Seasonal experiments

The resolution that only works in summer

Every January, millions of people set the same resolution: exercise in the morning before work. The logic seems sound. You read the research, designed the routine, set the alarm. On January second, the alarm fires at 6 AM. Outside, it is dark. It will remain dark for another ninety minutes. The temperature is hovering around freezing. Your bedroom is warm. Your circadian system, which has spent the past two months adjusting to shorter days and longer nights, is screaming that this is not morning — this is the middle of the night. You override the signal, force yourself out of bed, and drag through an exercise session that feels nothing like the effortless morning runs you remember from last July.

By February, you have missed more sessions than you have completed. By March, you have quietly abandoned the experiment. You file it under "lack of discipline" and wait for the guilt to fade.

Then May arrives. The sun rises before you do. The air is mild. One morning you wake naturally at 6:10, feel alert, and decide to go for a run. Nobody told you to. There was no resolution, no alarm override, no battle with the duvet. The run feels good. You do it again two days later. By June you are running three mornings a week without thinking about it. By August it feels like a part of your identity. By November it is gone again.

This is not a story about willpower. It is a story about testing a seasonally dependent behavior without accounting for the season.

The biology you are ignoring

The idea that human physiology and behavior remain constant across seasons is a modern fiction sustained by artificial lighting, climate-controlled buildings, and the cultural assumption that a good system should work in all conditions. The research tells a radically different story.

Till Roenneberg, the chronobiologist whose work on internal time has reshaped how we understand circadian rhythms, demonstrated that human chronotype — the biological timing of sleep, alertness, and peak performance — shifts measurably with photoperiod, the length of daylight exposure in a given day. In his studies of over 55,000 participants, Roenneberg found that people sleep longer in winter and shorter in summer, that the midpoint of sleep shifts later in winter months, and that these changes are not cultural preferences but biological responses to light availability. Your body is not the same machine in December as it is in June. It is a seasonally modulated system operating under different parameters.

Thomas Wehr's landmark research at the National Institutes of Health reinforced this finding. When Wehr exposed subjects to natural photoperiods — fourteen hours of darkness in winter conditions — their sleep architecture changed dramatically. Sleep expanded to an average of about ten hours, organized into two distinct blocks with a period of quiet wakefulness between them. Melatonin secretion extended, hormonal profiles shifted, and mood regulation changed in measurable ways. The subjects were not sick or depressed. They were expressing the seasonal biology that artificial light had been suppressing. Your January alarm at 6 AM is not just inconvenient — it is fighting a physiological system that has recalibrated for winter.

Norman Rosenthal's clinical work on Seasonal Affective Disorder further illuminated how profoundly seasons shape behavioral capacity. While full SAD affects a subset of the population, Rosenthal identified a much broader phenomenon he called "subsyndromal SAD" — a milder seasonal shift in energy, motivation, and mood that affects a significant portion of people living at temperate latitudes. You may not have clinical seasonal depression, but there is strong evidence that your motivational baseline, your energy availability, and your tolerance for effortful behavior fluctuate with the seasons. Designing behavioral experiments that ignore this fluctuation is like designing a bridge that ignores wind load — it might stand in calm weather, but the first storm will reveal the engineering failure.

Circannual rhythms and the experimental calendar

The seasonal shifts Roenneberg, Wehr, and Rosenthal documented are expressions of circannual rhythms — roughly year-long biological cycles that modulate physiology and behavior across months. Gregory Young's research on circannual rhythmicity showed that these cycles operate even in the absence of external seasonal cues, suggesting they are endogenous biological programs, not merely reactions to weather. Your body has an internal calendar, and it is running whether you consult it or not.

For behavioral experimentation, this means something specific and actionable: the same experiment run in different seasons is not the same experiment. When you pilot a morning routine in June (Piloting new routines), you are piloting it under one set of biological conditions — long photoperiod, early cortisol awakening response, high baseline energy, robust mood. When you attempt to continue that routine in December, you are running it under a different set of conditions — short photoperiod, delayed cortisol peak, lower baseline energy, contracted mood range. The routine did not change. The organism running it did.

