Artificial Intelligence

AI models don’t have to be black boxes. In the final article of our Expl(AI)ning series, discover how SHAP values reveal the contributions behind model predictions, enhancing transparency and trust in AI-powered investing. See how we use these insights within our AI Powered process.

Leaders in the video game industry, such as Tencent Holdings, NetEase, and Capcom, offer a distinct blend of predictability and growth through a combination of "Evergreen" titles and new game pipelines. These firms have generated recurring revenues and compounded earnings at approximately 20% annually over the past decade.

Can the top 2% make all the difference? The Quantitative Investments team tested an AI-powered stock picker across 20 years of data. By analysing hit rates alongside market concentration and dispersion, they uncovered fresh insights into portfolio management. Discover how AI reshapes investing in our latest article.

There is an emerging sense of unease in the markets around the scale and productivity of corporate investments in AI. As fixed income investors, not equity or tech managers, we will not aim to assess the longevity or possible applications of these nascent technologies, and nor should we opine on when or by how much equity markets might go up or down.

Backtests can mislead—or unlock real edge. In Part II of our mini-series, we show how to avoid common pitfalls, cut through bias, and turn backtests into a reliable tool for decision-making. Practical, actionable, and built for real-world investing.

AI-driven momentum trades and non-US value dominated markets in 2025. However, the sustainability of these trends remains uncertain, particularly given unproven ROI of AI investments. Amid this backdrop, quality stocks emerge as a prudent choice, offering both resilience and steady growth potential.

Backtesting can reveal powerful insights—or create dangerous illusions. In Part I of our mini-series, we show how biases and shortcuts can inflate results, using a simple momentum strategy as an example. Learn why careful design matters before moving from simulation to real-world investing.

How can machine learning ensembles be built in practice? From Bagging and Boosting to Stacking, this article explores different ensemble strategies and the role of model diversity. We also examine how ensembles are being applied in financial forecasting, drawing on both academic research and industry experience. Discover how combining models tends to deliver stronger, more resilient predictions in finance.

Oracle priced an $18bn six-tranche (5yr/7yr/10yr/20yr/30yr/40yr) bond deal which was increased from an initial $15bn on the back of exceptionally strong demand. It is the latest sign that the AI investment boom, long the focus of equity markets, is now spilling into credit.

Why rely on one model when many do better? By combining multiple forecasts, ensembles tend to deliver more accurate and resilient results. This first article in our two-part series explores the theory behind model ensembles, linking classic forecasting research with today’s financial machine learning. Discover why combining models consistently outperforms relying on a single predictor.