Madden Draft Tool

Run the NFL draft outside Madden so we can apply richer trade and pick logic, then export results in a shape Madden can re-import. Local web app — Python backend, JavaScript frontend, opens in any browser.

What it does

• Loads a draft year's data from Files/<year>/ (BigBoard, current rosters, pick order, GM traits, position needs, pick value chart).
• Lets you (controlling the Steelers) draft your own players, sim other teams' picks, sim to your next pick, sim to a chosen round, or sim to any overall pick number.
• AI teams will make picks (and eventually trades) based on their own big board, GM traits, and team needs — see backend/logic.py for the contract; the heuristics are placeholder shells today.
• You can offer trade-ups to any team's pick; when their pick is up, you can review trade-down offers from AI teams.
• Exports three xlsx files for re-import / inspection:
  • DraftPickOutcome.xlsx — every pick + the player drafted.
  • DraftPicks_updated.xlsx — same shape as input, with new CurrentTeam ownership and SelectedPlayer filled in.
  • Trades.xlsx — chronological trade log (for testing).

Run it

Requires Python 3.10+.

macOS / Linux (bash/zsh):

python3 -m venv .venv
.venv/bin/pip install -r requirements.txt
.venv/bin/python app.py

Windows (PowerShell):

python -m venv .venv
.\.venv\Scripts\pip install -r requirements.txt
.\.venv\Scripts\python app.py

If PowerShell blocks the activation script with a security error, you can either run the commands above directly (they don't need the venv "activated" — they invoke the venv's pip and python by path) or loosen the policy for the current session with Set-ExecutionPolicy -Scope Process -ExecutionPolicy RemoteSigned.

Then open http://localhost:5050. Pick a year on the setup screen (today only TestFiles is present) and click Start Draft.

Adding a new draft year

Drop the source files under Files/<year>/ matching the names in Files/TestFiles/:

Files/2026/
  BigBoard.xlsx
  Player.xlsx
  DraftPicks.xlsx
  GMInfo.xlsx
  PositionNeeds.xlsx
  DraftPickValue.xlsx
  DraftMaxPerPosition.xlsx
  all_colleges.json   (optional)

Restart the server; the year shows up in the dropdown automatically. If the requested year folder is missing, the loader falls back to Files/TestFiles/.

UI overview

• On the Clock (top left): which team is picking now. When it's the Steelers, you get Check Trade-Down Offers and Offer Trade Up buttons.
• Last Selection (left): most recent player drafted.
• Team Needs (left): position weakness ranking for whoever is on the clock.
• Current Round (center): 4×8 grid of the visible round; highlights the on-the-clock pick (pulsing green) and Steelers picks (gold). Round dropdown switches the view.
• Mel Kiper Best Available (center): public consensus board. When Steelers are on the clock, each row gets a Draft button.
• Team Big Boards (right): private per-team rankings. Mostly for testing how AI logic ranks players.
• Top bar buttons: Sim Pick (one AI pick), Sim to My Pick, Sim to Round…, Full Draft Order (modal showing the entire draft as a 4-wide grid), Export (downloads all three xlsx files).

Project layout

app.py                Flask entry + HTTP routes
backend/
  data_loader.py      Reads xlsx/json from Files/<year>/
  draft_state.py      Pick order, on-the-clock, drafted players, trades
  logic.py            Decision functions (placeholder shells)
  exporter.py         Writes the three output xlsx files
templates/index.html  App shell
static/js/app.js      All frontend behavior
static/css/app.css    Custom styling layered on Tailwind CDN
Files/                Per-year source data + Exports/ (gitignored)

See CLAUDE.md for engineering notes (data quirks, team-ID-space mismatch, what's stubbed vs. real).

Status

The plumbing — data load, draft state, API, UI, export — is real and end-to-end. The decision functions in backend/logic.py are placeholders with full docstrings describing the real algorithm. They will be filled in iteratively without changing their signatures.