Just for clarification, is the D value in hours or days?

Ok thanks, I think I now have it. I am basically searching for the best 10 coefficients that when applied to every vector give the smallest target error. I do the same for mse and mle and end up with 20 coefficients. Once found I apply those to the ...

Ok, after looking at the shark library I still am a bit lost on how your model works. I understand that using the input data you generate a regression line that gives you a slope and an intercept, but how does that apply to the MLE and the MSE?

It's good to see you back on this thread. You have posted details on the inputs and target outputs so that the rest of us can try to reproduce, but you have provided little detail on what you actually do to "build a machine learning model using ...

So just so I understand, you are trying to develop a system that triggers in the single digit seconds but you have not back tested this concept, so you are just theorising? Why don't you open a free demo Dukascopy account (I'm sure there are other ...

Hi algoTraderJo, Could you provide a few more details on your neural network so I can try to reproduce your results and move this conversation forward. I'm not entirely sure what you mean by inputs using the last 3 bars of 100 examples. Does this ...

Whist you need 58% prediction accuracy to make any profit, don't forget that as well as prediction direction, you also need to predict if the price will cover the spread. If you start doing some analysis on each timeframe and the percentage of bars ...

I'd like to see a comparison using different bar durations. For example, selecting a single test methodology, can you plot side by side the results using daily, 4 hour, one hour, 30 mins and 5 mins data as an example?

@PipMeUp I get the same losing results. I normally keep quiet on FF and only appear when there is something interesting being discussed regarding machine learning. This thread is only a couple of pages long but I hope that it does not turn into yet ...

A good graphical example of this is below.

An interesting take on the trend following turtle system. url Looks like it was profitable for the original turtles but the system has lost it's edge.

Yes this is one approach. Finding cointergrated baskets is the first problem, but it solvable. You then need to know when to enter and exit the basket, this is normally based on some std dev or some other measure. This is not so easy to solve and ...

I think you should be aiming for a system that uses many pairs over a longer time frame. Yes I know that by trading a basket you are actually only trading the currency exposure from that basket, but even so a basket will let you trade profitably ...

I think your are underestimating the stress you are causing some of us behind the scenes...

Give a man a fish and you feed him for a day; teach a man to fish and you feed him for a lifetime..

I agree that you need to be able to quantify the trend. You might find this interesting url

More details or info on the probabilities would be appreciated. I can understand your concept in parts, I have even calculated the probabilities but am yet to find a way of trading them successfully, I suspect I am missing something to filter the ...

I love these magic screenshots! So, for the first buy, how did you decide to buy there? At that point in time the price had just dropped a huge amount and you had no purple trend line so what would have made you buy at that point? For the exit, ...

You only need 52% to get rich if you pay no costs and have the tightest of spreads. Run a simulation using real world spreads and costs and you'll find that you need a lot more then 52%. Eg, as a staring point, look at the real world trading costs ...

Hi Killian, What do all the different colours mean on your chart? Eg what are the arrows and boxes? For binary options remember that you have to predict the next bar direction and also make sure that the next bar covers the spread eg if you use ...