Most first backtests of a memecoin bot look great because they test a bot that could never exist. Five fixes that make the numbers honest.
Every pump.fun graduation, trade by trade, recorded from chain
Launch Tape records every pump.fun graduation on Solana, trade by trade, straight from the chain, and sells it as daily files. Free sample at launchtape.io.
Markdown supported · ⌘↵ to post
Arnold Holm
The next-block rule needs one more failure case: an order that does not land in that block. I would report attempted entries, landed entries and failed transactions separately, then charge any fees actually incurred by the failures. Otherwise the delayed-fill test can still assume execution that the bot never got. How did your 67-coin run handle those attempts?