Hacker News Reader

Comment by tanderson92 | Hacker News Reader

Yes, and it's a worldwide phenomena. Embedded in this link is Thaler (A Nobel winner) using such an example as a counterexample to one view of the efficient markets hypothesis: https://ftalphaville.ft.com/2017/10/09/2194613/thaler-the-cu...

tl;dr price for the "CUBA fund" spikes following Obama announcement of normalizing relations, despite the CUBA fund having nothing at all to do with CUBA.

View on HN

What might be called the "homophone superstition" aspect of it is (apparently?) bigger in Chinese culture than in other language cultures, though.

http://www.nautil.us/issue/44/luck/is-the-chinese-language-a...

The efficient markets hypothesis seems Really Fucking Dumb™ to me. Like, trivially nonsense.

The EMH isn't necessarily that every stock is perfectly priced. It's that it's not possible to reliably predict where prices will go next.

But isn't the hypothesis that it's not possible to predict because it's perfectly priced already?

If you see EMH as a general principle: all present information is incorporated in the price and all Sharpe ratios trend to the market Sharpe ratio, it makes more sense.

No one actually believes in strong EMH, that is, it’s impossible to beat the market Sharpe ratio.

Not going to sign up to read that article, so apologies if this is off base or duplicative of what's in the article.

But, automated sentiment analysis gone wrong maybe? Bots see a bunch of negative things with "cuba" in it on twitter/the news and decide that cuba the fund is in trouble.

The excerpt is from this SSRN preprint, section III: https://papers.ssrn.com/sol3/Papers.cfm?abstract_id=2790606

The fund went up not down. That aside, I believe the point is that if a bot goes and does that, it is simply doing what a human told it to do; bots reflect the commands of their human masters in an environment (financial markets) created by and for human endeavors. Hiding behind "bots" is not a way to get insight into mechanics of how markets (mal)function, IMO.

Well sure, but (under my thoery) the humans rationally decided to use automated sentiment analysis, knowing that there is a error rate, but finding that it is low enough that they can make money. This just happened to be one of those errors. And substitute negative sentiment for positive sentiment in my original post to match the direction of stock movement.

Thanks for the direct link!

Reply on news.ycombinator.com