Regression is the most abused word in fantasy football. It often means “I do not think last season happens again,” delivered with a chart and no mechanism. A real regression case separates the work a player earned from the outcomes that bounced his way. This module declares a stored running-back chart but no current player projection, ADP snapshot, or year-over-year study, so it will not label present-day fades or publish remembered rates as evidence.
The chart shows volume and scoring, not destiny
The configured scatter reads carries and rushing touchdowns from stored weekly player stats rows. It can show that scoring varies around rushing volume. It cannot prove a player's future touchdown rate, establish a league baseline, or identify a current regression candidate. Those claims need a declared window, player cohort, and source-backed comparison across seasons.
The useful lesson is narrower: opportunity and outcome are not the same thing. Carries create chances. Touchdowns are outcomes attached to a smaller set of those chances. A projection should account for both without pretending last season's conversion repeats automatically.
Separate opportunity from efficiency
Opportunity includes snaps, carries, routes, targets, and work near the goal line. Efficiency includes what the player produced with those chances. Both matter, but they answer different questions. Opportunity describes the job. Efficiency describes how well the job paid in the observed window.
A strong regression case appears when the price assumes an extreme outcome will repeat while the underlying role stayed ordinary or weakened. A weak regression case is simply “the number looks high.” Good players can sustain better-than-average efficiency. The analyst still has to show which part of the result was stable and which part depended on a thin tail.
Touchdowns need role context
Do not erase touchdowns from the projection. Ask how they were created. Repeated short-yardage assignments are role evidence. A handful of long scores may say more about a player's explosiveness, but they also create a wider range. Team scoring opportunities, offensive line play, and the player's place in the red-zone rotation all matter.
The honest update uses current opportunity and regresses uncertain conversion toward a defensible baseline drawn from a declared dataset. This module does not include that baseline, so it does not print one. A number from memory is not improved by calling it regression.
Yards per carry can hide the shape of the work
An average can be lifted by a few long runs. That does not make the long runs fake; explosiveness is part of the player. It does mean the projection should inspect the distribution instead of treating every carry as equally productive. Median gains, success rate, contact context, and blocking can help describe whether the efficiency was broad or concentrated.
Again, the conclusion needs a source. “Big plays will disappear” is not analysis. State what changed in the role, what the historical comparison is, and what range remains plausible if the explosive outcomes cool.
Five questions before calling a fade
- Did the role grow, hold, or shrink? Start with snaps, routes, carries, targets, and high-value work.
- Which outcomes drove the finish? Separate repeatable participation from volatile scoring events.
- Has the surrounding offense changed? Personnel, coaching, and team quality can alter opportunity.
- Is the draft price charging for the peak? Regression matters only when the market has priced the repeat.
- What evidence would cancel the fade? A larger role can offset cooler efficiency.
Age is context, not a verdict
Older backs deserve an availability and workload review, but an age label alone is not a model. Career touches, recent health, current competition, and the offense's plan matter. Any age curve must name its data and window. Otherwise it becomes another way to turn a hunch into a decimal.
The same caution applies to young backs. Youth does not guarantee growth. A player can be young, efficient, and stuck in a narrow role. Price the job that exists, then leave room for change.
How to use regression in a draft
Compare the current role projection with the market price. If the room is paying for last season's scoring peak while the opportunity case is flat, demand a discount. If the role expanded, a lower efficiency forecast can still support the player. The ADP value-tier guide helps place that decision against comparable options.
Use the live rankings for current player inputs and the RB1 backtest method for the evidence standard. Regression is not a prediction that success must reverse. It is a reminder to stop billing the same lucky outcome twice.
