Last Updated on 13 January, 2021 by Samuelsson
The RSI indicator is one of the most popular and well-known trading indicators out there. It’s used by many to analyze the markets in search of profitable entries and exits. Before going any further, let’s just define what RSI is.
Relative Strenght Index (RSI) is a trading indicator developed by J. Welles Wilder in the 70s. It’s a momentum oscillator that measures the rate of change of up days and down days. The RSI then outputs a value of 0 – 100, where high values are considered overbought, and low values are considered oversold.
In this guide, as with everything we do at The Robust Trader, we aim to give you the most comprehensive and detailed guide possible. We are going to cover the basics of the RSI indicator, but also what we have found in our own testing. As a matter of fact, we disagree with the default settings for the indicator, and strongly recommend other settings that repeatedly have proven themselves superior!
We will also cover some RSI trading strategies that we have backtested. Many trading sites tell you that they have a trading strategy, but only present some made up examples. At the Robust Trader we have more than 20 years of experience in designing trading strategies that work, and we can say with certainty that very little of what you are presented with works! In this guide, we aim to be one of the very few sources that provide some real RSI strategies that work!
We will delve deeper into this discussion later, and it’s something that we urge you don’t skip! Reading and understanding what we are going to share with you will certainly put you ahead of most aspiring traders!
Let’s begin by having a look at the basics of the RSI indicator!
What Is RSI and How Does it Work?
RSI was invented by J. Welles Wilder and introduced in his book New Concepts in Technical Trading Systems, that was released in 1978. Since then, it has grown in popularity and become one of the most recognized, and in our opinion, useful technical indicators in the trading industry. Since then there have also been several spinoffs, like the Connors RSI, which we will also take a look in just a bit!
As we’ve already mentioned, RSI is a momentum oscillator that’s used to measure the speed and change of price movements. It fluctuates between 0 and 100, and by reading its value you can get a sense of whether the market is overbought or oversold. The traditional interpretation is that a reading of more than 70 is an indication of an overbought market, and 30 or less indicates an oversold market.
In trading, the terms oversold and overbought are terminology that describes the moment when a market has moved to much, and will soon revert. This tendency is called mean reversion, and is especially prevalent in equities, although it can be found in many more markets.
If the market is oversold, it means that it has moved too much too the downside, while overbought refers to the opposite condition.
However, in our testing, we have found that the RSI not only works for mean reversion, but also for momentum trading.
We will come back to this later in the article!
How Is RSI Calculated?
The RSI indicator works by comparing two different measurements.
- The first measurement is price gains on up days. Up days are defined as days closing higher than the previous day’s close.
- The second is price losses on down days. Down days are defined as days closing lower than the previous day’s close.
The RSI compares the relative strength of these two measurements and is calculated as follows:
RSI= 100- [100/1 + RS]
RS is the average of all positive changes within the look-back period, divided by the average of all negative changes.
RSI Basics: Oversold and Overbought Readings
Now that we have covered the basics of the RSI indicator, let’s go on to cover how traders use RSI to identify oversold and overbought conditions.
The most common way of doing this is to require the RSI to cross below or above a threshold. Usually, that’s 70 for the overbought threshold, and 30 for the oversold threshold. In the image below you can see what it looks like when the RSI crosses the overbought and oversold thresholds. The purple area covers the range from 30-70.
Will the Market Turn Around at Oversold Readings?
Even though the RSI can help to point out overbought and oversold levels, it’s always hard to know when the market indeed will turn around. Taking mean reversion trades many times is like trying to catch a falling knife. The price could very well continue to plunge after we enter the trade.
Of course, this is nothing spectacular. Every type of trading strategy will have its winners and losers. In that regard, seeing some trades just turn red with the falling market is completely normal.
However, where things are becoming a little tricky, is in that we are dealing with a type of trading strategy that gets its edge from that the market has fallen excessively. This means that when a trade goes against us, we’d better stay in it, since our odds just continue to get better. Once we’re stopped out by a stop loss, we know that the edge is stronger than it’s ever been since we entered!
Trading mean reversion strategies, be it with or without RSI, requires that we either use very large stop losses, or no stops at all.
How counterintuitive not using stop losses might sound, it does work, given that you are very careful about how much you risk on every trade. We have solved this by only allocating a small portion of our trading strategies to mean reversion strategies. This is possible through algorithmic trading, that lets us trade 100 or more strategies at the same time!
