How every player and draft pick earns a single number.
Every player and every draft pick on this site gets one number: a grade from 0 to 100. Higher is better. The best asset in dynasty football sits right up near 100. A dart-throw bench stash sits near 0.
We ask three of the most respected dynasty sources what they think every player is worth, line their opinions up, and blend them into a single grade. Three heads are better than one.
Two rules the grade always obeys:
That's it. If you stopped reading here, you'd already know more than Jeff.
Each grade blends three independent, well-respected dynasty valuations:
Each source is excellent, and each is a little different. More importantly, each publishes numbers on a completely different scale — FantasyPros runs 1–100, while KTC and FantasyCalc climb all the way to ~10,000. You can't just average those together; that'd be like averaging a golf score, a bowling score, and a credit score.
Instead of wrestling three different scales into agreement, we ignore the raw numbers and use each player's position in the list. Within each source we sort everyone best-to-worst and note where they land as a percentile — the very best sits at the top (0%), the very worst at the bottom (100%). A percentile means the same thing everywhere, whether the source counts to 100 or to 10,000.
For each player we average the percentiles from every source they appear in. Show up in all three? We blend all three. Listed in only one? We use the one we've got. This keeps the system sturdy — one source having a weird outlier can't blow up a player's grade.
Now we bend that blended percentile (call it p, running from 0 at the top to 1 at the bottom) into a 0–100 grade with an exponential curve:
p = how far down the list a player sits (0 = best, 1 = worst).
Two pieces do two jobs. The gentle 5.5p term spreads the whole field out from top to bottom. The 6p⁴ term stays asleep near the top and then wakes up violently toward the bottom — that's what drops the deep end of the list into the dirt. The net effect: the top is separated but not obscene, the upper-middle slides down gradually, and the bottom third falls off a cliff.
Here's that curve — grade on the way down as you move from the best player (left) to the worst (right):
Same thing as concrete numbers, this time by where a player falls in the pack:
| Where they rank | Grade |
|---|---|
| Best in football | 100 |
| Top 1% | 95 |
| Top 5% | 76 |
| Top 10% | 58 |
| Top 20% | 33 |
| Top 30% | 18 |
| Middle of the pack | 4 |
| Bottom 30% ↓ | <1 |
| Dead last | 0 |
This shape is deliberate. The top of the board is still where championships are won — but we eased off up there so your studs aren't absurdly far ahead of your merely-very-good players. The real teeth are at the other end: the bottom third of the league is pushed to basically zero, because a deep-bench dart throw isn't a trade chip. (Something to mention to Jeff the next time he counters your stud with four of them.)
The previous version of this system used z-scores — a statistics tool that measures how far above or below average each player is. It worked, but it had one fatal flaw: half of all players came out negative. A z-score is centered on the average, so everyone below average scored below zero.
That's mathematically tidy and practically nonsense. A bench player is worth a little, not negative something. Negative grades made trades confusing (adding a player could somehow lower your total) and made the whole board harder to read. Ranking-plus-a-curve fixes all of it: everyone lands between 0 and 100, the elite are properly rewarded, and the numbers finally read like real trade values.
The whole shape is controlled by the two numbers in the formula. The 5.5 sets how gently
the grade slopes across the field — smaller bunches the top together even more. The 6
controls the tail — larger drives the bottom third harder into the ground. We've set them so the
elite stay clearly ahead while the deep end is worth next to nothing.
Grades blend KeepTradeCut, FantasyCalc, and FantasyPros. No Jeffs were consulted in the making of this system.