I've been working on an academic-style report for a few years now called "Prediction Astrology."
The report aims to expose falsehoods in the two main strategies that are used in predicting financial markets.
The first is something that has always been false: Chart Analysis (or as the trader bros call "Technical Analysis")
The second is something that used to work but is now forever broken: Correlation Analysis.
Both strategies appeal to the human survival instinct of pattern recognition. Watch out for that stick looking thing in the bushes over there, some of them have fangs. See that big cat over there, it looks like the one that ate my cousin. Not sure if I should eat those purple berries - I nearly died last time.
That same instinct is deployed when looking at historical asset price movements. We see shapes and matches, and think they'll repeat or rhyme.
In the case of Chart Analysis, there is a serious lack of empirical data supporting it as a viable investment strategy. The main types of chart analysis are as follows:
- Trendline Analysis: Draw enough diagonal lines on a chart and eventually one of them “predicts” the future like a caffeinated cave painting.
- Support and Resistance: The belief that a stock somehow remembers the exact emotional trauma of a previous price level and will obediently bounce there again.
- Moving Averages: Smearing past prices into a smooth noodle so random chaos looks comfortingly intentional.
- Candlestick Patterns: Ancient rice-trader astrology where “Three Black Crows” and “Hammer Dojis” allegedly reveal the emotional state of millions of strangers.
- Fibonacci Retracements: Applying a sequence found in seashells and sunflowers to leveraged derivatives markets because… nature, apparently.
- Elliott Wave Theory: The conviction that markets move in elegant repeating psychological waves instead of occasionally behaving like a raccoon trapped in a casino.
- MACD (Moving Average Convergence Divergence): A mathematically elaborate way of discovering that prices went up recently and may continue doing that… until they don’t.
- RSI (Relative Strength Index): A momentum gauge designed to warn that something is “overbought” roughly six months before it doubles again.
- Bollinger Bands: Statistical volatility ribbons that make price charts look scientific enough to survive a PowerPoint presentation.
- Volume Analysis: Attempting to infer institutional intent from trading activity the way medieval priests interpreted bird flight patterns.
- Head and Shoulders Patterns: Humans spotting human anatomy in random squiggles.
- Chart Correlation Analysis: The sacred art of discovering that two lines once moved together and quietly assuming they’re soulbound forever.
- Ichimoku Clouds: Turning a price chart into a weather system dense enough that nobody asks difficult questions.
- Wyckoff Method: A framework suggesting “smart money” leaves detectable footprints everywhere while somehow remaining smart money.
- Gann Theory: Geometry, angles, numerology and occasionally what feels like financial freemasonry.
- Breakout Trading: Buying because the line escaped the box you personally drew around it five minutes earlier.
- Fractal Analysis: Zooming into charts until randomness begins resembling destiny.
- Market Breadth Indicators: Measuring how many things are rising or falling in the hope that collective participation contains hidden prophetic wisdom.
- Sentiment Indicators: Turning investor emotions into indicators, which is admittedly useful right up until the crowd becomes euphoric and rich simultaneously.
The main problems with Chart Analysis are data cherry picking and survivorship bias. Humans (especially men) love to celebrate their wins, but very few talk about their losses. The winners beat their chests on social media, write articles, go on podcasts, and even go on to teach "courses" on how to trade. The losers anxiously rehearse the best way to tell their wives where the house deposit went. This is precisely why the trading culture remains so strong despite the dismal results on aggregate.
Given the sheer enormity of the data set, there is a statistical certainty that some traders will generate extraordinary returns using Chart Analysis. After all, sometimes these methods do work. The problem is that they only work around 50% of the time. When you take into account the friction costs (costs of trading), interest on margins, and the fact that trading does not improve global productivity whatsoever (it merely transfers wealth from losers to winners), we have a negative sum game scenario. In the trading game, a tiny fraction win, and almost everybody loses.
The academic literature behind Chart Analysis is weak at best. There is more evidence criticising it. My advice to traders is: Don't bet more than you can afford to lose, because make no mistake - no matter how many "courses" you do, Chart Analysis is gambling.
