Keltner Channel - Better Than Bollinger Bands? Here’s How It Works

Hey everyone and welcome to our video! One of the coaches of Unger Academy here.

And today, we're going to be talking about the Keltner Channel. 

It's a price channel that is an alternative to the more classic Bollinger Bands. Although certainly less famous than the Bollinger Bands, in my opinion, this channel still deserves some attention.

Channels comparison

So let’s start with the video. First things first, I'd like to show you a chart on which we have already plotted the Bollinger Bands. And in this case, we calculated these bands with 20 periods and 2 standard deviations. The Bands here are applied to the Standard & Poors 500 chart with a 30-minute timeframe.

Now, let's plot the Keltner Channel on this chart, and build it in a similar way, using 20 and 2 as inputs, where 20 indicates the calculation periods and 2 is the multiplying coefficient associated with the width of the channel. As you can see, the output is very similar to the Bollinger Bands. 

So how is this Keltner Channel built?

Well, like the Bollinger Bands, this simple channel also starts from a simple moving average calculated on "n" periods. Then, instead of using the standard deviations to find the upper and lower sides of the channel, in this case, we use the Average True Range calculated on the same calculation period as the simple moving average, and we multiply it by a certain coefficient, which we will call "c", and in this case, the coefficient will be equal to 2, so the same number of standard deviations we used for the Bollinger Bands.

Now, if we look at the graphical comparison between these two channels, everything seems to be very linear and similar. However, I’d like to point out that this is not always the case. 

Let's take a look at this specific example. Up to a certain point, the market has been almost completely flat. And after that point, there was a small explosion, and then the market went back to a trading range area. And as you can see, in this case, the Bollinger Bands behaved in a very different way compared to the Keltner Channel.

But why? Well, the fact is that the Bollinger Bands work with standard deviations, and standard deviations tend to behave differently when there are some outliers, such as this one, for example. This outlier comes from a very low trading range and tends to behave differently than the mean of the average true range multiplied by two.

So this channel, the Keltner Channel, follows its amplitude, which increases if we consider the difference from a vertical point of view, but it doesn't do it as much as the Bollinger Bands do.

The strategies

Now let’s draw a comparison between a strategy that works with the Bollinger Bands and a strategy that works with the Keltner Channel. And for this purpose, we can use a strategy that we have showed you in a previous video. You can find the link here at the top right.

This strategy is very simple. Here we have the condition that limits the number of entries per session to 1, and then there is the condition that defines the upper and lower Bollinger Bands. To these two conditions, we added the instructions to define the Keltner Channel.

We added two different variables to define which channel to use: "KeltOk" or "BBOk". To set those variables in the script we'll use inputs: to use the Bollinger Bands, we need to set the "BBOk" input equal to 1 and the "KeltOk" input equal to zero. On the other hand, if we want to analyze how the strategy works with the Keltner Channel, we need to set the "KeltOk" input equal to 1 and the "BBOk" input equal to zero. 

The strategy is very simple and is based on the cross between the price and the channels. So let's go and start by evaluating how this system works with the Bollinger Bands as a condition for opening the positions. 

As you can see, we'll set the calculation period to 20 and use two standard deviations. So let's activate the strategy and go and see the performance report. 

Okay, the performance report is pretty good. We can see a rather abrupt movement in the post-Covid period from March 2020 onwards. However, as you can see, the equity line seems to be quite regular. The average trade is obviously very low, and this is due to the very high number of trades made by the system. Anyway, this is definitely an excellent starting point for further developments.

Let's check out what happens if we use the Keltner Channel. As a first thing, let's go and see what could be the best parameters to define the length of the calculation period and the multiplier associated with the average true range. To do this, we can simply launch an optimization for the period length from 6 to 30 and for the multiplier from 1 to 3 in small steps of 5 hundredths. Now let's replace the values, set the Bollinger Bands input to zero and the Keltner Channel input to 1, and launch the optimization. It takes very little time.

The results

And let's go and check out the results. Since we're optimizing two inputs, we can take a look at a 3D representation of the results based on the net profit. On the x-axis, we can see the length of the calculation period, and on the y-axis, we see the value of the multiplier. This sort of curve can show the location of the highs, so the more we increase the length of the period to calculate the Keltner Channel, the more the value of the multiplier should also increase. We can take a combination of values from this area. So, in this area, the length of the period is about 15, and the value of the multiplier is around 1.6. 

But let's see what the exact numbers are. The length is 14, and the value of the multiplier is 1.65. As you can see, the results we get with the values immediately before and immediately after are almost identical. If we take 14 for the length, the corresponding multiplier will be 1.7, which is immediately below. So I think that we can take the second result, for example.

Now let's go and see the equity line we would get with these values. The number of trades is lower. Before, we had 4,800. Now we have only 4,000. The average trade has increased from 18 to 24, and the equity line also looks better than before. There has been a sharp drop here, in the last period, resulting in a clearly out-of-the-range drawdown. However, it seems to be a good start.

Of course, now we have been looking for the best value for this channel. However, we still need to explore the results that we would get using the surrounding values and trying to see how they perform together with a stop loss or a take profit, and then we could also add some filters to further improve this entry trigger.

Anyway, the purpose of this video was just to show you the potential of this indicator, which in my opinion, can be as useful as the Bollinger Bands.

And with that, our video is over.

And if you want to dive deep into the world of systematic trading and learn more about how to build automated trading systems, I invite you to check out our free webinar in which Andrea Unger, the only 4-time world champion of real-money trading will show you the exact steps / to enter the world of systematic trading.

Thank you so much for watching! I will see you in our next video dedicated to trading systems! Bye-bye for now!