Varun Srinivasan
@v
The main thing I've learned from prediction markets is that no one understands how probabilities work
8 replies
12 recasts
87 reactions
Varun Srinivasan
@v
The intuitive reasoning that people have is that if p(X) is greater than p(Y) and Y happens, the model is bad. Most people know this isn't right The correct model is that if you're predicting an infrequent event like the election there is no deterministic way to figure out if the model is good or bad by looking at that prediction. You need to observe the model operating over a bunch of events to get any good understanding of it.
4 replies
1 recast
24 reactions
//trip
@heytrip.eth
Hence the mega REEE about that self absorbed clown silver doing it right and still getting absolutely mogged. "Most people" might be a stretch tho but I guess that's a function of p(x) too 😅
0 reply
0 recast
0 reaction
xh3b4sd ↑
@xh3b4sd.eth
Yeah, nobody knows what p(x) even means. Normal people do not understand what bids and asks and odds are. That is only one of the reasons why I am building Uvio on Base. Because this whole consumer trading thing is likely not where it's at.
0 reply
0 recast
0 reaction
dusan.framedl.eth
@ds8
"all models are wrong, some are useful"
0 reply
1 recast
6 reactions
atlasmoth
@osuji.eth
@v agree the predictions by themselves don't say much about the overall quality of the model, however simply observing how new events affect the model relative to the variance of the underlying distribution often tells you quite a lot https://warpcast.com/osuji.eth/0x3705dc70
0 reply
0 recast
0 reaction
