In a previous article, we explored the concept of spreads and the various market conditions that can affect them. We will now expand on those ideas and delve into a topic that can sometimes be the source of much frustration and confusion for traders: slippage. We return to our main character Bob.

It is Non-Farm Payrolls day and Bob is sitting at his computer smashing the refresh button waiting to see what the US Bureau of Labor Statistics has to say about the current state of the jobs market. He has a sneaking suspicion the news will be good and expects indices to pump across the board. Bob has been meaning to go long on the DJI for a while, but like any sensible trader, he waits for the confirmation he needs before placing the order. Any second now.

There it is! The results are in: the numbers are better than anticipated. A hive of automated bots spread the news around the financial web, economic calendars everywhere update to confirm the better-than-expected result, and social media is buzzing already. The coast is clear, it’s time to place the order and make some money. Bob sets up the trade, position size, leverage, price, all good to go. There is a slight tingle of euphoria as he clicks “buy”. Time to sit back and rake in the gains.

As the adrenaline subsides, Bob looks again at his screen and quickly notices something is wrong: that was not his buy-in price. He got in lower than that, he’s certain of it. His P/L should be higher than that by now. What’s going on here? Was there some kind of mistake? Is the broker trying to rip him off?

The answer to the previous couple of questions is a resounding “no”, but it is important to understand what is going on behind the scenes.

Simply put, the problem occurred because there was a delay between Bob clicking the “buy” button and the trade actually executing. It is an issue of latency and lag, that is to say the delay between the press of a button, and the consequences of that button press being applied. Much like attempting to purchase a concert ticket online only to find that by the time the order is received, the tickets are all sold out.

In Bob’s case, after he submitted his market order, the transaction was sent over the internet to a server, possibly situated very far away, handled by a matching engine responsible for dealing with millions of requests every second, before being settled to the best of engine’s ability. This type of software typically works on a first-come first-served basis, which creates an advantage for those with a faster connection to the service. The difference between old school copper wire and fibre-optic cables can make a difference, as can the distance from the server, as well as the capabilities of every single piece of hardware and software between Bob and the final settlement of the trade. Every single obstacle can slow down the passage of his transaction in a non-trivial way.

To give an idea of how vital a factor this can be, particularly in the insane world of high frequency trading, there are firms that spend millions of Dollars just to purchase office space as close as they can to the location of exchange servers, all to reduce latency by mere milliseconds. Make no mistake, this is what Bob is up against.

Intricacies of low-latency communication aside, the biggest factor to take into consideration when it comes to slippage is market volatility. Greater volatility leads to greater spreads and greater slippage. During times of important financial releases, such as the NFP example we used initially, there is a lot more going on than usual. There is a lot more chaos behind the scenes for the order matching system to sort out. Picture a million traders all shouting out their orders at once, some of them standing at the back of the room, others screaming straight into the brokers face. Poor Bob is at the back of the allegorical room, and by the time his buy order is processed, ten more have pre-empted it. They simply got there first. By the time it gets round to Bob’s order, the price has changed, sometimes dramatically. It is worth noting that in some cases, traders may encounter positive slippage, meaning the price they got is in fact better than the one they wanted.

Outside of poor price entries, perhaps a much worse scenario is the case of a stop loss not triggering at the correct level, causing a position to stay open for longer than intended. The consequences of this can indeed be quite severe, but the cause remains the same. During periods of high volatility, by the time the stop loss triggers there may be no orders to take the other side of the trade, forcing the order matching service to find the next closest match.

It can also have very serious consequences for those attempting to hedge. Unusually high volatility can lead to traders putting in a hedging trade at the wrong price, leaving them more vulnerable to margin calls and potentially exposing them to greater swap fees if they then have to keep the hedge open longer than they were expecting.

Because of the extra steps involved, the consequences of high latency for something like copy trading can also be disastrous, leading to completely different results for the master account compared to those under it. Latency is the reason VPS providers typically keep their servers in close proximity to the vast data centres responsible for handling global Forex trading. Incidentally, the NYSE and NASDAQ are actually hosted in New Jersey, not in New York, so a trader would be better off in suburban NJ than in lower Manhattan.

Like with any complex system, things are designed to work within tolerances. The vast majority of the time the system functions comfortably within those tolerances. There are times however, during which those tolerances are tested, such as periods of extreme price swings. The consolidation and execution of millions upon millions of trades require a colossal amount of infrastructure, the technical limitations of which sometimes require an element of conscious mitigation.

So what is the solution? As we alluded to earlier in the article, traders can avoid much of the damage by not trading during periods of intense market volatility. Unfortunately, this does not help in the event of stop losses, which were presumably set long before any risk of slippage came into play. In general, using limit orders is safer than using market orders, with the obvious trade-off that the desired price may never come, but this is by no means infallible. A more encompassing approach is to adopt the practice of factoring in potential slippage to any trade, accepting that the desired price may not quite be the one we get, and to implement a trading strategy with this in mind.