What gets saved
Every published board: redraft, dynasty, and rookie. Each player's rank and value, the source the numbers came from, and the date that source last pulled them. A snapshot is written once and never edited.
track record / logged before the outcome
Every ranking,
saved before
the games
DraftADynasty saves a copy of every board it publishes, on the day it publishes. Grading those copies against what actually happened is the next piece of work, and it is not built yet.
Logging has not started
Nothing has been saved yet. Saving is switched off on this deployment, so the log stays empty until that changes.
Grading
Not built yet. When it is, a board will only ever be graded against results that came after it was saved, so nothing here can be tuned to an outcome we already knew.
What comes next
Board-by-board results, misses included. That is the plan, not the page. Until it ships, this is a log and nothing else.
What this is not
There are no accuracy numbers on this page yet, and nothing here claims DraftADynasty ranks players better than anyone else. Today's boards are market values from FantasyCalc, priced off trades people actually made. They are not our own model.
Snapshots so far
The fingerprint is a checksum of the saved file. Edit a snapshot and its fingerprint changes, which is what makes the log worth reading.
No snapshots yet. Saving is switched off on this deployment.
The research program
A DraftADynasty model is in research. The tests it has to pass were written down before they were run: beat the strongest baseline on both scoring measures, hold that margin inside a 95% confidence interval, and let no position group go backwards. It has not passed them. Until it does, the site stays on market values.