Field guide
How points allowed is computed, and what it will not tell you
What the number is
Points allowed to a position is the plainest statistic on this site. Take every completed stat line from every game a defense has played. Keep the ones belonging to players at the position you are looking at, on the other team. Score each of those lines in your format. Add them up. Divide by the number of games that defense has played. That is the figure in the "Allowed / g" column, and there is no model, no adjustment and no weighting anywhere in it.
Read the rest
Read the rest
The scoring is the same scoring the rest of the site uses, computed from the raw components rather than copied out of anyone's fantasy column: 0.04 a passing yard, four a passing touchdown, minus two an interception, a tenth of a point a rushing or receiving yard, six a rushing or receiving touchdown, minus two a lost fumble, two a conversion, three a field goal and one an extra point. Receptions are what separate the three formats: none in standard, half a point in half PPR, a full point in full PPR. Switch the scoring chip and every figure on the board is recomputed from the same components, so you are reading your own format rather than someone else's relabelled.
The denominator is games played, not weeks that scored
A defense that held every tight end to nothing in a week allowed zero points that week, and that zero belongs in the average. It would be easy, and wrong, to divide by the weeks in which a position actually scored: doing that rewards a shutout by deleting it, and the defenses that deserve the most credit end up looking ordinary. So the denominator is the games the defense played, and a week with no points allowed to that position is carried through the weekly line as a real zero. Bye weeks are not games and never enter either side of the division.
Why last season is on the board in September
Three games is not a rate. Through the opening weeks of a season, one outlier afternoon moves a defense fifteen places, and a board that presents those numbers as a season average is telling you something it does not know. So any defense with fewer than four games this season is shown with its figure from last season instead, tagged on the row so you can see exactly which rows those are, and the band at the top of the page says how many games the sample is. When a defense clears four games its own current figure takes over automatically and the tag disappears. The game count travels on every row, in both seasons, precisely so you can judge the sample rather than trust it.
Last season's figure is a whole season of games against a schedule that no longer exists, played by a roster that has changed. It is the best available answer to "what has this defense been", and it is a poor answer to "what is this defense now". Treat it as a starting point that the current season overwrites, not as a forecast.
What this board cannot tell you
Points allowed is not defensive quality. A defense whose offense scores quickly and often faces more passing attempts than one that grinds the clock, and it will allow more receiving points for reasons that have nothing to do with its secondary. A defense that has already played three of the best offenses in the league will sit near the bottom of this board and may be excellent. Strength of schedule is a real confound here and this page does not adjust for it: see the strength of schedule board for the other half of that picture.
It is also a team-level number, and the matchup you actually care about is a player-level one. A defense can be generous to receivers as a group while a specific cornerback erases the specific receiver on your roster. It says nothing about who was healthy, who was suspended, whether a starting quarterback left in the first quarter, or whether the game was a blowout by half-time and the fourth quarter was played by backups against backups. Every one of those shows up in these totals as though it were a normal afternoon.
Kicker figures are the noisiest on the board by a distance. A kicker's points depend on how often his offense stalls in field-goal range, which is a fact about his own team's offense rather than about the defense opposite it, and a single afternoon of five field goals moves a season average that is only ever a few points wide. Read the kicker column as a tiebreak between two streamers you already like, never as a reason to move a kicker on its own.
How to actually use it
The board is at its most useful at the edges. The three or four defenses at the top of a position are worth a start you were unsure about, and the three or four at the bottom are worth a second thought about a flex player you were only going to start out of habit. Anything in the middle twenty is inside the noise of a fantasy week and should not change a lineup on its own. Rank one is the defense that has allowed the most to that position, which is the matchup you want; rank thirty-two has allowed the least.
Colour is a reading aid and never a verdict. A green figure is above the league median for the position on screen and a red one is below it, and the number itself is printed in every cell so the colour is never the only thing carrying information. Tap or click any row to open that defense's week-by-week line, which is where an average made of one enormous afternoon and three quiet ones stops being able to hide.
Pair it with the start/sit comparator when the decision is between two named players, and with the rankings board when the question is about the season rather than the week. The data behind this page is the same publicly licensed weekly record every other statistic on this site is built from, refreshed daily; the band at the top says when.