High frequency trading is not just about putting a lot of trades on in the fashion of a day trader. High frequency Trading is more about the analysis of real-time data that has a frequency that is much higher then we are traditionally used to. In fact it should be called Ultra high frequency. The follow statement I found on a blog just shows how little people do know about High frequency trading:

“HFT is not an inherently better way to trade the market. Indeed, given transaction costs, it is a relatively expensive way to trade. Furthermore, it emphasizes the importance of nano-second order transmission and smart order-execution routing.”                                –Finance Blog

High frequency can come in the form of pricing data that are consumed and acted upon by market participants. The original form of these prices is tick-by-tick data; each tick is one logical unit of information, like a quote or a transaction price. By nature these data points are irregularly spaced in time. Data vendors like Reuters transmit more than 275000 prices per day for a foreign exchange spot rate alone Ironically practitioners, most often, determine their trading decisions by observing high-frequency data but yet most studies published in financial literature deal with low-frequency, regularly spaced data. There are two main reasons for this. First, it is still rather costly and time consuming to collect, collate, store, retrieve, and manipulate high-frequency data. The second reason is somehow more subtle but still quite important: most statistical apparatus has been developed and thought for homogeneous time series. There is little work done to adopt the methods to data that arrives at irregular time intervals.

With the availability of these new data sets come new challenges associated with their analysis. Modern data sets may contain hundred of thousands of transactions or quotes in a single day for a single stock, time stamped to the nearest second. The analyses of these data are complicated by irregular spacing, diurnal patterns, price discreteness, and temporal dependence. These challenges are described below.