Research
Every research paper we have written up, newest paper first. New arrivals are screened automatically for topical fit with systematic trading; the ones that clear it are read, written up, and — where the idea is testable on our data — backtested. Particular focus on alpha hypotheses that can be formalised and tested. The Radar also covers portfolio construction, market risk and execution where the research is directly relevant to systematic investment processes. Follow new entries by RSS.
16,915 papers screened · 588 on the radar · 33 shown
Financial news can influence investment decisions long after publication. Timely news may reach an investor before prices fully adjust, whereas older news may already be reflected in prices.
PAPER REPORTS · LLM (GPT-4.1 mini) Market Data + Filtered News, 2023–2025 average of annual tests, net of 10 bp per dollar traded: -… · LLM Market Data + News + News Age, 2023–2025, net of 10 bp: 16.73% / 13.26% / 17.17% annualized at 0 / 1 / 4 week…
Optimal capital allocation in financial markets has long been a challenging task due to the dynamic nature of market conditions, risk factors, and the complexity of financial data.
PAPER REPORTS · Annualized Sharpe ratio 1.91, computed from daily portfolio returns with 252 trading days and a 2% annual risk-free… · Cumulative return 1.12. The test period is not stated beyond the 250 trading days in Fig 7, and it is unclear whether…
Estimating extreme risk in reinforcement learning becomes increasingly difficult as the confidence level increases, since only a small number of observations contribute to empirical estimates of Conditional Value-at-Risk (CVaR).
PAPER REPORTS · Simulated Merton jump-diffusion, K=140 call, 63 daily steps; transaction costs not mentioned. · Same setting, EVT-trained policy hedging error on 3,000 paths conditional on unhedged payoff > 0.999 quantile: mean…
Automated markets increasingly host autonomous, learning trading agents and are powered by formulaic automated market makers (AMMs).
PAPER REPORTS · Rolling 7-day annualized Sharpe ratio of net strategy returns ranges roughly between -20 and +20 for every agent…
Portfolio optimization is constrained by linear assumptions and limited integration of multi-modal information.
PAPER REPORTS · SBCA 2-asset (CAT, GILD), test 2021–2022, net of 0.25% one-way commission: final portfolio value 1.4392, annual return… · SBCA 4-asset (CAT, GILD, GS, KO), test 2021–2022, net of 0.25% one-way commission: final portfolio value 1.3564, annual…
OUR BACKTEST · Sharpe 2.62 · Return +54.1% · Max DD -10.1%
We develop a data-driven approach to dynamic portfolio selection that combines vine-copula simulation with reinforcement learning.
PAPER REPORTS · NN-vine-RL ensemble, 22 complete monthly holding periods within 31 Jul 2024–30 Jun 2026 decisions: CAGR 34.13%, net of… · Sharpe ratio 2.309 (zero annual risk-free rate), same sample and costs.
Market making under realistic order-flow dynamics is challenging due to the clustered, self-exciting nature of order arrivals, which are naturally modeled by non-Markovian Hawkes processes.
PAPER REPORTS · HSN mean cumulative episodic reward, simulated (no market data), multi-asset cross-excitation cs=0.7: 0.0548±.0007 over… · HSN, simulated power-law kernel ε=0.15: 0.0216±.0004 mean cumulative reward, +32.5% vs the best baseline (p<0.001).
Formulaic alpha mining seeks to discover executable expressions that capture predictive market signals in a vast, noisy search space.
PAPER REPORTS · AlphaRLM-8B CSI 300: annualized excess return 12.14% (0.1214±0.0011), MDD -6.15%, IC 0.0256, RankIC 0.0424, ICIR… · AlphaRLM-8B CSI 1000 (CSI 300-trained factor pool): annualized excess return 17.45%, MDD -11.98%, IC 0.0302, RankIC…
OUR BACKTEST · Sharpe 0.83 · Return +105.8% · Max DD -50.8%
PAPER REPORTS · AIM_PU_CVaR (real data, daily, 2020-01-01 to present, T=1471 days): Sharpe ratio 0.7718; annualization not stated; no… · AIM_PU_CVaR: total reward 1.5048, mean daily return 0.0010, total CVaR-like cost 0.0194, 23 crisis days (>5% daily…
OUR BACKTEST · Sharpe 0.98 · Return +61.8% · Max DD -22.0%
This paper studies exit timing for SPX short-straddle portfolios using a two-stage supervised and reinforcement-learning framework.
PAPER REPORTS · Supervised MLP exit policy, OOS 2024-2026 (T≈2.42y), stress-budget compounding sizing, bid entry/ask exit plus… · AC+MLP, 7 seeds, OOS 2024-2026, same costs: cumulative return 37.54% ± 0.92%, ARC 14.05% ± 0.32%, ASD 7.97% ± 0.14%,…
As option markets grow and AI advances, agentic systems for option trading are gaining increasing attention.
