The board

Lines nobody set by hand.

A play-money sportsbook for a 12-team dynasty league. Every spread, total and moneyline is priced by simulating the rest of the season a few thousand times — then betting runs through a Discord bot. No real money, anywhere.

Standings

units

Loading…

Futures

simulated

Loading…

Last settled

Loading…

Bettors

Who is actually any good.

Return on stake rather than units won, because a bankroll flatters whoever bets biggest. Average price is the tell — it separates a grinder on short favourites from someone hunting longshots.

Leaderboard

by net units

Loading…

By market

where the money goes

Loading…

Teams

Twelve rosters, ranked.

Title odds come from the simulation. The intangibles below them are, with one partial exception, measuring nothing at all — which is rather the point of having them.

Power index

simulated

Loading…

Intangibles

entirely made up

Loading…

Coaching is points above the projected optimal lineup — part real skill, part players simply going off. Defense is how far below projection opponents land, and home field is your boost at home plus how badly visitors play at your place. Your opponent's score has nothing to do with you, so those two are pure noise: simulate 200,000 games with zero real advantage and home teams still win 50.19%. Over one season the error bar is about ±5%, so anything from 45% to 55% means nothing. Measured consistently, though, which is all a bragging right needs.

Against the number

from closing lines

Loading…

History

The permanent record.

Every bet, line and result is archived before each season reset. Champions, all-time bankrolls and career records will live here.

2026

in progress

Seasons

archive

The 2026 season is the first on the book. Once it settles, champions and career records appear here — the archive is already collecting every line and bet in the meantime.

How it works

built solo

Pricing

Each team gets a scoring mean and variance — projections early, real results as they land. The season plays out thousands of times to produce every price on this site.

Correlation

Two legs on one game aren't independent. The book measures how often they land together and shortens the price to match, instead of multiplying odds and overpaying.

Settlement

Scores arrive Tuesday at 3am, bets grade themselves, bankrolls update and results post. Lines reopen at 9am. Nobody touches it.

The stack. A Discord bot on Cloudflare Workers handles betting. Google Apps Script and Sheets hold the ledger and run the simulation. Sleeper supplies rosters, projections and scores. This page reads a cached endpoint off the same Worker.

Why simulate. A real book moves lines against the money it takes. With a dozen bettors there is no money to read, so prices have to come from somewhere honest — a model of the league itself. Every number here is a probability first and a price second.

The part worth reading. Same-game parlays are where sportsbooks quietly make their margin. In fantasy the maths differ from real football: with no shared game script, a spread and a total are nearly uncorrelated when two teams have similar variance, and the correlation is driven by the gap in their volatility. So the book measures it per matchup rather than borrowing a number from the NFL — favourite-and-over gets shortened, favourite-and-under gets boosted. A fixed shading factor gets that backwards.