Xuanxue,
How do you mean by "the square root of time?" It may be a simple concept that you're trying to convey, but I'm missing it... Thanks in advance!
How do you mean by "the square root of time?" It may be a simple concept that you're trying to convey, but I'm missing it... Thanks in advance!
Quote from Xuanxue:
I left the door open too wide there.Subsequently as a result to save some time and effort I'll just go ahead and address everyone interested at once.
It sounds much more complicated than it is.
A standard deviation is simply the square root of a variance; and a variance is an average squared deviation from the mean. The mean is a simple, static average; add up all values sampled, then divide by the number sampled. The variance is found by subtracting the mean average from each individual sample (resulting sometimes in negative numbers), squaring each new value, adding all values together and dividing by the sampled degrees of freedom; which means one less than the population sampled. The reason for this is factoring infinite and dynamic room for variance to occur out of static and finite points of reference. The variance however doesn't have cohesion to the original data. Finding the square root of the variance takes care of this problem. There are however limits to this data coupling in practice when trying to forecast into the future how volatility will change, but someone more versed in implied volatility could if they wished expound. I only have a general knowledge of it.
But for measuring current moves based on past volatility (not historical volatility, that's found by finding the square root of time, and multiplying it to the standard deviation), this use of standard deviation will suffice.
No matter which used, past or historical volatility, you have to exponentially weight average the standard deviation to constantly reflect current fluctuations in volatility.
To do that first you need to inject a value to represent the sample. After deciding on the sample period you exponentially average the sample (finite). This forms the basis of the EMA. But let's backtrack. What's needed is a value we can weight to current prices or volatility, while by infinite degrees removing less weight from the oldest value. Those removed values once found in the moving average however always contribute to the current data set.
To find the exponent: you divide 2 by the period sampled, resulting in a negative number, and then round that number to the nearest hundreth.
The simplified equation is as follows:
Today's Exponential Moving Average=(current day's closing price (or whatever) x Exponent) + (previous day's EMA x (1-Exponent)).
I told some of you last night I'd give you my new target but I honestly can't tell which way this consolidation period will ultimately break to. If it breaks to the upside again, then a new volatility period begins at 4% still: 1312.48; then again it may be 1238.37.
That's the art part. Use it at your discretion.
Enjoy!