The expected value can be an irrealizable value. A dice roll has an expected value (and a median) of 3.5. Have you ever rolled a 3½? The expected value (and the median) of the return of a constant range bar is zero. It is impossible to have a doji ...

Voodoo Science-Statistical Analysis

I was not referring to any paper. The median separates the data into two sets of equal size: the small values and the large values. If your distribution isn't symmetrical your median is neither the highest probability (the mode) nor it is close to ...

Voodoo Science-Statistical Analysis

I workaround the problem by studying constant range bar charts to (almost) return to the normality. As soon as you try to solve things with fat tailed distributions you get untractable equations. I see a BB on your screenshot. Even if the returns ...

Voodoo Science-Statistical Analysis

Statistically it means the price is recurrent and it implies finite variance. Good news. It also implies a null expected trend. Bad news.

Voodoo Science-Statistical Analysis

Where is this thread supposed to go? It is known since the late 60's that the markets returns aren't normally distributed. Median assumes the distribution to be symmetrical i.e. no drift (aka trend/momentum). Mandelbrot has fit a Lévy distribution ...

Voodoo Science-Statistical Analysis

It makes sense

Diversified Trend Trading Approach

Putting the fear of pulling the trigger part apart I understand "periods of normal drawdown" to be "origin 1" drawdown. A trading system, even discretionary, is implicitely a model of the market. As with any model some part of the data is explained ...

Diversified Trend Trading Approach

No the definition of a trend is nothing subjective. What is subjective is the SCALE one choose to define the trend because of the fractal nature of the price movement. Should you use a baysian estimator or a SMA or a MACD or a regression line or ...

What are the big Ideas in successful trading?

I'm not sure I understand fully this sentence. I understand it in the sense that it brings the question about where is the point that separates "this is a trend" from "this is a range". It is like if you would try and define the limit this is black ...

What are the big Ideas in successful trading?

This picture is taken from Rasmussen & Williams, Gaussian Processes for Machine Learning. Let's forget all the maths around the GPs. image Let's imagine the 7 samples are a price series and we would like to smooth it with some MA to see if it is a ...

What are the big Ideas in successful trading?

I agree with leebsurag Aussie and Kiwi aren't cointegrated. They move together because big USD makes them dance. Try this dataset. It is CAC40 vs DAX daily.

Cointegration of currency pairs

My personal and subjective opinion don't make the market move. That is wishful thinking. At least as far as the accumulated evidence thus far can tell, in the picture the market is both and at the same time in a range and in a downtrend at two ...

What are the big Ideas in successful trading?

Bayesian thinking is also about acknowledging the ambiguity of the market and taking the optimal decision based on this. Does this chart show a range or a downtrend? If I'm currently short shall I stay with the down trend, take profit, take partial ...

What are the big Ideas in successful trading?

From the document Guruji Setup.doc This is a wrong statement. The R:R is only improved to 1:3.7 when all the steps are triggered and the trade wins. They aren't always triggered. Probabilities have to be taken into account. The trade can go straight ...

Money Management / Risk Management

From what I could find on the Internet (so I may be plain wrong) LTCM didn't fail because their system was bad. They bankrupt due to human mistakes. Their method was comparable to an arbitrage. Arbitrage work when you are discreet. They were trading ...

Diversified Trend Trading Approach

Sorry I didn't explain myself clearly enough. ELM can be interpreted as an approximation to a single layer feed forward network with infinite number of hidden units. The approximation is better when the number of neurons actually tends to infinity. ...

Machine Learning with algoTraderJo

A very interesting article about how the NN see what they want to see: url That was translated into a funny video presentation anybody can understand: video

Machine Learning with algoTraderJo

Yann LeCun seems to lead the detractors of this method. I think with reason. I read his article several weeks ago and some others. I had also watched the video you posted (which brings nothing). Clearly the ELM are mere overfitters with poor to no ...

Machine Learning with algoTraderJo

If 1. held true the price would continue in straight line over the week-end since there is no trading activity to act on it. If 2. held true you would only see rounded tops/bottoms and no pin bars. If there was inertia in the market sharp ...

Does it make sense to think of price movement in this way?

I don't know what volatility means to you. For lots of traders volatility means "big moves in the market" (=trend). For me it is more the academics definition "the standard deviation of the returns" = high ATR(.) usually without any significant ...

Methods for reducing the volatility of forex?