Up until now we have explored four small market effects, and developed five systems from those explorations. Each of the effects were researched to characterize what makes them a better or worse trade, and what features are stable over time. None of these systems we developed as part of this series are amazing on their own. And nobody should go out of their way to allocate hard-earned portfolio

I constantly see people rise in life who are not the smartest, sometimes not even the most diligent, but they are learning machines. Charlie Munger. As I mentioned in the piece Two years of Quantitativo, the newsletter went paid. In order to continue contributing to the broader community, I decided to share quick summaries of the recent papers Ive read over the past few weeks. These

Retail trading, especially in the options market, which has traditionally been the domain of institutional traders, has received relatively little attention. However, with the rapid growth of educational content, AI, social media, and commission-free trading platforms, this is no longer the case. Today, retail investors account for a significant share of options market volume and are changing

Last year I built a stock scanner in Python. The first backtest said my signals returned +40% over two years. I was, briefly, a genius. Then I fixed three bugs and the same signals returned roughly nothing. None of the bugs were in the strategy. All of them were in the measurement. The strategy hadnt changed; my ruler had. That experience turned into a validation harness that now gates every

You vary your parameters, watch performance hold up across the range, and conclude the system is robust. The old article "Parameter Stability Beats Best Parameter" told you to prefer the stable region over the lucky peak, and you did. The trap is that you can run a parameter sweep that holds up beautifully and proves nothing, because the sweep never tested the parameter space at all. It

This is post #10 in my 2026 series on portfolio optimisation. Time for a quick recap. I'm not going to revisit every post but instead summarise what I now think one should be doing when optimising forecast weights before costs (I haven't yet incorporated costs, nor thought about instrument weights). Pool all instrument returns together At a minimum use a 40 year EWM for SR estimates (and

The market changes all the time. Sometimes it trends, sometimes it oscillates, sometimes it goes sidewards. Trading systems that do not react on market regime change will bring uncomfortable times for their traders (and their wallets). In TASC 9/2026, Gaetano Di Prima and Fabio Baruffa provide a solution. Their market regime filter consists of three components for detecting trend, volatility, and

We investigate a Donchian breakout rotation strategy between QQQ (NASDAQ-100) and Bitcoin (BTC), with a cash fallback during consolidation, and test it across eight lookback horizons (550 trading days) and two priority variants over a seven-year sample spanning 20192026. The strategy consistently outperforms passive benchmarks on a risk-adjusted basis, achieving Sharpe ratios up to 1.69,

An autonomous research agent replicated nine published US-equity anomalies on clean, survivorship-free data. The question is not only what it found (out-of-sample decay is the rule, and on a faithful build none survive the lone apparent survivor turned out to be a construction error the discipline caught) but whether you can trust an agent to find it, and the checks that decide the answer. Can

This is the eighth post in a series I'm writing on portfolio optimisation. I haven't done one of these for a few posts, so here is the story so far: In the first post I showed that if you are optimising across forecasts from different trading rules and instruments, then you should first fit within; and then across, instruments. As I do anyway. In my second post I ran some experiments

The mega blog post series on portfolio optimisation continues! A couple of posts ago, here, I looked at using the idea of formal testing for structural breaks in parameter estimates. Important parameters like Sharpe Ratio (SR). Because stuff like this happens: This is the pre-cost performance of the momentum4 rule on CORN. The formal test found a structural break in 1989. It's fair to say the

The first three parts of this series built a trend-following program and took it apart: sixty-two futures markets replicated, distilled to one contract per sector, then measured for what trend actually adds to a risk-premia core. All of it sized to a volatility target, indifferent to the account behind it. This closing part asks the question that indifference skips what can a private investor

Recently, we shared a deep-dive on the importance of modeling transaction costs correctly, an exercise that inevitably forces us to confront the non-linear nature of market frictions. The Non-Linear Costs of Trading The Non-Linear Costs of Trading Concretum Research Jun 6 Read full story If you have ever worked on quant-trading desks, or been involved in advisory work, you already know that

Regression-based statistical learning is a convenient and transparent method for combining trading factors into composite signals. Sequential statistical learning considers only the data available at each time point to choose and parameterize the best model and to generate signals without hindsight bias. Yet assessing PnL potential in backtests also requires estimates of transaction costs as

Evidence-based investors have long debated the efficient market hypothesis (EMH), popularized by Gene Fama. In the new era of social media echo chambers, meme stocks, and information overload, it has become fashionable to argue that markets are growing less rational. BlackRocks William Ezratty, Gerald Garvey, Timothy McDade, and Andrew Robinson, authors of the study The Impressive Markets

Mean-Variance Optimization remains the intellectual cornerstone of modern portfolio theory, yet its real-world deployment via plug-in MVO often delivers unstable, over-leveraged portfolios that collapse out-of-sample. The core insight from VertoxQuants analysis is profound: raw plug-in MVO does not merely propagate estimation errorit systematically amplifies it. This error-maximization

TL;DR This project tests whether trend-following, a strategy family with decades of documented success in futures markets, transfers to spot FX. Three approaches (time-series momentum, moving-average crossover, and channel breakout) were backtested across the seven major currency pairs from 2003 to 2025, using 23 rolling walk-forward windows (3-year train, 1-year test), with parameters chosen for

Here's a nice picture from a lovely book written by a top bloke: It shows the cumulative p&l from different speeds of momentum over time (for portfolios containing 102 instruments) over 50 years of data. Notice how the two fastest speeds (2&4) get worse in the second half of the sample. I've called the line #2 here the 'second most famous hockey stick graph in history'.

