Let me try with a picture. Same initial price. Same final price. Same box. image It is THE question indeed. Long wicks? well yes but not always and non-choppy also have long wicks. Overlapping bars? Perhaps but steady ranges also have overlap... ...

Choppy market index: any good ideas?

The coast line algo trying to take the turning points into account: image The same oversampled using H4 data: image

Choppy market index: any good ideas?

Number of 10 pip range bars inside the daily bars (orange) versus the coastline/straight line (blue) image

Choppy market index: any good ideas?

If I use evenly spaced time based sampling then using the bars inside the bar (over sampling) is simply the same as running the indicator on the lower TF with a longer window (in # of samples) and the problem is unchanged. Actually I planned to use ...

Choppy market index: any good ideas?

@skenobi I want to MEASURE. I don't think it is even possible to predict. Anyway any attempt at predicting will need the ability to measure first to build the predictive model. @Adal Thanks for the ideas. I tried the coast line vs straight line ...

Choppy market index: any good ideas?

I don't choose the ticks I use them all. To build the chart for E/U all the 159982769 ticks I have in my files were used. EDIT: except that I posted A/U charts

Statistical mean of the market [quant corner]

My broker doesn't give me a time*price chart. It feeds me with ticks. The ticks contain a timestamp, a bid, an ask and the volume that I happilly drop. I build all my bars myself from the ticks. Gaps are taken into account. If the highs and lows of ...

Statistical mean of the market [quant corner]

That's the idea. If the price goes too far away from the MA I close the bar and the MA is updated. I also avoid closing a bar too near of the MA if it opened on the other side (somehow freezing the time) to avoid whipsaw.

Statistical mean of the market [quant corner]

Your chart gave me an idea. Instead of finding the MA with the lowest lag and highest smoothness why not adapting the price to the MA? What we dislike with MAs is their lag. When the price gets too far away from the MA I close the bar and start a ...

Statistical mean of the market [quant corner]

The vertical scale looks strange. Look on your very last screenshot. The time highest high double top appears to match the timeless highest high DT but the low in-between doesn't go as low on the timeless chart. The second next lower high around ...

Statistical mean of the market [quant corner]

There is something strange on these charts: the price goes straight up but goes down slowly following a line. Is the distribution symmetric over a long time? Can you generate and post the histogram of the price differences?

Statistical mean of the market [quant corner]

Perhaps I'm. I just try to be objective. I'm not living inside your head so I cannot know what you mean. I can only focus on what you write. Just explain yourself better. Write more clearly if what you write isn't what you mean. We only have access ...

Statistical mean of the market [quant corner]

In my mind a DSP classical filter is any filter you can describe in term of the lagged values on its input and output. These filters are used in more fields than just audio. EMA^N is such a filter. The frequency response is important in image ...

Statistical mean of the market [quant corner]

The EMA^N operator is simply the EMA of the EMA of ... of the price (applying N times the EMA). In terms of the lag operator the EMA formula is: image This being a time series we can apply the formula again and again on the result. So the ...

Statistical mean of the market [quant corner]

The 4th picture only shows the convolution formula not a recursive one. Anyway I found the paper by Zumbach. I attach it here for whoever is interested. I didn't finish reading it yet. FYI1 phase linearity in trading = all the frequencies lag the ...

Statistical mean of the market [quant corner]

Did you try those filters? Do they have a recursive formula? How do they compare with a Chebyshev or a Butterworth filter in term of cut off frequencies and phase linearity?

Statistical mean of the market [quant corner]

Here is the first daily chart posted by yaed but seen by my filter. The filter computes bar by bar so it couldn't recognize the range in the second fourth of the period. Instead it looks more like a double top. Despite this detail the two charts ...

Statistical mean of the market [quant corner]

image

Statistical mean of the market [quant corner]

You don't really explain what you're trying to do and how you do it so it's difficult to help or even give ideas. I thought you were using the 2nd moment of the distribution of the price return (variance) but it looks like not. You seem to focus on ...

Statistical mean of the market [quant corner]