The idea is pretty simple: public pages for stats and session history, plus a small admin area for recording buy-ins, gross cashout results, actual paid-out cash, and notes. The data sits in a CSV audit log, so everything is append-only and easy to follow later.

What it does

all-time leaderboard
per-session pages
per-player stat pages
admin login for recording events
open / closed session tracking
actual cash payout tracking with paid_out events
debt repayments and write-offs with debt_repayment / writeoff events
session and player charts
CSV import / export from the admin page
append-only entries.csv ledger instead of overwriting old rows

Stack

Built with a pretty lightweight setup:

Python
Flask
Jinja templates
Chart.js
plain CSS
CSV event log for storage

Ledger model

The app treats data/entries.csv as the source of truth.

Each row is an event, not a final snapshot. Instead of editing an old row, I append another one. That keeps rebuys, corrections, payouts, and session state changes visible in the log instead of hiding them behind edits.

Current event types:

buyin
front
debt_repayment
writeoff
cashout
paid_out
rollover_in
payout_carry_in
rollover_out
note
session_open
session_close

A few examples:

another buyin for a rebuy
another cashout if chip counts are corrected later
a paid_out event when someone is actually settled up
a debt_repayment event when a front is repaid outside poker
a writeoff event when a front will not be collected
a note event for bookkeeping context
session_open / session_close to mark whether a game night is still live

Accounting in the app keeps poker results separate from banker cashflow:

poker investment is buyin + front + rollover_in
poker net is cashout - poker investment
real cash in is buyin + debt_repayment
real cash out is paid_out
rollover_out settles the source session without counting as cash out
rollover_in enters play in the destination session without counting as cash in
payout_carry_in records prior-session value carried into a later payout; it increases the destination payout due without counting as poker investment or cash in
writeoff resolves a receivable without counting as cash in

Older ledgers may still contain paid, front_collected, or front_writeoff. The app reads those historical names as aliases for paid_out, debt_repayment, and writeoff.

It is still just a small side project, but I wanted the event model to stay clean enough that the numbers are easy to trust and the history is easy to read back through.

CSV format

Main file:

data/entries.csv

Header:

id,created_at,session_id,session_date,player_name,event_type,amount_cents,note,actor

Amounts are stored in cents to avoid floating-point issues.

Running it locally

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env
python app.py

Then open:

http://127.0.0.1:8000

Environment values

The app reads these from .env:

SECRET_KEY
ADMIN_USERNAME
ADMIN_PASSWORD

Example:

SECRET_KEY=change-this
ADMIN_USERNAME=admin
ADMIN_PASSWORD=change-me

Notes

A few choices here were deliberate:

no database for now
no user accounts, just one admin login
public-facing stats pages, admin-only controls
CSV backup before importing a replacement ledger

If I ever decide to take it further, the first real upgrade would probably be moving the storage layer to SQLite while keeping the rest of the app roughly the same.