Another way to look at it, is to consider each M1 bar in the last years and measure how far price has drifted away after 1 hour, 2 hours, etc. Here are the stats for AUDUSD over the last 4 years showing the 1 sigma range (p16-p84) and the 2 sigma ...

A Simple Mean Reversion Strategy

Note that I tuned the lot sizing of the AUDCAD backtest such that it exactly matched the dotted-line which is the Monte Carlo risk-adjusted return scaled at max. 35% drawdown at the 95th-percentile (which is just some arbitrary personal risk limit I ...

A Simple Mean Reversion Strategy

Good question. I have two answers to that. First, you never trade 1 strategy/setting, you trade a whole portfolio. Other settings or other symbols might already have given profit earlier. Here are some other symbols with the same EA: AUDCAD, 32+ ...

A Simple Mean Reversion Strategy

Ah, I see. I just thought that perhaps you used a novel way of smoothing the "spread" using A/X - B/X instead of using A/B directly. Correlation trading is also an interesting topic, but not for this thread of course. Thanks.

A Simple Mean Reversion Strategy

Can you please elaborate why you used these MAs rather than the EURGBP MA-200 versus its current price? It surely looks more smooth, but does it compare to @OP’s spread?

A Simple Mean Reversion Strategy

Inspiring topic! Thank you @Riskcuit! I wrote a simple Expert Adviser* following the basic rules from the first posts: - SMA200 - RSI14 - Scale-in only if SMA200 breached entry price of last trade - Close entire basket of trades when price crosses ...

A Simple Mean Reversion Strategy

They do offer daily Forex data, but not intraday. For example: url

Diary of a Pro Quant

Ouch... yes, that sort of thing can happen. With H1 data you can often see the timezone shift and correct for that, but with D1 data there is not much you can do other than look for a different data source. I normalize all my Forex data to UTC+2 or ...

Diary of a Pro Quant

Um... Same code in plain Python vs embedded in Jupyter notebook gives better results?! That's unexpected. Perhaps your system had outdated ML libraries that are now updated by installing Jupyter?

Diary of a Pro Quant

Thanks for the updates. I am following your experiments with great interest. Nice to see you consider many different ML techniques and not just focus on one. I am too lazy to try them all, so I am hitchhiking on your efforts :-) I don't have ...

Diary of a Pro Quant

I prefer Random Forests because: - very fast to train! - RFs simply ignore weak or useless inputs, while NNs can go crazy if you feed it weird inputs. This means I can freely feed it any indicator I can find (thanks, mladen ;-)) and later read back ...

Diary of a Pro Quant

In my experience the inputs you feed to the model (and how you pre-process them) are much more important than hyper-parameter tuning. As a basic example (which you probably already know): open[i] --> using absolute prices as an input causes the ...

Diary of a Pro Quant

I like how you add additional information sources to your model! Often people try to predict XYZ using only the historical data of XYZ (aka reading charts). Adding more information to the mix certainly helps. As for "trying next month": I hope you ...

Diary of a Pro Quant

Naughty... However, what does it matter? We can just transfer our money to a non-EU broker and trade there, right? I don't understand how these EU rules are going to help. Same for the US rules (FIFO ordering, no hedging, etc.): simply trade with a ...

This guy is a Mean reversion monster trader

Not within the EU. For example ICMarkets has two websites. The regular icmarkets.com website, and the EU version icmarkets.eu. When I visit their .com website from a European IP address, I am given this popup: image and get redirected to the EU ...

This guy is a Mean reversion monster trader

I see. I think what you call "paired" is what I call "basket", right? A set of trades in the same direction on the same symbol that are typically closed all together at the same time.

The Myth of Averaging

Such as limiting yourself to max. level 3 or 4 like my example? With "paired", do you mean opening an opposite trade in another correlated symbol? Wow, that's impressive! The examples I have shown in this thread are no where near these numbers. ...

The Myth of Averaging

What's next? I don't know. I have now experimented with averaging down and averaging up. The results proof that even a strategy without an edge (a strategy that just blindly enters the market without any logic) can have a positive expectancy if we ...

The Myth of Averaging

For reference, here is USDJPY with an averaging DOWN strategy: image Unfortunately both curves look correlated. Especially 2014-07 is a bad month for both with a strong peak in deposit load. Running both strategies together will not help us get a ...

The Myth of Averaging

Thanks for the suggestion! Here is USDJPY with an Averaging Up strategy: image

The Myth of Averaging