Content
@
0 reply
0 recast
0 reaction
Ryan J. Shaw βοΈπ©
@rjs
Found a bug regarding follower counts: replies/follower is now dramatically different. Newer users actually get *better* engagement per follower than the old users - presumably the result of auto-follow diluting the value of each follower? If you think these numbers seem way too high - you're not wrong. According to this, with my 20k followers I should get 160 replies on average to my cast, but I get maybe a 10th of that! My best guess is that this is because Warpcast hides a significant amount of replies to your casts (spam/low quality filtering) -- @pichi had documented this extensively. That or another bug in my analysis.
7 replies
4 recasts
23 reactions
Pichi πͺππΉπ© π‘πΈ
@pichi
I see this consistently. Warpcast vs Airstack What grinds my gears is this community is hypersub gated. Bots donβt buy hypersubs. But about half my subscribers comments donβt make the cut.
5 replies
0 recast
9 reactions
C O M P Ξ Z
@compez.eth
There is a huge difference in the number of likes and recasts. I think WC ignores a lot of real stuff.
2 replies
0 recast
7 reactions
Ryan J. Shaw βοΈπ©
@rjs
Hmm yeah looks like maybe 1/3rd reactions being accepted by WC and 1/2 the replies... Still my data predicting 5x more... Data analysis is hard π
1 reply
0 recast
2 reactions
C O M P Ξ Z
@compez.eth
π₯ you always do great analysis! thanks for doing everything in this regard man.
2 replies
0 recast
2 reactions
Pichi πͺππΉπ© π‘πΈ
@pichi
Hard agree. I feel gaslit some days and then @rjs or @saxophone55.eth whips up a Dune dashboard and itβs no longer my intuition telling me something is sus, I now have hard data. Appreciate you all!
0 reply
1 recast
3 reactions
Ryan J. Shaw βοΈπ©
@rjs
Still learning, trying to recall my stats classes from 20 years ago π€£
0 reply
0 recast
3 reactions