Differentials

The players almost nobody owns who are producing anyway.

Every dot is a player with 1,500+ minutes. Zone: under 15% owned with top-quartile output — hover any dot for detail.

How to read this, and why it matters

Rank is a race, not a score. If you and a million other managers all own a 60%-owned forward, his hat-trick moves nobody. Points from a widely owned player keep you level; points from one nobody owns are the only ones that actually move you up. That is the whole idea behind a differential.

The axes. Ownership runs left to right, output runs bottom to top. The interesting region is the top left: high output, low ownership. Everything in the bottom right is the opposite trap — widely owned and not producing, which costs you rank every week you hold it.

The shaded zonemarks under 15% ownership with top-quartile output. That threshold is a convention rather than a law, but it is a useful one: above roughly 15% a player is owned by enough of the field that his returns start being priced into everyone’s rank.

Switch the y-axis to Projected and the chart stops describing what already happened and starts showing where the model expects output to come from. That is the more useful view: a differential everyone can already see in the points column is usually about to stop being a differential.

On how many to carry. The common guidance is two to four at a time. The reasoning is straightforward — each differential is a bet that raises both your upside and your variance, and past a handful the variance is running your season rather than your judgement. Where you land in that range is a question about how much risk you want, and this page has no opinion on that.

2026/27 data. 1500+ minutes only, so a small sample cannot masquerade as a signal. Projections are estimates, not predictions — see how the model works for where it falls short.