The Edge of Technical Indicators

Classical technical indicators like RSI and Stochastic are commonly used to build algorithmic trading strategies.  But do these indicators really give you an edge in your market? Are they able to define the times when you want to be invested? This article will show you a way to quantify and compare the edge of technical indicators. Knowing the edge of the indicator makes it an easy task to select the right indicator for your market.

The edge of an indicator

Any technical indicator, let it be RSI, moving averages or jobless claims, has got a primary goal. It should signal if it is a wise idea to be invested or not. If this indicator signal has any value, on the next day the market should have a higher return than it has on average. Otherwise  the usage of no indicator and a buy and hold investing approach would be the best solution.

The edge of an indicator in investing consists of two legs.

  1. the quality of the signal
  2. the number of occurrences

The quality of the signal can simply be described as the average close-to-close market move after the signal occurred.

Reading an Indicator

To quantify the prognosis quality of a technical indicator we first have to have a short look on how to set up the test. Therefore I introduce a standardised set of rules on how to interpret the indicator. This will it make possible to compare different indicators based on the same rules, thus only testing the quality of the indicator and not the quality of the rules.

Slow Stochastic

As an example, the slow stochastic indicator (and any other indicator) can be described using 2 rules:

  1. the absolute level of the indicator
  2. the direction of the indicator

Using these 2 rules to describe the indicator, a possible test could be: If slow stochastic is between 0 and 20 (and rising/falling), should I be invested on the day after?

Annualised Indicator Edge

As mentioned above the edge of an indicator is defined by the quality of the signal and the number of occurrences. Let’s first concentrate on the quality of the signal only.

I like to use annualised readings, so I easily can compare the market returns to the quality of my indicator signals. Therefore the average percentage daily market move after a signal has occurred is multiplied by the number of trading days. This I will call the annualised indicator return.

Stochastic and S&P 500

Using the layout given above we can run a test of the edge of the 9-day slow Stochastic indicator (as an example indicator) using various settings:

Stochastic level edge test

Stochastic level edge test

The chart above gives you the annualised indicator edge depending on the absolute level of the indicator. The direction of the indicator was not taken into account and could be the basis of another test.

Comparing these results to the average buy&hold return it is clearly visible which levels of stochastic offer an edge in the S&P.

0-20 annualised Stochastic edge vs. average market returns

If the 9-day slow stochastic is between 0 and 20, the market has an annualised return of more than 30% on the next day (close to close). As the average annualised market return on any given day is around 10%, an investment in S&P in oversold areas seems to give excess returns.

Performance=Edge*No. of Occurrences

Beside the edge of the signal the number of occurrences is important to measure the quality of an indicator. If your 100% confidence signal only occurred every few years, you most probably will die as a poor man.

Stochastic level test sum profit

Stochastic level test sum profit

Summing up the percentage returns after indicator signals shows a clear correlation between the quality of the signal and the quality of the summed up returns. The setting with the highest edge (0-20) also has got the most favourable return curve. The chart above shows the total % return of the market within the given indicator settings.

Compare technical indicators and markets

Using the calculations given above I ran a short cross market test over 2 indicators: The 9-day RSI and the 9-day Stochastic. Both are tested for the signal quality in the oversold area from 0 to 30. Data from 2000-2019 has been used.

indicator-market test

First have a look at the annualised indicator return, defining the quality of the signal. Nasdaq seems to respond best to both indicators, showing the highest annualised return for RSI.

Also the spread between the annualised indicator return and the average market return is important. As it can be seen, all four markets had about the same average yearly buy&hold return since 2000 (5-6%). The spread between the  annualised indicator return and the average market return is the key number in this scan. It defines the absolute edge the indicator offers in a specific market for a long only investor. The scan also shows RSI with Nasdaq or S&P as the best combination for buying in the oversold area.

Takeaway

Separating the quality of the indicator signal from the number of occurrences makes it easy to compare technical indicators and settings. Showing the spread between average market returns and the returns after an indicator signal has occurred enables you to find the best indicator-market combinations.

 

 

If you like it please link it
error

Daily Extremes – Significance of time

Analysing at which time daily market extremes are established shows the significance of the first and last hours of market action. See how different markets show different behaviour and see what can be learned from this analysis.

