In this episode of Truth About FX, Walter gives away some very reliable points how to get the most of your backtesting data. He talks about market segments, market types, and currency pairs…so what do all these have to do with your backtesting?
Walter affirms the need to look at different point of views and the importance of forex testers. All this and more here in this episode…
Download (Duration: 04:16/ 4.89 MB)
In This Episode:
00:42 – hire someone
01:56 – pull sections out
03:20 – take snippets
What style of market are you in? [Click To Tweet].
Capture different styles of markets. [Click To Tweet].
Take snippets of different pairs that you’re trading. [Click To Tweet].
Announcer: Sometimes, forex trading is a wild and wooly place to be. That’s why Hugh is here, to post your questions to Walter, the naked forex guy. Hugh’s got questions and Walter’s got the answers. Here at the Truth About FX Podcast.
Hugh: Hi, Walter. I’ve got this question recently. This person says, “I want to backtest a day trading system” — I think it’s a scalping system, to be specific — “but there’s no way that I’m going to test 16 years of data for every currency pair. What do I do to make sure that I am getting a good backtest?”
Walter: Okay. Maybe you can hire someone else to do the backtest then hire them to do trading for you. I don’t know. It’s like saying, “Hey! I really, really want to be a brain surgeon but I just don’t want to go to medical school. What am I going to do here?”
I don’t know what to tell you you’ve got to earn it somewhere. Here’s what I would say — all silliness but that’s really kind of. This is what the person is saying, “I want to do the scalping system and I want to verify it but I don’t really want to do the work that is involved when I verify it.”
You could do an automated backtest but that doesn’t make a lot of sense because you have data problems in most cases and you also — if you’re going to be trading it manually which is what it sound like — you probably do not want to do an automated backtest because you probably going to be trading it differently than the computer would.
The other thing is, 16 years of data. So, let’s say for every currency pair, this is what you could do; you could take some random sample periods. Let’s say, you want to trade on the EUR/USD, the JPY/YEN, the CAD/USD and the AUD/NZD.
Let’s say those are the ones that you want to do. What you might do is you might pull sections out. You go, “Okay. We’ll, see what it’s like trading it on the EUR/USD from December 2013 to March 2014 and then see what it is like trading it, that same system on the EUR/USD July 2006 to December 2006.”
So, you take segments and what you want to make sure that you’re doing here is you are capturing different styles of markets like different seasons, so to speak, trending markets and directionless markets.
You want to see volatile markets and quiet markets, you want to make sure that you’re taking snippets of periods that you’re testing. The system is being put through the ringer. You want to make sure that seeing different styles of markets. That is what I would do.
The other thing is you do the same thing with all the pairs that you want to trade it with. I guess you don’t necessarily have to go through but I would definitely try to get… It seems there’s a difference like pre-2002. If you have pre-2002 data versus post-GFC data, there’s a difference there I found.
So, I would want to make sure that it’s fairly robust. Like, it doesn’t work in 2001 and doesn’t work in 2017, that sort of thing. I would want to do a little bit of that. In all of my samples aren’t just going to be in just 2005 to 2008, that sort of thing.
So, that’s what you could do. You could do that and take snippets for the different pairs that you’re trading. Just make sure that they’re far apart and that they include different types of markets.
Also, you could look at it from a number of trades point of view. Make sure that you’re getting 500 trades per pair rather than 16 years of data which I don’t know how many thousands of trades that would be. It sounds like it’s going to be a lot for this system.
Also, make sure that you’re testing it in forex tester. For example, when you’re sleeping and you’re not trading it. You’ve got to be careful. Sometimes, you’re taking signals from forex tester that you know you would’ve missed because you’ve been sleeping or whatever.
Those are the things that I would look at for but that’s how I would approach it.
Hugh: Cool! That’s an awesome advice.
Walter: Great! Thanks.
Hugh: Thanks, Walter.
Podcast: Play in new window | Download
Subscribe: Apple Podcasts | Android | RSS
Leave a Reply