Bots here are wallets that usually buy within 10 s of launch and were early on 10 or more tokens. Each rate counts only launches inside the bot's active period (its first to last early buy); a launch counts as bought when the bot bought within 60 s. Ticker, socials, description and starting market cap made no consistent difference. The base token, the creator's launch buy and whether the bot was running did.
Share of launches each bot bought, split by what the token is priced in. A bot near 0% on "Other" only trades ICP-priced tokens.
ICP-priced tokens only, grouped by the share of the curve the creator bought at launch. Darker cells mean the bot bought more of those launches; n is how many launches fell in the cell.
Each dot is one early buy of an ICP-priced token, by launch time and ICP spent, colored by the creator's launch buy. Flat bands mean a fixed buy size; a band that moves at a date means the bot's settings changed.
Every ICP-priced launch in each bot's active period. Filled dots were bought within 60 s; hollow dots were missed. Long runs of hollow dots are stretches when the bot was likely switched off.