Probability of Extremes

A day of trading usually starts with a lot of fantasies for the future, then we try to survive the day and end it with a lot of hope for tomorrow. This psychological pattern can also be shown when analysing intraday market data. A high level of fantasies usually leads to a strong market movement, and thus market extremes can often be seen near the beginning or the end of the trading session. Continue reading

If you like it please link it
error

S&P500 – when to be invested

The stock market shows some astonishingly stable date based patterns. Using a performance heat map of the S&P500 index, these patterns are easily found.

Date based performance

The chart below shows the profit factor of a long only strategy investing in the S&P500. Green is good, red is bad. The strategy is strictly date based. It always buys and sells on specific days of the month. Continue reading

If you like it please link it
error

Noisy Data strategy testing

Algorithmic trading adds noise to the markets we have known. So why not add some noise to your historic market data? This way you can check if your algorithmic trading strategies are fit for the future. Learn how to generate noisy data and how to test your strategies for stability in a noisy market.

Synthetic market data?

Continue reading

If you like it please link it
error

Factor investing in portfolio management

Factor investing has been around in portfolio management for some years. Based on algorithmic rules it became the big thing in trading and the ETF industry. But is there still some money to be made? Is small beta still smart or just beta? This article will give you a Tradesignal framework to test the factor investing ideas by your own. Continue reading

If you like it please link it
error

Backtesting Market Volatility

If you want to trade volatility, you can place a bet on the option market. Just buy an at the money put and call, and at expiry day you will either win or lose, depending on the actual market move since you bought the straddle and the price you paid for the straddle. To put it simple, if the market moves more than you paid for the two options you will win, otherwise you will lose. This article is about a back test of volatility. Continue reading

If you like it please link it
error

Demystifying the 200 day average

The 200 day average is considered as a key indicator in everyday technical analysis. It tells us if markets are bullish or bearish. But can this claim be proved statistically, or is it just an urban legend handed down from one generation of technical analysts to the next? Let’s find out and demystify the 200 day moving average. Continue reading

If you like it please link it
error

Statistics of VIX

The CBOE volatility index VIX  measures the market’s expectation of future volatility. This article will show you some key statistics of VIX and help you to decide if it is better to buy or to sell volatility.

Statistics of VIX

The spikes to the top and the long phases of relatively low volatility are reflected in a left-leaning distribution diagram and a long tail towards the higher levels. The median value is 17%, meaning 50% of the prices are above (below) this level.

The next chart shows the distribution of returns over 25 trading days. The median price movement being slightly shifted to the negative area shows the mean reverting characteristics of volatility.

Buy or sell volatility?

Analysing the level of VIX and the returns afterwards yields an even more interesting picture:

The green line gives the 25 bar percentage returns of VIX, with VIX noting above 25, the red line gives the returns with VIX below 15. Observe the median of the two lines:

The median 25 bar return with VIX above 25 (green) is around -15%, only 20% of the returns are positive when VIX is currently above 25. Sell volatility.

The median returns with VIX currently below 15 (red) is above 0% and with a fat tail to positive returns. Buy volatility. (data from 2004-2018)

Adverse movement of VIX

The above chart suggests that going short on volatility, if VIX is above 25, seems to be a good idea. But it is not without risk. The chart below shows what can go wrong during the next 25 days. The distribution diagram gives the maximum adverse movement of the VIX, with VIX currently trading above 25.

The green line, VIX currently above 25, shows a +10% median maximum up movement over the next 25 days. So do not expect a short vola position to be without risk. Some adverse movement has to be expected.

On the other side, the distribution of the maximum loss of the VIX during a 25 day period shows a median of below -20%. This represents the profit potential of a short volatility position.

Conclusion of VIX statistics:

If you plan to short volatility wait until VIX is trading above 25. If you want to buy volatility, do so if VIX is trading below 15.

 

If you like it please link it
error

Machine learning: kNN algorithm explained

I always thought that inspiration and experience are key factors in trading. But every time my chess computer beats me without any inspiration, just by brute force, I get my doubts. This article will be about a brute force approach in trading. The kNN algorithm.

