Who should I start comes down to three inputs: projected scoring rate, expected volume, and the gap between them. Start the player with the higher expected fantasy points, not the higher ceiling. Expected points equal per-touch or per-opportunity value multiplied by projected opportunities, adjusted for opponent and game script.

That answer works as a search snippet because it is actionable and falsifiable. The rest of this article explains the machinery behind it, when it breaks, and how to apply it in minutes rather than hours.

The mechanism behind a start/sit call

Every fantasy start decision is an expected-value problem. Expected fantasy points equal the sum, across every possible outcome, of that outcome's point value times its probability. A player projected for 14 points is not "better" than a player projected for 12 in any mystical sense; the 14-point player simply has a probability distribution whose mean is higher. Start decisions should maximize the mean, not the maximum.

Volume drives the mean more reliably than efficiency does. Targets, carries, and passing attempts are sticky week to week; yards per carry and touchdown rates are noisy. A receiver who earns 9 targets per game will usually outscore one who earns 4, even if the second player is more explosive per touch. This is why professional projection systems weight recent opportunity share heavily and efficiency lightly.

Opportunity quality is the second lever. A target near the end zone is worth more than a target at midfield, because touchdown probability rises steeply as the offense approaches the goal line. The same logic applies to carries: a running back who gets 15 carries between the twenties may score less than one who gets 10 carries with four inside the 10-yard line. Depth of target and red-zone share are the two opportunity-quality inputs that matter most.

Finally, game script modifies volume. Teams trailing by multiple scores pass more; teams leading run more. A projection that ignores the expected point spread will misprice both passing volume and rushing volume. This is why the same player can be a strong start in one matchup and a weak one in another without any change in talent.

Plain version: start the player the math says will touch the ball more, in better spots, on a team whose game plan feeds him.

Quantified context: thresholds that separate starts from sits

The numbers below are general decision thresholds used in projection work, not guarantees. Treat them as starting points that shift with league size and scoring format.

A useful rule: in a 12-team league with standard scoring, the replacement-level starter at each position produces roughly 10 to 12 points per game at quarterback, 9 to 11 at running back and wide receiver, and 6 to 8 at tight end. If a player's projected mean sits at or below that replacement line, the start decision is close to a coin flip and matchup factors should decide it.

Opportunity thresholds worth memorizing:

Wide receiver: 7 or more targets per game is a stable starting baseline in most formats. Running back: 12 or more carries plus 3 or more targets signals a three-down role. Tight end: 5 or more targets per game places a player in the startable tier. Quarterback: 30 or more pass attempts plus rushing volume separates streamers from starters.

Efficiency thresholds are weaker signals. A yards-per-carry figure above 5.0 over a full season is rare and tends to regress; a single-game mark above 8.0 is almost always noise. Do not start a player because of one explosive game.

Red-zone usage is the highest-leverage number. A player with 30 percent or more of his team's red-zone targets or carries has a touchdown pathway that volume alone does not capture.

Edge cases that break the simple model

The model fails when the inputs themselves are unstable. Injuries are the clearest case: a player listed as questionable with an unspecified designation has a probability distribution wide enough that his mean is unreliable. In that situation, the correct move is usually to start the healthy player with the lower ceiling.

Weather is a second edge case. Wind above 20 mph suppresses passing efficiency and deep attempts; heavy rain increases fumble risk and reduces downfield throwing. These effects are real but smaller than most managers assume, typically shifting projections by a few percent rather than several points.

Vegas-implied team totals are the most underused input. A team with an implied total of 27 points will score more often than one implied at 17, and fantasy points follow scoring. A player's individual projection should always be anchored to his team's expected scoring environment, not his season average.

The counterintuitive detail: in close matchups, the correct start is often the player with the lower ceiling and the higher floor. When your opponent's lineup is projected to score within a few points of yours, variance helps the underdog and hurts the favorite. If you are the favorite, bench the boom-or-bust player for the steady one. Most managers do the opposite and lose close games they should have won.

Red flags that should override any projection:

A player downgraded in practice on Friday or Saturday. A reported snap-count restriction after injury. A teammate returning from injury who historically splits the role. A late-week trade or coaching change altering the depth chart. A game with a weather forecast of sustained wind above 20 mph.

Common misconceptions about start/sit decisions

The biggest misconception is that matchups dominate. Opponent defensive rankings are computed over a full season against a schedule of varying quality, and they regress toward the mean. A defense ranked last against wide receivers in September may be average by December. Matchup matters, but it is a modifier on volume, not a substitute for it.

The second misconception is that "start your studs" is always correct. It usually is, because studs have high means. But a stud with a questionable designation, a snap restriction, or a brutal game script can have a mean below a healthy replacement. The slogan is a heuristic, not a law.

The third is that past performance in a specific matchup predicts future performance. One game is a sample size of one. A player who scored three touchdowns against a defense last season did so under different coaches, personnel, and game conditions.

Practical actions: a start/sit checklist

Run this sequence in order. It takes about 10 minutes per lineup and replaces most of the guesswork.

Confirm availability. Check the final injury report and any pregame reports. A player not confirmed active is not a start. Set the baseline. List each starter's projected mean from at least two independent sources and average them. If the sources disagree by more than 3 points, treat the projection as uncertain. Check the environment. Note each player's implied team total and the expected point spread. Fade passing volume in games with a spread above 10 points in either direction unless the player is the primary target. Apply thresholds. Compare each player against the position thresholds above. A player below threshold needs a strong matchup or red-zone role to justify a start. Adjust for risk preference. If you are favored, prefer the higher floor. If you are an underdog, prefer the higher ceiling. Quantify this by comparing the 25th-percentile outcomes, not just the means.

The threshold reconciliation matters here: the replacement-level line, the opportunity thresholds, and the red-zone share threshold often disagree. When they do, weight opportunity thresholds first, because volume is the most stable input. Use replacement-level lines to decide close calls and red-zone share as a tiebreaker.

Limitations of projection-based starting

Projections are estimates with real error. Published fantasy projections typically carry a standard error of several points per player per week, which means a 12-point projected player and a 14-point projected player are statistically close. Do not treat small projection gaps as decisive.

The model also assumes stable roles, and roles change without warning: a coaching change, a trade, or an injury to a teammate can reshape a depth chart in a week. Projection systems lag those changes by one or two weeks.

Finally, no framework removes the inherent variance of a single game. A well-reasoned start can lose to a poorly reasoned one on any given Sunday. The goal is to make decisions that win more often over a season, not to win every week. General educational information only; this is not a diagnosis, treatment plan, or investment recommendation, and no outcome is promised. Consult a qualified professional for personal decisions, and contact local emergency or crisis services if you are in immediate danger.