When a bank wants to move a big quantity of shares, for example, it doesn't want everyone to know what it is doing. If news of a big buy leaked out before the big buy could be completed, the price may go up. To hide their motions, they employ the same technique as stealth planes: they use algorithms to break their giant trade into thousands of little ones, and do so in such a way that they look random. Their sizes and timing are scattered.

In order to identify this stealthy algorithmic movement, competing banks hire other mathematicians to write other algorithms that monitor trading and look for clues of these bigger trades and trends. The algorithms actually shoot out little trades, much like radar, in order to measure the response of the market and then infer if there are any big movements going on. The algorithms are, in turn, on the lookout for these little probes and attempt to run additional countermoves and fakes. The algorithmic dance—what is known as black box trading—accounts for over 70 percent of Wall Street trading activity today.