Recent content by Hugin

Ok. I guess that only some simpler transformation will work when including state decision variables in the model. And if they have a heavily non-linear impact then it will probably not work. One other potentially good thing with not having market state included in the model and instead use...

To me this is the most important question together with how to define the trading model you are going to optimize. The optimizer will make anything to fulfill what the fitness function says and use any hole in the logic. As you say, using net profit will probably not do it (at least in my...

Interesting idea using Hurst exponents. I sometimes find the relationship between market state and the trading model a bit challenging. Should one create a number of models and let the market state enable/disable them or should one try to include the variables describing the state into the...

I my opinion this is valid for any selection/optimization process, even manual ones. To some degree this is what we are looking for - adapting a set of parameters to the data. Personally I do not like the term "curve-fitting" since in most cases (and definitely for the GA/GP) it is actually...

Even if I understand why they exist (easy to understand, apply and at least give some thinking about risk into the process) simple rules of thumb like ”never risk more than 2% in any position”, I think they often miss the point. First, they only refer to the risk of one position whereas...

Interesting, but quite technical, paper. Haven't read all of it and I'm not sure I fully understand the things I've read, but doesn't he just prove that provided that you know the conditional distributions of returns given a certain event, and these distributions are IID, you can find an optimal...

Good to hear from you as well, I check your blog quite frequently. The post on MINE is bookmarked for later reading. No, it goes back to norse mythology and history classes in school, but since Hugin Expert is a Danish company I guess they choose their name for the same reason.

Yes, it is definitely so that defining the model is the trickiest part. I wish it wasn't so, so you could just throw technology at the problem. I have always found it hard to find the right balance between the number of parameters (implicit and free) in a model and the expected walk-forward...

We have done some work with CART like classifiers within non-trading applications. They are easy to use and you can easily understand the results, which is not necessarily true of other classifiers. One problem is that they are sensitive to the nature of the dependencies between input...

Generally I agree with this. Money management is important but it is inferior to having an edge. If you do not have an edge and participate in a real market you will most likely lose money. IF you have an edge money management will help you exploiting your edge in an optimal way (given the risk...

Agreed. Maybe linear regression can be seen as a first order approximation. The only benefit is that you basically only estimate one parameter (the slope) which makes it sort of robust (which of course is of little value if it cannot be used for other reasons). One problem with a non-linear...

OK I agree. Have done some basic tests with this indciator and it gives good indication whether the market HAS been trending, at least my eyes seems to agree. So you could calim it solves some of the problem that other trend indicators have. However, as I said before, I really doubt whether...

Well, trend indicators needs some kind of window over which the trend is calculated. They just state that the trend estimate is calculated for the center point of the interval. It is trivial to adjust to only use historical data points and use the regression estimate as an indicator. The big...

One problem with many of the methods commonly used to identify trends is that risk confuse trends with volatile markets. The method described in the paper below claims to avoid such problems. http://www.eurekahedge.com/news/attachments/Amplitude_Examining_Trend_Characteristics.pdf

This discussion gives me the déjà vu feeling. It's been around ET more than once before. Selection bias is indeed not the same as curve fitting, but it is perfectly possible to describe the created model as an instance of a function (or more correctly as a relation) with parameters from a...