In Friday’s ‘Market Profile, the Value Area and Trading’, I introduced the 3 ways to calculate the mean and standard deviation. In today’s piece, I evaluate the 3 methods.

The first is the traditional mathematical approach. This approach assumes that every data set is random and therefore each data set will exhibit a classical bell curve. In a classical bell curve, the following are at all the same price, located at the centre of the bell curve:

  • Mode - the price containing the greatest number of occurences. 
  • Mean - the average of the data set and
  • Median - the mid data point e.g. if are data set is 10, 7 5, 3 1, the median is 5 being the middle number of the data series.

The problem with traditional maths theory lies with its assumption - that every data set is a bell curve: not every data set fits into a ‘normal bell curve’ pattern because the data set is skewed. To the extent of the skewness is the extent the maths formula will distort the probabilities of occurrence. For example if the skewness is toward the left of the data set, the mode will be towards the left of the data set but the traditional maths formula will treat those occurences as part of the standard deviation of the mean.  

Pete recognised this and came up with a simple way to deal with the problem.

In Friday’s Blog post, I attached a piece from Market Delta setting out the two methods used in Market Profile Theory. Method A was the first formulation, Method B came later.

Here’s the difference between the two. Let’s take a data where the Point of Control (the mode) is at 300. At 310 and 311 we have 10 & 4 occurrences (TPOs) and at 290 and 289 we have 10 and 6 occurrences. The data set will kick over the 70% we subtract 14 or more to the current sum. So, taking lines 290 and 289 will satisfy the requirements of the Value Area (1st Standard Deviation).

But wait a minute, if we look at the 4 lines, 310 -311 and 299 -289, we see that in each, 10 occurrences appear. If we follow the traditional Market Profile calculation, we’d ‘favour’ the bearish ‘10′. That did not make sense since a ‘10′ is a ’10′. Hence, the approach I use is to compare one line at a time once a set of kicks over the required number to constitute the 70%.

Thus, in the example give, both 310 and 299 would be chosen as part of the Value Area.   

What difference does the method selection make to the trader? A great deal if the trader uses stats as part of his set of tools. In my case, I use stats to objectively determine overbought and oversold levels. If I use the maths method rather than Method ‘B’, I’ll end up with a less reliable assessment.

Related Posts