사 모드: What It Means to Work Hard — Effort Is Not a Virtue but a Problem of Allocation

Introduction — The Blank in the Advice to Work Hard

Work hard is the safest and the emptiest piece of advice there is, because it says nothing about how much, when, on what, or at what cost. And that blank tends to get filled in one direction. Longer, more often, further past the limit.

Why that direction in particular is the first thing to look at. Of the several components of effort, time is the only one anybody can count. The quality of judgement, the density of concentration, the degree of recovery are all hard to measure, but hours sitting in the chair are visible to you and to everyone else. So, along the familiar path by which the measurable becomes the target, hard is naturally translated as long. Where that translation creates losses is the subject of this post.

If the previous post dealt with the structure of not waiting for drive, this one deals with the effort you actually spend on top of that structure. To state the conclusion first, effort is not a virtue that improves the more you raise the total but a problem of allocating a limited resource. The point at which people who work hard and people who last diverge is usually not quantity but allocation.

1. Sprinting and Middle Output — Output Is Not Proportional to Time

Underneath the view of effort as a total sits a simple arithmetic. Work twice as much and twice as much comes out. There is a classic analysis showing that this arithmetic breaks in actual data. In the economist John Pencavel's study of records from British munitions factory workers during the First World War, output was proportional to hours worked up to a point, but past a threshold the rate of increase bent. There plainly exists a region in which pouring in more time does not increase output by the same measure.

This is a single-source analysis with the limitation that it is manual labour data from a century ago. It cannot be transplanted onto modern knowledge work as it stands. That said, in knowledge work the curve is likely to bend earlier, because a misjudgement made in a fatigued state is discovered later and costs more to repair than a defective part made in a fatigued state. Anyone who has torn out code written at 11 p.m. the following morning knows this curve in their body.

The principle of distributing force that I wrote about in the table tennis post applies directly. Good players do not use full power on every ball. They handle most balls with a force of 60 and spend 100 on the one decisive ball. A structure of sustainable middle output with selective maximum output laid on top comes out ahead in the total against a structure that attempts maximum output every day and collapses on the third.

Let me sketch a quarter as a constructed example. Two people take on the same project. One gets ahead early by stitching nights together. A few weeks later the graph bends. Fatigue seeps into judgement, design mistakes multiply, and the time banked earlier goes right back into removing those mistakes. The other sustains a similar volume every day and schedules the hard parts only for the few clear-headed mornings each week. Even if the cumulative output of the two is similar at the end of the quarter, their states are different. One is in no condition to start the next quarter; the other is in a condition to repeat the same method. Score effort by the single quarter and there are days when sprinting looks like the winner; score it across several quarters and allocation wins.

2. Recovery Is Not Outside the Work but Part of It

Seen from the angle of allocation, recovery is not time spent not working but time that produces the next round of output. The problem is that the body leaves the office and the head does not.

Sonnentag and Fritz, while building an instrument to measure the experience of recovering from work, organized recovery into four experiences (their 2007 study, validated on 930 respondents). Psychological detachment, actually getting away from thinking about work; relaxation, the loosening of tension; mastery, learning something in a domain unrelated to work; and control, the sense of spending your time as you intend.

The item on this list that matters most in practice is the first. Psychological detachment is not the same as physically leaving work. An evening spent lying on the sofa checking the work messenger is a state in which only the body has clocked out. If recovery is part of allocation, then detachment is also something to be designed. Setting a time at which work notifications go off; closing the day on the commute home with three lines about it and a note of tomorrow's first task. The head cannot let go of unresolved work, so you make a form of closure for it in order that it can let go.

The mastery item is also worth noticing, because it means recovery does not have to be the sofa. Learning something in a domain unrelated to work — an instrument, a sport, a language — can be a recovery experience while still spending energy, thanks to a different kind of concentration occupying the space where thoughts of work would enter. For the person who finds that doing nothing when they are supposed to rest only brings on more thoughts of work, a different absorption often works better than lying down.

And at the bottom of recovery there is one non-negotiable layer. Sleep. Sleep is the first thing cut when deadlines pile up, and the order is exactly backwards. For someone who works with their judgement, cutting sleep is like burning the principal because the budget is short. The case that the body is the foundation of mental state and performance was made separately in the body comes first, and if stopping at all brings on guilt, the burnout post and a day of doing nothing come first.

3. The Price Tag on Working Long Hours

A strategy that ignores allocation and runs on totals carries one more price tag besides the output curve. Health.

In a 2015 meta-analysis published in The Lancet by Kivimäki and colleagues, pooling data on more than 600,000 people, people working 55 hours or more a week had a 13% higher risk of coronary heart disease and a 33% higher risk of stroke compared with people working 35 to 40 hours a week. Stroke risk climbed in steps along with working hours: 10% at 41 to 48 hours a week, 27% at 49 to 54.

These are relative risks at the level of populations, so they are not to be read as a personal chart of destiny. If your baseline risk is itself low, even a 33% increase is small in absolute terms, and because this is a synthesis of observational studies, the possibility remains that factors other than working hours are mixed in. Even so, what this data tells us is clear. Long hours are not free, and the bill keeps arriving after the output curve has already bent. What a culture that defines hard as a total pays is written down here.

