That's a long question... Uncorrelation of strategies and/or instruments you trade is the key to reducing risk. It's in my view the most important thing you need to do I would do both and use as diverse systems and markets/instruments as possible. ...

I don't trade spot FX anymore for tax reasons. Just futures + stocks. There were times when I was trading much less and doing mostly tech startup stuff, now again more involved in the markets.

Here is an out-sample backtest curve of my current system (long/short equity). Live is not much worse :-D

Very, very long time since I posted here. Are any of the old beards in this thread still active? I'm trading quant long/short equity strategy nowadays. Post here if you are alive, it would be nice to hear what everyone is up :-D I have been posting ...

This is so excellent paper and probably so interesting to people reading this thread that I'll post it after a looong silence in this thread. Enjoy! 101 Formulaic Alphas Abstract: We present explicit formulas - that are also computer code - for 101 ...

Here is some really interesting research stuff from Bundesbank url

Can you please advertise your trading somewhere else, this thread is about discussion of market randomness, not your trading. Start a new thread where you can brag with your 2 months of trades.

And it's also important to remember that there are participants in the market who want to trigger the stop levels just to get commission for the stops.

That might be true in equities but not in fx (there is no book on fx) - also, with dark pools there is no real "level 3" in equities either anymore and even if there would be, hedge funds would not see it (there are rumours but that's it). Also, if ...

The thing with Osler paper is that what she's talking about is stop cascades, ie - placements of stops triggering cascades. And people (even the big players) will put their stops to psychological levels like above/below peaks, round numbers and so ...

You can't model the market based on simple probability density function. Hell, you can't even model the market with simple markov process (because of volatility clustering). If it would be that easy, trading volatiltiy options would be trivial. See ...

Read the Osler paper. When you understand it, it will be a true eye opener. Microstructure is the key to knowing shor term price movements and short term price movements will sometimes (with larger percentage than 50% that is) lead to larger price ...

Does the hypothesis hold really? For example here is one from Osler (I really like this one) url Here is a book that every random walk advocate should also read url

That was kind of my point, futures market has [limited] order book for fx.

You can use futures for estimating volumes to some extent or if you are looking for longer term picture (check COT report for some non-chart information *). Other than that there is no way to estimate flow as the fx is not centralized and FX is one ...

Quantum random generator isn't random either, photon particles give the numbers! Or thermal noise! Now we know what some quantum random generators are based on photons, can we predict the results?

Ensembles always work better than single estimators. There is a lot of research about different estimators but the problem always almost (especially eith trading IMHO) comes to how you build the feature set for your regressor/classifier. Here is a ...

By the way if someone is interested in more basic systematic stuff, here is some really excellent stuff url

This might be a wrong forum for this question but has anyone tried using restricted boltzmann machines to predict time series data? PM me if you don't want to post to thread.

What are you trying to say, that doesn't make any sense.