--- title: "Pedestrian Crosswalk Mystery" description: "Winfree's outdoor group lab: watch a real pedestrian crossing, record what happens before explaining it, pool the class's logs, and turn the oddities into questions that can be tested." type: Activity tags: [course, student-facing, problem, section-3, observation, fieldwork, hypothesis-testing] status: stable problem: section: 3 session: 12 identification: unknown kind: lab generated: by: "claude/fable-5-1" at: "2026-09-16T00:00:00Z" sources: - id: winfree-handout-2001 resource: "https://web.archive.org/web/20030111064604/http://eebweb.arizona.edu/Faculty/Winfree/handout_479.htm" title: "ECOL 479/579 The Art of Scientific Discovery: course handout with retrospective syllabus (archived)" author: "A. T. Winfree" - id: fhwa-hawk-2010 resource: "https://www.fhwa.dot.gov/publications/research/safety/10042/10042.pdf" title: "Safety Effectiveness of the HAWK Pedestrian Crossing Treatment (FHWA-HRT-10-042)" author: "Kay Fitzpatrick and Eun Sug Park" - id: nyt-placebo-buttons-2004 resource: "https://web.archive.org/web/20101110072424/http://www.nytimes.com/2004/02/27/nyregion/for-exercise-in-new-york-futility-push-button.html" title: "For Exercise in New York Futility, Push Button (The New York Times, 27 Feb 2004; archived copy)" author: "Michael Luo" - id: aosd-syllabus resource: "https://github.com/tyson-swetnam/aosd/blob/main/docs/assets/aosd_syllabus.pdf" title: "The Art of Scientific Discovery: original course syllabus (PDF)" author: "Arthur T. Winfree" --- # Pedestrian Crosswalk Mystery

This work is licensed under a Creative Commons Attribution 4.0 International License. *[Section 3](../section3.md), session 12, where the lab begins, alongside [Sums of Integers](sums-of-integers.md); the observations are pooled and discussed in session 13.* !!! abstract "The problem" Reconstructed from the syllabus. Winfree's own text is lost; what follows is built from the lab's name and its place in the schedule. Nothing records what the "mystery" was, or even whether the crossing had a signal. Go, as a group, to a pedestrian crosswalk on a busy street. Spend at least half an hour there over the coming week, at more than one time of day, and keep a log in your GamesWorth book. 1. **Describe it before explaining it.** Write down what is actually there: the markings, the signs, the kerbs, whatever machinery there is — and the people. Time events to the second, and note what happened just before each change. 2. **Watch the people as closely as the hardware.** Where do they cross relative to the painted lines? Who waits and who steps out? How do two opposing streams of walkers get through each other without colliding? When do drivers stop, and when do they roll through? 3. **Find the mystery.** Somewhere in your log there should be something you cannot yet explain. It may be mechanical, it may be human. Let the log tell you which. 4. **Pool the data.** Next session, combine everyone's timings and counts. Look for regularities no single observer could see, and for disagreements between observers, which are data too. 5. **Question, hypothesize, test.** Write your rules for the crossing as if-then hypotheses, say what would refute each, and go back and look. The product is not "the answer" but a clean record of what happened, a list of the questions it raises, and the explanations you tried to eliminate. ## Why it is in the course The lab straddles a seam in the schedule. It begins in session 12, the last of Section 2 on creative blocks, and is discussed in session 13, the first of Section 3, "Observations and Questions". It is the first exercise whose raw material is not a puzzle on paper but the street outside. The habits it drills are named elsewhere in the syllabus. Session 4 asks for "facts before explanations of facts" and for "distinguishing things we know vs only imagine"; a street corner demands both at once, because almost everyone arrives believing they already know how a crossing works. Pooling matters as much as watching: nobody can watch the traffic, the crossing and the walkers at once, so the logs come back incomplete and contradictory, and Section 3 is where disagreements become evidence. ## Where it comes from This is Winfree's own field exercise, not a puzzle borrowed from anyone. It survives as two lines of the schedule: the lab's name in session 12, and a line in session 13 telling the class to discuss its outdoor observations of the crosswalk. The problem sheet students were handed was never published. The archived course handout that carries the schedule is a plain reading copy with no working links, and Winfree's other surviving pages — his lab site and his eighteen "Adventures in Discovery" columns — say nothing about a crosswalk. The syllabus marks group labs "on the day they begin in class", which is why this one spans two sessions. ??? tip "Hints" - Time everything. A log without seconds cannot tell an effect from a coincidence. - Include controls. Compare a quiet hour with a busy one; if there is something to press, compare the intervals when someone presses it with those when nobody does. - Count people, not just machinery. Where they walk and who ignores the markings are observations as good as any timing. - Keep three columns: what you saw, what you inferred, and what you assumed because a sign said so. Everyone else acts on the third column too. - Ask what is not there. An event that never happens, a car that never stops, a stretch nobody uses: absences are observations. ## Sources - **A. T. Winfree**, *ECOL 479/579 The Art of Scientific Discovery: course handout with retrospective syllabus* (2001), archived copy — [Wayback Machine](https://web.archive.org/web/20030111064604/http://eebweb.arizona.edu/Faculty/Winfree/handout_479.htm){target=_blank} 🔓 - **Kay Fitzpatrick and Eun Sug Park**, *Safety Effectiveness of the HAWK Pedestrian Crossing Treatment*, FHWA-HRT-10-042 (2010) — [PDF](https://www.fhwa.dot.gov/publications/research/safety/10042/10042.pdf){target=_blank} 🔓 - **Michael Luo**, "For Exercise in New York Futility, Push Button", *The New York Times*, 27 February 2004 — [Wayback Machine](https://web.archive.org/web/20101110072424/http://www.nytimes.com/2004/02/27/nyregion/for-exercise-in-new-york-futility-push-button.html){target=_blank} 🔓 - **Arthur T. Winfree**, *The Art of Scientific Discovery: original course syllabus* (2001) — [PDF](https://github.com/tyson-swetnam/aosd/blob/main/docs/assets/aosd_syllabus.pdf){target=_blank} 🔓 !!! note "How sure are we that this is Winfree's problem?" The identification is **unknown**. That the class went outdoors to a pedestrian crossing and pooled the results in the next session is certain from the syllabus. The location is not recorded, the "mystery" is not described, and the syllabus does not even say the crossing was signalized. All four readings below are reconstructions. - **Watching the people rather than the machinery** (medium confidence): the exercise is named for the *pedestrian* crossing, and this fits the course's rule that its exercises "depend as little as possible on knowledge of any particular subject area" and "require no lab setup". - **Inferring an unfamiliar signal's rules** (low-medium confidence): Tucson developed the HAWK — High intensity Activated crossWalK — beacon in the late 1990s, and its sequence was novel in 2001. But nothing ties Winfree's lab to a beacon, and traffic engineering is the specialist knowledge the course says it avoids. - **Testing whether a push button does anything** (low-medium confidence): a 2004 New York Times report found more than 2,500 of New York City's 3,250 walk buttons deactivated. That reporting postdates the course by three years. - **A rhythm-and-timing observation** (low confidence): two signals drifting in and out of phase, in the spirit of Winfree's work on coupled oscillators. The syllabus speaks of one crossing, and nothing supports this reading. --- *Back to [Section 3](../section3.md) · [All problems](index.md) · [The schedule](../syllabus.md#section-3-observations-and-questions)*