Exactly. Let's assume that Day 0 = day of IPO, and Day -1 = the day before the IPO. Let's assume that the (unobserved) value of the stock on Day -1 is the 'true

Exactly. Let's assume that Day 0 = day of IPO, and Day -1 = the day before the IPO. Let's assume that the (unobserved) value of the stock on Day -1 is the 'true' underlying value of the stock. The only way that a glitch on opening day would have a permanent affect on the stock price is if the stock price on Day n (for all n > 1) is conditionally independent of the price on Day -1 given the price on Day 0. In other words, the best predictor of tomorrow's price is today's price, and yesterday's price adds no new information. This is a reasonable assumption, but if you believe that, it means you also believe that stock prices are essentially a random walk, in which case there is never any convergence to the true underlying value, even if the price starts at that true value (which it may not!) Let's assume that stocks are not a random walk and are some reflection of the underlying value of the company (however you define that). Even if there had been a glitch, on the first day, as soon as the real quotes start getting published and the information is symmetric all around, then the prices should start to approach the fundamental value, even if they don't converge. If a week later, Facebook is still trading below the IPO price, we can assume either that the convergence process takes time (unlikely, given the volume of stock being traded) or that the underlying value of Facebook as a company has changed in the last week (unlikely, since there has been relatively little news about Facebook-the-company, as opposed to Facebook-the-stock).