Sprint 3 — vW3 — Automate, and Show Your Work
Last week you did it by hand. This week two things change: at least one manual task gets automated, and — starting now — your team's actual Scrum process (backlog, Definition of Done, task status) has to be visible on the shared FigJam board, not just in your repo.
Sprint 1 and 2 didn't require board activity, and that wasn't penalized. From Sprint 3 onward, it is. This isn't decoration — it's how Scrum actually works: a backlog and a Definition of Done that only exist in someone's head, or buried in a GitHub file, aren't visible to your team or gradable as process. They have to live on the board.
The Sprint Goal
The Backlog — All Six Roles, In Concert
Write your team's Sprint 3 backlog as individual cards in your team's To Do column on the shared FigJam board (not a duplicate — your section of the one master file). Each card needs a one-line Definition of Done. Move cards through To Do → In Progress → Review → Done as the week progresses. At Monday's presentation, log attendance and who spoke to which role.
Script the lookup of at least one calendar/seasonality pattern or macro data point instead of finding it manually.
Write a script that fetches weekly OHLC for some or all of the 16 instruments from Sprint 2 via an API or library, instead of reading Yahoo Finance by hand.
Take your team's Sprint 2 manual figures and your new Sprint 3 automated output for the same data, and check them against each other. Where they disagree, say why.
Give a human score (1–5) to how much you'd trust this week's automated output vs. last week's manual output, with one sentence of reasoning. Then build a page that reads from R2/R3's automation output (a committed data file) rather than hand-typed numbers — if the automation reruns with new figures, the page should update without anyone touching the HTML. Include at least one chart (Chart.js via CDN, or hand-rolled SVG — numbers in a table alone isn't new this sprint), and show this week's automated figures side-by-side with last week's manual ones so the score has something concrete to react to.
Own the actual commit/PR for this sprint's automation code.
Prediction Challenge — New This Sprint
prediction_vW3.md in your repo by end of day Tuesday, timestamped by the commit itself so nobody can quietly edit it later.
Saturday morning (Singapore time), not Friday: the NYSE closes 4pm ET Friday, which lands around 4–5am Saturday SGT — so the real close isn't visible here until Saturday. Check then, compare your prediction to the actual close, and compute your error as a percentage: |predicted − actual| / actual × 100. Commit the result to the same file. Closest team wins bragging rights at Monday's Sprint Review — no grade attached, this one's just for fun and to make the automation feel real instead of academic.
Source of record — use this one, so every team is scored against the same number: the official S&P 500 close as reported by Yahoo Finance's ^GSPC historical data page. If it's unreachable when you check, fall back to Google Finance's SPX figure — and if you do, say so in your commit message, since the two can differ by a few cents.