I saw an interesting chart over the weekend and want to share it with you so I recreated it.
There are Bollinger Bands, moving averages, trend lines, stochastics, RSI and MACD. I think the chart is saying to buy but it’s not easy to tell. This chart is clearly an example of too many indicators. In fact, all three of the indicators at the bottom use the same information and will almost always give the same signals. We can confirm that with another chart.
In the chart above, all three indicators have been scaled to show the same range. They all generally move in the same direction at the same time. From the chart, we can see that stochastics, RSI and MACD appear to be nearly identical.
This shouldn’t be surprising. Stochastics, RSI and MACD are all just different calculations involving the closing price. This is true of most momentum indicators. Analysts have always sought an edge in trading and many have designed indicators using the closing price in pursuit of an edge. Some used moving averages and others decided they could get an edge with faster moving averages like exponential moving averages, which became popular in the 1960s. The result is hundreds of momentum indicators that all seem to provide the same information.
Looking at the chart of the three indicators, we can see slight differences in the indicators.
RSI seems to be the most volatile. This is because of the way the indicator is calculated. Indicators usually smooth the closing price to reduce the day to day volatility of the price action. MACD and stochastics use two levels of smoothing by calculating moving averages of the initial moving averages that are used in the formulas. RSI only uses one smoothing step in its calculation and is therefore a more volatile indicator. That should make it more responsive to the market action.
We can test this idea. The test can be fairly simple. We will use standard default settings for all indicators and apply standard trading rules.
For stochastics, we will enter the trade when the fast stochastic crosses above the slow stochastic and exit four weeks later. For RSI, we will enter when the indicator crosses above 30 and exit four weeks later. For MACD, we will buy when the MACD differential turns positive and sell four weeks later. These are standard interpretations of the indicator although it’s possible to add a number of subtleties to the rules. For example, we could take buy signals in stochastics only when the indicator is below 20 and sell only when the indicator becomes overbought and is above 80. We will explore tweaking the indicators in future articles, but our purpose here is simply to see if it’s worth following multiple momentum indicators on a chart. To focus solely on the indicators, we are using the exact same rule to measure the effectiveness of the signal.
Our test will include the stocks that are currently in the S&P 500 and cover the past 25 years. Commission costs of $5 will be deducted for each trade. The results are shown in the table below.
The results show fairly similar results but overall there seems to be a slight edge for RSI. This is to be expected based on how the indicators are calculated.
In technical terms, we lose information when we smooth indicators and excessive smoothing decreases the usefulness of an indicator. In practical terms, when an indicator calculation includes taking a moving average of a moving average the results are likely to be disappointing. MACD and stochastics take moving averages of moving averages in their calculation and the second step results in later signals. Test results for these two indicators is almost exactly the same so there is definitely no need to look at both of them.
From testing, we can draw several conclusions about momentum indicators.
MACD is the least useful indicator using default parameters. It’s possible we could find a way to improve the accuracy of the indicator. But that might not be the best use of time. The indicator is simply not among the best performers and we might be better off working to improve one of the better performing indicators.
Stochastics offers only a small improvement on MACD and again, it might not be worth the effort to improve on the signals.
RSI seems to need the least amount of fixing. At first glance, we have a profitable trading strategy. Of course there is more to the story than just the percentage of winning trades and the average gain per trade. We would also want to look at the largest losses and how the indicator fares in bull and bear markets. But the test results indicate that this is a project that could be worth the time. We will look at improving RSI in a future article.
Although RSI delivers better results, the performance is fairly similar to the other two momentum indicators. It seems safe to conclude that momentum indicators all duplicate each other. When adding indicators to charts, it is important to consider how they are calculated. There is no need to add multiple indicators that provide redundant information.
It’s probably best to only use one indicator that smooths the closing prices to form an oscillator, since adding more will generate trading signals at about the same time. Some traders will claim these different signals add to their confidence in a trade but they don’t understand the math. RSI, MACD and stochastics are all designed to give the same signals within a few days of each other. Normally, we see RSI signal first followed by the other two within a few days. This should not boost confidence in the strength of the trend because it is not providing new information – it is merely telling us that all momentum indicators using closing prices are turning bullish.
To obtain new information, consider combining a momentum indicator with a trend following indicator like a moving average. For example, only take RSI buy signals when the stock is above its 50-day moving average. This way you are buying stocks that are experiencing short-term pullbacks in long-term up trends and that has been proven to be a winning strategy over time. Historically, buying pullbacks in up trends is known as “buying the dip.” Using one momentum indicator and one moving average is a simple way to implement a dip buying strategy.
But, again the key will be to avoid cluttering the chart with meaningless information. Don’t complicate the chart with multiple indicators that are calculated in a similar way. A 50-day simple moving average will be similar to a 50-day exponential moving average or a 50-day triangular-weighted moving average.
Perhaps the best way to think of what should go on a chart is to set a goal of maximizing the data-to-ink ratio. Each data point you add will require additional ink to print. You should think through whether or not the additional ink provides valuable information.
Turning back to our original indicators, MACD uses the most ink to print and provides the least amount of useful information. Stochastics will require twice as much ink as RSI to print but doesn’t deliver twice as much information. In fact, stochastics offers less information than RSI if information is defined as profit. To maximize the data-to-ink ratio, use only RSI based on this test.
I know ink isn’t very expensive and most traders rarely, if ever, print charts. But the data-to-ink ratio should be an important consideration when adding indicators to charts and testing confirms the value of this tool with momentum indicators. One momentum indicator is enough and two are actually never needed.