Buxton and Marcelli's epidemiological work on seasonal variation in physical activity confirmed this at the population level. Analyzing accelerometer data across thousands of participants, they found that physical activity levels peak in summer months and trough in winter, with the magnitude of seasonal variation depending on latitude, age, and baseline activity level. The pattern held even among highly motivated exercisers. This is not a discipline problem. It is a biological-environmental interaction that no amount of motivational self-talk will override.

The implication for your experimental practice is that you need a seasonal testing protocol — a deliberate structure for running behavioral experiments across seasons rather than assuming that a result obtained in one season will replicate in another. A behavior that passes a two-week pilot in April has been validated for spring conditions. It has not been validated for July heat, October darkness, or January cold. Each season represents a distinct experimental context, and responsible experimentation requires testing in each one.

Designing the seasonal experiment protocol

A seasonal experiment protocol builds on the pilot framework from Piloting new routines but adds a temporal dimension. Where the standard pilot asks "Does this routine work under current conditions?", the seasonal experiment asks "How does this routine perform as conditions change, and what adaptations does it need to remain viable?"

The protocol has three layers. The first layer is seasonal awareness — explicitly identifying which environmental variables your behavior depends on. Every behavior has a dependency profile: outdoor exercise depends on temperature, daylight, and precipitation; morning waking depends on photoperiod and circadian phase; social behaviors depend on schedule density, which often varies seasonally with holidays, school calendars, and work cycles; creative practices may depend on what Mihaly Csikszentmihalyi described as the environmental conditions most conducive to flow states, which for many people shift with seasons as ambient light, thermal comfort, and the availability of unstructured time change across the year. Before you run a seasonal experiment, list every environmental variable your target behavior touches. This is your seasonal risk register.

The second layer is variant design. For each season — or more precisely, for each major shift in your environmental conditions — you design a variant of the behavior that accounts for the changed context. The goal is not to maintain identical behavior across seasons. That is the mistake most people make, and it is the reason January resolutions fail. The goal is to maintain the function the behavior serves while adapting its form to seasonal conditions. If your morning run serves the function of cardiovascular exercise, mood regulation, and transitioning from sleep to alertness, then a winter variant might substitute indoor cycling or a high-intensity bodyweight session that serves the same functions without requiring you to run in darkness and freezing temperatures. The function is stable. The form is seasonally adaptive.

The third layer is transition testing. The most dangerous moment for any seasonal behavior is not deep winter or peak summer — it is the transition between seasons, when conditions shift gradually and your awareness lags behind the change. Kahneman's peak-end research helps explain why: you remember the peak of your summer performance and the endpoint of your autumn decline, but you are largely blind to the slow environmental drift that connects them. By the time you notice the behavior is struggling, the seasonal shift has been underway for weeks and you have accumulated a string of missed sessions that your brain interprets as personal failure rather than environmental mismatch. Transition testing means scheduling explicit re-evaluation checkpoints at the equinoxes and solstices — four times per year — where you assess whether your current behavioral variant still matches your current environmental conditions and switch to the next variant if it does not.

Memory distortion and the seasonal self-comparison trap

There is a psychological trap embedded in seasonal behavior change that deserves its own examination, because it is responsible for more abandoned experiments than any single environmental factor.

When your morning routine collapses in November, you do not compare your current performance to an objective baseline. You compare it to your memory of peak performance — which, thanks to the peak-end rule Kahneman documented extensively, is a distorted highlight reel of your best summer days. You remember the morning you ran five miles effortlessly and felt invincible at your desk by 8 AM. You do not remember the three mediocre summer runs that surrounded it, the Tuesday you skipped because it was too hot, or the week your schedule disrupted the pattern. Your memory has compressed an entire season of variable performance into a single idealized image, and you are comparing your struggling November self to that fiction.

This comparison produces a gap — between remembered peak and current reality — that your brain interprets as evidence of personal decline. "I used to be able to do this easily. Now I cannot. Something is wrong with me." The seasonal experiment protocol breaks this cycle by replacing memory-based comparison with data-based comparison. When you have tracking data from each season, you can see that your summer performance was not the effortless peak you remember but an average with variance, and that your winter performance, while lower, may still meet appropriately calibrated winter success criteria. The problem was never your November self. The problem was comparing your November self to an imaginary June self that never existed.

