Four Futures Exchanges Crack Down on Algo Trading; Saba Takes 29% Stake

Today's Roundup

01 Rule Revisions Target Programmatic Trading

On October 12, the Shanghai Futures Exchange (SHFE), Zhengzhou Commodity Exchange (ZCE), Dalian Commodity Exchange (DCE), and Guangzhou Futures Exchange (GFEX) simultaneously released draft revisions to their violation handling rules for public comment, with a deadline of October 19.

The revisions implement the Futures and Derivatives Law and programmatic trading regulations by explicitly incorporating programmatic trading oversight into violation rules. Failure to report as required,违规 high-frequency trading (HFT), frequent order placement and cancellation, self-matching, and evasion of surveillance are all classified as violations. Penalties range from orders to correct and warnings, to forced liquidation, suspension of new positions, and market bans.²

02 Re-inventing the "Tool to Smooth Economic Fluctuations" 2026 Nobel Prize in Economics Announced

The 2026 Nobel Prize in Economics has been awarded to American macroeconomists Robert Hall and Michael Woodford.

The two scholars were recognized for "laying the microeconomic foundations for macroeconomic stability." Their work focuses on individual expectations and corporate pricing behavior, helping global central banks more precisely smooth economic fluctuations through policy.

Following the announcement, Woodford described the award as a delightful surprise, while Hall humorously thanked the California and U.S. federal taxpayers who funded his education, hoping they felt the investment was worthwhile.

Woodford is currently a professor at Columbia University, and Hall is a professor at Stanford University and a senior fellow at the Hoover Institution. Both hold Ph.D.s in economics from MIT.

Big Tech & Market Moves

03 270 Billion GBP Trust Industry Under Siege by Hedge Funds

Baillie Gifford, a Scottish asset management giant, is embroiled in a battle for control. U.S. activist hedge fund Saba Capital has acquired 29% of the voting rights in Baillie's American Growth Trust and plans to dismiss existing directors at the October shareholder meeting, attempting to seize control. This conflict has rippled through the entire 270 billion GBP UK trust industry.

The backgrounds of the opposing sides are prominent. Baillie Gifford is a century-old British top-tier asset manager, famous for long-term investments in high-growth tech giants like SpaceX and Tesla. Saba Capital, led by Wall Street "activist arbitrage" boss Boaz Weinstein, specializes in exploiting trust market discounts, using aggressive tactics to secure short-term windfalls.

The clash between an old-school gentleman and a Wall Street shark—a showdown between "long-term investing" and "arbitrage dismantling"—is attracting global financial attention. ¹

04 Goldman Sachs Q3 Net Revenue Hits 18.68 Billion USD, Second-Highest Quarterly Record

Goldman Sachs reported net income of 5.29 billion USD and diluted earnings per share of 17.01 USD for the period ending September 30, with an annualized average return on common shareholders' equity of 18.7%.

Chairman and CEO David Solomon stated that in a complex operating environment, clients' continued demand for advice, execution, and financing support drove year-to-date net revenue to an all-time high.

The company will hold a conference call at 9:30 AM ET to discuss financial results and outlook. ⁴

Quant Research

05 Can Empirical Finance Be Reproduced? AI Provides the Answer

Can published empirical financial research be reproduced using original samples? Does it hold up with new data?

Dmitriy Muravyev’s 2026 working paper uses an automated LLM reproduction pipeline (with each result independently audited) to reconstruct the main results of 1,328 papers published in the Journal of Finance, Journal of Financial Economics, and Review of Financial Studies between 2000 and 2020, of which 1,005 had out-of-sample test data.

The results show two distinct layers: in-sample, approximately 75% maintained their original sign and statistical significance, with the median reproduced coefficient being 0.93 of the original value. Out-of-sample, the success rate dropped to 42% (48% after adjusting for test precision), with the median coefficient ratio falling to 0.45.

Yearly re-estimation shows coefficients dropping sharply by approximately 0.28 right at the end of the original sample, remaining relatively stable before and after—this pattern aligns more with data snooping/sample selection than investor learning or gradual market changes; papers with stronger original t-values showed greater persistence (survival rate ~31% for the weakest quartile vs. ~72% for the strongest); publication selection bias explains only about one-fifth of the total decay.

This is a 2026 working paper, interpreted and published by the quant blog Alpha Architect (founded by Wes Gray, a well-known quant author). ¹

06 Loss Function vs. Model Selection: The Role of Forecast Levels in Crypto Volatility Prediction

How do the overall level and daily changes of volatility forecasts jointly influence downstream decisions, and does the impact of loss function selection in original comparisons stem from forecast level differences or the superiority of the models themselves?

This study validates a forecast level alignment method based on validation sets by comparing seven loss functions and five models on major cryptocurrency samples.

Results show that before alignment, score differences caused by loss functions were significant. However, after level adjustment, model selection became the larger source of forecast variation, and cross-loss-function Value-at-Risk (VaR) breach rate differences narrowed significantly. This indicates that when explicitly considering forecast levels and downstream risk, the original influence of loss functions is overestimated, and model architecture is the key.

This work has been accepted by the ADMA 2026 conference and is currently available as an arXiv preprint, authored by Andrzej Tokajuk and Jarosław A. Chudziak. ⁶

Quant Career

07 Never Underestimate the Power of Human Connection

On October 13, The TRADE interviewed Kevin Flood, Head of Trading and Execution Analytics at Royal London Asset Management (RLAM).

Appointed in 2021, he leads a multi-asset team covering fixed income, equities, FX, ETFs, and OTC derivatives, guiding the team to win the Leaders in Trading Awards 2025 Best Trading Desk.

Flood started in trading support at Deutsche Bank, then served as a credit trader in the loan portfolio management group, experiencing the 2008 financial crisis and the European debt crisis firsthand. In 2016, he moved to the buy-side, joined RLAM in 2018 to build a centralized fixed-income trading desk, and was named a Rising Star that same year.

He positions trading as a strategic partner to investing: "On the buy-side, you are a true strategic partner to portfolio managers, helping shape trading ideas, build workflows, and protect alpha." Facing the wave of technology, he insists that the essence of the industry is a people business: "Never underestimate the power of human connection; pick up the phone and proactively build networks." In an era dominated by screens and algorithms, direct dialogue and genuine trust are significant differentiating advantages. ¹