LIGHT NIGHTS
Hockey Streamers
The next fourteen nights, games league-wide
Tap a category to add or remove it. Punt keeps it on the list, greyed, and takes it out of the value. The same picker as the category board, so a link from one opens the other on your settings.
well above the poolabovebelowwell belowfigures are per game, from our projection model
FIELD GUIDE
Streaming in hockey is a nightly trade, and the light night is where it pays
Fantasy hockey lineups lock one game at a time, every night of the week, and a roster of 16 or 18 skaters never has every seat busy on the same night. Streaming is the habit of filling those idle seats for one evening: drop the bench player who is not playing, add a free agent who is, and take whatever he puts up. Over a matchup week the extra games add up to a real edge, and the edge is largest on the nights when almost nobody plays. A Tuesday with fourteen games gives every manager a full lineup; a Wednesday with three games leaves most benches dark, and the third-line winger on one of the six clubs dressing that night is worth more to you than the first-line centre sitting at home.
This page does the counting. It reads the league calendar, finds the next light night, and lists the skaters on the clubs playing it who are rostered in fewer leagues than your ceiling, ranked by what each is projected to do in one game in the categories your league counts. The goalie tab does the same, but only for the goalie we expect in net.
Read the rest
Read the rest
Why the light night matters more than the player
A light night is a date when the league plays fewer than eight games, which is the same threshold the schedule planner marks in red. Through most of the season those are Mondays, Wednesdays and Fridays, with the heavy slates on Tuesday, Thursday and Saturday, though the calendar bends around holidays and the all-star break. On a heavy night your open seat is competing with your own starters for a place in the lineup, and a streamer only helps if he beats a player you already own. On a light night the seat is simply empty. Anyone who plays is better than nobody, so the question stops being whether the streamer is good and becomes which of the few skaters playing is best, and that is a question a per-game projection can answer. The strip of nights along the top shows the count for each of the next fourteen days so you can see the week's shape at a glance, and the page opens on the first light one.
The roster-percentage ceiling
A streamer has to be available to you, and the closest thing to a league-wide measure of availability is the share of fantasy leagues that already have a player rostered. The slider sets a ceiling, 50 percent by default: a player rostered in more leagues than that is left off the list because in a typical league he is probably already gone. Drop it to 20 percent for a deep league where the waiver wire is thin and only the genuinely unowned are left, or push it up to 80 percent for a shallow ten-team league where useful players go unclaimed. The figure on each row is the current share, and a player the roster data does not list at all is shown as unlisted and kept on the list, because the most likely reason nobody tracks him is that nobody has him.
A goalie stream needs the starter, not the club
Skaters play when their club plays. Goalies do not. Every club carries two and announces which one starts at the morning skate, so a goalie you add for tonight is only worth anything if he is the one in net, and a backup added on the strength of his club's schedule scores nothing while he watches. The goalie tab therefore lists only the goalie we expect to start for each club playing that night, with the tandem's split beside him, from the same rule the goalie tandem guide uses: a rested club goes to its starter, the second half of a back-to-back goes to the backup unless the starter took both halves more than half the time last season, and a goalie designated out or on injured reserve is removed from the rotation. The expected starter is a default, never a confirmation. Check the club's announcement on game day when a start decides your matchup, and treat a row marked assumed, past the seven-day window, as the tandem's starter rather than a call about that game.
How the value is scored
Every figure in the category columns is a per-game rate from our own projection model, the same one behind the projections page, and the value at the end of the row is that line scored the way your league scores it. In a category league each rate is turned into a z-score against every skater the model projects, or every goalie for the goalie categories, and the z-scores are summed over the categories you keep; a punted category stays visible and leaves the sum, and goals-against average is turned so that a lower figure counts as better. In a points league the same line is multiplied by the default point weights instead and shown as points per game. Because the scale is per game rather than per season, a player on a two-game week and a player on a four-game week are compared on equal terms here, which is right for a one-night add and wrong for a draft. The category board is the season view.
Acquisition limits and the nights that are not worth it
Most leagues cap how many players a manager can add, either per week or over the season, and the cap is set by the commissioner rather than by the platform, so this page does not know yours. Check it before you build a habit around this list, because a week of nightly streaming can spend a season allowance fast. Where the cap is tight the light night is still the night to use it on: one add on a three-game Wednesday buys a game nobody else in your matchup is getting, while one add on a fourteen-game Tuesday buys a game everyone already has. Where there is no cap, the list rewards checking it most days.
What this page does not know
It does not know your roster, your league's free-agent list or your opponent's, so a player shown here may already be owned in your league and the ceiling is only a proxy. It does not know tonight's line combinations or power-play units, which move a skater's real chances more than anything a season projection can see, and it does not know a healthy scratch until the club announces one. The projections are ours alone, not a consensus, and before the season opens they rest on last season's game logs and an ageing curve, so a rookie or a player who changed clubs over the summer is projected from less than the rest. Before the first puck drop the page shows the first week of the schedule and says so; after the last game it says the season is over rather than showing an empty night. The night, the ceiling, the categories and the punts are all written into the address, so the link you copy opens the same list for anyone in the group chat.
The schedule and the roster percentages are refreshed on our own feeds; the expected starters come from the goalie feed, updated daily. If a feed is unavailable the page says so rather than showing a stale list.