Meb Faber published A Quantitative Approach to Tactical Asset Allocation in 2007. It became one of the most influential investment research papers of the past two decades. The rules are simple: five asset classes, one trend signal per asset, monthly rebalancing. The original backtest ran from 1972 to 2005 and produced a Sharpe ratio of 0.81, a CAGR of 11.7%, and a maximum drawdown of 9.5%. We

This is a very preliminary result (snapshot June 15, 2026) It rests on 21 earnings events from a single three-week window, and every number below should be read as a first sighting, not a finding. We are publishing it now to describe a pattern and to set a baseline we can check against as the sample grows. Thanks for reading! Subscribe for free to receive new posts and support my work. What this

My work on covariance estimation has recently been published as an Advanced Review in WIREs Computational Statistics, a highly regarded, peer-reviewed journal in the field. It feels remarkably rewarding to see a decade of my curiosity finally bound together in one place. The writing process started about 4.5 years ago on evenings, weekends, and holidays as a side-project. But I actually wrote my

Diversification is a key principle in portfolio construction, yet equal-weight portfolios often fail to deliver true risk diversification. This study shows that capital-based allocation can mask strong concentration in a small number of underlying risk factors. We analyze a simple multi-asset portfolio of ten ETFs spanning equities, bonds, commodities, credit, private equity, and Bitcoin. Despite

Last month, we published two articles touching on a topic that over the past year has been, and continues to be, quite central to our research efforts: short-term trading opportunities in single-name equities. You can access the first two articles by clicking on the banners below. Identifying Stocks to Fade Identifying Stocks to Fade Concretum Research May 23 Read full story When Short Sellers

This is the sixth (!) post in a series I'm writing on portfolio optimisation. A quick reminder of the story so far: In the first post I showed that if you are optimising across forecasts from different trading rules and instruments, that the rules within an instrument cluster naturally together, suggesting you should first fit within; and then across, instruments. Luckily, this is what

Maintaining the status quo and searching for information that confirms its sufficiency are fairly well established human biases. Hence, producing research that satisfies these biases is an easy way to make it popular among many, although it does not contribute anything new scientifically and is often directly anti-scientific. I have seen many examples of this in finance and economics academia. In

This study investigates short-term price reversalstemporary retracements following adverse daily returnsand develops a systematic trading framework to capture this effect across multiple asset classes. Using daily data from six liquid ETFs spanning equities, fixed income, currencies, gold, and commodities over the period 20062025, the strategy applies a long-term trend filter based on a

TLDR: This is a post about something that doesn't work. So don't read if you only care about cherry picked delightful backtests. This is my fifth post in a rapid fire intense series on portfolio optimisation. In my last post I looked at the optimal amount of shrinkage to use with real data, when running a bayesian methodology for mean variance optimisation. I found two things. Firstly,

In a previous blog post of this series, the main univariate Value-at-Risk (VaR) estimation methods were described. Among these, and for scenario-based VaR estimation like historical VaR or Monte Carlo VaR, the most widely used [non-parametric] estimator is the corresponding order statistic of the empirical quantile of the portfolio return distribution, or a linear combination of two subsequent

This week we backtested 2,000,000 trading strategies. Every one of them was pure noise. We generated the returns ourselves with a random number generator, so we know, with complete certainty, that the true Sharpe ratio of every single strategy is exactly zero. The best ones still looked brilliant. Run 1,000 of these noise strategies over ten years of daily data and pick the best, and you get a

Commodities represent a vital but highly volatile asset class, characterized by pronounced cyclicality, lack of yield, and susceptibility to severe macroeconomic drawdowns. While cross-sectional (relative) momentum is a well-documented anomaly, its application in commodities often forces portfolios to hold the least declining assets during broad-based bear markets, resulting in unacceptable

It won 5th place at the Quantpedia Awards 2026. The strategy compounds at 16.3% per year over a 17-year backtest, delivers a Sharpe ratio of 1, and keeps equity market correlation near 15%. This article shows how to build an automated VIX volatility trading strategy using Python and Interactive Brokers. Based on our award-winning research, the strategy seeks to capture the volatility risk premium

This is my fourth post in my summer 2026 mini series on portfolio optimisation. It will very much follow the format of (also with a sports alluding title) blog post number two, so it might be worth rereading that. A reminder if you can't be bothered, I used random data to compare some optimisation methods: monte carlo (random, parameteric) bootstrapping (random, non parametric) double

