Pod Documentation Packet

Read the pod system without going through the website repo

This packet is the public guide to how pods run. AggieQuant's main GitHub is for the website and shared club systems. Each pod keeps its own repo, write-up packet, and handoff trail for pod-specific work.

Distribution Model

Public packet here, pod ownership elsewhere

The model is simple: the site explains the system, pod teams own their own repos, and polished pod write-ups live in Overleaf, PDFs, or another pod-owned docs surface.

Website repo for site maintenance Pod repos for pod code Overleaf or PDF for polished packets Handoff docs for continuity

System Guide

How the pod system actually works

A pod is a small team that inherits a starting point, ships one semester-bounded artifact, and leaves the next cohort a cleaner restart path.

01

Pod family

The family is the broad domain: prediction markets, volatility, microstructure, data infrastructure, or another durable lane of work.

02

Starter kit

The starter gives members a first runnable path so the semester begins from inherited structure instead of blank-page paralysis.

03

Semester deliverable

Each pod still needs one bounded artifact: a memo, demo, notebook, dashboard, simulator, or benchmark result that proves the pod moved.

04

Handoff

The handoff is the real continuity mechanism. Without it, even a strong semester output gets lost on turnover.

First Semester

How underclassmen ramp without getting lost

New members should not be asked to invent structure on day one. The right first semester is inheritance, one narrow contribution, and a better trail for the next person.

01

Read the current state

Start with the README, the last handoff, and the last successful run before proposing new directions.

02

Own one narrow slice

Good first work is one importer, one metric, one baseline, one chart, one setup fix, or one cleaner demo surface.

03

Update the trail

Research logs, README fixes, and short assumption notes are what keep the pod from resetting next semester.

04

Leave a restart kit

By semester end, a new member should be able to see what worked, what failed, and what to do first next time.

Repo Model

What should live in the main repo, and what should spin out

AggieQuant's main repo owns the website and shared club infrastructure. Pod teams own sibling repos for pod-specific code, research, tests, and documentation.

Main Repo

Website and shared club systems

The AggieQuant repo owns the public site, shared member-facing tools, and lightweight starter patterns that explain the system clearly.

Examples: website pages, shared UI surfaces, club infrastructure, public-facing docs packet.

Pod Repos

Separate sibling repos for real pod work

Pod teams create sibling repos similar to the prediction-markets repo model. Pod-specific work does not live permanently in the website repo.

Examples: `aggiequant-prediction-markets`, `aggiequant-options-vol-lab`, `aggiequant-market-microstructure`.

Doc Surfaces

Packets, PDFs, and Overleaf write-ups

Polished pod write-ups live in Overleaf, exported PDFs, or another pod-owned docs surface. The site links to those packets instead of sending members into the website GitHub repo to read them.

Rule of thumb: public index on the site, polished write-up in the pod's chosen doc surface, code in the pod's own repo.

Worked Example

The prediction-markets pod is a strategy example, not the product roadmap

Use prediction markets to teach scoring, calibration, contract wording, and event-market discipline. Keep the online prediction exchange separate as its own AggieQuant product.

Prediction Markets

Why this pod is still a strong model

The starter is useful because the evaluation standard is obvious. Members learn proper scoring, compare model probabilities against market probabilities, and work on systematic event-market research without confusing that work with the club's separate exchange product.

Good first question: are our probability forecasts better calibrated than the market prices we collected?
Good first artifact: a public demo with Brier score, log loss, calibration bins, and one honest memo.
Good long-term path: real market importers, category-level calibration, and cleaner tradable-edge analysis.
Separate feature boundary: the online prediction exchange is its own product. It can use ideas from this pod without becoming part of pod scope.
Stage 1

Teach the scoring loop

Start with question wording, probability forecasts, resolution discipline, and honest evaluation against a baseline or market.

Stage 2

Improve research depth

Add real data, better importers, and cleaner separation between paper edge and tradable edge.

Stage 3

Spin out mature work

When the pod has real depth, it should live in its own sibling repo with its own tests, write-up, and roadmap.

Documentation Flow

The packet is the public index

The site introduces the system, pod packets hold the polished write-ups, and sibling repos hold the pod-specific code. That is the default documentation model.

Back to Pods