Question about RSI implementation about indicator HOT 4 CLOSED

You got me convinced! :) I will take a look and update the calculation here.

from indicator.

Hi! So sorry for my very late response. I was on a trip, and just had a chance to take a look at this. First of all, thank you very much for taking the time to write such a descriptive bug report. Very much appreciated definitely.

I looked at your calculations document, and I noticed that on Column F, you are calculating the SMA with =SUM(D3:D16)/14 but afterwards, you are using =((F16*13)+D17)/14 for the remaining rows. I guess that is the reason for the difference in the results.

I copied your calculations document, and added the columns J and K to calculate the SMA through the =SUM()/14 method, and I was able to get results aligned with the ones calculated by the page at rows 92 to 101.

Could you take a look? I hope it didn't miss anything.

https://docs.google.com/spreadsheets/d/1e3r1mBAQNXd3yI2l-CiEls0Pfn3mRGv_jxTVJ8kNT9Q/edit?usp=sharing

from indicator.

@cinar Thanks for responding!

you are calculating the SMA with =SUM(D3:D16)/14 but afterwards, you are using =((F16*13)+D17)/14 for the remaining rows

The logic here is that on the 15th closing I calculate the average of the previous 14 values, after that, I calculate the weighted average using =((F16*13)+D17)/14.

I would be glad to take a look if you allow the document for reading :)

from indicator.

@cinar Looks like I found the answer, to why my calculations are the same as on the Binance but differ from those provided by this library.

The Binance exchange uses the TradingView plots. If we take a look at the TradingView RSI documentation we will see, that in their RSI formula they use the Rolling Moving Average:

which in its turn has exact the same formula that I used in my calculations (screenshot is taken from here):

While in this library for the RSI calculation the Simple Moving Average is used.

During my further research, I found out, that the first average gain and all subsequent should be calculated slightly differently. Here's what I found on the stockcharts.com:

The same is told on quantinsti.com:

Not an expert, but it's clear to me that the current RSI implementation differs a bit.

What do you think about it?

from indicator.