This is the second installment in a multipart series on selecting Tactical Asset Allocation (TAA) strategies based on recent performance. Read Part 1. We advocate combining multiple TAA strategies together into Model Portfolios to limit the risk of any single strategy underperforming. In our previous study we selected strategies for our portfolio with the highest recent return. In this

From the trading desk to the portfolio committee, investors face a familiar question: how should alternative stores of value fit into a diversified portfolio? This research explores that question through a systematic dual-momentum framework comparing Bitcoin and physical gold in a rules-based tactical allocation model. Rather than debating ideology, we focus on practical portfolio construction and

A strategy you have been trading for years has just had a terrible month. Looking at the market environment, you think these trades should not have been taken. You make some small changes to your strategy and now the backtest shows that the terrible month is OK and the overall strategy statistics improve. Should you keep that change in your strategy? For the longest time, I would keep that

For much of last week I was at the National Association of Active Investment Managers (NAAIM) Uncommon Knowledge conference. NAAIM is a terrific organization that I have become more involved with over the years. NAAIM has published its NAAIM Exposure Index since 2006. I did some research a few years ago on the index to determine whether the numbers might be valuable as part of a model. I

Most sentiment research treats the question as one thing. Take a corpus of news articles. Classify each as positive, negative, or neutral. Aggregate to the daily level. Correlate with next-day returns. Report a coefficient. Argue about which classifier is best. This is tidy. It is also wrong, in a specific way that took us a while to see clearly. The articles being classified are not the same kind

In an era of increasingly fragmented market microstructure, the emergence of cross-asset connectedness between Crypto and public equity markets presents a critical challenge for modern portfolio construction. This blog post examines the recent working paper by Harin de Silva, The Attention Factor: The Speculative Risk You May Already Own, which identifies a previously underappreciated

A self-contained synthetic benchmark of a small mask-conditional CNN against calendar-projected linear interpolation and a per-slice SVI fit. A volatility surface marker is rarely a clean rectangle of quotes. Strikes go unobserved during illiquid hours, wings get crossed and then erased, broker stripes drop out across an entire maturity, and weeklies arrive at the desk with random missingness on

This post presents an accessible introduction to the Exponential GARCH (EGARCH) modela widely used tool in financial econometrics for modeling time-varying volatility in asset returns. Unlike standard GARCH models, EGARCH captures both volatility clustering and the leverage effect, whereby negative shocks tend to increase future volatility more than positive shocks of equal magnitude. The post

May got off to a positive start. But a strong start to May has typically been followed by a dip in the next few days. This can be seen in the study below, which was featured in this weekends subscriber letter. Of the 25 instances that rose on the first day in May since 1987, 17 of them closed lower 4 days later. Below is an equity curve that shows how it has played out over time. Ill note

The risk of overfitting is serious and can lead to significant losses. It has been discussed in previous posts. In this edition, we revisit the topic, given its continued relevance to quantitative strategy development. Formal Study of Overfitting in Trading System Design A serious problem when designing a trading system is the overfitting phenomenon, wherein the system is excessively tuned to

Why I'm publishing this I wanted to build a trading bot like a lot of people did once Claude integrated with TradingView. Took the leap, my strategies kept failing, and the backtests kept being way too optimistic compared to what happened when I actually ran them. Started digging into why. This post is what 10 days of running 22 popular strategies on real Binance fees with real L2 spread

The previous article extended the inverse-volatility allocation across SPY, TLT, and GLD back to 1968 using a synthetic price construction. Over fifty-eight years the strategy delivered a CAGR of 7.1%, volatility of 7.5%, a Sharpe of 0.97, and a maximum drawdown of 22%. The volatility-targeting overlay, justified by the persistence of volatility across the same window, kept realised vol close to

Several years ago, I had explored accuracy and performance of different ways to imply the Black-Scholes volatility. Jherek Healy proposed some improvements over my naive algorithm on his blog. Recently, a Linkedin post mentioned a new paper from Wolfgang Schadner which presents an almost explicit formula for the implied volatility. Almost because it actually relies on some implementation of the

