RSI-Mathematicians and senior members please see

Dear Rahul,

Divergence is measured for Higher Highs or Lower Lows of prices. It's not measured at the close of every day or every tick. Thus RSI/MACD/for that matter any other indicator is calculated for only higher pivotal highs or lower pivotal lows of prices with the RSI/MACD/Other Indicator for the previous pivotal high or low, when divergence or convergence is to be measured.

Now an indicator's value is affected by its base period. For example, a 14-day RSI takes into account closing prices for last 14 days only and discards all other prices.

For example, let's assume that Stock X made a pivotal high of Rs. 100 on December 14th. The 14-day RSI at that point was, say, 95. For the sake of simplicity, assuming that there were no holidays in the month of December, the RSI would be computed using closing prices of X from 1st Dec-14th Dec. On 16th, prices retrace and on 20th Dec., the upward journey is resumed once again. Another pivotal high of Rs. 115 is made by stock X on 25th January. Now at this point of time, 14-day RSI will be calculated taking into account the closing prices from 12th Jan-25th Jan (assuming no holidays) and not from the date of start of uptrend to 25th Jan. Thus, it’s very well possible that RSI for 1st Dec-14th Dec. may be higher/lower than the RSI for 12thjan-25th Jan as the series of closing prices which formed the basis of calculating previous RSI is independent from the series used in computing current RSI.

So a divergence will take place, if RSI is lower than the RSI for the previous pivotal high.

This is the only mathematical explanation. I might have liked to give a real-life example, but short of time at the moment. Will suggest you to please calculate RSI on a spreadsheet for the divergences of NIFTY shown in the chart of Tanewbie and it’ll help u in understanding the concept more clearly if I have not been able to do the job.

Best Regards,
--Ashish

P.S.: This is my understanding only and not an expert opinion, so i apologise in advance if I am wrong at any place.

Hi ashish,

NO you have made it very clear. And this would also show that the momentum in the period has slowed down and more the number of divergence on the pivots the more chances of reversal. It is indeed very clear and I will try to calculate it. Thank you so much.

Rgds

Rahul