academic Ledger · research note
Pediatricians read a child's height against percentile curves for their age. A paper's citations can be read the same way — but with a twist a growth chart doesn't have: where a paper sits today is not where it will end up. That gap is the whole problem QaL is built to measure.
Half the field is never cited. Here is the cohort's own distribution — the percentile bands of cumulative citations as papers age. The median line never leaves zero; the 99th percentile climbs to 122. And one real paper (blue) sits at 1,759, so far above the field it needs a log scale to share the page.
This extreme skew is exactly why an absolute count can't be compared across papers, let alone across fields. The fix is to stop reading the count and start reading the rank.
Now the y-axis is a paper's standing within its field — its percentile among the same 24,791 peers. The shaded tiers are the percentile bands (like the curves on a pediatric chart). Some papers track a band; others cross them. Toggle the archetypes and watch the sleeper climb from below the median to the 99th.
Band-crossing is the story. A child mostly tracks their percentile; papers do not. The sleeper is uncited for two full years, then overtakes almost everyone — the phenomenon a single early cutoff (accept/reject, an early citation count) would miss entirely.
QaL is a belief about where a paper will eventually stand, formed from the evidence so far. The green band is the 90% interval for the eventual (10-year) percentile, re-estimated at each age from how similar papers turned out. It starts wide and narrows as the paper matures — the honest picture, and the case for deciding late.