Most trend-following research focuses on signal construction: how to detect trends better, faster, or earlier. The paper asks a different question, and arguably a more important one for investors: once a market regime has been identified, what is the optimal portfolio exposure in that regime? That is the central novelty of the paper which is available here. Traditional time-series momentum

The shape of yield curves in developed swap markets reflects the state of growth, inflation, and credit supply. This is primarily because central banks adjust short-term policy rates in response to evolving economic conditions, while their credibility helps anchor longer-term forward rates. In monetary policy regimes committed to price stability, and when short rates are above the zero lower

This is the first of a multipart series examining the selection of Tactical Asset Allocation (TAA) strategies based on recent performance. We are proponents of combining multiple TAA strategies together into what we call Model Portfolios to limit the risk of any single strategy going of the rails. In this study we ask, what if, each month, we selected strategies for our portfolio that had

Why the path to making money in trading runs through work youd better find interesting Data mining and vibe quanting are essentially the same thing. Both fundamentally and philosophically. Fundamentally, data mining says: Ill try enough rules until something sticks. Vibe quanting says: Ill get AI to try enough rules until something sticks. Same thing, different packaging.

The growing dominance of passive investing has fundamentally altered the dynamics of equity markets. A substantial share of trading volume is now driven by index-tracking strategies, which mechanically allocate capital based on index membership rather than company-specific fundamentals. This raises an important question: can predictable flows associated with index rebalancing be systematically

A research team in Zurich has shown that the financial sentiment classifiers running inside many automated trading and risk pipelines can be flipped quietly, undetectably, and for pennies by anyone with access to GPT-4o. Thanks for reading! Subscribe for free to receive new posts and support my work. The paper has been out a few weeks. It deserves more attention than its getting. The

Cohen, Malloy and Nguyen's Lazy Prices paper found that small year-on-year changes in 10-K filings predict large negative returns. Here is what the paper actually says, and how Snowflake Cortex AI and Semantic Views collapse the original eight-year engineering pipeline into an afternoon's work. On 23 February 2010, Baxter International filed its annual report with the SEC. The stock did

In Beyond 60/40 I argued that the classic balanced portfolio rests on an assumption that stocks and bonds will hedge each other and that the assumption fails when the macroeconomic regime changes. The argument was built on the post-2005 ETF era, the only window where clean real-price data exists for the three assets needed to test it. Twenty years made the case. Fifty-eight years sharpens

New research challenges a long-standing rule in momentum investingand reveals surprising insights about when to use it For decades, investors using momentum strategies have followed a simple rule: ignore last months returns. This skip-month convention has been standard practice since the 1990s, designed to avoid short-term reversal effects where stocks that jump up one month tend to

In the previous blogs, we examined supervised learning algorithms like linear regression in detail. In this blog, we look at what unsupervised learning is and how it differs from supervised learning. Then, we move on to discuss some use cases of unsupervised learning in investment and trading. We explore two unsupervised techniques in particular- k-means clustering and PCA with examples in Python.

Who Wins and Who Loses In Prediction Markets? Evidence from Polymarket Pat Akey (ESSEC Business School), et al. April 2026 We study pricing efficiency in decentralized prediction markets by comparing marketimplied probabilities from Polymarket with benchmarks derived from option-implied riskneutral distributions extracted from the derivatives market. We study Bitcoin and Ethereum prediction bets

One of your first thoughts when looking at a strangers backtest is probably that its overfit, or that there is some look-ahead somewhere. When you go a step further, you are probably constantly worried about overfitting your own backtests too! In this article, we will introduce a framework that allows you to identify both! Its a two-stage approach introduced in D. Nikolopoulos (2026). We

Forward records for strategies people can actually inspect. TradeLock helps managers and signal providers turn live strategy intent into a forward-tracked public record that is harder to fake than a backtest, PDF, or spreadsheet.