If you want to learn more about how to trade like this, why not read our complete guide to algorithmic trading!
What Settings are Best for RSI?
The most common settings for the RSI indicator is a 14-days lookback period with the oversold threshold set at 30 and the overbought threshold set at 70. However, the RSI can be tweaked, and the inputs can be changed to work better with the specific market you’re working with. Sometimes, different trading styles may require different RSI settings. We encourage everyone to experiment to see what seems to work the best. If a 10 day lookback period consistently appears to be better than the default 14-day, there is no need to stick with the latter.
The Threshold Settings
However, you may also look at adjusting the threshold values to account for rising and falling markets.
- In a market that trends upwards, it could be useful to adjust the RSI threshold values to make them a little higher. What can be said generally about rising markets is that they tend to not recede as much as flat or falling markets before they make the next great swing. In such a scenario, with too low RSI threshold values you run the risk of getting very few trading signals
- Conversely, when in a falling market, it could be beneficial to adjust down your threshold values. Falling markets tend to fall for longer periods of time before there is a pullback. You should make sure to reflect this in your parameter settings.
But all this is very general advice. You may wonder what we have we found in our own testing.
Well, in our experience, the 14-period RSI tends to perform quite poorly. It’s too slow to react to price action in any useful way, and leaves us with too little room to profit in the markets.
To get better results, we have found that it’s better to use a shorter period RSI. Most often this means using a lookback length set to around 2-7. That seems to be the sweet spot where it’s short enough to react swiftly to what the market does, before it gets too long and lags too much.
And for the oversold and overbought thresholds, these may also have to be adjusted for the increase in responsiveness that comes with a shorter RSI. That means moving the oversold threshold lower (maybe to 15) and the overbought threshold higher(perhaps to 85).
Which Exact Settings Should You Use?
As much as every trader wish that there was a definite answer, there isn’t. The best settings vary with the market, trading strategy, and timeframe. In order to find out what works well for your particular scenario, backtesting is a must!
RSI as a Momentum Indicator
Some markets simply don’t work too well with mean reversion strategies, since they are not mean reverting by nature. However, that doesn’t need to mean that the RSI cant be used for that market. It could be that we just need to think a little differently, and turn the logic upside down. That is, instead of buying at low values, we buy at high values.
What we then effectively have done is to take on a trend following approach, which is the opposite of mean reversion. In other words, we ride the trend instead of going against it in the anticipation that it will turn around.
In our own testing, we have found that the RSI shows a lot of merit in this field as well, in certain markets. So It’s certainly something that’s worth taking a closer look at!
RSI Trading Strategies
In this part of the article, we are going to explore some of the strategies that are popular among new traders. However, you should keep in mind that most traders are unsuccessful, with the corollary that common trading strategies also tend to not work so well.
Upon a visit to some of the greater trading sites online, you will find a lot of “trading strategies” that actually don’t work at all. In most cases, the author has just put together some nice looking indicators or logics, and presents them as working trading strategies. We know from experience, after having backtested many of these strategies, that close to every strategy of this kind is garbage.
Below, we will present some examples of popular concepts involving the RSI indicator. However, don’t take these for truth, but let them serve as inspiration for other trading ideas, that eventually could become fully working trading strategies!
However, after we’ve covered these non-proven strategies, we will actually show you two strategies that have worked for a long time, and that we believe will continue to work into the future!
Let’s get to it!
1. Buy Low Sell High
This is the most common trading strategy that’s proclaimed online, and as a matter of fact, it at least has some merit to it. At least it does work, although perhaps not well enough for us to trade it ourselves.
The rules are simple: Enter if the RSI crosses below 30, and exit when it crosses above 70.
So, we wait for a market pullback and then buy in anticipation of a coming correction, that will take the security to new heights.
Here is an example of a successful trade using this logic.
The yellow circle at the bottom signals our entry, and then we wait for the RSI to rise above 70 to exit the trade.
As you see, the market has to move quite a bit for us to exit the strategy. This will lead to that quite a few trades will start to go in your direction, but revert before your exit level is reached.
One way of addressing this issue could be to lower the sell threshold. For instance, if we sell at 50 instead of at 70, then more trades will end up as winners. Of course, with the trade-off that the winners will be smaller.
