A Year of Night Sky Data

I have a spreadsheet with 274 rows. Each row is a night. Each night has 18 columns. That's 4,932 data points about one woman, one deck, one bay, one dog, and one year of chronic pain on the Oregon coast.

This is the article I've been building toward since the first one. Nine articles of mechanism — why cold works, why darkness works, why sound works, why the dog works, what happens when none of it works. All of it supported by peer-reviewed research, all of it framed by Toni's experience, all of it honest about the limitations.

Now the data. Not "here's what it felt like." Here's what happened, measured, over time.

I want to be upfront about what this is and what it isn't. This is a personal dataset — one person, unblinded, uncontrolled, and collected by someone who deeply wants the intervention to work. Every bias you can name applies. I've tried to account for what I can. What I can't account for, I'll name.

The Year in Numbers

Toni went outside on 274 of 348 nights — a 78.7% adherence rate across twelve months, including Oregon winter. The 74 missed nights break down as: 31 high-wind nights, 18 multi-day flare nights (day 3+), 14 rain-too-heavy nights, and 11 nights she chose not to go for reasons the data doesn't capture.

Is −1.7 Points Real?

This is the question that matters. Not "is it statistically significant" — with 274 paired observations, almost anything is significant. The question is: does a 1.7-point average drop on a 0-10 pain scale represent a meaningful change?

The answer, according to the clinical literature, is yes.

A landmark study by Farrar et al. (2001), analyzing over 2,700 patients across ten clinical trials, established that a reduction of approximately 2 points or 30% on the 0-10 Numerical Rating Scale represents a "clinically important difference" — the threshold where patients report feeling "much better."¹

A 2004 study by Salaffi et al. refined this further for chronic musculoskeletal pain: a 1-point reduction (or 15%) represents the minimal clinically important difference (MCID) — the threshold between "no change" and "slightly better." A 2-point reduction corresponds to "much better."²

Our mean delta of −1.7 sits between "slightly better" and "much better." And 58.4% of sessions crossed the 2-point threshold — the "much better" line. This is not a cure. But it's in the range where clinical trials of approved medications declare success.

The Seasonal Pattern

Oregon gives you four different decks across a year. Here's what happened on each:

Fall is the best season. Not because of any single variable but because fall on the central Oregon coast delivers the optimal combination: cold enough for vagal activation, dry enough for clear skies, calm enough for the bay soundscape, and dark enough early enough that 10 PM sessions work as well as 3 AM ones. Summer is the weakest — warmer temperatures reduce the cold exposure effect, and the late sunsets mean true darkness doesn't arrive until after 10 PM.

The Five Variables That Matter

Eighteen columns. A year of data. And the honest conclusion is that most of the variables I track don't predict the outcome very well. Moon phase: nothing. Moon visibility: nothing. Humidity alone: weak. Cloud cover alone: weak.

Five variables account for nearly all the explainable variance in pain delta:

• Wind speed. The strongest single predictor. Below 5 mph: mean delta −2.2. Above 12 mph: mean delta −0.4. Wind isn't just uncomfortable — it disrupts the sound environment, triggers startle responses, and converts cold exposure from analgesic to aversive.
• Temperature. Sweet spot: 36-44°F. Warm enough to stay for 30+ minutes. Cold enough to activate cold-induced norepinephrine release and vagal pathways. Below 32°F, session duration drops and the cold becomes the dominant sensation. Above 52°F, the cold mechanism weakens.
• Duration. The effect inflects between 20 and 30 minutes. Under 20: mean delta −0.8. Over 30: mean delta −2.3. Something happens in that window — possibly the time needed for parasympathetic dominance to fully engage, for the cortisol/norepinephrine shift to stabilize, or for respiratory entrainment to complete.
• Baseline pain. Sessions starting at 4-7 show the most consistent drops. Below 4, there's less room to improve. Above 8, the system is too sensitized to respond to the intervention. The deck works best in the middle range — exactly where most chronic pain lives on most days.
• Kona contact. Pressed: −2.0 mean delta. Near but not touching: −1.3. The 0.7-point difference held across all seasons. The dog isn't optional.

What Didn't Matter

I tracked moon phase for twelve months. New moon, waxing, full, waning — no correlation with pain delta. None. Moon visibility didn't matter either; clear-sky sessions performed the same whether the moon was up or below the horizon.

Humidity showed a weak effect — higher humidity nights had marginally better deltas — but this confounds with fog, which confounds with temperature and wind. I can't separate it cleanly.

Barometric pressure affected Toni's baseline pain (as the research predicts³) but didn't meaningfully change the session delta. Falling-pressure days start higher and end higher, but the drop is similar. The intervention works on top of whatever the barometer is doing.

Tide state showed a real but small effect — incoming tide nights averaged 0.3 points better than outgoing — consistent with Toni's observation about rhythmic pulse versus continuous draw. Not enough to plan around, but enough to notice.

Putting It in Clinical Context

I keep asking myself: how does this compare to actual treatments? Not because the deck is a treatment — it's not. But because context matters. If a 1.7-point drop is trivial compared to what medication achieves, that changes the story.

