[carsonbaker]

No, I wouldn't say pricing is fundamentally an infrequent activity, although analyzing the "idiosyncratic risk" may well be. That's not all that's going on though. For example, investors in, say AAPL, do not live in a Plato's cave where nothing else but Steve Jobs' latest keynote matters (believe it or not!). There's the rest of the market to consider.

In fact, a lot of algorithmic trading is about quickly assessing the value of a company in the context of a larger portfolio. If you accept that the market index can (and should) fluctuate second by second as the world tuns and news gets reported, then you must accept that companies' stocks will be traded alongside it (usually in rough proportion to the amount of liquidity available).

Here's how that can work. If you're comparing two stocks that share fundmental drivers, then you can make "pair trades" when they fall out of correlation. If Pepsi (PEP) does really well one day and CocaCola (KO) does terribly for no apparent reason, then it may be a good trade to buy KO and sell PEP on the supposition that the stocks have moved out of equilibrium.

This is basic arbitrage, and if you extend the practice across all the traded securities you'll find you need some pretty serious computing power. I suspect this accounts for much of the silicon in the basements of banks and hedge funds.

Notice that I'm just talking about the pricing function here. You can price a stock all day long and not have to trade on it.

And I'll agree that there's way too much volume, churn, and speculation in the market. I'm not sure, but I feel like it's becoming more and more a behavioral system. I wish people and institutions would better understand long-term risk and the costs of commissions and slippage when making investments. If "managed" funds became a dirty word, there may be a lot less volatility in the market.