A season total says how much a skater scored. It never says how. Two forwards can both finish on 70 points and hand you completely different seasons. Every number here is counted off individual game logs, one row per game.
Point-game rate is the one measure a season total cannot reconstruct. Sixty games is the floor here, so nobody arrives on a hot fortnight.
CaveatThis describes 2025-26. It is not a 2026-27 forecast. Read the games-played column beside every rate; the stability table below tests how far these rates move in a year.
Loading the below-median group.
How to read itA head-to-head week is won by a total, not an average. Four blanks then a hat trick with five shots and three hits still wins the week. These players concentrate value instead of spreading it.
CaveatConcentration cuts both ways. The weeks the boom misses are lost outright, and these boom rates run from single digits to the low twenties. A 50-game floor applies.
Loading the longest 2025-26 droughts.
Each run is consecutive games played with no point, inside the top 100 of the desk's ESPN blend board. Read the average-run column first: a seven-game worst with a 1.2 average was one bad month, an average over two is a habit.
CaveatRuns count games played, not calendar days. A three-week injury inside a cold stretch reads as one continuous run, because the game log has no gap. Treat the longest run as an upper bound.
Loading the power-play share board.
Each bar splits a skater's 2025-26 points into even-strength and power-play halves. At or above half on the power play prints in coral.
Half his points arrive on the power play. That half belongs to a deployment chart, not to him. It needs the club to draw penalties and needs him on the first unit when it does. Even strength needs only the line he already plays. The even-strength-per-game column is what survives a reshuffle. Discount, do not avoid.
CaveatPower-play share uses the skater's 2025-26 points on his 2025-26 club. Anyone who moved in the offseason carries old deployment into this table; the crest shown is the new club. A 20-point minimum applies, so a fourth-liner with three points does not top the board at 67 percent.
Every number above describes one season. Here is the test: for skaters who cleared 40 games in both 2024-25 and 2025-26, the two point-game rates side by side, six biggest climbs and six biggest falls.
Consistency is not a universal good. What it is worth depends on how your league turns production into standings points.
Every week is its own contest on a small sample. A high floor wins more weeks than the same total delivered in spikes, because spikes overshoot weeks you had already won. Here the metronomes board is a draft board.
A fixed lineup with few moves makes the boom-bust names live again. You cannot guess their nights, so you roster them and take the variance. Free streaming favours the floor players instead.
A roto season is one very large sample. Eighty-two games land in the same place smooth or spiky. Read the power-play and blank-run columns as risk. Skip the floor rate.
Season-long points behaves like roto. Weekly points behaves like head-to-head. Check whether your scoreboard resets on Monday, then pair this with schedule edges: a high floor is worth more on a four-game week.
CaveatFantasy figures use the published ESPN points weights. Yahoo or Fantrax scoring would shift the floor and boom rates, most of all for shot-volume and hit-volume players. Point-game and multi-point rates are format-free.
Public rankings run on season totals, because that is what the rate sheets publish. A total throws away the ordering of the season. Game rows give it back.
Point-game rate first. Inside the top 250 the best skaters clear four games in five with a point; the median sits near half. A points-per-game column averages a metronome and a spike machine into the same figure.
Blank streaks next. The longest pointless runs inside last season's top 100 reached eleven straight games. That is about a month, and every one of those players finished useful. A five-game blank from a good player is ordinary. Check the average run before you react.
Scoring mix last. Half power-play share does not make a skater overrated. It makes his production conditional on a line that can change. Worth knowing in the seventh round, not guessing at.
None of this is a ranking. Take the rank from the Fantasy Draft Desk, then check whether the player behind it is a floor, a ceiling, or a power-play unit in a jersey.
