RSI.

RSI – or relative strength index is one of the most commonly used indicators out there.
It is loved for its simplicity and its remarkable efficiency and correlation with price.

However, because it is one of the most commonly used indicators, it also means that it is one of the
most abused indicators with a lot of people not understanding its many intricacies and trading off
the signals that every man and his dog is familiar with (we’ve all heard of reversal divergences and all
that).

This small PDF cannot do such an amazing instrument justice for all it can deliver, but in this PDF I’ll
try my best to explain the intricacies of RSI so you see price and RSI the same way I do.

I reference heavily to John Hayden’s “The Complete RSI Book” which I think is an amazing piece of
writing, J. Welles Wilder’s (creator of RSI) book “New Concepts in Technical Trading Systems” and
certain chapters in Constance Browns “Technical Analysis for the trading professional”.

As analysts and to not be sheep, we need to understand how and why the RSI is calculated and see
what it measures.

I created a PDF spread sheet in #add-your-own-materials where I created an RSI/MACD calculator

based off price.

The reason why I did that was to understand how RSI reacts to price and to really understand the
innards of RSI.

So in this PDF I’ll go into detail with regards to RSI and it’s intricacies.

The truth is I’m not expecting a lot of people to read this PDF, but for those who do I hope to share
something and hopefully improve your analysis on the RSI.
RSI. WHO and WHAT?
RSI is a fucking amazing tool created by J Welles Wilder Jr and first introduced in his book “New
concepts in technical trading systems” which was first published in 1978.

He identified that a lot of traders used some sort of a momentum indicator to help them decide the
direction or potential reversals coming in the market. However he established 3 flaws in a traditional
momentum oscillator.

1) Extreme volatility (no smoothing at all, just looked like an ECG)

2) No objective value on the volatility (how do you know something is oversold/overbought if
there is nothing to measure it up against? There was no real scale with defined parameters)
3) A lot of data to deal with (less of a problem now tbh with TV and all that)

You’ll see in the PDF that RSI is calculated from this formula.
100
𝑅𝑆𝐼 = 100 − [ ]
1 + 𝑅𝑆
Where

RSI = Relative strength index (on a scale of 1-100)

𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑔𝑎𝑖𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑡ℎ𝑒 𝑙𝑎𝑠𝑡 (𝑥)𝑝𝑒𝑟𝑖𝑜𝑑𝑠
And RS = 𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑙𝑜𝑠𝑠 𝑜𝑓 𝑡ℎ𝑒 𝑙𝑎𝑠𝑡 (𝑥)𝑝𝑒𝑟𝑖𝑜𝑑𝑠

How it’s calculated isn’t that important but it is handy to know. It’s also handy to know what values
of RSI are meaningful and what the price implication is at certain RSI values.

Just out of curiosity I plugged this function into an excel sheet to see what it would do, RS is the X
value and RSI being the Y value. This is what I get.

RS vs RSI
100
90
80
70
60
50
40
30
20
10
0
0 2 4 6 8 10 12

What do we notice?
We notice an asymptotle relationship between RS and the RSI and what that means is, for the RSI to
move higher it needs to RS to increase exponentially. For RSI to move lower it needs RS decrease
exponentially (towards 0).

This tells us, that for higher RSI values price needs to become more and more vertical and for lower
RSI values price needs to drop more and more vertically.

Therefore, in parabolic moves up we see exceedingly high RSI’s for such long periods of time.

Everyone yacks on about certain RSI points that are of great importance, namely

- 20 (oversold region in bear market)

- 30 (normal oversold region)
- 40 (oversold region of bull market)
- 50 (crossover point)
- 60 (Resistance of bear market/1st support of bull market)
- 70 (normal overbought region)
- 80 (overbought region in bull market)

But what do these points mean?

I’ll break it down for you.

Reverse engineering the RSI formula

100
𝑅𝑆𝐼 = 100 − [ ]
1 + 𝑅𝑆
Let’s plug in those values into the RSI to derive an RS value.

RSI RS Interpretation (over x number

of periods)
20 0.25 (1/4) The average rate of loss of
prices is 4 times that of the
gain.
30 0.428572 (3/7) The average rate of loss of
prices is 2.33x that of the gain
40 0.666666(2/3) The average rate of loss of
prices is 1.5x that of the gain
50 1/1 Average gain = average loss
60 1.5 The average rate of gain of
prices is 1.5x that of the loss
70 2.333(7/3) The average rate of gain of
prices is 2.33x that of the gain
80 4 The average rate of gain of
prices is 4x that of the losses

So next time when we look at the RSI chart we can sort of estimate what’s going on.
Overview of the indicator:

Welles originally outlined that RSI values below 30 were considered “oversold” and values above 70
was “overbought”.

This is generally reflected in the price too.

HOWEVER, we tumble upon a roadblock, almost immediately.

“but if it’s overbought then why is prices still going up?”

This is because of the inherent flaw of the use of the indicator and the common misconception that
you sell when we are overbought and buy when we are oversold.

Research performed by Brown and Cardwell had them conclude that RSI is rangebound in different
market conditions and that simply using “oversold” and “overbought” regions is far from ideal.

So I challenge you guys to throw the idea that whenever prices are above 70 to be “overbought” and
there I should enter a short/shell and the vice versa for values less than 30 right outta the bloody
window.
What should we be looking at then?
We know from our experiences and from previous market analysis that there are 2 kinds of markets;
bull and bear.
As a result it would seem reasonable to determine a bull and bear RSI range.