An overview of Cochrane Reviews on non-pharmacological interventions for fibromyalgia — the highest level of evidence synthesis — found low certainty evidence that exercise and CBT produce "clinically relevant" improvements, defined as a minimal important difference of 0.5 on a 0-10 scale or a standardized mean difference above 0.2.⁴

A 2023 review of multimodal non-pharmacological therapies confirmed that combined interventions — multiple modalities stacked together — outperform individual therapies for chronic pain. CBT plus exercise beats either alone. Acupuncture plus physiotherapy beats either alone. The pattern is consistent: combinations work better than single interventions.⁵

What Toni does on the deck is, by any reasonable definition, a multimodal intervention: cold exposure + darkness + natural acoustic environment + companion animal interaction + sensory reduction + outdoor nature immersion. Each component has individual research support. Combined, they appear to produce an effect (−1.7 points per session) that exceeds the MCID threshold and sits within the range of what approved pharmacological and non-pharmacological treatments achieve in clinical trials.¹²

But — and I cannot say this clearly enough — this is one person's data. Clinical trials test interventions across hundreds of people with randomization, blinding, and control groups. We have none of that. The comparison is illustrative, not evidential. It tells you the effect size is in the right range. It does not tell you the effect is real in the way a trial would.

The Nature Dose

A 2025 meta-analysis on nature dosage and mental health outcomes found that for people with diagnosed conditions, interval nature exposure of up to 600 total minutes showed a dose-response effect — more time, better outcomes. The minimum threshold for benefit was approximately 120 minutes per week.⁶

Toni averages 37 minutes per session, 4.7 sessions per week. That's roughly 174 minutes of outdoor nature exposure weekly — comfortably above the 120-minute threshold and within the dose-response curve. She's not an extreme case. She's right in the range where the population-level research predicts benefit.

Nature prescription programs — where doctors literally prescribe time outdoors — are emerging across the country, supported by evidence that time in nature lowers blood pressure, reduces nervous system arousal, and improves immune function.⁷ Toni's deck sessions are, functionally, a self-prescribed nature dose. What she's doing isn't unusual. The only unusual part is that she's doing it at 3 AM.

What I Got Wrong

A year of data collection teaches you what you didn't know when you started. Here's my list:

• I overweighted the stars. The first articles were about the sky. The data says the sky doesn't matter as much as the ground-level environment — temperature, wind, sound. Stars are beautiful and they produce awe, but the analgesic effect correlates more strongly with what's happening below the horizon than above it.
• I underweighted the dog. Kona's contact is the second-strongest predictor after wind speed. I didn't even start tracking her position until month three. Three months of missed data on the most consistent variable in the dataset. That's my biggest methodological regret.
• I assumed linearity. More cold ≠ more benefit. More time ≠ more benefit. The relationships are curvilinear — there's an optimum, and past it, the effects flatten or reverse. The deck isn't a dose-response ramp. It's a window.
• I didn't track the next day. I have before-and-after for each session. I don't have pain at noon the following day, or sleep quality that night, or activity level the next morning. The acute session data is clean. The downstream effects are invisible. If I could start over, I'd add one more data point: pain at noon tomorrow.
• I was afraid of the null results. Moon phase: no effect. Humidity: marginal. Half the variables I track don't predict anything. I resisted this for months. A good dataset includes null results. They tell you where not to look, which is as valuable as where to look.

What the Numbers Can't Show

The spreadsheet tracks pain on a 0-10 scale. It does not track what changed inside Toni's relationship with her own body over the course of a year. I can't put a number on that. But I can describe it.

A year ago, Toni's pain was something that happened to her. She endured it. She fought it. She debugged it. She tried to fix it like broken code. The pain was the enemy and her body was the battlefield and every day was either a win or a loss.

Now, the pain is something she lives with. Not accepted in the passive sense — she still tries things, she still tracks data, she still has bad nights and bathroom-floor mornings. But the relationship has changed. The pain is no longer in charge of whether she goes outside. She goes outside with the pain, not despite it or to escape it. That's what psychological flexibility looks like in practice, and no column in my spreadsheet captures it.

I started this blog to document an intervention. What I documented was a person learning to live in her body again. The deck was just where it happened.

The 2024 Cochrane overview concluded that the most important outcomes in fibromyalgia may not be pain intensity at all — they may be disability, mood, and quality of life.⁴ The research community is moving toward the idea that pain acceptance, not pain reduction, is the better target for chronic conditions. Our data shows a consistent pain reduction. But what it can't show — and what matters more — is that Toni has a practice now. A relationship with the night and the bay and the cold and the dog that didn't exist a year ago. A reason to go outside. Something that is hers.

That's not in the spreadsheet. It doesn't need to be.

The bay is still there. The deck is still there. The stars come out when they come out and don't when they don't. Kona is always at the door first. Samba supervises from the window. The weather station logs another row. And tomorrow night, if the wind is calm and the air is cold and the tide is coming in, Toni will say a number at the back door, and we'll add it to the data, and she'll go outside to lie in the dark and listen to the water and hold onto the dog and wait for the shift that comes when you're still long enough, cold enough, and quiet enough to let your body remember what safe feels like.

274 nights. One year. One deck. One bay. One dog. One person.

The numbers say it works. The person says it matters. Both are true. Neither is the whole story.