The way to move these numbers into practice is policy, not fear. Set a ceiling on weekly working hours and treat it not as a mark of laziness but as risk management. If there is a week in which the ceiling has to be exceeded, that is an exceptional situation, and an exception stays an exception only when it is settled by a reduction the following week. An exception that repeats without settlement is not an exception but a new default.

4. The Uncomfortable Part — Invisible Effort Does Not Get Assessed

If everything so far has been the arithmetic of allocation, now comes a more uncomfortable structural point. For the same amount of effort, effort that is visible to the organization and effort that is not are rewarded differently. This is a problem that can be stated as structure rather than by citing empirical research, and it will match most people's experience of organizational life.

Assessment operates only on top of observation. Assessors, from where they sit, have no choice but to judge on what is observable — remarks made in meetings, documents shared, incident responses that caught the eye. So an outage quietly prevented over several nights can end up recorded as smaller than a fire fought spectacularly. A prevented problem by definition did not happen, and what did not happen is not visible.

Neither resenting this structure nor being cynical about it as politics gets you anything. By the same logic as the table tennis post, where I wrote that skill is a signal that pays the cost of verification on your behalf, making work visible is not exaggeration but an interface. It is not inflating the work; it is widening the definition of the work to include making it observable.

Concretely, three things are enough. First, a weekly note. Writing down briefly what you did this week and what you will do next week, and sharing it, means that come assessment season you hand over records instead of rummaging through memory. Second, a decision record. Leave a paragraph on why you chose this approach and you can later explain what those outwardly quiet weeks actually avoided. Third, a prevention report. A problem stopped before it blew up becomes an event that never existed unless you say so yourself, so share it afterwards in a paragraph. All three together take thirty minutes a week, and what separates them from showing off is form. Facts with no adjectives make a record; adjectives piled up with no facts make theatre.

And this is not purely a matter of individual conduct either. An organization that knows rewards are determined by what is visible has to build a procedure that deliberately seeks out and counts the invisible kinds of work, because an organization that counts only incident response ends up cultivating the people who create incidents.

5. The Allocation Procedure — A Weekly Effort Budget

Lowered into a weekly procedure, the view of this post looks like this.

There must be exactly one maximum-output block. Saying everything is important is the same as saying you have given up on allocation. For running a focus block a simple tool like a pomodoro timer is enough, and if where your time goes is itself the problem, the diagnosis covered in the illusion of time management comes first.

Having used the word budget, the rules of a budget have to follow. When there is an overrun, something has to come out somewhere. The most common failure is erasing the recovery block first when new work arrives, and that is not a budget adjustment but a loan taken from the future. The order should be the reverse. Stop or postpone the lowest-priority item in the base output block, and if that is still not enough, change the target of the maximum output block rather than increasing the count. Ten minutes each weekend, check how far this week's allocation differed from the plan and fix next week's table. Allocation is not something you set once but something you settle every week.

What Has Not Been Verified

• Ego depletion — the theory that willpower wears down like a muscle when used may look similar to this post's budget metaphor, but its evidentiary status is different. In a 2016 preregistered multi-lab replication involving 23 laboratories and 2,141 participants, the ego depletion effect was effectively close to zero (effect size 0.04, with a confidence interval including zero). The budget this post speaks of is not a total quantity of psychological willpower but the allocation of observable resources — time, sleep, attention — and it does not lean on ego depletion theory.
• Pencavel's output curve — a single-source analysis of a particular period and a particular industry. The direction agrees with common sense and with modern observation, but how many hours the threshold sits at cannot be generalized.
• Recovery experience research — a validation study for a measurement instrument based on self-report questionnaires; it did not prove experimentally that the four experiences cause recovery.
• The visibility discussion — section 4 is structural reasoning, not an empirical result. Observation systems differ from organization to organization, so the scope of application differs too.

Closing — What People Who Last Have in Common

This is not an argument for not working hard. It is an argument for redefining hard. Not giving everything every day, but handling most of it at a sustainable intensity and saving maximum output for what matters. Putting recovery in the calendar. And counting the work of making that effort visible as part of the work.

The language of totals is simple and therefore attractive, but the cost of that simplicity is paid out of the body and out of judgement. The language of allocation is less heroic, and it lasts. Allocation also becomes a practice. A wrong allocation this week becomes material for fixing next week's table, and after a few quarters you will have become someone who knows the rhythm of their own output. And when it comes to building skill — as the next post will discuss — lasting is very nearly the whole of it.

Self-development fundamentals series — Previous: How Not to Wait for Motivation · Next: The Structure of Being Good

References

All sources below were accessed on 2026-08-16.

• Pencavel (2014), IZA Discussion Paper — The Productivity of Working Hours, the nonlinear output curve in munitions factory data
• Sonnentag & Fritz (2007), Journal of Occupational Health Psychology — the Recovery Experience Questionnaire, psychological detachment, relaxation, mastery and control
• Kivimäki et al. (2015), The Lancet — long working hours and coronary heart disease and stroke, meta-analysis of more than 600,000 people
• Hagger et al. (2016), Perspectives on Psychological Science — preregistered multi-lab replication of ego depletion, effect size 0.04