Thoughts of high-frequency trading, electronically induced volatility, dark pools, errant algorithms and latency arbitrage scare the bejesus out of the market. Doomsayer headlines, questions on gaming and discussions about nanosecond latencies have traders worried that speed has become the most important aspect of the business at a time when it is so low it is difficult even to measure.

The challenge with algorithmic trading is that the world is anything but simple; there are more than 50 venues on which to execute U.S. equities. Complicating matters is the fact that each venue offers multiple matching models driven by multiple order types.

Trying to understand who is trading in these venues and how your orders are being executed can stymie even the smartest, most experienced traders. But who provides your liquidity absolutely can impact the quality of your execution. Historically, market makers provided liquidity at lit venues, such as exchanges; now, however, it is not uncommon for market makers and other high-frequency traders and arbitragers to have direct connections to dark pools.

The problem with an increased number of high-frequency types in dark pools is that, while the hit ratio will increase, so will the odds that your order is being seen by an HFT shop before it gets to the exchange. This information can be used to influence the overall market, as well as any subsequent trades associated with your initial order. In effect, the more HFT flow there is in dark pools, the greater the amount of information that is being leaked.

Now, not all high-frequency liquidity is bad, and not all buy-side liquidity is good. Much of it depends on the trading and routing strategies employed -- certain routing strategies give up more information than others. Traditionally, algorithms flooded the various markets with orders, for example. But sending transactional waves, even (especially!) into dark pools, can notify traders of an impending large order.

As a result, a number of brokers are taking a more in-depth look at trading algorithms and dark pool and routing strategies with the desire to remake these tools safer and less able to be gamed. There seem to be three major strategies being employed to improve execution quality: synchronization, proximity and analytics.

Synchronization strategies take into consideration the various market latencies to ensure that all orders reach their destinations simultaneously, reducing the ability of liquidity providers to pull their orders as liquidity is taken. Proximity solutions leverage various colocation strategies and divide larger orders into smaller blocks that are executed independently within each colocation center, eliminating the latencies involved with obtaining, normalizing, analyzing and trading against multiple market data feeds. Finally, firms are employing more sophisticated analytics to determine if they are trading against a toxic flow and, if the flow is toxic, to limit its impact, make it less toxic or kick the toxic stream out.

While it would be nice if everyone traded in ways that didn't impact others, this is both naive and probably impossible. Yet this is how a market should work -- customers who feel disenfranchised complain; if their service providers don't respond, customers shift their flow to those that do. While this change process may not occur overnight, insight and service typically win out over lethargy, and that's the beauty of competition.