Week 10 Lab: Name the objective

Validate a supplied two-car decision set, rank it for the team, then re-rank the same outcomes for car 14 and account for the trade

10 May 2027

Two hours, and you submit inside the session. This lab carries 2% of the course, one of the eleven. You prepare beforehand; validation, both rankings and the objective-change report happen here.

What leaves the room is one decision set ranked twice over unchanged evidence: first by expected constructors’ points across both cars, then by car 14’s expected race points. You also report what the change costs the team, car 14 and car 22.

Before the session#

Install the supplied codebase and run its public tests until they pass. Read the plan, path, objective and report schemas. Make sure you can identify the two car records inside one plan/path pair, the probability attached to the path and the field where the objective names its stakeholder.

Then trace one public path by hand. Look up each car’s points from its finishing position and add them. Bring that working with you. The locked plans, joint outcomes and rankings are not available before the checkpoint.

What is supplied#

The points lookup models one ordinary race reaching at least 75% of scheduled distance under FIA 2026 Section A, Issue 03. Every plan, path, probability, finish and expected value is synthetic. The FIA supplies neither this fixture nor the objective used to rank it.

In the session#

1. Validate the decision set (25 minutes). Require exactly car 14 and car 22 once in every plan/path pair. Accept only P1–P10 finishes, reject two cars in the same position on one path, require every plan to cover the same complete path set, and require the path probability mass to equal 1.00. Prove that invalid input produces no ranking.
2. Rank for the team (30 minutes). Translate each finish through the points lookup, add both cars’ points on every path, weight those totals and rank all three plans by expected constructors’ points. Retain the complete scorecard, not only the winner.
3. Change stakeholder, not evidence (25 minutes). Rank the same immutable paths by car 14’s expected points. Add a test that fails if any path, weight or joint outcome changes between objectives.
4. Account for the trade (25 minutes). Compare the team-selected plan with the car-14-selected plan. Calculate the changes in expected team points, car 14’s expected points and car 22’s expected points, then say which named interest gains and which loses.
5. Unlock and report (15 minutes). Run the locked tests, reproduce both rankings and emit the regulation version, synthetic boundary, candidate set, objectives, stakeholders, selections, deltas and limitations.

What you hand over#

• validators and tests that prove malformed or incomplete input cannot produce a ranking;
• expected-points functions for the team and each named driver;
• two complete rankings over unchanged evidence;
• the objective-change comparison and final report.

The final instrument carries this decision forward: it must state what it optimises and whose interest that serves.