Why Technical Indicators Are Used as Model Inputs
Summary
The document discusses why researchers use moving averages, MACD, RSI, and related transformations as inputs instead of raw financial time series. The responses offer several possible roles: smoothing noisy observations, representing momentum through moving-average crossovers, and testing whether an indicator has predictive value. They caution that common indicators may behave differently across market cycles, volatility conditions, and price patterns, and that flexible interpretation can weaken research conclusions.
The main methodological point is to treat indicators as hypotheses that require clearly defined statistical evaluation, rather than assuming their conventional interpretations are reliable. One response contrasts indicators with tools such as Kalman filters, which have specified statistical purposes, while another recommends evidence-based testing to distinguish predictive signal from folklore. The discussion is opinionated and supplies no datasets or reproducible empirical results; its claims about indicator performance should therefore be read as cautions and perspectives, not as a systematic review.
Key ideas
- Moving averages can smooth noise in financial time series.
- Moving-average crossovers may also encode momentum information.
- Common indicators may not perform consistently across market conditions.
- Researchers should define and statistically test an indicator's predictive claim.
- The responses provide perspectives rather than a reproducible comparison of indicator strategies.
Tags
Full text
# The Basis of Using Technical Indicators as Inputs # The Basis of Using Technical Indicators as Inputs In the process of my research I very often come across academic papers regarding modelling and trading strategies that in one way or another incorporate some technical indicators. For example in some papers, rather than taking the raw time-series as input to the model, things like MACD, SMA, EMA and RSI of the time-series is used instead. I have however never come across a single paper that states the basis for applying these technical indicators. My question is why are these technical indicators so commonly seen in researches in this area even though technical analysis has a bad rep for being a pseudo-science. Are the above papers simply lacking in rigor or are there some deeper reasons for this phenomenon? ## Answer by Matt Wolf (score 7, accepted) https://quant.stackexchange.com/a/7057 - The only "indicators" that I believe add value in academic research are time series smoothing functions. ( I don't call them indicators because they are all lagging thus do not indicate anything into the future). - There is clear empirical evidence and a number of academic papers have been published that show that none of the common indicators (common defined maybe in pseudo-bibles such as Murphy book "Technical Analysis of the financial markets") in isolation and even in combination perform equally well throughout different cycles, market dynamics, volatility regimes, or price patterns. Thus, for many it becomes a self-fulfilling prophesy. They are very silent when such indicators do not work and point to the "technical analysis is an art" aspect and remind every last soul that indicators "just work" during regimes when, for example an RSI indicator indicated oversold conditions and price action indeed picked up. So, yes I believe not few in the academic community already make wrong choices in the tool selection of their research efforts. Please let us not confuse "financial technical indicators" with sound statistical modeling techniques and tools. A Kalman filter is doing exactly what it is supposed to do at ALL TIMES. An RSI indicator actually very rarely does what it is supposed to do, which is to indicate "oversold" or "overbought"conditions. In fact I ran tests a long time ago and showed that a strategy which shorts the asset on which an asset's RSI reaches oversold territory for the first time as well as the reciprocal almost always outperforms a strategy in which someone uses RSI in the general prescribed way. (this incorporated "prudent" trading, meaning realistic risk-management and stop loss levels). I ran many such tests in my junior quest for the holy grail which I soon figured does not exist. Unfortunately, technical financial indicators suggest to many that they are second best thing after the holy grail. In summary, your question can lead to the never-ending discussion of pro and cons of technical indicators. I believe they do not belong into any serious academic research for the simple reason that they can be interpreted in all sorts of ways. A research tool should be applicable to a specific approach and it should output results that can be relied on with certain margins for error that are themselves well defined. ## Answer by vonjd (score 7) https://quant.stackexchange.com/a/7046 I think the answer to your question is very dependent on the respective indicators. It can be argued for example that moving averages not only smooth out time series but because they are a shifted version of the original series signals on crossovers make use of the momentum factor. In general you might want to check out the book Evidence Based Technical Analysis by David Aronson. In it he applies statistical techniques to determine whether certain technical analysis indicators have any predictive power. It's an interesting read and should equip you with some ideas on how to differentiate between folklore and statistical rigor. You can find a good overview and summary of the book on CXO Advisory. You can also find further material on the webpage of the author. ## Answer by IgorS (score 3) https://quant.stackexchange.com/a/7045 Typically "average" lines are used to get rid of noise in the original data. It seems pretty logical to smooth intra week fluctuations when working with a year of data.
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.