Financial markets produce mountains of data, spanning simple price movements to limit order book dynamics. A common misconception assumes a larger dataset guarantees superior predictions. Reality proves the opposite. Excessive variables introduce noise, and invite to overfitting. Every input added to a model represents a specific market hypothesis. Including a volatility metric implies price

Its a Tuesday morning, you pull up a name youve been watching, and its 30-day implied vol is printing 48, sitting right near the top of where its traded all year. Vol is mean-reverting, everyone knows that, so you do the obvious thing. You sell the straddle, perhaps even an iron condor. Two days later, the print is 61, the short is deep underwater, and the reversion you were promised is

This is the fourth article in the small market effect research series. The first looked at the holiday effect on SPY. The second looked at the turn of the month effect, also on SPY. The third looked at the holiday effect on gas and energy assets All three sets of research resulted in tradable systems that I have now implemented as overlays on my personal systematic trading portfolio. If you missed

Today, AI and machine learning techniques are evolving at a rapid pace, making the development of trading systems increasingly accessible. Generating signals, building models, and testing ideas is easier than ever. As a result, the challenge is no longer simply developing a trading strategy, but determining whether it is genuinely robust or merely the product of overfitting and data mining. In

Asset pricing research often focuses on risk, valuation, and macroeconomic forces. But this paper highlights another surprisingly powerful driver of returns: the timing of dividend payments. Across 44 international equity markets, the authors uncover a large and persistent dividend premium. Dividend-paying stocks outperform non-payers by a meaningful margin, even after controlling for

The piece we present today stems from some internal exchange within the Concretum team ahead of the highly anticipated SpaceX IPO, an offering that has dominated headlines for a string of firsts in recent market history, from its record valuation (~$1.75 trillion) to the one that interests us most: the prospect of a fast-track inclusion into the Nasdaq 100. A few numbers can frame why this

Academics have identified hundreds of factors that supposedly explain stock returns. New research shows most of them are telling the same story in different words and only a few truly distinct forces actually drive the market. The problem: too many factors, too little meaning. Over the past few decades, academic researchers have proposed more than 400 factors characteristics or variables,

Yes, this is my first article without a podcast attached. The first of many, I hope. In time, Id also like to produce more YouTube content that isnt strictly interview-based. Ill admit it: Ive been biting off more than I can chew. Courses, podcasts, trading, farming (yep!). As you know, Im working on courses with Market Wizards, but Im also building my own courses, which I plan

Every retail trader eventually asks the same question: do candlestick patterns predict where price goes next? I decided to measure instead of believe. I built a small pipeline on SPY that discretizes each bar into one of twelve candle states, tests whether the next bars direction depends on that state, and then asks what survives once you account for sample size, transaction costs, and the

Monte-Carlo Simulation is the most straightforward way to price an option, and if you dont care about speed, its a solid choice. The moment you care about speed, like when quoting live, or when calibrating a pricing model where you need to reprice options thousands of times, Monte Carlo quickly becomes unusable. This article presents three Fourier-based methods that solve this problem,

This is my third post in a series about optimisation and fitting. In my previous post I used random data to calibrate and evaluate many portfolio optimisation techniques. It's worth quoting in full from that post: Random data is not real data: Well duh. But why is this important? Because random data is drawn from a fixed and well behaved distribution. This means the optimiser only has to

A Tale of Two Prices (the core idea of stat arb) Moneyball (finding undervalued pairs using unconventional metrics) The Winter of our Pairs Trading Discontent (problems, limitations, frustrations) The Metamorphosis (from pairs to portfolio) When is a Mispricing Not a Mispricing? (something looks mispriced why?) At its simplest, stat arb is a bet on convergence: two related stocks drift

Combine a three-asset risk-premia portfolio with a trend-following program and the Sharpe ratio jumps from 1.1 to nearly 1.5. It looks like free diversification. But a trend program is long equities, bonds and metals and a risk-premia core is equities, bonds and metals. So before we accept the free lunch, we should ask what part of it we are paying for twice, and what part is genuinely new.

As I said in my last post I'm currently in the process of a mega-sized research project on fitting. In the first post I examined the correct way to cluster combinations of trading rules and instruments. This next post is rather meatier, and is about evaluating and calibrating some portfolio optimisation techniques. We might call this 'meta optimisation', since we want to find the

Two different sources each tell you something about where a stock is going. The first is insider trading filings: SEC Form 4, which executives, directors, and large shareholders are required to submit within two business days of buying or selling shares in their own company. The second is the news cycle: analyst reports, earnings coverage, breaking stories. Each source has its own academic

How much do we really know about corporate bond returns before the modern data era? Until recently, the answer was: not enough. Most empirical work in corporate bond pricing has relied on relatively short samples, especially the post-2002 TRACE period, leaving open the question of whether observed risk premia are robust over longer horizons. Ghaderi, Plante, Roussanov, and Seo (2026) Ghaderi,