Every fantasy site has a trending list. Sleeper has one, your league app has one, the big content sites all have one. They tell you the same thing: here are the players a lot of people added this week. What none of them tell you is the only thing I actually want to know — is adding him a good idea?
So we built our own. It shows the ten biggest risers and ten biggest fallers, and every single one gets a short written take that ends in a verdict: target, hold, or fade. It's live now on the site.
Quick disclaimer up front, because this one needs it more than most: the takes on this page are written by a language model, not by our projection model. It's marked Experimental on the site on purpose. I'll get into exactly what that means and what we did to keep it honest, but I'd rather you know that in paragraph three than find out in paragraph thirty.
The problem with every other trending list
A raw add/drop count is a popularity contest. It cannot tell the difference between the two situations that actually matter.
Here's the example that convinced me this was worth building. On the board I'm looking at right now, Alvin Kamara is the No. 1 most-dropped player in fantasy — 34,480 drops on Sleeper. If all you see is that number, the read is obvious: everybody's out, get out too.
But Kamara finished RB9 and RB11 in the two years before this one, and the reason he's being dropped is an MCL scare, not a collapse. He's fallen to round 13. That's not a decline, that's a discount.
Now compare that to Dylan Laube, who spiked into the top ten adds this week. His two career seasons are RB121 and a negative-point RB156. He's the third back in Vegas and there isn't even a news item attached to explain the spike.
Same kind of signal. Opposite conclusions. A count can't separate them — you need to know what the player has actually done. So that's what we fed it.
How it works
Three signals feed the board, and they are not created equal, so let me define them since we'll be here all season:
- Sleeper adds/drops — real add and drop counts, league-wide. This is the only source with true volume, so it drives the headline number on a row whenever it's there.
- ADP movement (ours) — how far a player's Average Draft Position moved between our own daily snapshots. Real magnitude, but it needs two pull dates to say anything.
- Most added / most dropped — a 20-name ordered list per direction from our data vendor. Order only, no counts. You'll never see a number next to one of those rows, because there isn't one.
Then, before anything gets written, each player gets a facts block assembled entirely from our own data: current ADP, where he sits on his team's depth chart, the injury feed, recent headlines, and — the part that does the real work — his fantasy finishes going three seasons back. That's what turns "the market is dropping him" into an argument. RB9 and RB11 is a buy. RB121 and RB156 is not.
The take is written from that block and nothing else. It isn't allowed to reach for outside knowledge, and the block gets saved word for word alongside every take we publish. That last bit matters to me: when a take says something wrong, the only question worth asking is whether the model made it up or our data told it that, and if you don't keep the block you can never answer it.
What the board looks like today
Pulled 8/24. A few that show the range of what it does:
Alvin Kamara — RB, NO — TARGET. "Buying the dip on back-to-back top-12 finishes." Round 13 for a guy who went RB9 and RB11, with a genuinely messy MCL report attached — reports range from a few weeks to a month and he's questionable for Week 1. It says that out loud rather than pretending the injury isn't real.
MarShawn Lloyd — RB, GB — FADE. 60,453 adds off one preseason touchdown and a "wants him more involved" report. He is still RB3 in Green Bay and his career line is a three-point cameo in 2024. "At round 13 you're paying real draft capital for buzz, not a role." That's the sentence I wanted this page to be able to write.
Malik Davis — RB, DAL — TARGET. The single most-added player in fantasy this week at 122,616 adds. ADP 322 overall. He was RB73 in ten games last year, which sounds like nothing until you remember what else is available at pick 322.
Keaton Mitchell — RB, LAC — TARGET. He's on the fallers list, and his ADP moved from 145 to 148. Three spots. It correctly called that a rounding error rather than a warning sign, which I appreciated more than I expected to.
Jayden Higgins — WR, HOU — FADE. "There's nothing to debate here — he's out for the year." Torn ACL, season-ending IR. Sometimes the right take is four words long.
Houston's defense — HOLD. Fell from 181 to 270 overall in a single day and tops the most-dropped list, with no news attached to explain any of it. The take says so: "Hard to have a strong opinion on a move like that without context." I'd rather it tell you it doesn't know.
One number that says more about the page than any individual blurb: of the 20 players on the board, only 6 got a target verdict. Seven were holds and seven were fades. A trending page that told you to go get all 20 would be worthless, and the whole reason I let this thing write prose at all is that it's willing to say "this is noise."
Where it's thin, and what we already got wrong
I'm not going to pretend this shipped clean.
The first live runs produced genuinely bad takes. It called a player the 19th-best running back off a list that turned out to be a pool of free agents ranked against each other, not a depth chart — so a guy who'd just been released read as a likely camp cut for the wrong reason entirely. It also turned an injury feed entry reading "Knee - Acl" into a confident "torn ACL." Both are fixed, and both are the reason we store the facts block. But that's two wrong takes that reached a page before I caught them, and I'd assume there are more like it in here.
ADP movement is barely contributing right now. It needs two snapshot dates to compute anything, and our vendor's ADP has been coming back byte-identical day over day — zero rank changes across a thousand players. That's not our bug, but the practical effect is that the ADP rows on the board are thin until they start republishing. It's why you're seeing team defenses moving nine spots in the riser slots; there isn't much else moving.
Rookies are the same problem they always are. A player with no NFL snaps has nothing in his fantasy history to reason from, so those takes lean on news and price alone. Barion Brown at ADP 649 gets called "pure name recognition off two vague news items" — which is correct, but it also means the page can't do much for you on rookies beyond telling you the dart is cheap.
And it's a market signal, not a projection. This page is about what everyone else is doing and whether they're right. It is not our ranking of these players. If you want the model's actual number on somebody, that's the rankings page, not this one.
How I'd actually use it
Draft season, I'm using it two ways. First, the fallers list is where the value is — a guy dropping on an injury scare with a top-12 finish behind him is the single most reliable discount in fantasy, and that's exactly the Kamara case. Second, I'm using the fade verdicts as a do-not-reach list. Lloyd at round 13 is the kind of pick I've talked myself into a hundred times, and having something tell me flatly that I'm buying buzz instead of a role is worth more to me than another sleeper recommendation.
Once the season starts it flips to a waiver tool: the adds are what your league is about to do, and the takes are a second opinion on whether to beat them to it or let them have it.
The board refreshes daily. Take the verdicts as a starting point rather than gospel — same as everything else we publish here.
Thanks for reading, and if you want to see our top player projections each week, create a free account by clicking this link.