> I still fail to understand what does being one millionth of a second faster accomplish. A simple example: suppose AAPL is currently quoted $95.00 x $96.00. A

> I still fail to understand what does being one millionth of a second faster accomplish. A simple example: suppose AAPL is currently quoted $95.00 x $96.00. A reasonable estimate for the "fair value" is the midpoint, $95.50. Now, a new sell order for $95.01 arrives at the market. AAPL is now quoted at $95.00 x $95.01. In the absence of other signals, a trader might reason that there's something wrong with the guy selling at $95.01 when everyone else thinks fair value is ~$95.50. If so, he'll rush to trade against the new order. First one to get there gets shares cheap; everyone else gets nothing. Hold the position for as long as you like. There are a lot of details to get right, but this is an actual HFT strategy, not a toy example. Note that your speed matters only insofar as it lets you trade ahead of everyone else. All the infrastructure investment is required to stay ahead of other participants, but it does not confer a durable advantage: the arms race will quickly erode the edge. The investment is best-viewed as a tax imposed by other HFT firms, paid to technology vendors and exchanges. The big winners are those who sell the equipment and fast lines that certain HFTs desperately need. > What if the regulators decided to ban all HFTs. The market would look pretty similar to how it does today. The big change would be somewhat wider spreads: market makers need to compensate for the additional risk they take on by leaving their orders out. Any other traders here have an opinion?