Rule based trading

Rule based trading – algorithmic trading, is just a name for a set of if..then rules which will define the machines trading decisions. e.g. if the market crosses below the 200 day line, then short 100 contracts. If the market rises by 2% then exit the position.  Easy stuff like this… (for the beginning)

This article will be a short introduction to machine learning. I will use a classic algorithm of machine learning to let my computer find a prediction for tomorrows market move. In the meantime I’ll have a glass of wine with some friends and let the machine do the job; At least that’s the idea, but can it be that simple in real life trading?

Unsupervised machine learning – kNN algorithm

The kNN algorithm is one of the most simple machine learning algorithms. Learning, in this case, is only a nice sounding label, in reality kNN is more of a classification algorithm.

This is how it woks:

The scatter chart above is a visualisation of a two dimensional kNN data set. For this article I used a classical indicators of technical analysis to do the prediction: a long-term and a short-term RSI indicator. The dots on the two dimensional scatter chart represent the historic RSI values at a given point of time.

Now have a look at the fat circled point. This point represents today’s value. It means, that today’s RSI1 has a value of 63, and RSI2 got a value of 70.

Additionally to the position on the chart the dots have got colours. A green dot means that the market moved up on the following day, a red dot shows a falling market on the day after.

We already know what has happened in history, so it is easy to colour the historic dots. But we do not know the colour of today’s dot, as it is not known where tomorrow’s market will end.

Based on the chart above, will it be a red or green dot? Will tomorrow be up or down?  Should I go long or should I go short?

kNN – k nearest neighbours

To do a prediction of tomorrow’s market move, the kNN algorithm uses the historic data shown on the scatter plot above and finds the k-nearest neighbours of today’s RSI values. As you can see, our current fat point is surrounded by red dots. This means, that every time the two RSI values have been in this area, the market fell on the day after. That’s why today’s data point is classified as red. Wish it would be that easy all the times…

Call it classification or prediction, the two dimensional kNN algorithm just has a look on what has happened in the past when the two indicators had a similar level. It then looks at the k nearest neighbours, sees their state and thus classifies today point.

kNN as Tradesignal Equilla Code

In this article I would like to show you an implementation with the Tradesignal programming language Equilla.

To implement the algorithm in Tradesignal we first have to do the shown scatter plot. The algorithm stores the values in an array.

8/9 calculates the value of the fast and slow RSI indicators

12/13 looks what will happen on the day after (for the training data set)

16/17/18 stores everything in an array.

The next task to complete is to calculate the distances of today’s RSI point to all the historic points in the training data set.

23/27 calculates the euclidean distance of today’s point to all historic points, line 29 then creates a sorted list of all these distances to find the k nearest historic data points in the training data set.

Nearly done. The next step is just to find out what classification (colour) the nearest points have got and use this information to create a prediction for tomorrow. This is done in lines 33 to 35

Have a look at the scatter chart at the beginning. If this would be the data stored in our training data set, the prediction, using the 5 nearest neighbours, would be -5. All the 5 nearest neighbours of our current data point are red.

Now that we got a prediction for tomorrow, we need to make use of this prediction and trade it. The returns then will show if everything works as predicted.

Over here I just do a simple long/short interpretation of the prediction, but of course you could also use the quality of the prediction (+5 or +1?) in some sort of way. Position sizing…?

kNN algorithm performance

The next chart shows 2000 bars of daily Brent data. It uses a 14 and 28 day RSI to predict the next day’s move in the Brent oil market. The training was on the first half of the data set, and the 5 nearest neighbours did the classification.

Underneath the chart the returns of this test are shown. (strategy equity). On the bottom of the char you see the two RSI indicators used for the generation of the prediction / buy-sell command.

kNN algorithm – conclusion

The kNN algorithm offers a framework to test all kind of indicators easily to see if they have got any predictive value. Judging on the shown graph it seems to work. It seems to be possible to use these two RSI indicators to predict tomorrow’s Brent move.

But unfortunately this also could be just completely useless curve fitting. It is you who has to select the indicators and their periods and you will have to define if you like the outcome of a selected parameter set. To many degrees of freedom to be sure. The kNN algorithm is useful, but its application in finance has to be treated carefully. Otherwise bad surprises are guaranteed

Not everything can be done by brute force, inspiration and experience are key factors in finance…

 

 

 

If you like it please link it
error