But the market price reflects an average, an expectation. And in certain conditions, it doesn't require much correct information to tip your expectation favorably.
Imagine a theoretical market for a stock where the company will either succeed and be worth a billion dollars, or fail and be worth nothing. Right now the company is trading at a valuation of $10MM. So the "market" thinks the company has a 1% chance of succeeding, right?
It seems to me that any belief you have which a) has some effect on the success probability of this company and b) differs from the average belief of everyone else can be used to update your probability on the overall success of this stock away from 1%, and cause you to want to buy or short the stock.
For example, let's pretend the company were Tesla. You think Elon Musk is a badass, and you think badass founders have a substantially higher probability of their companies succeeding than the base rate. Further, you also think most people don't take this into account as much as you do. Shouldn't you be able to use this information to raise Tesla's success probability above baseline and therefore cause you to want to buy Tesla?
To view this idea another way, imagine the market as a bimodal distribution of people who think the company will succeed and people who think it will fail. Most people fall into one of these two buckets. The market price is "obviously too low" if you're looking from the success bucket, and "obviously too high" if you're looking from the failure bucket. Now, if you have no information you should probably assign the "prior" which is the probability of success as determined by the market, but any information which updates you away from the prior makes the stock a good buy or good sell, I think.