A small bet is a bounded action that makes uncertainty easier to see. A long horizon is a commitment to a question that deserves more than one attempt. I need both. Without the horizon, experiments become restless novelty. Without the small bets, a worthy intention can stay abstract for years.

The current question is simple: can I make a useful neighborhood atlas without turning other people's lives into material? A neighborhood atlas is the container. One walk, one sketch, or one invitation to correct a caption is a bet. None needs to prove that the whole project will succeed.

The two clocks

The short clock asks what I can learn by Friday. The long clock asks what kind of practice I want to have by spring. These clocks should talk, but they should not be forced to agree on every tick. An awkward first sketch can be a good weekly result if it exposes a problem the seasonal project needs to understand.

My seasonal promise is to stay with the question for twelve weeks. My weekly promise is to make the next action small enough that real evidence can arrive. “Explore community” is not an action. “Ask two people whether this map omits a place to sit” is one, provided the invitation is optional and the answer is not treated as representative of everyone.

Expected value is not the whole story

For a simple reversible decision, I sometimes write E[V]=pG−(1−p)L−CE[V] = pG - (1-p)L - C, where pp is the chance of a useful result, GG its value, LL the loss if it fails, and CC the cost of trying. The symbols force me to say what I mean by “worth it.” They do not turn uncertain guesses into measured facts.

The more important value can be learning. A trial that makes no attractive artifact might reveal that my research question was badly framed. I record that separately rather than inventing a number to make the equation look complete.

Vtrial=Voutcome+Vlearning+Voption−C V_{\text{trial}} = V_{\text{outcome}} + V_{\text{learning}} + V_{\text{option}} - C

Here an option is the ability to take a better next step without being obliged to take it. Keeping a project reversible often matters more than maximizing its first result. This is where Enough is a design constraint meets experimentation: a budget can protect the room needed for learning.

Decide the stopping rule first

Before starting, I write down the time limit, the evidence I hope to encounter, and the condition under which I will stop. Otherwise an interesting activity can quietly become an endless obligation. A stopping rule also prevents me from interpreting persistence as proof that the idea is good.

For the atlas, I will pause if I cannot explain what a participant gets from contributing. More data is not an answer to that question. Braiding Sweetgrass has made me more alert to the difference between receiving and extracting, though I do not want to flatten its richer account of reciprocity into a project-management slogan.

A record, not a scoreboard

My review has three lines: what happened, what changed my mind, and what I will try next. I do not tally a winning streak. The point is to accumulate better contact with the world, not evidence that I am an unusually productive person.

The current checklist lives in A season of small bets. When I lose the larger thread, I return to Walking as research: the place gets a vote. That is the horizon worth keeping.