PAPER REPORTS · Total return 18.32% over 2025-03-03 to 2025-08-29 (six-month out-of-sample), 10 underlyings; fills at bid-ask midpoint… · Annualized Sharpe ratio 1.60 (same period and costs)
Reinforcement learning (RL) is increasingly used for financial optimal-control problems when complex dynamics make analytical strategies difficult to obtain.
PAPER REPORTS · Proposed method (frozen analytical policy + residual PPO), execution cost halved from 0.0012 to 0.0006, σ_ξ=100, full… · Same method at a_true=0.0012: change 0.00019 [-0.00041,0.00079]; at a_true=0.0018: change 0.00019 [-0.00034,0.00071]…
We propose an interactive robo-advising framework that learns personalized risk preferences from scores provided by clients.
Deep reinforcement-learning policies for order execution can vary substantially across training seeds, so apparent architectural gains may reflect favourable training realisations rather than reproducible properties of the architecture.
PAPER REPORTS · Mean implementation shortfall, out-of-sample Jan 16–20 2023, frictionless replay with no commissions or market impact,… · Per-run CVaR-95 under the same setup: DDQL 110.8, MoE K=8 122.2, DDQL-cap8 104.1 bps.
Reinforcement learning for trading often struggles to balance upside participation with drawdown control.
PAPER REPORTS · SPY, test 2020–2022, seed 42, 0.1% proportional transaction cost: total return 27.62%, annualized return 8.48%, Sharpe… · SPY, test 2020–2022, mean ± std over seeds {7,11,19,23,42}, 0.1% cost: total return 0.2725 ± 0.0109, Sharpe 0.6219 ±…
OUR BACKTEST · Sharpe 0.89 · Return +42.3% · Max DD -15.6%
In this paper, we extend earlier findings of supra-competitive outcomes in optimal-execution games by identifying a learned punitive mechanism that deters deviations and provides behavioural evidence of collusion.
We investigate the systemic macroscopic dynamics emerging from Limit Order Books (LOBs) populated exclusively by autonomous reinforcement-learning agentic traders.
We present SAiFE_gym, a Python module that provides a collection of simulation environments for studying trading problems in Constant Product Markets (CPMs) with Concentrated Liquidity (CL).
PAPER REPORTS · Nominal PPO LP agent, mean terminal PnL (gas included, 2 units X per rebalance) = +55.43 at training domain… · Domain-randomized PPO LP agent, mean terminal PnL (gas included) = +55.38 at (sigma=0.03, alpha1=450); grid values span…
Reinforcement learning trading systems published in the academic literature overwhelmingly rely on price-aggregate state representations (OHLCV bars) or limit-order-book depth features, leaving microstructure pattern theories from the practitioner literature,…
PAPER REPORTS · TSLA: +163.6% ROI, -27.5% max drawdown, 27,019 trades, held-out final 12 months of Jan 2020-Feb 2026 data, zero… · NVDA: +116.5% ROI, -47.8% max drawdown, 12,892 trades, same held-out 12-month partition, zero leverage, $0.0002/share…
In recent years, when traditional portfolio optimization (PO) models deal with highly volatile and irrational market environments, they often face problems of deviation in return prediction and lag in risk response.
PAPER REPORTS · Annualized return 0.196 (out-of-sample test set within 2018-2023, daily rebalancing, no trading costs) · Annualized volatility 0.187; Sharpe ratio 1.048 with risk-free rate assumed zero (no trading costs)
Classical market-making strategies based on stochastic control, such as the Avellaneda-Stoikov and the Guéant-Lehalle-Fernandez-Tapia (GLFT) extension, provide closed-form quoting rules, but rest on assumptions that break down at realistic microstructure…
PAPER REPORTS · Stationary flow: DQN risk-return frontier lies above the GLFT frontier throughout the observed risk range (N_sim = 1000… · Regime-switching (tau_r = 60 MO events, omega = 0.30, N_sim = 100 paired episodes of 100,000 LOB events): Algorithm B…
Formulaic alpha discovery is a pool-dependent symbolic search problem in which informative feedback is observed primarily when a complete expression is evaluated.
PAPER REPORTS · No Sharpe, annualized return, alpha, t-stat, hit rate or drawdown is reported anywhere. · ICIR 0.379±0.159 (CSI500), 0.307±0.058 (CSI300), 0.339±0.027 (CSI800), 0.505±0.003 (CSI1000), test period 2022-2024.
We study deep hedging in the context of dynamics risk measures, where sequential decisions are time-consistent.
PAPER REPORTS · Terminal hedging loss CVaR95% at 1-year maturity, 10,000-path test set with initial state perturbation, 0.1%… · Mean terminal P&L, same setup (alpha=95%): log -1.1143 (std 1.7095), static +0.2161 (std 2.6423)
Modern portfolio management increasingly demands a balance between traditional risk-adjusted returns and strict Environmental, Social, and Governance (ESG) mandates.
