← Course home · Sprint 9A brief

Scrum Scoreboard

Shows whether each team follows the Scrum process on the way to its Product Goal. It is not a ranking.

Checked Sun 8 Nov 2026 00:07 SGT

Sprint 9B milestones

Brief: Sprint 9B: Did our change work? Results appear here shortly after each deadline.

TeamM1Tue 10 Nov 23:59M2Wed 11 Nov 23:59M3Wed 11 Nov 23:59M4Sat 14 Nov 23:59M5Sat 14 Nov 23:59
Team 1⏳ Not due yet⏳ Not due yet⏳ Not due yet⏳ Not due yet⏳ Not due yet
Team 2⏳ Not due yet⏳ Not due yet⏳ Not due yet⏳ Not due yet⏳ Not due yet
Team 3⏳ Not due yet⏳ Not due yet⏳ Not due yet⏳ Not due yet⏳ Not due yet
Team 4⏳ Not due yet⏳ Not due yet⏳ Not due yet⏳ Not due yet⏳ Not due yet
  1. M1 · Tue 10 Nov 23:59 SGT · Sprint 9B backlog on FigJam and in docs/backlog-9b.md
  2. M2 · Wed 11 Nov 23:59 SGT · docs/change.md (one rule per market)
  3. M3 · Wed 11 Nov 23:59 SGT · predictions/vW10_old.json and predictions/vW10.json
  4. M4 · Sat 14 Nov 23:59 SGT · vW10 rows in data/errors.csv (old, new, baseline) and the A/B chart on ui/index.html
  5. M5 · Sat 14 Nov 23:59 SGT · docs/did_it_work.md

Sprint 9A milestones

A result is fixed at its deadline and never changed. Arrival times are GitHub’s own server record of when the file reached main, in SGT.

TeamM1Tue 3 Nov 23:59M2Wed 4 Nov 23:59M3Wed 4 Nov 23:59M4Sat 7 Nov 23:59M5Sat 7 Nov 23:59M6Sat 7 Nov 23:59
Team 1✅ On time✅ On time❌ No submission✅ On time✅ On time✅ On time
Team 2✅ On time✅ On time✅ On time✅ On time✅ On time✅ On time
Team 3❌ No submission❌ No submission❌ No submission✅ On time✅ On time✅ On time
Team 4❌ No submission✅ On time✅ On time✅ On time✅ On time✅ On time
  1. M1 · Tue 3 Nov 23:59 SGT · Sprint 9 backlog on FigJam (cards for milestones 2–6, each with owner and “done” line)
  2. M2 · Wed 4 Nov 23:59 SGT · predictions/vW9.json in the common format (numbers only)
  3. M3 · Wed 4 Nov 23:59 SGT · data/old_predictions.csv (rows checked against original sealed vW4–vW8 files)
  4. M4 · Sat 7 Nov 23:59 SGT · engine/check_error.py + data/errors.csv (every error recomputed independently)
  5. M5 · Sat 7 Nov 23:59 SGT · error charts on ui/index.html (open and screenshot)
  6. M6 · Sat 7 Nov 23:59 SGT · docs/what_went_wrong.md (paragraphs quote their own error rows; one specific change)

Sprint 8 close-out (due Mon 2 Nov)

Required: docs/live-test.md written from the live test, and docs/peer-check-team-N.md for the team you checked (Practical A: Team 1 checks Team 2, Team 2 checks Team 3, Team 3 checks Team 1; Team 4 has no peer check).

Teamdocs/live-test.mdPeer check
Team 1❌ Not valid: first reached main Sun 1 Nov 03:16 SGT, before the test❌ No docs/peer-check-team-2.md
Team 2❌ No submission❌ No valid docs/peer-check-team-3.md (a blank template was committed under another file name)
Team 3❌ No submission❌ No docs/peer-check-team-1.md
Team 4❌ Not valid: first reached main Sat 31 Oct 18:41 SGT, before the test— Not required (Practical B)

No team has a valid Sprint 8 close-out record.

Team cards

Team 1

🟠 Drifting from Scrum

Milestones 4–6 on time, but three error rows use the wrong week’s close and the TNX review is missing.

Submitted work

M1 · Sprint 9 backlog

M2 · vW9 prediction (locked Wed 4 Nov 23:59)

Arrived on main Wed 4 Nov 12:57 SGT.