RSI features in a bullish market (generic rules) RSI features in a bearish market (generic rules)
Ranging between 40-80 Ranging between 20-60
Presence of simple bearish divergences Presence of simple bullish divergences
Generally higher highs and higher lows Generally lower lows and lower highs
MDRP Up’s (positive divergences) MDRP Downs (negative divergences)
Higher highs/higher lows Lower highs/lower lows

However I think it would be appropriate to assume that we can have long, medium and short term
directional biases that RSI ranges can form. For instance, we can have 4 hourly bearish RSI structure
inside of a 3D bullish range.

Let’s go into an example.

Some notes on this chart.

1) Note every time we bounce off 40 and continue back up, most importantly breaking through
60 and push into “overbought” regions, we are bullish.
2) When we enter bear range is when we get rejected from 60 and proceed to form lower
lows. Note that the 40 level does not hold in a bearish environment.
3) Now in today’s there is 2 possibilities that I have for bitcoin. Notice that despite being
bearish for a while, we still have NOT broken through the 40 level. That, strictly speaking is
still considered bullish. That said, an argument can be made for being rejected from 55 and
that would confirm our bearish bias. Until 40 is broken through, I think its safer to assume
bull.
4) Note in extreme bullish environments, we bounced off the 55-60 region. That is a very tell-
tale sign of an extremely bullish market.
Some things I have noticed while studying RSI range:

Here are general ranges that Hayden references as the lower and upper limits for a bullish and
bearish case.

Bullish scenario Bearish scenario

Lower Limit – 33-40 Lower Limit – oversold region (<30)
Upper limit – overbought region (>70) Upper Limit – 60-67

When market reversals happen USUALLY this happens.

Example going from Bullish to bearish.

1) RSI moves from “overbought” to neutral (30-70)

2) RSI bounces off lower bullish limit (depends on TF, each TF has its own special tendencies)
3) RSI gets rejected at upper bullish limit (usually confirmation of bearish trend)
4) Lower bullish limit RSI does not hold for a second time, confirmation of bear trend
(momentum lows usually form after this)
5) RSI then finds resistance at lower bullish limit. The bullish limit now becomes resistance (not
to the exact number, but the bullish limit range is now resistance)
6) If the bearish trend is consistent, the upper bearish limit rejects RSI each time. This process
repeats (3-5) until the limit is breached

Bitcoin weekly example ^^^

Daily bitcoin example ^^^

Bitcoin daily example ^^

And going from bearish to bullish.

1) RSI moves from “oversold” to neutral (30-70)

2) RSI after struggling with lower bullish limit, pushes through to get rejected at 60
3) RSI bounces at lower bullish limit or near the median line
4) RSI goes through 60
5) RSI then finds support at upper bearish limit
6) “oversold” conditions/SFP formed
7) Bounces from either 50-60 (in extreme bullish markets) or lower bullish limit after which 4-6
repeat themselves until this cycle is broken.

Bitcoin weekly example ^^^

Bitcoin daily example ^^

From my testing, I have found the 4hourly, 12 and daily timeframes make great directional bias for
the next couple of days.
The 3D and weekly are great for looking at slightly longer-term trends.

Another thing that the RSI is brilliant at spotting is divergences. Divergences are one of the most
powerful signals there are, provided they are used in correctly and in confluence with other factors
as well.

I’ve already made a PDF on divergences, which I highly recommend you look at, but here is a quick
summary of the divergences that are available.

Bullish divergences Bearish divergences

Bullish SIMPLE divergence, appears ONLY in Bearish SIMPLE divergence, appears ONLY in
bear trends and indicate some sort of a reversal bullish trends and indicate some sort of a
(not guaranteed for a full reversal, but a relief pullback (not guaranteed full reversal, pullback
rally is possible). is possible). Again, another signal that every
This is the divergences that every retail trader retail trader knows and loves.
knows and loves
Positive divergences, appears ONLY in Negative divergences, appears ONLY in
continuation trends AND at the end of trend continuation trends AND often at the start of a
reversals. Works great on lower timeframes, on trend reversal. Great on lower timeframes, on
higher timeframes, you need to be careful. higher timeframes they need to be used with
due diligence.

I’ve outlined the 2 major kinds of divergences there are for RSI (and any oscillator for that matter).
From my study of price charts and backtesting over various timeframes, I find the presence of the
positive and negative divergences to be the most profitable. Simple divergences are great, but you
never know if it the real reversal or not. I’ll go into detail explaining how to spot trend reversals.

When is a good time to be using the RSI?

RSI on its own and if used incorrectly is a disaster in the making. These are indicators and should not
be used as a sole means of trading, though you can, but prepare to lose your money.
What you SHOULD be doing with RSI, is marrying it with other strategies and finding confluence.
Personally, I use RSI to confirm:

1) Horizontal support and resistances

2) Fibonacci levels
3) Trendlines
4) Elliot wave (rarely now, I’m phasing away from this)
5) EMA momentum.
6) Confluence across MULTPLE timeframes (ideally all timeframes from Daily down to 4 hourly)
References:

1) New concepts in Technical Trading Systems, J Welles Wilder

2) Andrew Cardwell’s RSI course (pdfs, no audio)
3) RSI – The Complete Guide, John Hayden
4) Technical analysis for the trading professional – Constance Brown