What makes you believe your idea is any good? I don't mean to be negative, pejorative or whatever. Nothing personal against you. But I'm really pissed off by such a thread. It is clear you have a looooong way to home. There are lots of things that ...

Good idea for an EA - 'hedging'

I thought you were still using the maximum positive and negative excusion to estimate the direction; what you call MLE/MSE. In which post I missed did you changed this?

Machine Learning with algoTraderJo

The maximum likelihood of the TP is smaller for a small horizon but its variance is also smaller. At longer horizon the maximum likelhood of TP is greater but the variance is much greater. The variance grows linearly with time but the TP only grows ...

Machine Learning with algoTraderJo

FYI 1-3% per day is 22-81% per month or 1090-12049% per year. Not exactly small. Risk reward is not as important to me [...] to build positions If you build positions isn't it because a big RR is important to you?

What is your secret and edge?

Could you please elaborate on this a bit more? Do you mean an indicator which gives a vector of N characteristics of the price action instead of N indicators each in charge of estimating one of them?

Cognitive indicators and neural-networks

To me it still doesn't make a lot of sense. I also thought it was about the bias/variance tradeoff but algoTraderJo replied me it wasn't. I thought it was about the determination and relevance of the independent variables but algoTraderJo replied me ...

Machine Learning with algoTraderJo

That's why I prefer using the name of the timezone "Europe/Paris". The library (joda time) does the rest. The trading fees models are different. FXCM adds a premium on top of the spread with no commission. Dukascopy adds no premium to the ...

Machine Learning with algoTraderJo

I don't get the same thing as you do. But I think the difference comes from the lines 39-40 in your script: mle(j) = hi - op; mse(j) = op - lo; I use the % return: mle(j) = (hi - op)/op; mse(j) = (op - lo)/op; I'm happy the see that you also get ...

Machine Learning with algoTraderJo

Ah ok. English language tends to be too weak sometimes. Like the word "random" for which I found more than 10 different translations in French (!). Do I understand correctly? 1. Curve fitting bias is about the "degree of freedom" (complexity) of ...

Machine Learning with algoTraderJo

Can you explain us how you make this miracle? Machine learning is about infering a model from the data. Per definition you cannot measure how wrong your model is (=data mining bias) since you don't know the real underlying model; not even if there ...

Machine Learning with algoTraderJo

On E/U with SL=30% ATR from 01/01/2001 to 27/05/2015, changing one parameter at a time: A B C D 9 110 4 8 147.24% 8 110 4 8 -81.52% 10 110 4 8 -24.50% 9 109 4 8 198.52% 9 111 4 8 131.50% 9 110 3 8 113.25% 9 110 5 8 2.81% 9 110 4 7 192.90% 9 110 4 9 ...

Machine Learning with algoTraderJo

If I generate 160000 variants of the lin.reg. algo I will always find one that outperform all the others regardless of the dataset used. Trivially argmax k { Sharpe(algo(A_k, B_k, C_k, D_k)) } always exists. And it will certainly outperform almost ...

Machine Learning with algoTraderJo

What do you mean by "distribution of expected candidates"? Is it the distribution of the 4-vector or the distribution of the equity curves?

Machine Learning with algoTraderJo

I wish I'm wrong but I'm afraid you overlook the problem. Do you mean using the very same system, i.e. the vector (A,B,C,D) is fixed, ran on the N random samples or searching a new vector (A,B,C,D) for each of them? The second option shall be ...

Machine Learning with algoTraderJo

This paper is incomprehensible to me. I thought West's notations were awful this guy beats him by a mile! Is there an unciphered version somewhere? From the few I could get he proposes to test the hypothesis that the strategy performs no better than ...

Machine Learning with algoTraderJo

I think this is the reason FXCM avg spread 2.4 pips <-- commission is included Dukascopy avg spread 0.85 pips <-- commission is not included

Machine Learning with algoTraderJo

FYI when A=0 the spread is more around 7.

Machine Learning with algoTraderJo

To be honest I don't understand the test you propose. If I detrend the dataset (EUR/USD returns you mean?) and shuffle them I'll simply get a random walk of some sort. If I apply the trades at the same points of time for the same duration and with ...

Machine Learning with algoTraderJo

Datamining biais. I think it is time to talk about it. I can now reproduce the experiment. KaBo will certainly do as well soon. But is this system a real one or a lucky realization? AlgoTrader said that he found A, B, C and D using a genetic ...

Machine Learning with algoTraderJo

It also shows that I can perform a linear regression by hand with no third party framework.

Machine Learning with algoTraderJo