Hi all! I’ve been trading for about 5 years - algo trading on the majors, doing 5 to 50 intraday trades per day per symbol, flat at the end of each session. I’ve been analyzing my trades - profits and losses, and recently I have figured out the exact impact of slippage on my trading results. Below is the explanation of my measurements and calculations. Pretty substantially not in my favor I must say. I built a script to calculate the differences between the prices my algo sees when sending trades and actual execution prices. My algo is hosted on a VPS CNS and I’m trading on a Pepperstone Razor account, so the latency to broker’s server is not more than 1 ms. I’ve collected data for more than 1.5 months and now I have stats on 4621 trades and got following results: This means on average I lose about 1.71 pips (4th decimal) on slippage each trade . Having 10 trades a day means that I lose on slippage 17.1 pips a day -> 376.2 pips per month -> 4,514.4 pips a year. When I calculate the sum in USD the figure really scares me - it’s 4 times more than I have in my account!!! If I could keep this money I’d be way closer to retiring. At your brokers - what slippage do you see? What do you do to reduce your losses from slippage?
Thx for sharing these interessting results with us. I've no experience with the FX markets yet, but I just wonder: wouldn't limit orders (instead of market orders) help here?
Something wrong with your methodology. The spread costs both ways, opening and closing should have symmetrical slippage.
Slippage doesn't have to be symmetric. It' probably a short term momentum system, which chases the market immediately after sizable price movement or even widening spread. When you chase a fast market like that, of course you are going to get a larger slippage.
Thanks for providing quantitative analysis on a statistically significant sample size of something most of us just anecdotally complain about, this is great info! Do you mind telling us the broker, if that is allowed? I don't know that this is good or bad on them, since you're the first person to provide such detailed info, but it would be great to get some comparisons that are as rigorous as yours.
Unless there is a structural bias in the traded instrument slippage should have a basic symmetry on entry and exits. Slippage ratios of 5:1 to 8:1 between entries and exits is not normal. More information is needed, such as a volatility measure before entries and exits.
Use stop limits for entry. Otherwise, just chalk up the slippage to cost of doing business. Or you could try a different broker. Try two or more brokers, give them the same orders and see who consistently gives you the best prices.
Limit orders guarantee the price, but not the execution. So I can not always go using limit orders. That’s actually the reason why the slippage problem understanding arose.
Slippage we see depends on available Market Depth (volumes, spread) and quote/trade/execution report processing time and volatility during entering or exiting the trade. So opening and closing slippages can be different (especially for momentum driven trading strategies).
The stats were collected on a Pepperstone Razor account. In the meantime yesterday I’ve tried out the following combination: an FXCM standard account + TradeMUX trading platform. FXCM underlines a great pool of competing liquidity providers, TradeMUX talks about trades processing time of 100 nanoseconds. So here are the stats: Looks cool to me so far... 40,19% less round turn slippage - really cool... I’ll collect more stats - let’s look at the figures during the week…