[Yizahi]

Exactly this. I'm observing how a midsized company shares fluctuate completely unrelated to the progress of our development efforts and actual relations and sales to customers. Product A was frozen and discarded after 2 year intense dev cycle (a failure), product B is not faring well in the competition, but we announced product C that may or may not succeed in 1+ years on a major tech show - shares go up. Next year product D is succeeding immensely, record contracts are signed with biggest customers in the world, product is ahead of competition - shares are down, because nobody knows.

tl;dr - PR is everything, often shadowing actual hw or sw "stuff" people make, and HFT capitalizes on this.

[nradov]

PR is nothing. The institutional investors who actually move stock prices barely even look at it. Humans try to make sense of random events by looking for causality, but ascribing stock movements to single factors is often like ascribing weather changes to animal sacrifices.

[TuringNYC]

You have a good point but remember that there are other factors that affect stock price completely external to the company -- the biggest one is interest rates.

Also, external happennings in other companies such as competitors, purchasers, suppliers have an impact on your company. So those should technically feed a model some factors (and in fact they do at hedge funds.)