The importance of trading costs over the long time
Trading costs are something that few people takes into account on automated trading even when it affects a lot on the profitability.
Depending on the kind of system you are running, trading costs can have a very important impact. On systems such as scalpers, that targets just a few pips on every trade, trading costs are incredibly high, but even if you trade only long term profitable systems (which should be more or less spread independent and broker independent) the impact can be quite significant.
Trading costs obviously increases the risk too, as it adds a handicap for a trade to become profitable reducing the real reward/risk ratio.
What are the trading costs?
Trading costs refers to the cost of opening a buy or sell in a given instrument. The most important trading costs are the following ones:
- Spread: Represents the difference between what the market gives to buy from a trader and what it gives to sell to a trader. All brokers will charge you the spread on any trade.
- Commissions: Some brokers will charge you an additional commission above the spread on any trade.
- Swap: Is the interest paid (or received) for owning a given amount of the underlying instrument. Usually the interest you pay is much higher than the interest you receive.
- Slippage: It is the difference between the price your strategy want’s to open a trade and the price that this trade is actually filled at. This includes broker execution time, volatile markets, lack of liquidity or simply slow network or big distance to the broker servers.
You have direct or indirect control over some of this costs, and some others are outside your control. Obviously it’s our responsibility to try to reduce trading costs as much as possible if we want to succeed.
How much am I really paying?
To make it easy, let’s talk just about the spread and we will add the rest of the costs later. Suppose your broker is charging you on average 1.5 pips on every trade. Is it too much? It all depends on the systems you are running:
- For a scalper that targets 3 pips on each trade, the broker is charging you 50% of the estimated profits! Both on winning trades and losing trades!
- For a system that targets 50 pips on each trade such as RobinVOL, the broker is charging you a 3%
- For a trend following system that targets 300 pips on each trade, the broker is charging you a 0.5%
Now we need to add slippage that, depending on the broker, could be between 0 and 5 pips (although negative slippage -that favors you- is possible, is uncommon). This is very important: take into account that slippage is not present on backtests, it only applies when trading live.
Now we need to add the swap, which is an overnight cost of all trades opened. Can be positive or negative, but as the slippage, the positive swap is usually smaller. The swap changes over the time (the interest rates changes).
For example, RobinVOL in a whole year can make about 500 trades and return an average of 2500-3000 pips. If delays, spread, slippage … if a bad broker adds just one pip on every trade, then you are loosing 500 pips/year, which is a 20% of your yearly profits.
How can I reduce the impact of trading costs on my results?
One obvious way is to use strategies that trades less. But that is not the point. We need to run the best strategies based on our analysis and then understand and try to reduce trading costs at maximum.
The spread (and commission) is one obvious thing that we can manage by choosing brokers that charges low spread on each trade. In the example above, if Broker-A charge on average 1 pip and Broker-B charges 2 pips, RobinVOL will make at least 20% more pips on Broker-A.
When talking about execution time (which becomes slippage) the quality of the broker is important, but you can help reducing it much more by running your strategy near the broker machines. If you run your strategy on your home computer with a ping above 100 ms, you will probably pay another 300-500 pips/year or so due to slippage compared to someone that runs the strategy on a VPS with a ping less than 10 ms.
Obviously downtime is very important too, the same as using brokers that runs Virtual Dealer plugins, but I will not consider them trading costs. Just consequences of bad choices.
A profitable strategy that makes on average 2500 pips / year, executed on a good broker and running on a VPS near the broker server will perform much similar to simulations.
But the same strategy on a broker that charges 2.5 or 3 pips spread, that have slow execution times and run on a home server might become a 500 pips/year or even become a breakeven or loosing strategy, all due to trading costs.
And this has nothing to do with broker dependency.
There are some aspects of trading costs that you cannot control. But at least it is wise to improve those aspects that you can control considering that choosing a good broker is free and there are very cheap VPS providers available.14