Most fantasy analysis discusses which players to select. The larger effect comes from how a roster is distributed across positions, which is decided before any individual choice matters.
Points are only useful relative to a replacement
A position where every starter scores similarly produces little advantage, no matter how many points the top player accumulates.
What determines value is the gap between a player and the worst player who could plausibly start in his place, since that gap is what a manager actually gains by owning him.
Positions differ enormously in that gap, and the differences are stable enough across seasons to plan around rather than react to.
Scoring settings redraw the map
Small rule changes move whole positions in value. Awarding points for receptions rather than only yardage changes which player types are scarce.
Roster requirements do the same. Requiring an extra starter at a position immediately deepens the pool that gets used and shrinks the gap between the best and the ordinary.
This is why advice imported from a differently configured league is unreliable, and why the first step in preparation is reading the settings rather than the rankings.
Scarcity is created by the roster limit
Every manager has a fixed number of slots, so acquiring depth at one position necessarily means accepting less elsewhere.
Concentrating early selections in one position leaves later rounds to fill the rest, and the quality available late is what determines whether the concentration was affordable.
Structural strategies are essentially bets about how fast a position's quality declines, which is a question about the pool rather than about any player in it.
Variance is a structural property too
A roster built entirely on consistent producers has a high floor and rarely posts the score needed to beat a strong opponent.
A volatile roster loses more often against average opposition but wins some weeks nothing else could have won.
Which is preferable depends on the format, since a league that rewards total points across a season treats variance differently from one decided by weekly head-to-head results.
In-season decisions inherit the structure
A manager who allocated well has surplus at a position and can trade from it, while one who spread evenly has nothing anyone wants.
Injury exposure is also structural: concentrating value in few players makes a roster brittle in ways no waiver activity fully repairs.
The consequences of construction therefore persist for months, which is why it deserves more attention than the marginal choice between two similarly ranked players.