Now onto Correlation Analysis.
Correlation Analysis became enormously influential because unlike Chart Analysis, many historical market relationships genuinely did possess practical utility for extended periods of time. Equities and bonds frequently diversified one another during disinflationary environments, gold often behaved inversely to the US Dollar, liquidity expansion frequently supported risk assets, and oil shocks often pressured growth-sensitive sectors. These relationships were useful enough that entire institutional portfolio frameworks were built around them.
The problem emerged when investors gradually stopped treating these relationships as conditional characteristics of a particular macroeconomic regime and started treating them as permanent structural laws of markets. The traditional balanced portfolio is probably the cleanest example. Over the last 50 years, investors became accustomed to an environment where bonds frequently cushioned equity drawdowns, reinforcing the assumption that diversification through stock-bond negative correlation was structurally reliable. Then inflation returned and in 2022 both sides of the portfolio fell simultaneously, with the S&P 500 down -18.04% and US 10-Year Treasury bonds down -17.83%. The relationship crumbled under the weight of money-for-nothing financialisation - the modern era of reckless money printing and financial engineering, designed to get money out of the working economy and into the coffers of elites.
The same instability increasingly appeared across other supposedly reliable relationships. Gold stopped behaving like a clean anti-dollar instrument during multiple periods where both assets strengthened simultaneously. Bitcoin, heavily promoted as both “digital gold” and a debasement hedge, increasingly traded like a meme. Rolling correlations between Bitcoin and equities strengthened materially during multiple risk-on/risk-off periods, undermining the idea that it functioned as a stable uncorrelated reserve asset. The "debasement trade", to the extent one ever existed, is no longer a reliable predictor of gold, silver or Bitcoin prices. All correlations between these assets and money supply, DXY and derivatives thereof has been broken. You'd have better odds on the roulette table.
These divergences are not random. Modern markets are operating in conditions radically different from those that shaped many of the historical relationships investors still rely upon. Central-bank intervention, passive index concentration, algorithmic trading, sovereign debt saturation, reflexive ETF flows, geopolitical fragmentation, and structurally distorted liquidity conditions have created a financial environment where historical correlations have permanently detached. The issue is no longer whether a relationship existed. The issue is whether that relationship retains meaningful predictive utility moving forward.
The most dangerous word in finance today is “historically”. Historically, bonds protected portfolios. Now, they amplify both gains and losses. Historically, gold moved inversely to the dollar. Now, it can move in lockstep one period and opposite the next. Historically, diversification reduced volatility. Now, traditional diversification amplifies volatility.
Historical observations remain useful, however investors and institutions increasingly make the mistake of equating success in the "everything bubble" as evidence that Correlation Analysis works. The recent crashes in Bitcoin and Gold should serve as a warning sign to Equities and Bonds. Markets do not care about historical correlations. Markets only care about the conditions operating right now. This time really is different, bad different.
Bryce Jenkins
Founder & CEO | Financial Planner since 2008
Phone 07 3184 8149
Email bryce@thevco.com.au
Head Quarters 28 Charlton Street, Southport 4215
Bryce Jenkins is an Authorised Representative (AR# 332683) of The Virtuous Licensee Pty Ltd AFSL 526892
***Privacy & Confidentiality Notice and General Advice Disclaimer*** This email (which includes any attachments) may contain information which is privileged, confidential and exempt from disclosure under applicable law and is intended only for the use of the person or entity entitled to use it. If you are not the proper recipient of this email, please delete this email and advise us immediately. Any unauthorised use, dissemination, distribution, or reproduction of this email is strictly prohibited. We expressly exclude liability for viruses contained in this email. The information in this email is not to be considered personal financial advice. Please do not act upon any document unless it is titled 'Statement of Advice'. Neither Virtue Planning Pty Ltd t/as The Virtuous Collective, nor any of its directors, shareholders, support staff nor associated entities will be held liable for any action arising from, or influenced by any part of this email.
