Released = publicly announced and available to at least some form of private beta, e.g. consisting of ChatGPT Plus members.
I will compare the most recent closing price a week after release to the most recent closing price before the release date.
The motivation is to test the hypothesis that markets will predictably underrate the capabilities of GPT-5, as they may be currently underrating the speed of AI progress.
I won't trade in this market.
@NoaNabeshima Yeah I agree. There’s also the argument that price slowly goes down then goes predictably up on release. If you have ideas for a better market go for it or lmk :)
Also even if the market was really calibrated about AI, the price should go up on release date because the expected release date will be longer than the true release date immediately before it's released
I think there are other arguments kind of like this to expect the stock to go up when it's released without the assumption that the market is really underrating AI progress
edit: (sorry started this before your comment loaded or something)