SPXNDXGOLDTNX
7620.082800039604.9

M3 · Old predictions table

No file at data/old_predictions.csv. A table was committed on Wed 4 Nov 12:57 at data/old_predictions/predictions.csv instead; it also puts DOWN and NaN in cells that should be blank.

M4 · Error script + data/errors.csv

First reached main Fri 6 Nov 01:12 SGT.

21 rows. I recomputed every error: 18 match. The three vW9 rows for SPX, NDX and TNX use Thursday 5 Nov closes instead of Friday 6 Nov (SPX 7,353.73 instead of 7,376.86; NDX 26,629.81 instead of 26,989.99; TNX 4.837 instead of 4.859). The file was last updated Fri 6 Nov 20:27 SGT, before Friday’s US session opened.

M5 · Error charts on ui/index.html

Opened from main as it stood at the deadline; screenshot below.

Team 1 error charts on ui/index.html, opened from main at the deadline

M6 · docs/what_went_wrong.md

First reached main Fri 6 Nov 13:17 SGT.

SPX and NDX paragraphs quote their own error rows and name a change (start from the latest close before the lock). The GOLD section lists errors without units and ends in a general list of factors rather than one specific change. There is no TNX section.

Product Goal

Stage 1
Data SourcesPhase 1
Stage 2
Agent AnalysisPhase 1
Stage 3
LLM SynthesisPhase 1
Stage 4
Human ScorePhase 1
Stage 5
Locked PredictionPhase 1
Stage 6
Delta MeasurementM3–M5
Stage 7
PrescriptionsM6

Sprint 9A builds Stage 6 (Delta Measurement) and Stage 7 (Prescriptions) of the Architecture Deep Dive.

Evidence: main at 4fbbd58 · FigJam section · Discord channel

Team 2

🟠 Drifting from Scrum

All six Sprint 9A milestones on time; two members were absent from the Sprint Review.

Submitted work

M1 · Sprint 9 backlog

M2 · vW9 prediction (locked Wed 4 Nov 23:59)

Arrived on main Wed 4 Nov 23:16 SGT.

SPXNDXGOLDTNX
75802805038904.95

M3 · Old predictions table

Arrived on main Tue 3 Nov 18:04 SGT. Every number matches the original sealed files. The vW4 row lists predictions/vW4.json as its source, but that file is at evidence/R1/sprint 4/predictions/vW4.json on main.

weektarget_dateSPXNDXGOLDTNX
vW42026-10-027640.030700.0
vW52026-10-097810.5311435.22
vW62026-10-167760.030450.04115.05.27
vW72026-10-2377502980041305.12
vW82026-10-3077502980041494.99

M4 · Error script + data/errors.csv

First reached main Wed 4 Nov 23:40 SGT.

21 rows. I recomputed every error from the team’s own numbers and checked every actual close: all 21 match (TNX in basis points).

M5 · Error charts on ui/index.html

Opened from main as it stood at the deadline; screenshot below.

Team 2 error charts on ui/index.html, opened from main at the deadline

M6 · docs/what_went_wrong.md

First reached main Wed 4 Nov 23:33 SGT.

All four markets quote their own error rows and name one change each. The GOLD change starts from Thursday’s close, which is not yet known at the Wednesday 23:59 lock, so it cannot be applied as written.

Product Goal

Stage 1
Data SourcesPhase 1
Stage 2
Agent AnalysisPhase 1
Stage 3
LLM SynthesisPhase 1
Stage 4
Human ScorePhase 1
Stage 5
Locked PredictionPhase 1
Stage 6
Delta MeasurementM3–M5
Stage 7
PrescriptionsM6

Sprint 9A builds Stage 6 (Delta Measurement) and Stage 7 (Prescriptions) of the Architecture Deep Dive.

Evidence: main at 992b701 · FigJam section · Discord channel

Team 3

🔴 Not following Scrum

Milestones 4–6 on time with the best-documented error file, but the vW9 prediction was not locked on main by Wed 4 Nov 23:59.

Submitted work

M1 · Sprint 9 backlog

No Sprint 9 backlog by the deadline; cards were added afterwards.

M2 · vW9 prediction

No predictions/vW9.json on main at the lock (Wed 4 Nov 23:59). Only a TNX evidence file under evidence/vW9/R6/ was merged.

M3 · Old predictions table

No data/old_predictions.csv on main at the deadline.

