The Census · Learn
Your rating
Two numbers, two jobs
Every session you complete produces one small piece of evidence: how strongly your description favored the image that turned out to be real. Over time, many of those pieces build into a single number — your rating. They work at different scales, and it helps to keep them separate.
Outcome Evidence: one session
Each reveal shows you that session’s Outcome Evidence — how your description compared to the winning image, framed against pure chance. It’s a snapshot, not a verdict. One striking session, good or bad, doesn’t mean much on its own.
Your rating: the pattern
Your rating — shown 0 to 1000 — is what all of your sessions add up to. It updates slowly, on purpose. It’s built to reward a real, sustained pattern of favoring the winning image across many sessions, not a single good run.
Why luck can’t win
A lucky streak nets close to nothing on the rating scale. That’s deliberate: the scale is discounted for chance from the ground up, so it takes real volume — not a good day — to move it meaningfully. 500 sits at the chance line; moving well above it, or falling well below it, both take sustained sessions, not a fluke.
The 600 line
Around 600 sits the earn line — the point where a rating is considered strong enough evidence to earn points for sessions, rather than spend them to commission your own. Reaching it was never the point of any single session; it’s what a long, honest pattern eventually shows. More on what that unlocks in the next chapter, The Order.
The leaderboard
Standing ranks every established seat by rating. A newer seat on a hot streak won’t jump the board — the same luck-discounting that protects your rating protects the ranking too.