A player's raw projection tells you what they will produce. It does not tell you what that production is worth, and those are different questions. What matters is the gap between a player and the one you could have had instead — the freely available guy at the end of the draft.

That gap is DV VORP — our value over replacement — and it is the backbone of every ranking we show. We set replacement level from the league-wide pool: in a 12-team league with 13 roster slots, roughly the 156th-ranked player. Every projection is expressed relative to that line.

It is also why our board disagrees with ADP, and why that disagreement is the point. Average draft position is a real signal and we show it on every row — but it measures consensus, not value. Drafting to ADP cannot beat the field, because ADP is the field. Your edge lives entirely in the gap between what a player is worth and where the room is taking him.

Chasing ADP means paying a first-round price for a player whose production says third, and missing the player whose production says second while the room has him in the fifth. Both numbers sit on the same row so you can see the gap rather than take our word for it.

Every projection system estimates games played. That part is table stakes, and we will not pretend otherwise. What is less common is treating risk and growth as first-class signals you can sort the board by — rather than a footnote, or an adjustment folded invisibly into one score.

Each is its own model, built and validated separately, and each sits alongside production on every row:

• Durability — how much to trust that a player will be there, as a forecast rather than a recap of last season.
• Availability — what you can realistically expect to get, given where a player stands the day you draft.
• Upside — the case for a step forward, weighed against the odds it does not come.

The part that matters is that they stay separate. Durability, availability and upside are three models, kept apart on purpose rather than folded into one composite you have to take on faith. On the board they read as plain words — Breakout, Wildcard, High, Low — so you can see what each says about a player and disagree with any one of them on its own.

Three players the same age, projected within a few points of each other per game. One reads Breakout, one Stable, one Low. Age is not what separates them — same age, three different futures — and the projection alone says they are the same player. If you are taking a swing in the fourth round, that is the whole decision, and it is invisible on a board that only shows production.

Most tools stop at the ranking: here is the list, good luck. But nobody wins a league by reading a list. You win by making the right call with the clock running — and that is where the signals have to show up, not in a spreadsheet you check afterwards.

So strategy lives inside the board, not beside it. Choose your punts and the board re-prices on the spot — tap the categories you are giving up, and every ranking re-sorts around the team you are actually building.

Real punts are rarely one category. Punting free-throw percentage tends to pull turnovers along with it, and committing to points and boards can mean conceding assists and threes. So you can punt as many as you commit to — and the value number recomputes rather than adjusts: punted categories drop out of the calculation entirely, so the board reshuffles rather than nudges.

Punt free throws and turnovers together and a certain kind of dominant, poor-shooting big becomes dramatically more valuable — while some early-round guards are worth materially less to you than to everyone else in your league. Same players, different prices, because your team is built differently.

It is easy to add a signal that sounds smart, and easier still to repeat a rule of thumb because everyone else does. The discipline is testing it against real data — and being willing to find out the rule is wrong.

Take one of fantasy's oldest rules: fade last year's breakout. The logic sounds shrewd — a big jump is probably a fluke, so let someone else pay for it. We checked it against eleven seasons of NBA data, covering more than two hundred breakouts.

The rule is mostly wrong. The typical breakout player kept nearly nine-tenths of his gain the following season. About one in four faded badly — but the other three held on, and many kept improving.

Fade every breakout and you hand your rivals the players most likely to keep producing. That is the standard we hold every idea to: compare it against doing nothing, and assume a striking result is luck until it survives being tested properly.

We would rather show you six numbers you can act on than sixty you cannot. If we cannot explain in a sentence why a player moved, we do not ship the number — and if a signal cannot beat a simple baseline, it does not go on the board at all.