2. RSI Breakouts
As we already mentioned, the RSI often works great for detecting market strength that’s worth acting on. That’s also in line with the main objective of this trading strategy, that attempts to identify those times when the market is strong enough to continue in the direction of the momentum.
This is done by spotting breakouts in the RSI indicator. If RSI breaks out to the upside, we follow along by going long, and if it breaks down to the downside, we instead go short.
In some markets, this may work better than using the traditional, mean-reverting approach. It all has to do with the characteristics of the markets. If you, for instance, are playing with the stock market, it’s going to be much easier to find a mean reversion system that works, than a trend following. Conversely, if you want to trade other, more trending markets like energies, you might be better of choosing a trend-following/momentum approach.
But let’s get back to this particular strategy. As you probably remember, we previously stated that the best settings for RSI usually is somewhere around 2-7. As such, that is the setting we will go with here as well.
When it comes to the breakout level, you could adopt either of the following approaches
- Using a previous high or low in the RSI indicator for the breakout level
- Using a fixed breakout level. (Such as 80 for going long, and 20 for going short)
Let’s have a look at each of these approaches!
Using a Previous RSI High or Low
With this approach, we look at the previous highs and lows in the RSI to find appropriate levels where the security, if breaking that level, will continue in the direction of the short term trend.
For example, if the RSI made a reversal at 80, we might watch the RSI break 80 to take a position to the long side. Conversely, if the RSI made a reversal to the upside at 20, we may watch the 20 – level, and go short if RSI crosses below 20.
Here you see an example of how we enter when the RSI breaks the previous high.
In other words, we are using the concept of support and resistance, but applied to the RSI instead of the price chart.
The fact that the breakout level moves depending on where the previous RSI highs and lows are located, means that the breakout level becomes dynamic. That’s very different from the fixed breakout level, where we assume that the breakout level remains constant!
Using a Fixed Breakout Level
This approach is easier than the previous one. Instead of spotting RSI highs and lows, we just set our breakout levels beforehand. In this example, we have set them to:
- 20 for going short
- 80 for going long
So as soon as the RSI goes below 20, we short the market, and when it goes above 80, we go long in the market. Here is an image demonstrating the concept
Now, this means that we don’t take into account price action that could be relevant the placing of the breakout level. However, how unintuitive this may sound, it sometimes could prove to work better than employing a dynamically adjusted break out level!
How About the Exit?
Here are two exit methods that you could have a look at:
- Exit when the RSI retraces back to the breakout level again. So if we entered at 80, we will exit when the RSI crosses below 80 again.
- Use a profit target and stop loss. The first of the two that is hit will be our exit.
Divergence trading is a well-known concept in trading. In short, it refers to when two data streams that normally go in the same direction converge or diverge.
So what we are looking for is that the RSI indicator starts going in another direction than the price. Normally you divide divergences into two categories depending on in which direction they occur. Those are:
- Bullish Divergences
- Bearish Divergences
A bullish divergence means that the price makes a new lower low, while the RSI makes a new higher low. Like in the image below:
A bullish divergence is an indication that the speed and strength of the downtrend is abating. Therefore, bullish divergences are thought to signal an imminent reversal of the trend to the upside.
A bearish RSI divergence, on the other hand, is the opposite of a bullish RSI divergence. In other words, we need to have two higher highs in the price chart, coupled with two lower highs in the RSI. Again, this indicates that the strength of the trend is abating, and that a reversal to the downside is imminent.
One of the drawbacks of using divergences of any kind, is that they can persist for a long time. Thus, they often become a sort of self-fulfilling prophecy, since a reversal is very likely to occur at some point over the life span of the divergence. In other words, divergences need to be used together with other entry timing techniques.
3. RSI Support and Resistance.
When we used the highs and lows of the RSI to set the breakout levels, we effectively made use of support and resistance levels. By noting when the market broke through a resistance level, we ensured that there was no resistance in the way for the market to get stuck at. Put differently, the upside was cleared from any potential obstacle.
For those who don’t know, support and resistance refer to levels in price that a market or security has found hard to break through in previous attempts. When the security then reattempts to get past that level, it usually takes some extra effort, since market participants are standing by and defending that level each time it’s reached.
