NASARAVI - here is my quick attempt to answer your question shown below .. I am about to head out for an outing, so will continue thinking about the root of your question
"Question
Is it possible to use a ema 2,3,4 discriminator on BB width to find the precise top or at least an impending top with less ambiguity?"
my Answer
if you look at the basic four A-D line charts on Page 1 of my Public List, you will see:
1. a fast moving average placed on the daily Advance-Decline values of each index ( FYI - the A-D line is a cumulative version of all these daily values beginning at the start date selected in the chart settings )
AND
you will notice these daily Advance-Decline values did NOT reach the upper positive levels that were printed on the chart in the week ended July 13, 2018...
no ambiguity here because the shouting loudly at you for nearly two full weeks is the Negative Divergence by the most important internal which drives all these other Advance-Decline tools ...
there was much less internal Net Advancing going on for all four indices after July 13th while each index price level attempted to keep advancing into the July 26, 2018 evening when Amazon reported earnings
2. and you will see the McO
* vs. its Bollinger Band and * vs. its zero line
for assessing directional shifts which are growing more serious
3. use #1 and #2 above combined with the changing slope AND the degree of slope by the McSum for the newest direction of the McSum
your best odds for being accurate in my opinion are using the data in #1 and #2 above for fine tuning your trade entry timing based on the McSum changing behavior as these #1 and #2 chart elements begin to worsen, which may occur after a hesitation of one or more days ... in this case where you are asking how best to time your trading action as early as possible before the full price damage starts in earnest
further Answer - I will do some research on your question regarding applying a moving average to the BB BandWidth.
And get back to you once I am satisfied I have the full picture of what is technically accurate. Could be a while before I have this answered for you.
Good work, NASARAVI. Thanks for your serious efforts to understand how you can benefit from these A-D tools.