Inviting nine universities to contribute lectures does not automatically create a nine-week curriculum.

The schedule may look impressive while the student experience remains fragmented. Topics can overlap. One problem set may be much harder than the next. Solutions arrive in different forms. A late contribution can leave an empty week that nobody owns.

The proposed NAPC-Online 2027 program distributes advanced training across nine universities. Each university would contribute one subject area during the weeks before the North America Championship.

The value of the program will depend on the system connecting those contributions.

Give each university a complete unit of ownership

A participating university should own a usable training module: one or two recorded lectures, six to ten curated problems, solutions or editorial references, and a trainer who can answer questions or support the review.

That boundary is large enough to preserve the trainer’s expertise. A geometry specialist can choose the sequence that makes the subject coherent. A university teaching combinatorial optimization can decide which results require proof and which techniques deserve implementation practice.

The boundary is also specific enough to verify. Seven days before publication, the organizing team can ask whether the lecture, problems, and solutions are ready. “We are participating” is not a deliverable.

Keep one learning cycle across all nine weeks

Teams need a stable rhythm even as the subject and trainer change.

During the week, students watch the lecture and work through the related problems on their own schedule. On Saturday, they enter a five-hour team contest that draws from a broader range of algorithms. The topic label should not reveal which technique solves the next problem.

After the contest, a short review examines selected solutions, common mistakes, and decisions that cost the teams time. The remaining work belongs to the teams through independent upsolving.

This cycle separates exposure from transfer. A student may understand a suffix automaton while watching a lecture. The mixed contest tests whether the team can recognize when a string representation is useful without being told the chapter name.

The organizing team owns the seams

No university sees the full program from its local module.

Someone has to remove unnecessary overlap, notice a missing prerequisite, and compare problem difficulty across weeks. The organizing team also has to select the Saturday contests, assign review leads, publish materials, and track late work before it reaches students.

These seams are where distributed programs usually become fragile. Every contributor can complete their own piece while the combined experience still fails.

Ownership should therefore be explicit. University coordinators track commitments and deadlines. Curriculum reviewers inspect topic coverage and problem quality. A platform owner handles publication and technical logistics. The operating group resolves exceptions that do not fit one module.

The names matter less than the absence of gaps between them.

Design the recovery path before a module is late

A nine-week program has nine opportunities for a dependency to slip.

Trainers have research deadlines, teaching responsibilities, and travel. A strong operating plan assumes that one contribution may become unavailable. It identifies backup trainers, permits topic adjustments, and sets a confirmation deadline early enough to replace a missing module.

The wrong recovery is to let uncertainty continue until publication week. That transfers the cost to students and to the person who has to assemble emergency material.

A credible deadline names the next action. If a university cannot confirm its trainer, topic, and week by the decision date, the program activates a backup. The schedule remains intact without pretending the original commitment still exists.

Separate delivery from learning

The program first has to ship. Nine confirmed owners, completed modules, usable solutions, eight contest simulations, and on-time publication are delivery outcomes.

They do not prove that teams improved.

For learning, I want to examine weekly participation, problems solved during the contests, problems completed through upsolving, and changes across simulations of comparable difficulty. Participant feedback can identify lectures that were clear but poorly connected to the practice set, or problems that were challenging without teaching the intended judgment.

The final simulation is a progress checkpoint. It is not a clean experiment: contests differ, teams miss weeks, and the strongest students enter with different baselines. The measurements can still tell us where to inspect.

Nine university names are inputs. One training program exists when a team can move through every week without seeing the organizational boundaries that produced it.