Most often we look at the price in search of these levels, but we could also attempt to use the RSI. The principle is the same. We look for levels that have been reached and defended one or preferably several times. Once the RSI once again gets to the support or resistance levels, we look for more indications that the price won’t get past that level. One very common way to increase the accuracy of the entry signal is to use candlesticks.
3. Failure Swings
Failure swings could be said to be a more advanced version of RSI divergences, where we add additional criteria to complement the divergence. As we touched on before, divergences signal that a change in the trend is coming, but is less suited to point out the exact turning point.
The RSI failure swing tries to solve this issue, by adding one more signal to the mix, that will show when it’s time to enter the position.
Here is what an RSI failure swing could look like:
As you see, we have the typical divergence. The RSI is making new lower highs, while the price is busy making new highs. However, soon, the RSI crosses below its last low, and that’s when we enter a short position.
What is important to note here, is that the first RSI high must be in oversold territory(typically over 70), while the second RSI high preferably is below the overbought threshold.
The Two Types of Failure Swings
What we just showed you, is a negative signal, also called a failure swing top. However, there is one more type of failure swings, which is named failure swing bottom. The failure swing bottom occurs at the bottom of the range, and signals a reversion of the trend to the upside. It consists of
1. An oversold RSI reading
2. A new lower high
3. A new, higher low
4. At the same time, the price must have performed two lower lows (we have a divergence)
If these four conditions are true, we enter when the RSI crosses over the lower high.
4. Trend ID
The trend ID strategy is based on two assumptions; that a bull market will have shallow pullbacks and prolonged upswings, and that a falling market will have deep pullbacks and shortlived upswings.
Following these assumptions, Constance Brown, who is the inventor of the Trend Id strategy, draw the conclusion that the RSI thresholds need to be adjusted based on the current market trend. In bull markets, he found that the RSI tends to oscillate between 40 and 90, with 40-50 becoming support levels. And for the bear markets, he instead discovered that the RSI tends to oscillate between 10 and 60.
So basically, the Trend ID covers the same rules as the conventional interpretation of the RSI, with the difference that the oversold and overbought levels adapt to the current market phase. For those who don’t remember the conventional interpretation, it was to buy when the RSI crosses below the oversold threshold.
Adjusting the thresholds with the market is also in line with that we covered under “best parameters for RSI”. There we recommended that you adjust the oversold and overbought thresholds with the current market trend in mind.
5. RSI Trend Lines
Remember that we used RSI to find support and resistance levels. Well, you could draw trend lines as well. You just connect the highs and lows of the RSI line with each other, and then you have the trend line. And as with any other trend line, it could act as support or resistance.
Here you see how we connected several lows to create a rising trend line.
The Cutlers RSI is a variation of the original RSI. It’s quite the same, except that it uses an exponential moving average instead of a simple moving average. To understand the difference better, let’s just remind us of the default calculation.
RSI compares two measurements, which are:
- The price gains on up days.
- The price loss on down days.
Then we apply an average to the up days, and down days, before we divide the average of the up days by the average of the down days.
This is where Cutler’s RSI uses an exponential moving average instead of a simple moving average. The result becomes that Cutler’s RSI adapts quicker to recent price changes.
The Connors RSI is a somewhat more advanced version of the RSI indicator, and gets its value from three separate components. The RSI as developed by J. Welles Wilder plays an important role in the calculation but is completed by two more measurements. These are:
The Up/Down-Lenght- This component gets its value from the number of consecutive up or down days. Each up day is represented by a positive value, and each down day is represented by a negative value.
A short term RSI is then applied to the up/down value. This RSI reading becomes the second component.
The Rate Of Change (ROC) – The ROC component uses a user-defined length and calculates the percentage of percentage changes for each day that are below the current percentage change for the day.
To get the RSI value, these three values are then added to each other, and divided by three. In other words, the Connors RSI is the average of these three components.
Connors RSI vs Normal RSI
We have found that Connors RSI works a little better than the normal RSI. However, the differences are not very significant, so feel free to test for yourself!
RSI Strategies That Work!
Now that we have covered some theoretical examples of strategies, I wanted to show you what you don’t get to see on close to any other trading site.
Strategies WITH backtests.
You probably hear about good profitable strategies all the time, but how often are you presented with proof that they work?
Not very often!
Here we will show you two trading strategies on the S&P-500 that work.