The decomposition of risks and returns into overnight and intraday components is an emerging area of research. In this post, we examine how these components differ in terms of volatility clustering and the variance risk premium, and what this implies for forecasting, risk management, and strategy design. Breaking Down the Volatility Risk Premium: Overnight vs. Intraday Returns The decomposition of

The fact that mean-variance optimizers are highly sensitive to changes in expected returns [] is well known in investment practice1, with a couple of practical solutions already described in this blog, for example using near efficient portfolios or subset resampling-based efficient portfolios. In this blog post, I will introduce another approach originally described in Sharpe2 and known as

What if a meaningful part of a usual trading strategys performance has nothing to do with your signalbut simply when you rebalance? A recent paper written by Carlo Zarattini & Alberto Pagani highlights a largely underestimated risk in systematic investing: rebalance timing luck (RTL). For practitioners running rotation or factor strategies, this is not noiseits a structural source

A long-bond ETF sits in almost every balanced portfolio. Ours included TLT is one of the three core holdings in the risk-parity base of our portfolio architecture. And yet when TLT lost 48% between 2020 and 2024, most holders experienced it as a shock. It should not have been. The mechanics were entirely predictable from the yield level at which investors bought in, and the historical

This study examines the profitability of mean-reversion trading strategies applied to binary outcome contracts on Polymarket, the worlds largest decentralized prediction market platform. We analyze three distinct contracts representing varying risk profiles: a quasi-risk-free instrument (No to Will Jesus Christ return in 2025?) and two high-yield speculative contracts (No to Will China

Simple inflation-based trading factors have proven their predictive power in global financial markets over the past decades. Excess inflation ratios measure the average difference between CPI growth and a countrys effective inflation target (relative to that target). Inflation pressure ratios combine excess inflation ratios with recent CPI growth surprises. Both can be calculated for headline

By the Fourier theorem, any price curve is a mix of many long-term and short-term cycles. Once in a while a dominant market cycle emerges and can be exploited for trading. In his TASC 5/2026 article, John Ehlers described an algorithm for detecting such dominant cycles, using them to tune a bandpass filter, and creating a profitable trading system. Heres how to do it. Ehlers Easylanguage

People often criticise how ML models are just black boxes that take in some features and spit out a prediction. While some models (like linear regression) are naturally a lot more interpretable than others (like neural networks), its wrong that you cant figure out why a model made a certain prediction and how the different features affect the prediction. In this article, we will look at some

It starts with a basic problem every quantitative researcher faces: you have a universe of stocks, a universe of news, and a question. Which of todays headlines actually matter for tomorrows price? Step One: Finding the News Before you can classify sentiment, you have to collect articles. The naive approach search for a company name and take everything produces a lot of noise. An

RayforceDB is a recently open sourced time series database that offers blazingly fast performance. It is built with inspiration from kdb+, which is also known for its fast performance and minimal application size. RayforceDB offers similar benefits, being written in pure C and having a binary size of less than 1MB. Another benefit of RayforceDB is that it offers Python bindings with minimal

This is how I started last years update: "Mad out there isn't it? Tarrifs on/off/on/partially off/on… USD/SP500/Gold/US10/Bitcoin all yoyoing like crazy." Well the orange peril is still at it, and as I write this the global supply of oil has been severly curtailed for several weeks now; with a certain amount of reaction in oil futures (which some of it perhaps supressed since

The most common question we hear from advisors is whats the optimal stack? So we ran the optimizer bootstrapping 10,000 simulated 25-year histories across five asset classes to find the portfolio that would have maximized return at 60/40 volatility. The answer is mathematically elegant and practically unusable. In this piece, we walk through why the optimal portfolio would have been

April 15th is tax day. Tax day has historically been a good day for the market. A reason tax day may be bullish is that it is the last day that people can make IRA contributions to count for the previous tax year. This can create a last-minute rush and you will often have an inflow of funds heading into the market right around and on April 15th (or whenever tax day ends up falling, since it is

Stock momentum has long been a workhorse idea. Buy recent winners. Sell recent losers. Critics argue those profits mostly come from riding factor trends like value, size, or industry tilts. This paper pushes back. It shows there is a durable, stock-specific momentum component tied to how prices react to firm news around earnings dates. The result is a cleaner, lower-risk way to capture momentum