THE WEEK UNDER THE ADD CAP
Waiver Planner
The nights left in the week, 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. FG% and FT% are weighted by attempts; turnovers count against.
One streaming seat, the nights left, at most your cap. A player added on a night plays that night and every night his club plays until the next add replaces him. The total is the seat’s production summed over those games; the planner finds the best sequence exactly and uses the fewest adds among equals.
COURT GUIDE
The add cap turns streaming from a nightly habit into a weekly plan
The streamers page answers a nightly question: who is the best free player on the floor tonight? This page answers the weekly one that sits above it. On the host most category leagues use, a manager gets four acquisitions a week, and four is not enough to chase the thinnest night every night. Spend two adds on Monday and Tuesday for players whose clubs are done by Wednesday and the back half of the week is played with an empty bench seat. Hold every add for Saturday and the light Tuesday slate goes unexploited. The cap makes the order of the adds matter as much as the names.
So the unit here is not a night but the rest of the week: every unrostered player is scored on the games his club has left before Sunday, and the planner lays the adds across the nights.
Read the rest
Read the rest
What one add buys: games left, not games a week
The schedule grid tells you a club plays four times this week; that figure is only useful on Monday. By Thursday the same club might have one game left, and an add on Thursday buys one game, not four. The board on this page counts each player’s games from today to the end of the fantasy week, today included, and the weekday letters beside the count show which nights they fall on. A second night of a back-to-back is drawn in gold, because that is the night a coach rests a veteran with a sore knee and the night a streamer’s projected line is least reliable. The rest-of-week total at the end of the row is the per-game value multiplied by those games, and the board is sorted on it, so a modest player with four games left can sit above a better one with two. That ordering is the whole point of the page, and it is why it exists apart from the streamers board.
How the plan is chosen, exactly
The planner models one streaming seat on your roster, the nights left in the week, and at most as many adds as the cap allows. A player added on a night plays that night and every night his club plays until he is dropped for the next add; the seat’s score for the week is the sum of the per-game value of every game it hosts. The planner searches every sequence of adds by dynamic programming over the nights, the adds remaining and the player in the seat, and returns the best one. It is exact for that model: no sequence of adds inside the cap scores higher under these values, and among equal sequences it uses the fewest adds, because an add you did not need is an add you still have on Saturday when a player goes down. The rows it chooses are drawn in the plan above the board with the nights each one holds the seat and what he produces over them, and the night strip marks the nights on which an add is made.
What the model leaves out, on purpose
One seat is a simplification. Most rosters have two or three players idle on a given night, and a manager with four adds might stream two seats. The planner works one seat because the answer is then readable, a short list of adds in order, and because the second seat’s plan is the same computation run on what is left of the pool after the first. It does not know your roster, so it cannot tell you whom to drop or whether the player it names fits a lineup slot you have open; the position chip narrows the pool to a slot if you know which one is idle. It does not know news after tonight: a player ruled out is left off, a day-to-day player stays on with the flag, and a rotation change on Wednesday is not in Monday’s plan. And the per-game values are our own projection model’s, the same figures the streamers page shows, so the two pages never disagree about a player’s night, but they are projections, not a promise.
Categories, punts and the empty seat
In a category league each player’s per-game value is a z-score in every category in play, summed, with the two percentages scored on their impact rather than the raw percentage so a high-volume shooter moves the column the way he moves a real lineup, and with a punted category left out of the sum entirely. A z-score is measured against the average projected player, which means most streamers score below zero: they are, by definition, below the average of everyone with a projection. The planner does not treat a below-zero player as worse than an empty seat, because in a real week an empty seat scores nothing in every category and any game played beats it. It measures each game against what an idle night costs, a line of zeros scored the same way, so a streamer with a negative z-sum is still added when nobody better is free. In a points league the arithmetic is simpler: value is points per game, an empty seat is zero, and the total is points.
The ceiling, the cap and the week
The roster ceiling is the same proxy for availability the streamers page uses: a player rostered in more leagues than the slider says is treated as gone, and a player the roster data does not track is kept. Set the cap to your league’s number; the host most category leagues play on defaults to four, the other major host to unlimited or a commissioner’s figure, and a cap of eight here is close enough to unlimited that the plan becomes one add a night. The week picker lets you plan next week before it starts, which is what Sunday night is for, and before the season opens it shows week one and says so in the band. Everything you set, the week, the cap, the ceiling, the format, the categories, the punts and the position, is written into the address, so the link you copy opens the same plan for anyone in your league.
Reading it alongside the other pages
Use the schedule grid on Monday to see which clubs play four and which play two; use this page to turn that into an order of adds; use the streamers board, linked from every night in the strip, for the full list on the night an add falls due, when the injury report for that evening is in. The next man up page tells you which free players just inherited minutes, and those are the rows to look for here, because a lift in minutes that the projection model has not caught up with is where a streaming edge comes from.
The schedule, the projections, the roster percentages and the injury report are refreshed on our own feeds. If a feed is unavailable the page says so rather than showing a stale plan.