I'm on record saying classic technical analysis - trendlines, chart patterns, volume and open interest, etc. - has grown obsolete as algorithm-driven trade continues to evolve.
If the key driver of markets - algorithms - aren't paying attention to such things then neither should we.
From an analytical point of view, our attention turns to the mathematical side of technical analysis, though we still don't know what goes into the thousands of different trading programs.
Last Friday, I had the opportunity to be the guest on Barchart’s Market on Close program, hosted by the company’s Senior Market Strategist John Rowland. I’ve had a few conversations with John over the past number of years, and it is interesting how we can arrive at similar conclusions while taking two separate paths. John was correct, in a way, when he said we don’t always see eye to eye, but that’s to be expected whenever two people talk market analysis, economics, politics, or basically any subject.
Our discussion this time around was sparked by my piece on Barchart from January 14, “Who Are the Funds in Commodity Markets and What Might They Do in 2025?”. What captured John’s attention was my point that, due to the evolution of market’s, classic technical analysis has largely become irrelevant. This struck a chord with John because he is a technical analyst, a Chartered Market Technician (CMT) at that. What does this mean? According to Investopedia, “A CMT is a professional credential that demonstrates an individual has deep knowledge and capability in technical analysis and trading.” I’ve been asked before, given my nearly 40 years of studying charts of every type, why I’ve not taken the steps to become a CMT. My response is as it has always been, “Why would I?”
Back to our discussion though. The point I continue to make is if Watson, my name for the computer algorithm trading industry that generally sets market direction, doesn’t pay attention to things like trendlines, classic technical patterns (e.g. head and shoulders, price gaps, trendlines, etc.) then why should we? After all, if our job is to make money in the markets, whatever sector it might be, then it will likely be more profitable to understanding what makes Watson move and be able to recognize when a potential change in direction is drawing near.
This brings a couple key discussions into play: First, Newton’s First Law of Motion applied to markets, “A trending market will stay in that trend until acted upon by an outside force”. My argument with the industry, literally, has been that Watson is that outside force. Or in other words, a trend setter rather than a trend follower. The subject was touched on during my discussion with John, though there was so many other things to talk about we didn’t dive deep into the debate. I remember he asked if Watson could be viewed as both the trend setter and trend follower. This is certainly a possibility, but if Watson follows trends, who, or what, is big enough to actually change price direction over time (simplest definition of trend)?
Before we go down that rabbit hold, let’s return to the subject of understanding what makes Watson tick. Twenty years ago, roughly, when computer algorithms started taking over markets, the key factor was words in headlines that were classified as either bullish or bearish, and when fed into the machine would trigger the expected response. (Yes, this is a gross oversimplification of the process but is what I observed while working in the newsroom at the time.) The game then became to write certain words into headlines, regardless of what the actual story was about. Algorithm writers soon caught on and we could watch as those same words were largely ignored. Then came the US presidential election of 2016, and the following four years of nothing but attempted market manipulation through the use of media, both traditional and social, by randomly throwing out key words like “tariffs”, “trade deals”, and “covfefe”. As that US administration ran its course, Watson learned to ignore nearly all of the nonsense, returning to its roots of…
Mathematics. When I talk about a new branch of analysis, outside of the classic technical, fundamental, seasonal, etc., the reality is the understanding of what makes Watson tick isn’t new at all. It’s the old as time study of mathematics.
During our discussion, John started listing some of the things he watches: Moving averages, Bollinger bands (a technical analysis tool that helps gauge the volatility of a financial instrument and determine if it’s overvalued and undervalued), volatility, and Fibonacci retracements. The common theme among these indictors is they are all determined by mathematical equations.
Those of you following along with my analysis (daily, weekly, monthly, etc.) know I use many of these same elements. I’ve never been big on moving averages, though I was fascinated by the section devoted to the topic in John J. Murphy’s classic book, “Technical Analysis of the Futures Markets”. However, I use a variation of the idea to track momentum with the Four-Week Rule, a system described by the author as “simplicity itself” (pg. 268). For overvalued and undervalued I created price distribution studies, though Watson isn’t concerned about previous price ranges, as we can see in the recent moves in cattle (both feeder and live) and the ongoing rally in gold (GCJ25). Volatility has long been a key component of my market rules (part of #3: Use filters to manage risk), with the VIX (CBOE Volatility Index) ($VIX) near the top of the list of market’s my son Ben tracks for our in-house algorithms.
Which brings us to Fibonacci, the famed Italian mathematician from the Middle Ages. I’ll go into his work in more detail next time, but a quick summary is his ratios are found to hold true in the natural world, so it is interesting they work in a system (algorithm trading) driven by mathematics. I’ll close this piece with what I considered the key takeaway from last Friday’s discussion: The more advanced new trading programs have become, the more our analysis depends on the old work of greats like Isaac Newton and Fibonacci.
