The World of High Frequency Trading: We can’t live with 'em, can't live without 'em

High Frequency trading (HFT), also known as algorithmic trading, is a procedure for automated trading which is mainly focussed on detecting market idiosyncrasies/irregularities and exploiting these all within a fraction of a second. This is done by dedicated powerful computers constantly analysing the market data and driven by algorithms which utilize the trading opportunity. These computers operate in a cycle of milliseconds, outperforming the human brain which usually cannot cope with very low-latency information, digest it, accurately resolve and act upon it.

HFT operates on powerful computers co-located by (sitting next to) the trading platforms in the stock exchange.  These non-human traders process millions of orders at lightning speed and reap huge profits.  Since computers seldom get ill, are easily replaceable, have no emotions and have become really cheap, HFT is a highly lucrative business. It is the hottest thing in the equity market and governs almost 60% of volume; in essence HFT drives the market.

Trading at the highest frequencies with well-designed and implemented strategies produces double-digit Sharpe ratios while real-life Sharpe ratios for well-executed daily strategies tend to fall in the 1-2 range.

High-Frequency Trading Strategies:

Each quote, or “tick,” carries unique information about market conditions. High-frequency traders look for small gains at tick level on short-term market arbitrage, mostly trying to take advantage of temporary market inefficiencies. The traders move in and out of the positions several times in a day.

Broadly speaking, HFT strategies make their profits through:

• Market-making: makes money on the "bid-ask spread" by selling and buying on both sides of the trade. Exchanges offer discount of about one-third of a penny per share as these firms are willing to step up and provide shares when needed.
• Ticker tape reading
• Event arbitrage
• High-frequency statistical arbitrage

Programming: 

Mostly C++/Python/Matlab and some Java

Firms:

Getco, Lime Brokerage/Tower Capital, Goldman Sachs, Citadel Investment Group LLC, Renaissance Technologies LLC, Jane Street Capital LLC, Hudson River Trading LLC, Wolverine Trading LLC and Jump Trading LLC.

Issues with HFT:

HFT practitioners claim to have lowered fees, boosted liquidity and massively increased volume, but the problem is that, unlike a traditional market maker, they have no requirements.  No minimum size to display, no minimum time to display a quote and no capital commitment to a client.  The market is unregulated and most high-frequency traders feel they have no obligation to help maintain an orderly market.

HFT’s simply shut down their computers and walk away when their model corrupts,  creating a panic in the market (Biggest unexplained crash in the market 6^th May 2010);  if these HFT firms decide to shut down, what will happen to that 60% of the volume that they now control?

Traditional traders are at a disadvantage, in that unless until you have instant access to data across markets and the ability to place your own trades in a nano-second, you cannot compete.

It’s an arms race with only the biggest, fastest computers making money consistently. HFT causes mass-correlation and, ironically, encourages more volatility as opposed to more stability. 

Summary Position:

The problem is not that we can't live without HFT liquidity; the problem is that the traders using HFT eat up all other sources of liquidity and when they shut down their shops, basically no one is left.