How the Acceptance Outlook Works (And Why It Is Not a Percentage)
feature · methodology · trades
Open a two-team trade in the calculator and you'll see a panel about the manager on the other side: how often they trade, what they tend to buy and sell, and a five-tier Acceptance Outlook from Cold to Hot with a few receipt lines under it.
This post is the long version of what that score is made of. The short version: it is a fit score describing how well your offer matches what this manager has actually done, and it is deliberately not a probability.
Why There Is No Percentage
The honest answer is that the data to compute one does not exist.
Sleeper's API returns completed transactions. A trade that was proposed and declined, or countered, or left sitting in someone's inbox until it expired, is never returned by any endpoint. That means there is no denominator. To say "there is a 64% chance they accept this" you need to know how many offers they turned down, and nobody outside Sleeper has that number.
Plenty of tools will show you a percentage anyway. We would rather show you a label backed by numbers you can check than a decimal that looks precise and is invented. So the headline is a tier, and every line underneath it is a fact about trades that actually happened.
What It Reads
Everything comes from executed trades involving that manager, valued as of the date each trade happened rather than at today's prices. That distinction matters: a manager who traded for a running back in 2023 made that decision at 2023 values, and grading it against 2026 values would tell you about the market, not about them.
The panel aggregates by Sleeper user id rather than by team name or roster slot, because display names change and roster ids get reshuffled between seasons. It follows your league backward through every previous season it is linked to.
The Part That Actually Matters: Their Window
Here is the problem with reading a manager's whole history as one number. The same person behaves like two different people depending on where their team sat at the time.
When they are contending they sell picks and buy proven production. When they are out of it they do the exact opposite. Average those together and you get a manager who does nothing in particular, which describes nobody. A flat career aggregate is close to useless for predicting the next deal.
So every past trade is weighted by how similar that manager's situation was then to their situation now. We reconstruct where their team stood at the moment of each trade from the league's game results: for an in-season trade, their record going into that week, which is exactly what they could see when they made the call. For an offseason trade, how the previous season finished. A trade made when they were contending counts fully if they are contending today, and fades if they are not.
Three other factors adjust the weight:
- Recency. A deal from four seasons ago counts, but less than one from last month.
- League format. Superflex changes quarterback valuation so completely that a quarterback preference from a superflex league is not weak evidence in a one-quarterback league, it is misleading. Those leans are dropped outright rather than discounted.
- Which league. History from the league you're actually trading in counts more than history from another league you share with them.
Other Leagues
If you and your trade partner are both in more than one league together, the panel offers to fold that history in. Most managers have far more trades across all their leagues than in any single one, and more evidence means a sharper read.
Two deliberate limits. It only looks at leagues you share with them, never at leagues you are not part of. And it is opt-in, behind a button, because analyzing another full league's history is real work and we are not going to do it on every panel load on the chance you wanted it.
The Four Components
The score is additive out of 100, and the panel shows you the tier rather than making you do arithmetic:
- Surplus fit, up to 45. Where your offer's value surplus falls within the distribution of surpluses this manager has historically accepted. A manager who has only ever accepted deals that clearly favored them is a harder sell than one who trades near even, and this is the term that knows the difference.
- Window fit, up to 20. Two things: whether the shape of what you're sending matches what a team in their position wants, and whether they have a track record of dealing with teams in your position. Sending three first-round picks to a team fighting for a title is a bad offer even when the raw value works.
- Shape fit, up to 20. Their revealed appetite: do they accumulate picks or spend them, consolidate into fewer better pieces or spread into depth, and do they buy at the positions you're offering.
- Activity, up to 15. How often they trade and how recently. Deliberately the smallest block. It is the only component that has nothing to do with your actual offer, so letting it dominate would mean a bad offer to a busy manager outscoring a good offer to a quiet one.
What It Does When It Does Not Know
Small samples are the normal case, not the exception. A manager with four trades on record does not support a confident read, and pretending otherwise is how you get a score that sounds authoritative and is noise.
So the surplus comparison blends the manager's own history against a league-wide baseline, weighted by how much comparable evidence actually exists. With one or two relevant trades the read leans mostly on what the league does; as their record fills in, their own behavior takes over. There is no threshold where the answer suddenly flips. When a manager has almost no history at all, the panel says so and treats that reluctance as the finding, because a manager who never trades is genuinely hard to trade with.
Known Limits
Stated plainly, because you should know where the edges are:
- No declined offers, ever. Everything here is inferred from deals that closed. This is a floor on how good any tool of this kind can get, ours included.
- Their past situation is read in three bands (contending, on the bubble, out of it) rather than the five-tier window we show for current rosters. Splitting the bottom three requires reconstructing who owned which draft pick on a given date in the past, which the data does not cleanly support.
- Value surplus depends on your selected value source. Switch sources and the numbers move. Every receipt line names the source it used.
- People are not models. A manager who has hoarded picks for two years can wake up and decide to win now. The panel reads evidence, not intent.
Use It As A Starting Point
The most useful thing in the panel is usually not the tier. It is the receipt lines and the split showing what this manager does when contending versus when out of it. That tells you which deal to build before you send anything, which is a better use of the information than a number telling you how a deal you already built is likely to land.
Open the trade calculator, pick a two-team trade, and the panel appears under the trade summary.