Interestingly when you datamine for systems you'll notice that those with out of sample edges tend to have high Risk:Reward ratios. You could therefore say that although a high RRR does not guarantee an edge, it is true that a system with a higher ...

Oh mate, how careless! You're right. I made a mistake on my post, I mean an edge where you have 35% winning trades (clearly an edge with 66% winning trades would be gargantuan). One in a million is equal to 0.0001% success probability (100*1/1*10^6) ...

Depends on many factors and the way you do it. Complex money management on each system is not fundamental, you can simply pile up genetically generated approaches that have been properly screened for edges and you can obtain very complex "smart ...

If you already have good knowledge about basic statistics I would suggest researching about data-mining on synthetic data (artificial data created with the statistical properties of a given financial time series), that should keep you busy for a few ...

Recently Hanover suggested the blog mate, but I don't know the fellow. Yes, but not sharing trade secrets mate. I won't do your homework. This comes from my experience with exploitation of edges in backtesting data and then trading them live, trade ...

very good way of doing this. Part of our analysis involved Kalman estimators. However Hurst exponent and auto-correlations are key. Take ideas from here: url .

I can give you a few facts here: Currency pairs tend to trend, their distribution of returns shows fat tails and their price series show autocorrelation. I advice you to run statistical tests before you "believe" a trend exists. There are many ways ...

Nice to be quoted! I had never seen the blog or this guy, I'll check it out to see if he is as you say. I just saw this article while performing a quick google search.

Nice work compared to the rest of the "herd" guys... I've been working on automatic system generation based on genetic programming for several large institutions for the past decades (being independent and successful using automatic generation ...

You need to think about entries/Exits in term of the mathematical expectancy of trades -- in terms of statistics. The maximum favorable excursion (MFE) and the maximum adverse excursion (MAE). Your entries determine your MFE and MAE, therefore your ...

You had a short trade which is exited at the Ask, what you see on your charts is the Bid. This is the reason why you don't see this price being reached on your charts, because your charts do not depict the Ask price (which is where your position was ...

I disagree with the horse racing analogy, it assumes that data-mining has to be dumb and can have no insights into causality. You can be successful with a pure data-mining approach (I have 15+ years of experience doing this) The problem is that you ...

Simple price action based rules but you can do more complex rules without any harm (I tried much more complex rules at some point with the same success). Be careful however to always perform many in sample Vs out of sample runs to ensure that the ...

Searching for a negative edge is not an intelligent idea because there are millions of combinations that give down-slopping equity curves, from which you will have a very hard time distinguishing which have a "negative edge" and which lose because ...

Reversing a losing system doesn't give you an edge. Most people don't lose money because they have a "negative edge" a more than random chance of losing money above the spread, but because they are following a negatively biased random walk (negative ...

Your problem will be that the probability to have +2 consecutive losses is very high. On a Monte Carlo simulation of a system with a 2:1 reward to risk ratio with a 66% winning percentage, the probability to have +2 consecutive losses in twenty ...

There are a few things I would like to contribute here, from an objective and quantitative manner: You can achieve great profits in Forex with only a few trades. You can theoretically make millions from 50 USD in less than 20 trades if you have ...

All these are just opinions about whether price is random or not. Does price follow a random walk? You can test for this!!!!!! For example you can use this statistical test : http://ideas.repec.org/a/eee/econom/v151y2009i2p140-149.htm. If you run ...

What I see are tons of opinions about what "change" means to different people. Zero quantitative insight This is quite sad as something like market change can be easily defined and calculated. Does anyone on this forum care about analysing the ...

Well, what does it mean for the market to change? If you mean change in the sense that it becomes more or less like its past trading action then this is something you can calculate. You can measure the distribution of returns from a market ...