I've written a program to back test my futures trading ideas. One of the test variables is transaction cost for which I have been using a $100/rt default value. I have been wondering about the reasonableness of the default value and the potential benefits of varying the transaction cost based on volatility, tick size, tick value, volume or other factors. Any thoughts on this topic would be appreciated. Thanks.
not sure what time frame your system is looking at but at $100/RT transaction cost, it must be intermediate term swing or long term trend following. Please do tell?
I apologize. The ideas I am testing are intermediate to longer-term trend following systems. Trades last anywhere from 1 day to many months. Thanks.
I must misunderstand your question. Assuming trading in normal hours, I would have thought the variables are the product and the number of contracts being traded. If it is the SP you are trading, $100 is too low. For <5 contracts of the products I trade, I would suggest: -ES $30 -YM $20 -TF $25 -CL $30 -ZG $50 I use $5 for commission and the rest is slippage. You'd need a higher number for slippage if you're looking at >4 contracts (except on the ES). Of course, having a "profitable" test after transaction costs doesn't mean you have found a system that will actually work. I'm interested in what others think about my transaction cost suggestions.
I trade ES primarily and I think 5 dollars for commish and 2 ticks slippage is correct for RT costs. 2.50+1 tick in and the same on closeout of trade. I think the other costs look good as long as they are traded electronically, otherwise I would double or triple them.