Prototyping FX strategies in Tradestation

L

lolatency

I'm using EL to test some theories.

For FX, Tradestation is basically printing off the bids and asks that are quoted from GAIN Capital, the owner of forex.com. However, if you run a sub-10 minute strategy, there's a serious problem with backtesting because the bids and ask charts are different and the spread is a serious factor.

So for example, if you sell short and expect to book a profit of 5 pips, your strategy won't work in the real market because the ask price could be a full 5 pips up, since it's reflected as [pairname].A.

How do you FX traders using this platform get around this? The strategy reports are not going to merge trades done on the bid chart with the trades done on the ask chart. Take the generated trades, print them to a file, and then go run the strategy against real market bid and ask values that you picked up off of ECNs?

I'm looking for an easy sleazy hack here.
 
Quote from lolatency:

I'm using EL to test some theories.

For FX, Tradestation is basically printing off the bids and asks that are quoted from GAIN Capital, the owner of forex.com. However, if you run a sub-10 minute strategy, there's a serious problem with backtesting because the bids and ask charts are different and the spread is a serious factor.

So for example, if you sell short and expect to book a profit of 5 pips, your strategy won't work in the real market because the ask price could be a full 5 pips up, since it's reflected as [pairname].A.

How do you FX traders using this platform get around this? The strategy reports are not going to merge trades done on the bid chart with the trades done on the ask chart. Take the generated trades, print them to a file, and then go run the strategy against real market bid and ask values that you picked up off of ECNs?

I'm looking for an easy sleazy hack here.
More...

Use the closing price and add 5 pipis (or whatever) of slippage to each trade. Anyway, the spread changes all the time. During fast markets you may get wild openings of the spread.

A more complicated way of doing it is using two timeframes, 10 minutes for the system and tick for the orders.

You sound like a newbee:D Good luck to you! Forex is tough but there is a lot of money in it.
 
Quote from intradaybill:

Use the closing price and add 5 pipis (or whatever) of slippage to each trade. Anyway, the spread changes all the time. During fast markets you may get wild openings of the spread.

A more complicated way of doing it is using two timeframes, 10 minutes for the system and tick for the orders.

You sound like a newbee:D Good luck to you! Forex is tough but there is a lot of money in it.
More...

5 pips slippage, I guess will work; however, let me ask you this: How do I model passive strategies where I don't remove liquidity? Supposedly there are new brokers offering straight up commissions and allow you to throw orders out into the ECNs.

Do I get the bid/ask on the tick chart? The biggest reason I like tradestation is the reports it gives me and that optimization feature. That is a fantastic feature.
 
Intraday Bill,

Just for thoroughness, would you mind telling me what you have placed on this screen? I get wildly different results if I go with per trade vs. per share/contract, and I'm unclear as to what "contract" means in FX spot. 1 100k lot? 1 10k lot?

If you wouldn't mind taking a moment to tell me how you parameterized the backtest. For me, it's intuitive for equities. For FX, I am somewhat confused -- particularly when I am trading a pair that has no connection to the dollar.

Here is the screen I am talking about:

<img src="http://media.chartvibe.com/backtestscreen.jpg">
 
You must log in or register to reply here.
Back
Top