Scouted

How a call gets recorded

Two things have to be true, and the second cannot be known on the day of the first. Both are applied by one query, over the daily ranking the boards are drawn from and the raw counter readings underneath it.

1. The flagThe first day a game’s Velocity reaches 0.20 with a confidence of at least 0.50. Confidence is the share of the signal’s weight we could measure multiplied by how much of a fortnight we had been watching the game, so 0.50 is a gate on both at once: a whole ladder read over one week only just clears it, and three fifths of a ladder read over ten days does not. The methodology gives the formula. A game that sits above that line for a fortnight is one flag, dated to the first of those days, not fourteen.
2. The outcomeAt least 30 days after that flag, the game’s ratings on itch.io or upvotes on Roblox read at least 4 times what they read on the flag day, and have passed 200 or 1,000 outright.

The outcome is measured on the last rung of the ladder on purpose. Velocity leans hardest on the counters that move first, so grading a call against Velocity would be marking our own homework: the number that raised the flag would also be the number that cleared it. Ratings and upvotes are the least-weighted of the counters Velocity reads, and they are what a scout watching the platform directly would have been looking at instead of us.

This is a list of calls that landed. A flag that went nowhere is not on it, so read the page as a record of what a flag looks like when it works, not as a hit rate. Nothing here is written by hand or dated backwards: every row is a game you can open, on a day you can check against that day’s board. The rule above is written out on the methodology page as well, and both pages take its numbers from the same constant the query is built from, so the two cannot drift apart.

The log