Where We Disagree, August 6

15 players we rank far above the consensus, 70 we rank far below.

Everyone publishes the consensus. The only thing worth saying against it is where our own board departs from it, and by how much. These are the 85 players the engine and the expert consensus disagree about most sharply today, with injured players removed from both sides.

We are higher

The engine ranks each of these inside the top 300 and above the value of the last rostered player in a 12-team league, while the consensus has them past 400, outside where anyone is rostering. Consensus rank is the closest thing to an availability signal we have, so treat it as a proxy, not a promise about your league.

We are lower

The consensus treats these as startable, inside the top 150, and the engine does not have them in its top 400. None of them are hurt; that filter is applied before this list is built, because "the engine discounts a pitcher who is not pitching" is not a disagreement worth publishing.

_Showing 12 of 70._

What this means

Luis García Jr. is the one I would act on, and it is not close. The engine has him 33rd overall at 408 projected points; the consensus has him 627th, which in a 12-team league means he is sitting unrostered in most of them. Nothing else on the high side combines that much projected production with that much availability. Bryan Baker at 407 points is the same order of value, but he is a reliever, and relievers get docked by consensus for reasons that are at least legible: the role can vanish in a week. García's gap has no such excuse. He is dual eligible at first and second, and 408 points is not a fringe-starter number, it is a middle-of-your-infield number.

What is producing the disagreement? The engine is pricing playing time and the market is pricing name recognition, and the two lists together make that obvious. Look at the low side. Edwin Díaz is the consensus 53rd pick and the engine's 916th at 62 points. Jordan Beck: 60. Brenton Doyle: 80. Luis Robert Jr.: 76. Those are not bad-player projections, they are not-playing projections. A full season of even a replacement outfielder clears 80 points. So the engine is forecasting that several of these guys accumulate a fraction of a season, and the market is forecasting that they accumulate all of it.

Which means Díaz is where the engine is most likely wrong, not the market. A closer's counting stats are already thin, so any haircut on innings crushes the total, and 62 points is a haircut down to almost nothing. The market's read is straightforward: he is a leverage arm on a contending roster and he will pitch. The engine has to be right about something quite specific, a lost job or a lost season, for 916th to be defensible, and it does not show us what that something is. The driver is not visible in this table, and I would rather say that than invent one. Ignore this row. The cost of being wrong is that you passed on a top-five closer to make a point about a model.

The strongest case against García runs through the same machinery, in reverse. If the engine is overshooting playing time for him the way it seems to be undershooting it for Díaz, 408 points is a full-season projection for a player the market expects to lose at-bats to a platoon or a prospect. That is the honest counterargument, and the consensus rank of 627 is not the sort of number a whole panel of analysts produces by accident. Somebody knows something about the depth chart. Nick Gonzales at engine 133 and consensus 468 is the same shape of bet at a smaller size, and Brayan Rocchio at 305 points against a 486 consensus is the same again.

Here is the test. Watch the lineup card, not the box score. If García is in it five or six times a week against both handedness splits through the first two weeks, the engine is right and the 627 is stale, and you have a top-40 asset for a last-round pick. If he is sitting against right-handers, the playing-time case is dead and the 408 comes down to something like 250, which is a bench bat.

So: add García now, in every format, and pay nothing to do it. Add Baker and Hicks if the roster spots exist. Leave Díaz alone. And treat every projection under 100 points on the low list as a claim about games played, which is the softest thing a model predicts.

How to read this

Consensus rank and its spread come from FantasyPros. Spread is the standard deviation across the experts who ranked him: a small number means they agree with each other, which makes a gap against our board more interesting, not less. Players ranked by only one source carry no real consensus and are excluded entirely.

Injured players are removed from both directions. Without that filter, more than a third of the "we are lower" list is simply players on the IL, and the top of it is injured aces the consensus still ranks highly as keepers, a comparison that says nothing about either opinion. 56 players were dropped for that reason today.

This is a disagreement report, not a waiver-availability report. We cannot see your league, and no roster-percentage feed backs this site, so consensus rank is used as the proxy for whether a player is gettable.

All articles · Redraft rankings · Dynasty rankings