farcasterunion

To indicate the effect of the changing spam labels on the data with active FIDs, I have tabled the data from April based on the last three spam label datasets.

Share of spam label 0 decreased from 66.6% to 54.5%
Share of spam label 2 increased from 16.1% to 40.6%

The data with spam label flows (courtesy of @mvr) clearly shows that the netflow in the dataset with spam labels is in favour of spam label 2.

hamst.art $hamster /hamster 
martinvr.eth 
deploy.degen.eth

Interests: crypto, economics, TA & blockchain related innovations | 🔗 https://linktr.ee/cryptonjal 🔗 https://dune.com/nhejyht

Wow. This is good to hear. I hope it means more well meaning casters are getting upgraded.

Pi-chi.eth Artist /animeoutcasts
Hosting: /sakura /kyoto /anime-manga /farcasterunion

https://portrait.so/pichi

Do you think pro folks are getting upgraded for optics?

No, it's a side effect of a broader change benefiting active FIDs. Relative small changes in the dataset leads to significant changes in the active FIDs.

Spam label dataset  from May 9th had together 44.0% label 1+2, active FIDs with label 1+2 was 21.8%
Spam label dataset  from May 23rd was 48.6% label 2, active FIDs with label 2 was 34.1%
Spam label dataset  from June 5th was 52.3% label 2, active FIDs with label 2 was 40.6%

I think they will strip everything that is a quality judgement from the spam labels. Every rating system makes positive and negative mistakes, and it seems that they would rather give more spammers a label 2 than wrongly give a non-spammer label 0.

In the image, active FIDs, and all the same behaviour. Some are getting label 0, and some label 2. You can click on any of these (example has spam label 2), Why is this not a label 0?

@v seems like some really bad accounts are getting marked positively.

Ahhh so maybe cast more original content is key?