M4 · Error script + data/errors.csv

First reached main Fri 6 Nov 17:11 SGT.

21 rows. I recomputed every error and checked every actual close: all 21 match (TNX in basis points). Each row records its Yahoo source and when it was fetched, and marks entries that were not locked on time.

M5 · Error charts on ui/index.html

Opened from main as it stood at the deadline; screenshot below.

Team 3 error charts on ui/index.html, opened from main at the deadline

M6 · docs/what_went_wrong.md

First reached main Fri 6 Nov 17:16 SGT.

All four markets quote their own error rows and keep late predictions separate from on-time ones. SPX, NDX and TNX each name a change. The GOLD section proposes several changes at once, which makes it hard to tell which one worked.

Product Goal

Stage 1
Data SourcesPhase 1
Stage 2
Agent AnalysisPhase 1
Stage 3
LLM SynthesisPhase 1
Stage 4
Human ScorePhase 1
Stage 5
Locked PredictionPhase 1
Stage 6
Delta MeasurementM3–M5
Stage 7
PrescriptionsM6

Sprint 9A builds Stage 6 (Delta Measurement) and Stage 7 (Prescriptions) of the Architecture Deep Dive.

Evidence: main at c51cabe · FigJam section · Discord channel

Team 4

🟠 Drifting from Scrum

Milestones 2–6 on time; milestone 1 was missed.

Submitted work

M1 · Sprint 9 backlog

No Sprint 9 backlog by the deadline; cards were added afterwards.

M2 · vW9 prediction (locked Wed 4 Nov 23:59)

Arrived on main Wed 4 Nov 08:43 SGT.

SPXNDXGOLDTNX
75602830040504.92

M3 · Old predictions table

Arrived on main Wed 4 Nov 22:59 SGT. Every number matches the original sealed files; blanks are where the original file had no number (vW8 holds none).

weektarget_dateSPXNDXGOLDTNX
vW42026-10-0230550
vW52026-10-0978355.28
vW62026-10-167800302005.26
vW72026-10-2377202965041504.15
vW82026-10-30

M4 · Error script + data/errors.csv

First reached main Sat 7 Nov 22:06 SGT.

14 rows (weeks with no number are left out). I recomputed every error and checked every actual close: all 14 match (TNX in basis points).

M5 · Error charts on ui/index.html

Opened from main as it stood at the deadline; screenshot below.

Team 4 error charts on ui/index.html, opened from main at the deadline

M6 · docs/what_went_wrong.md

First reached main Sat 7 Nov 11:34 SGT.

The NDX and TNX sections quote their own error rows; NDX names one specific, testable change. The SPX section lists predictions rather than errors, and the SPX, GOLD and TNX changes are general rather than one specific change.

Product Goal

Stage 1
Data SourcesPhase 1
Stage 2
Agent AnalysisPhase 1
Stage 3
LLM SynthesisPhase 1
Stage 4
Human ScorePhase 1
Stage 5
Locked PredictionPhase 1
Stage 6
Delta MeasurementM3–M5
Stage 7
PrescriptionsM6

Sprint 9A builds Stage 6 (Delta Measurement) and Stage 7 (Prescriptions) of the Architecture Deep Dive.

Evidence: main at 681df9f · FigJam section · Discord channel

How the board is checked

Only three places count: the team repository’s main branch, the team’s own sections on the shared FigJam board, and the team’s Discord channel. Work on other branches, in other boards or in other chats is not seen.

Every “Team N - …” section on the FigJam board is read, not only the newest one. Arrival times come from GitHub’s server-side record of when work reached main, not from commit dates. Each update is a snapshot taken at the time shown at the top, and every evidence link points to the exact commit and FigJam section that was checked.

Scrum light: 🟢 Following Scrum = every check passes · 🟠 Drifting from Scrum = one check fails · 🔴 Not following Scrum = two or more checks fail, or a missed prediction lock, or no GitHub commit for 3 or more days.

Checks: Sprint Backlog visible (Sprint 9 cards on FigJam) · Daily Scrum (at least 10 messages in the team Discord channel in the last 48 hours) · Increment on main (team work reaching main during Sprint 9) · Sprint timebox kept (every milestone due so far on time) · Whole team at the Sprint Review (at most one absent at the latest practical) · Whole team involved this week (at least half the team active on GitHub or Discord).