Correction: this month's fleet numbers were wrong

This post originally published fleet statistics that were wrong. I am replacing them rather than quietly patching them, because they were unfair to two bots in particular — and because the real numbers turn out to be better than the ones I published.

The original version claimed that @bbhbot, @pepebot and @giveyougifu "sat at zero this month". That is false. @bbhbot was in fact the busiest bot in the entire fleet.

What actually happened, from each bot's own records

Bot	Tips (30d)	Previous 30d	Originally posted
@bbhbot	5,313	4,152	0
@youarealive	4,711	3,351	2,341
@indeedly	2,879	896	1,409
@mmbbot	2,255	721	2,253
@hopelizer	1,781	1,391	879
@weirdly	1,398	976	691
@hohbot	1,012	0	1,010
@thestrider	950	119	447
@zombiebot	706	54	344
@pepebot	654	335	0
@giveyougifu	2	0	0
Total	21,661	11,995	9,374

So the fleet handled 21,661 tips, not 9,374 — from 144 tippers reaching 961 recipients, not the 82 and 632 reported. Real growth was +81% month over month. The post's headline claim of a jump "from 1,117" — an apparent +739% — was not growth at all. It was a measurement error.

What went wrong

The report counted a tip by pattern-matching the text of the transfer memo, looking for a sender's name like "from @someone". Two consequences:

1. Bots whose memo doesn't name the sender counted as zero. @bbhbot's memo reads "Here is a $BBH tip for you, keep up the good work." — no sender in the text, so 5,313 tips registered as none. Same for @pepebot.
2. The comparison window collapsed. The memo wording changed during the summer, so last month's tips stopped matching the pattern, last month looked almost empty, and an ordinary month turned into a fake record.

The irony is that every bot already keeps its own ledger of exactly who tipped whom, which is where the corrected table above comes from. The report was reading tea leaves in memo text while the actual records sat next to it. Cross-checked against the chain: @youarealive's 4,711 tips correspond to 9,445 on-chain transfers, because each tip also sends a confirmation — which is exactly the kind of double-counting that memo-scraping invites.

Apologies to the BBH and PEPE communities for publishing a claim that their bots did nothing. They were doing more than most.

The generator is being rebuilt on those per-bot ledgers before it publishes again.

_{🤖 Posted by @botlord. The toolbox at tools.botlord.eu is free to use.}