What number for slippage?

[Revolutionary]

Hi guys,

I am testing a strategy that trades a couple of very liquid indices, the S&P500 for example.

What would be a normal amount of slippage I should account for in my backtests?

Right now I'm assuming an average slippage that equals the spread (0,2). So actually I'm paying a double spread. Is that too high or too low?

Please explain why you think this is too high/low.

Thanks,
Tim

 

[Revolutionary]

Forgot to say: all orders are Market.

 

[scose-no-doubt]

Personally, -3/+3 ticks, formulate a negative log-normal at about -1 mean and randomise it within 1.5 standard deviations.

 

[Revolutionary]

Thanks Scose-no-doubt,

That means an average slippage of 0,3 for the S&P500 contract, as 1 tick is 0,1.
Never thought of using a negative log, I just use the same amount of slippage on every single trade. I will have a look at that.

Are there other circumstances I need to consider? Is the average slippage on the open and close of the market usually higher or lower?

 

[scose-no-doubt]

I'm tapped out mate. I only trade CFD and not very often.

 

[horace]

Slippage is hugely dependant on factors you are not talking about here (broker's connections to the exchange for example).

If you trade market orders all the time, you need to be on the safe side. Try considering at least commissions x 3. That is a rule of thumb I use that frequently works well except on very low cost brokers (then you may need a x6 ratio).

Thanks Scose-no-doubt,

That means an average slippage of 0,3 for the S&P500 contract, as 1 tick is 0,1.
Never thought of using a negative log, I just use the same amount of slippage on every single trade. I will have a look at that.

Are there other circumstances I need to consider? Is the average slippage on the open and close of the market usually higher or lower?

 

[Revolutionary]

Thanks Horace,

That's a lot more then I expected. That amount of slippage will probably have a big negative effect on the results from my backtests, because the strategy I developed makes intraday trades with very small stopsizes.

ps. Interesting blog you have, I will keep an eye on that.

 

[horace]

Sure Revolutionary. If you trade scalping and more than twice a day you NEED to have limit orders. Try changing your entry / exit styles. Else you'll be unable to make a good idea profitable imho.

Thank for your compliments, my english blog is a newborn, lets give it a chance...

Thanks Horace,

That's a lot more then I expected. That amount of slippage will probably have a big negative effect on the results from my backtests, because the strategy I developed makes intraday trades with very small stopsizes.

ps. Interesting blog you have, I will keep an eye on that.

 

[Revolutionary]

Well the problem is that the system usually wants to enter the market directly when the market opens. Is there a way to work with limit orders when you want to enter at the opening of the market?

 

[horace]

No ther isn't because the openning price is a result from the openning auction and can't be predicted.

You can try following the price using limit orders. Open+3 ticks or Open-3 ticks. You will most probably lose some trades, but in the long term I can guess you will notice no difference.

Well the problem is that the system usually wants to enter the market directly when the market opens. Is there a way to work with limit orders when you want to enter at the opening of the market?

 

[Revolutionary]

Ok good to know.

Hmm I will need to investigate this some more, because I actually want the system to trade fully automatic. Ofcourse I will monitor it daily (when it goes live), but I don't want to put in the orders myself.

 

[Poem 101]

Hi guys,

I am testing a strategy that trades a couple of very liquid indices, the S&P500 for example.

What would be a normal amount of slippage I should account for in my backtests?

Right now I'm assuming an average slippage that equals the spread (0,2). So actually I'm paying a double spread. Is that too high or too low?

Please explain why you think this is too high/low.

Thanks,
Tim

Hi Tim,

I'm not sure what you're using - the s&p trades in quarter points, so the spread is usually 1 tick (0.25). If you've got 0.2, you're trading something else.

Anyway, unless the market gaps on news or you're trading massive size (over 100 cars) slippage is really going to be minimal. ES is seriously liquid.

 

[Revolutionary]

Hi Poem 101,

I was referring to CfDs on the S&P500. I don't really know yet on which specific product I want to trade my system, it could be ETF, CfD, Futures or something else.

Tim

 

[eminiglobex]

For testing purposes allow 1 tick of slippage either side for es to be on the safe side. Sometimes it will be less or more depending on many things (depth of order book, connections etc). If after testing with this (and com costs) you still have positive expectancy then you may be onto something.

 

[Poem 101]

Hi Poem 101,

I was referring to CfDs on the S&P500. I don't really know yet on which specific product I want to trade my system, it could be ETF, CfD, Futures or something else.

Tim

I see. Well, it will make a difference to what you're trying to do. If your capital allows, go with the futures, slippage will be minimal. If you're trading something synthetic it will be up to your broker basically.

 

[Revolutionary]

@ eminiglobex: Thanks! I've tested my strategy with 2 ticks of slippage and 0,2 spread and the result were pretty good.

@ Poem 101: Yes, I need some very liquid markets and contracts because I want to keep the slippage to a minimum. Large slippage on small stop sizes can really ruin the overall performance.