# Dumb Luck?

> A thinking toy for separating a decision from its outcome. Describe one everyday call - your own, a friend's or relative's, a contestant's on a show, or anyone's - write one to six notes on what was known or done before it, and say whether it went well or badly. One metered run scores the call on four factors by what was knowable at the time, then crosses that score with the outcome: good call, dumb luck, bad beat or own goal.

URL: https://dumb-luck.skillsafe.ai/
API: https://dumb-luck.skillsafe.ai/api.html

## What runs where

- Free, in the browser: a scratch-off ticket (four scratch panels with clovers for each factor's score, a prize box with the total, the outcome and the verdict, and a prize table of the four verdicts), your own scoring of the four factors by hand, a "what if it went the other way" flip that shows the verdict move while the decision score stays put, the page's free read of your notes, SVG, PNG, Markdown, CSV and JSON downloads, and a share button that hands the ticket PNG to the device's share sheet where it supports files. A re-scored call names which factor moved since the last run.
- Metered, one task (`ticket`): each factor (information, time, options, odds) scored 0 to 3 from your notes with the note behind it quoted word for word, the total out of 12 (sound at 7 or more), the verdict, the weakest factor (ties go to odds, then information, options, time), a short reason, the part that was luck, one habit for next time and a closing line. Runs on gpt-terra.
- The page re-checks every reply: four factors in order, quotes word for word, a factor with no note scores 1, the total is the sum, the verdict follows the total and the outcome, and the weakest factor really is the lowest.

## Contract

- Input (all strings): task (`ticket`), who (`me`, `friend`, `contestant`, `someone`), tone (`wry`, `warm`, `deadpan`), call, notes (`n | note` lines), outcome, went (`well` or `badly`), page_facts, optional clip_note and retry_note.
- Reply: one JSON object with status (`ok` or `declined`), reason, call_label, factors (four, in the order information, time, options, odds; each with factor, score, note, note_quote, reading), decision_total, decision (`sound` or `shaky`), went, verdict (`good-call`, `dumb-luck`, `bad-beat` or `own-goal`), verdict_reason, luck_part, weakest, next_time, closing.
- Declined: calls about someone hurt or ill, medical care, crime, weapons, drugs or drink, and money staked or risked (betting, gambling, lotteries, trading, investing, loans or debt).

## Sources

- Built on the SkillSafe skill @pmprompt/thinking-in-bets: https://skillsafe.ai/skill/@pmprompt/thinking-in-bets (judging decisions on process, not results).
- Saved tickets live in the person's own SkillSafe account only when they press save.

This is a thinking toy for fun, not professional, financial, medical or legal advice.