Seasonal experiments as epistemic humility

There is a deeper principle operating here that extends beyond behavioral logistics. Designing seasonal experiments is an act of epistemic humility — an acknowledgment that you do not fully know yourself in all conditions. You know your summer self reasonably well, because that is the self that produces your peak experiences and most confident self-assessments. But your winter self, your transitional-season self, your disrupted-schedule self — these are versions of you that you have spent less time studying and more time judging.

The seasonal experiment framework treats each version of you as a legitimate experimental subject deserving of its own hypothesis, its own success criteria, and its own evaluation. Your winter self is not a degraded version of your summer self. It is an organism operating under different environmental parameters, and it may need different behavioral protocols to achieve comparable functional outcomes. When you design a winter variant of a behavior rather than trying to force the summer version through a hostile season, you are respecting the biological reality that Roenneberg, Wehr, and Young documented — and you are converting seasonal variation from a source of shame into a source of experimental data.

The Third Brain

An AI assistant is particularly valuable for the seasonal experiment protocol because it does not experience seasons and therefore does not share your seasonal biases.

When you are designing seasonal variants, describe your current behavior, the environmental conditions you expect in the coming season, and the function the behavior serves. Ask the AI to generate three to five variant designs that preserve the function while adapting the form to anticipated conditions. "My current behavior is a thirty-minute outdoor run at 6:15 AM. In three months, sunrise will be at 7:45 AM and average morning temperature will be minus five Celsius. The function is cardiovascular exercise, mood elevation, and a transition from sleep to work readiness. Generate five winter variants that preserve these three functions." The AI will produce options you might not have considered — timed indoor circuits, stairwell intervals, dawn-simulation light exposure paired with indoor cycling — because it is not anchored to the form of the behavior the way you are.

At equinox and solstice checkpoints, share your tracking data from the current season alongside historical data from previous seasons. Ask the AI to identify seasonal patterns in your performance data that you might miss due to the memory distortions discussed above. "Here is my exercise tracking data from the past twelve months. Identify any seasonal patterns in frequency, duration, or self-rated difficulty. Compare my actual winter performance to my actual summer performance and tell me the magnitude of the seasonal effect." The AI provides the longitudinal perspective that your peak-end-biased memory cannot, turning seasonal review from a subjective self-assessment into a data-driven analysis.

You can also use the AI to build a seasonal transition calendar — a forward-looking schedule of variant switches, checkpoint dates, and anticipated environmental shifts — so that seasonal transitions happen proactively rather than reactively. The goal is to switch variants before the current one starts failing, not after you have accumulated two weeks of missed sessions and the accompanying self-criticism.

From seasonal to social

You now have a framework for designing behavioral experiments that account for the most pervasive and least acknowledged source of experimental variation: the seasons themselves. You understand that your biology shifts across the year, that behaviors validated in one season need re-testing in others, that variant design preserves function while adapting form, and that memory distortion makes seasonal comparison unreliable without tracking data.

But you have been running all of these experiments alone. Every hypothesis, every pilot, every seasonal variant has been designed, executed, and evaluated by a single experimenter — you. The next lesson examines what happens when you run behavioral experiments with a partner or group, where shared data, mutual accountability, and the collision of different experimental results accelerate learning beyond what any solo experimenter can achieve.

---

Sources

Roenneberg, T. (2012). Internal Time: Chronotypes, Social Jet Lag, and Why You're So Tired. Harvard University Press.

Wehr, T. A. (1992). In short photoperiods, human sleep is biphasic. Journal of Sleep Research, 1(2), 103-107.

Rosenthal, N. E. (2006). Winter Blues: Everything You Need to Know to Beat Seasonal Affective Disorder (Revised ed.). Guilford Press.

Csikszentmihalyi, M. (1990). Flow: The Psychology of Optimal Experience. Harper & Row.

Young, G. (2012). Circannual rhythms. In Encyclopedia of the History of Psychological Theories (pp. 210-213). Springer.

Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.

Buxton, O. M., & Marcelli, E. (2010). Short and long sleep are positively associated with obesity, diabetes, hypertension, and cardiovascular disease among adults in the United States. Social Science & Medicine, 71(5), 1027-1036.

Tucker, P., & Gilliland, J. (2007). The effect of season and weather on physical activity: A systematic review. Public Health, 121(12), 909-922.