**Participant 01 · Opening report · October 3, 2026**  
**Sample size: 1 · Protocol status: provisional · Publication status: participant-approved opening report**

## Reason for observation

**Participant report.** The participant is approaching 40. The participant has previously run ultramarathons and now wishes to return to consistent running as soon as practicable, with ambition, sustained effort, and continued enjoyment. The near-term stated event goal is the Seattle Marathon on November 29, 2026, a few days after the participant turns 40.

The participant presents with a question: how can this be done?

**Evidence status.** The material reviewed to date[1] does not establish a tested training progression for this particular combination of prior experience, reduced recent running, and a near-term marathon goal. In recognition of this limited evidence status, the study has several aims: document the participant's return to running; examine which evidence applies; record decisions made under uncertainty; and make the limitations of the searches and tools available for inspection. Curiosity is a working method: questions will be recorded, assumptions checked, and the plan revised when observations warrant it.

**Design.** One participant will be observed while following and revising a participant-selected training plan. Running, planning decisions, literature searches, and AI-agent involvement will be recorded. The laboratory tone is an editorial device. No institutional or clinical supervision is yet claimed.

## Questions entered before prospective results

1. How should the participant progress when the available evidence does not settle the participant's question?
2. Which findings alter the plan, and which apparent findings dissolve when their sources are inspected?
3. What prompts the participant to increase, repeat, reduce, or otherwise change planned activity?
4. Can ambition and enjoyment be maintained while the plan is repeatedly reconsidered?
5. What do AI agents contribute to the investigation, and what errors do they introduce?

**Temporal limitation.** Training has already begun. The initial activity entries will be reconstructed retrospectively. Their footers will identify the writing date and distinguish later interpretation from contemporaneous observation. An early run will not be made to look like a prospectively designed experiment.

## Baseline specimen: the available record

**Recorded activity.** Retrieved Strava summaries contain 179 running miles in 2023, 112 in 2024, and 59.1 in 2025. January through August 2026 contains 40.7 recorded running miles. These are account totals. They are not a complete inventory of everything the participant did.

**Recent records, with an unresolved classification discrepancy.** Weekly summaries report approximately 4.4, 9.4, 13.4, and 15.9 running miles during the recent September build. The final week is unfinished at the time of writing. The September 28 walk has been excluded. The September 23 classification conflicts between sources and remains under review; affected totals are provisional.

**Historical self-report.** The participant's 2017 account, [“the bigfoot proof”](https://danielsgriffin.com/2017/08/19/the-bigfoot-proof), describes completing the Bigfoot 200 and observing the participant's own responses during the event. This establishes a relevant history of ultramarathon participation and self-observation. It does not establish current readiness. Other public race records bearing the participant's name remain unattributed where identity is uncertain.

**Missing observations.** No specific current injury or pain detail has been supplied for this opening report. No finding of symptom absence is inferred. Missing notes will remain missing rather than being manufactured from pace, distance, or enthusiasm.

## Proposed procedure

**Judgment; participant-confirmed provisional status.** The existing plan proposes increasing long-run distance and weekly mileage, a cutback week, and a taper. It changed after a nine-mile run on September 22. A twelve-mile run was recorded on September 30. Completion of either distance is an observation, not validation of the remaining schedule.

The procedure is incomplete. Total mileage, cross-training, strength, routines, recovery, fueling and calorie requirements, and stretching require further consideration. The proposed peak long run and its corresponding weekly target also require reconciliation. No final protocol is implied by the presence of a table.

**Search procedure.** Each consequential recommendation will be examined for its basis: research, coaching convention, or judgment. Directly read findings will be distinguished from descriptions of findings supplied by another source or an agent. Applicability will be stated before a finding is used to support a decision.

**Access limitation.** If a source cannot be opened, the record will say so. Failure to access a source will not be reported as absence of research. Failure to find an applicable trial will not be repaired by substituting a study of a different question. The searches conducted so far do not constitute an exhaustive review.

## Instrument behavior

**Reported agent errors, with verification status retained.** Earlier accounts describe an agent declaring that the participant's base began only weeks earlier without inspecting the full history, assigning weekly days without asking, and building a recommendation on a study that did not test the relevant question. The precise study-to-recommendation error remains to be identified. The initial [Strava MCP partial-history bug report](https://danielsgriffin.com/2026/09/19/strava-mcp-partial-history) documents the retrieval problem: whole-history claims made from a partial activity history.

**Document-observed error.** The training log classified a September 28 walk as running and asserted that pace confirmed the classification. The participant confirmed it was a walk. The resulting mileage calculation has been corrected.

Agents are part of the apparatus under observation. Their errors will be retained alongside their corrections. The apparatus will not be granted exemption from the record.

## Recording and release

**Proposed measures.** Entries will retain activity date, distance, moving duration, elevation gain, calculated moving pace, stroller flag, participant-supplied comments, and planned versus actual activity. Plan overrides will include their stated reasons and the relevant plan version. Measures will remain distinct from participant accounts and agent interpretations.

**Release boundary.** Drafts require participant review before publication. Routes, GPS traces, start locations, clock times, child names, and family routines are excluded. Personal health details will be considered as specific proposed inclusions, not inferred from private records or approved in the abstract.

## Conditions of interpretation

**One-person observational evidence.** This record can describe decisions, expose search limitations, and generate questions. It cannot establish that the plan caused an outcome or demonstrate what works for other runners. Public observation may itself influence the participant's behavior.

Other returning runners are invited to record comparable fields and the questions those records raise. Agreement between logs will not be treated as proof. Differences may be informative without being resolvable.

**Formal engagement.** Readers who wish to formally engage with this inquiry are invited to [submit questions here](https://danielsgriffin.com/contact). Questions about the evidence, methods, missing observations, and alternative interpretations are welcome. This invitation opens a correspondence; it does not constitute enrollment in an institutional study.

The participant remains ambitious. The protocol remains provisional. Observation proceeds.

---

**[1] Scope of review.** the provisional training plan and training log, retrieved Strava summaries, the participant's earlier writing, and the search and sourcing difficulties recorded during planning. No systematic literature review has been completed. A recommendation previously attributed to a study remains unverified because the exact study and claim have not yet been identified. This statement describes the material actually reviewed, not a finding that no relevant research exists. Subsequent entries will record the question searched, sources opened, findings read directly, applicability, and access failures.

*Prepared with AI assistance. Initially written October 3, 2026, after training had begun; revised and approved by the participant for publication October 4, 2026. Earlier activity entries are retrospective and will retain explicit backdated footers.*