How do I decide who to start in fantasy football each week?

Start the player with the higher weekly projection unless you have a specific reason to override it. Reasons that qualify: news the projection has not seen, such as an injury ruling or depth-chart change; your need for a high-variance or low-variance outcome given your record; or a role change you can name concretely. Reasons that do not qualify: last week's score, draft position, or a general sense that a player is due.

How much does matchup matter for start/sit decisions?

Less than most people adjust for. A strong defence turns a good player into a slightly less good one; it rarely turns a startable player into a bench player. Matchup carries real weight mainly in extreme cases and in the split between what a defence concedes through the air versus on the ground. Overriding a three-point projection gap on matchup alone is usually the wrong call.

Should I start a player coming off a bad game?

Usually yes, if the projection still supports it. A single game is one noisy observation, and weekly fantasy scoring varies enormously for the same player in an unchanged role. Reacting to last week's score is one of the most common ways lineups get worse. The exception is when the bad game reflects something structural you can name — a reduced snap share, a lost starting job, or a role that has genuinely changed.

How often should I change my fantasy lineup during the week?

Set it early from the weekly rankings, then change it only for new information. The information that matters arrives late: injury designations toward the end of the week and the inactives list shortly before kickoff. Revisiting close decisions repeatedly does not add information, it only gives recency bias more opportunities to act.

How to make weekly start and sit decisions

Start the player with the higher weekly projection unless you have a concrete, specific reason to override it — and the reasons that genuinely qualify are narrower than most people assume. The most useful thing to know about start/sit decisions is that the majority of them are between players whose projections are within a point or two, where the choice barely matters and no amount of deliberation will improve it. The decisions worth your time are the few each week where the gap is large or where your situation makes the projection the wrong question.

Most start/sit questions are not real questions

If two players project within roughly a point of each other, the projection is not telling you to start one over the other. It is telling you they are indistinguishable at the resolution anyone can measure. Weekly fantasy scoring is extremely noisy — the same player in the same role can post 4 points and 24 points in consecutive weeks without anything having changed.

This is worth internalising because the alternative is spending an hour on a decision worth a fraction of a point in expectation, and then feeling responsible for the outcome. You were not responsible. You made a coin-flip and the coin landed.

The place to spend attention is the decisions with a real gap — three points or more of projection — and the ones where something about your specific situation changes the calculation.

Reasons that genuinely justify overriding a projection

News the projection has not seen. Projections refresh on a schedule; injury reports, inactives and depth-chart changes do not. If a player's backfield partner was ruled out an hour ago, you know something the numbers do not, and you should act on it. This is the single most common legitimate override and it is why checking inactives before kickoff is worth more than any analysis.

Your record and your position in the standings. If you are chasing a playoff spot and need a large week rather than a good one, the player with more variance is the better start even at a slightly lower projection. If you are protecting a lead, the reverse. Projections give you an expectation; only you know which shape of outcome you need.

A role you believe has changed and the numbers have not caught up with. This is legitimate but it is also where most bad overrides come from, so hold yourself to a specific claim: a new starter, a changed snap share, an offensive line returning. "He is due" is not one of these. Neither is last week's score, which is one noisy observation and worse information than the projection you are overriding.

Reasons that feel compelling and are not

Recent scoring, on its own. A player who scored 22 last week and 6 the week before has told you very little. Single-game results move people far more than they should, in both directions, and chasing them is the most reliable way to make your lineup worse.

Matchup, beyond a point. Defensive strength matters, but it matters less than the size of the adjustment most people make for it. A strong defence turns a good player into a slightly less good player; it rarely turns a starter into a bench player. Where matchup does carry real weight is in extreme cases and in the pass-versus-run split of what a defence gives up.

Name recognition and draft position. What you paid for a player in August is a sunk cost and tells you nothing about this week. This is the bias that keeps underperforming early picks in lineups long past the point where the projections stopped supporting it.

A workable weekly routine

Set your lineup from the weekly board early in the week, so you have a default that is not made under time pressure. Note the two or three spots that are genuinely close, and leave them alone until you have to decide.

Check the injury designations late in the week and the inactives list before the first kickoff of each slot. This is where nearly all of the real information arrives, and where a lineup gets meaningfully better or worse.

Then stop. The remaining decisions are close because they are close, and revisiting them repeatedly does not add information — it just gives recency and worry more chances to change your mind. Our weekly boards publish projected points and the value-over-replacement figure for every projected player, so the size of the gap between two options is visible rather than implied, which is usually enough to tell you whether the decision deserves any more of your evening.

Common questions

How do I decide who to start in fantasy football each week?: Start the player with the higher weekly projection unless you have a specific reason to override it. Reasons that qualify: news the projection has not seen, such as an injury ruling or depth-chart change; your need for a high-variance or low-variance outcome given your record; or a role change you can name concretely. Reasons that do not qualify: last week's score, draft position, or a general sense that a player is due.
How much does matchup matter for start/sit decisions?: Less than most people adjust for. A strong defence turns a good player into a slightly less good one; it rarely turns a startable player into a bench player. Matchup carries real weight mainly in extreme cases and in the split between what a defence concedes through the air versus on the ground. Overriding a three-point projection gap on matchup alone is usually the wrong call.
Should I start a player coming off a bad game?: Usually yes, if the projection still supports it. A single game is one noisy observation, and weekly fantasy scoring varies enormously for the same player in an unchanged role. Reacting to last week's score is one of the most common ways lineups get worse. The exception is when the bad game reflects something structural you can name — a reduced snap share, a lost starting job, or a role that has genuinely changed.
How often should I change my fantasy lineup during the week?: Set it early from the weekly rankings, then change it only for new information. The information that matters arrives late: injury designations toward the end of the week and the inactives list shortly before kickoff. Revisiting close decisions repeatedly does not add information, it only gives recency bias more opportunities to act.