So I’ve finished up my research on the breakout strategy described in the previous post. While I think it has some merit, I couldn’t find a configuration that produces a sufficiently high expectancy for me to run it long time. I did find acceptable configurations for AUDUSD, AUDJPY, GBPAUD, and GBPJPY. If you email me, I can let you know the ranges I used.
However, all of them make most of their money on big trends. Given it takes days/weeks before a new high is created, it is quite a laggy strategy so there are a lot of mini-trends that usually don’t pay off because we come in too late and ultimately get only a small gain/loss. So taking into account the lag, it makes sense it makes most money on big trends. Those kind of trends don’t come very often, and you could go through very long periods of small loss after small loss before a strong enough trend comes to erase those losses.
For example: The most profitable configuration for GBPJPY for 2006 to 2013 for a $7000 account with a 10k notional had the following changes final balance over 2 year periods.
2006-2007: $3000 Win
2008-2009: $10600 Win
2010-2011: $2600 Loss
2012-2013: $1200 Win
Now that’s not bad you might think, at the end of the 8 year period with quite modest betting I increased my account balance by 170%. Certainly not bad, but let’s look more closely at this.
The max fall in equity during the 8 year period was ~$6000. That’s 85% of the original account size – could you stomach that?
What about not making any money for the entire 2010-2011 period – is that acceptible? Well, it’s not for me. Ultimately, there two major things I don’t like about this kind of slow strategy.
Firstly, there isn’t enough feedback, or it is too slow. When I optimized this strategy over the last 8 years of data, the most winning strategies had less than 50 trades total. That isn’t really enough data. Also, once you are running the strategy, with such infrequent trading going on you really have no idea if it is still good?
Secondly, this strategies success comes from rare events. We are waiting for that strong trend to come along (like in 2008/2009) to give us the big payment. While I’m sure there will eventually be a trend, the problem is you don’t know when it can come. A financial crisis seems to come every 5-10 years or so, but still, I don’t want to go 5 years with any reliable source of income.
I’m trying to find something that can generate money to live off. Even if I could improve the strategy such that it at least didn’t lose money during non-optimal periods, 2 years of no income isn’t really acceptable. You could try to run a fleet of strategies across multiple pairs, but usually strong trends are done across the board in varying degrees.
So while the idea seems okay, and the strategies seem to work, they aren’t really what I’m looking for. If I had a few grand to dump in account and let it run without depending on it, then it might be a good strategy to run.
I wonder if these kind of strategies worked better in the old days before HFT and computers came along and changed the way the markets work. Anyone have any thoughts on this?
Anyway, I need something that trades more frequently, and provides more steady income… and I’ve spent a few days researching for ideas again and I’ve discovered something called Grid Trading. It’s genius, and suicidal. I’m going to look into this next so stay tuned.