Preconference Day: Summits & WorkshopsMonday, 16 November 2026 - GMT (Greenwich Mean Time, GMTZ)

A big picture look across asset classes. How are allocators and managers adapting to market shifts, evolving risk dynamics, and the rise of quant integration in multi-asset portfolios?

Every manager now claims an AI edge and very few can show one. This session gives the room the buyer's side of that conversation. It covers what the research shows once returns are split into market, style and selection, why so much of the claimed edge turns out to be beta in disguise or a tool everyone is renting, what the world's largest sovereign and pension funds have decided to build in-house instead of buying, and the four questions those allocators now put to any quant manager claiming an AI edge. Managers in the room leave knowing the test they will be judged against in 2027.

How are investors implementing AI successfully? A deep dive into model development, data infrastructure, back testing frameworks, and real-world performance attribution.

There are different vantage points for studying market microstructure. In order of increasing level of detail, these are: (a) public market data, (b) analytics datasets based on proprietary exchange data, (c) participant-internal data, and finally (d) exchanges’ internal data. This talk will present a range of empirical findings of market microstructure on Xetra and Eurex using (a), (b), and sufficiently aggregated (d). Topics range from the status quo of ULL trading, the behaviors of different categories of participants, and insights into dark order books (auctions, midpoint, retail).

How are quants adapting to tail risk in a year of sharp market moves? Lessons in stress testing and model resilience.

A practical exploration of how firms are constructing portfolios across asset classes while optimising for risk, cost, and return.

At mid-frequency and with meaningful capacity, a systematic crypto strategy is effectively trading a single asset, BTC or ETH, with only a few years of usable history. There is no cross-section to average away noise, so overfitting becomes the central problem. This talk walks through a strategy step by step, from idea to live trading, using the single-asset setting as the hardest possible test of whether an edge is real. We look at where an idea and its data come from; how a deliberately simple signal is built and sized, including the practical choices that quietly move PnL; how to tell whether a Sharpe is genuine, through parameter sensitivity, behaviour across coins, and a Monte-Carlo test on a single price path; and what changes in production. The methods are shown on real signals and carry over to any market with a short history and few instruments, in or out of crypto.

High-frequency trading has profoundly reshaped the dynamics of equity markets. This talk presents two proprietary strategies developed and tested on the constituents of the FTSE MIB, the main benchmark of the Italian equity market.

The first is a market making strategy in which quotes include a premium that accounts for expected market impact. A hedging offset controls exposure to systematic risk factors, which are extracted through principal component analysis.

The second focuses on the closing auction and is designed to capture the opportunities specific to this phase. The optimal quantities to submit are found by solving a convex quadratic optimisation problem that balances expected profitability against the variance of the residual portfolio. The problem is solved efficiently with an interior point method based on Newton iterations.

Both strategies are evaluated in a proprietary event-driven backtester that reproduces order-by-order market microstructure. The results show consistent profitability, robustness across stocks and market conditions, and effective control of inventory imbalances.