Now, nobody can know if these trading strategies are going to work in the future. Markets change all the time, and with those changes, we risk being left with strategies that are not fit to the current market conditions. This is something that we cover more in-depth in our article on algorithmic trading.
However, these strategies indeed have worked in the past, so let’s get to it!
1.The Connors RSI Trading Strategy
This first trading strategy is one that’s quite similar to the conventional view of how the RSI should work.
The rules are simple. We buy if the 2-day RSI crosses below 20, if the market is above its 200-day moving average.
Then we sell the position if the close crosses above the 5-day moving average.
Here are the backtest results for this strategy!
You see that it struggles during some periods, but we have a general upwards-sloping curve! And that’s much more than you can say about a majority of the strategies that are taught online!
2. The Relative RSI Strategy
This strategy uses the RSI in a rather uncommon way. Actually, here we are not concerned with the actual RSI reading, but more where RSI is in relation to its previous closes.
Here is the backtest for the strategy. It looks quite good!
Here are the rules.
- Today’s RSI is lower than Yesterday’s RSI
- Yesterday’s RSI is lower than RSI two days ago
- RSI two bars ago is lower than RSI three bars ago
So in effect, the buy condition is that we have three consecutive bars with lower RSI readings.
Then, to filter out some bad trades, we also have a filter applied to the strategy. This filter ensures that we only enter during optimal volatility levels.
To filter out trades, we use the ADX indicator with a ten-period lookback, and require it to be higher than 20. This strategy seems to benefit from higher volatility levels, and by requiring the ADX to be higher than 20 we filter out low volatility environments.
By the way, just look at how bad the strategy performs when the ADX is LESS than 20! We certainly don’t want to take these trades!
If you are interested in more trading ideas like this, we recommend that you have a look at our library of edges!
How to Improve RSI Trading Strategies
When we build a trading strategy, we usually start with the raw idea, and then improve on that idea. The improvements could be done by adding filters or additional conditions that remove bad trades, and make the equity smoother.
Since this article is about the RSI, the raw idea could be some of the conventional RSI signals that we’ve covered, such as:
- Oversold or overbought Readings
- RSI divergences
- Going long on high readings and short on low readings (momentum)
In this part of the article, we are going to show you some of the filters and conditions that we often use when we build trading strategies for our algorithmic trading.
Measuring volume is sometimes a great way of gauging the market sentiment and the underlying strength of the move. You could say that adding volume is like adding a new dimension to your trading. And with some strategies, it can provide a significant performance boost!
Of course, you can use volume in an endless number of ways, but the simplest approach, probably is to look for bottoms and peaks. For example, if the market breaks out from a range, you might want to see how many market participants actually fuelled that move. In that case, a big volume spike could be a good sign that there was enough support behind the move for it to last going forward.
Moving averages can also be used to analyze the price of the security. By smoothening the price action, they help to clarify the general trend direction and could be great as filters.
Here are some ways that you could use moving averages to improve on an RSI strategy:
- Close is Greater Than the Moving Average- Here we demand that the close of the bar has to be higher or lower than the moving average before we take a trade
- Shorter Moving Average is Above or Below Longer Moving Average- Here we use two moving averages; one longer and one shorter. Then we may decide to only take a trade if the short term moving average is above or below the long term moving average.
- Rising or Falling Moving Average- By looking at the slope of the moving average, you can get some clues about the trend direction, as well as the trend strength. For example, a steep slope means that the trend is strong, while the opposite is true for a flat moving average.
In addition to using these three conditions, you could also experiment with different types of moving averages. In our own testing, we’ve found that the exponential moving average tends to work better than the simple moving average.
In our article on moving averages, we take a closer look at different types of MA’s and how to use them!
The smaller the timeframe, the more noise you get. That’s why adding a second, higher timeframe could add a lot of value to a trading strategy!
Once you’ve added the second timeframe, you just apply conditions to it as you would have done to the first timeframe.
WE have now come to the end of this article, and hope to have provided some valuable tips on how you can use RSI in your trading.
Before we end we just want to once again stress the importance of being skeptical of what you read and hear! The trading field is littered with false information and concepts that don’t work, and you need to be able to sort them out, before you go live!
And (nearly) the only way of knowing what works and not, is through backtesting! Never trust what other people say. They will mostly be wrong!