PAPER REPORTS · Europe persona, out-of-sample 2015–mid-2016: Sharpe drop −0.144 (±0.335) vs Λ=[1,0,0,0] baseline; ESG score gains… · Asia persona, out-of-sample 2015–mid-2016: Sharpe drop −0.045 (±0.068); ESG gains +13.12%/+28.75%/+46.60%.
Abstract A key puzzle in finance is why algorithmic traders with advanced neural models sometimes fail to beat simple traditional strategies, while in other cases they clearly outperform them.
PAPER REPORTS · Cluster 0 (most efficient), test 2025-2026, net of 0.1% one-way costs: DDPG annualised return 34.72%, cumulative… · Cluster 0 best classical: HRP Sharpe 1.591, Calmar 2.3562, max drawdown -13.07%, annualised return 30.80%
OUR BACKTEST · Sharpe 0.78 · Return +25.3% · Max DD -11.7%
Automated market makers (AMMs) are a cornerstone of decentralised finance (DeFi). Constant product markets with concentrated liquidity, such as UniswapV3, are now a well-established design.
PAPER REPORTS · PPO_narrow, risk-neutral, sigma=0.01, g=2: mean PnL 49.91 +/- 0.38 USDC, 5% CVaR 9.15 +/- 0.65, over 1000 evaluation… · PPO, risk-neutral, sigma=0.01, g=2: mean PnL 42.31 +/- 0.97 USDC, 5% CVaR 6.08 +/- 0.97
We introduce a reinforcement learning framework for market making in a limit order book.
PAPER REPORTS · Normalized cash flow (20), noise market: LN mean 6.03, sd 2.81 (M=2); mean 4.66, sd 1.41 (M=20); 10,000 test episodes,… · Normalized cash flow, noise+tactical market: LN mean 9.11, sd 2.20 (M=2); mean 5.68, sd 1.03 (M=20)
We develop simulation-based policy iteration for continuous-time portfolio choice with predictable returns and convex constraints. Each outer step re-evaluates a fixed-latent OL-BPTT adjoint after deployment and solves the constrained update.
PAPER REPORTS · Tilted Monte Carlo log-certainty-equivalent gap to the reference on the three-factor, fifty-asset constrained…
Reinforcement learning has gained increasing attention as a data-driven approach for stock trading. However, learning a policy that is both profitable and stable remains challenging due to non-stationary market behaviour and noisy reward signals.
PAPER REPORTS · DJI (test 1 Jan 2024 - 31 Mar 2025, 0.1% transaction fee both sides): annual return 21.785%±1.42, cumulative return… · FTSE (same period and costs): annual return 19.164%±1.36, cumulative return 24.596%±1.79, Sharpe 1.124±0.08, max…
The authors present a rigorous empirical evaluation of three distinct optimization paradigms for institutional factor portfolio construction: an entropy-based photonic quantum annealer (Dirac-3, Quantum Computing Inc.), a commercial mixed-integer programming…
PAPER REPORTS · Dirac-3, best overall configuration (beta1=0, beta2=1): Sharpe 0.760, Sortino 0.841, Calmar 0.567, MDD -3.47%, CVaR5%… · Dirac-3 (beta1=0, beta2=0.5): Sharpe 0.721, Calmar 0.538, MDD -2.63% (lowest in both sweeps), CVaR5% -1.107%, annual…
OUR BACKTEST · Sharpe 0.25 · Return +11.4% · Max DD -18.7%
Risk-aware Q-learning (RaQL) provides a model-free, two-timescale estimator for dynamic risk objectives, but its finite-budget behavior remains fragile: fixed inner-loop hyperparameters can produce unstable value estimates, persistent Bellman residuals, and…
PAPER REPORTS · Scheme 6 out-of-sample (918 daily obs, chronological test set, after 5bp turnover costs, mean over 20 seeds): Sharpe… · Scheme 0 fixed-parameter baseline out-of-sample (same test set, after 5bp costs, 20 seeds): Sharpe 0.5628 (sd 0.2281),…
OUR BACKTEST · Sharpe 0.34 · Return +16.6% · Max DD -10.4%
For a trading desk, residual climate hedging valuation adjustment (HVA) is the climate cost left after its inherited hedge and any admissible overlay have been taken into account; it therefore cannot be inferred from a stand-alone stress loss.
PAPER REPORTS · Residual climate HVA (own method): entropic climate charge reduced to 0.831 from a 0.906 post-inherited-hedge residual,… · Residual Dyna mean exact regret 0.00757 after 30 updates (6,000 gradient trajectories), vs 0.10863 for observed replay…
Recent advances in Generative AI have substantially improved financial sentiment analysis through post-trained financial large language models (LLMs).
PAPER REPORTS · FinSMART (static): cumulative return 264.9%, annualized return 91.5%, Sharpe 1.97, Sortino 2.40, Calmar 4.23, RankIC… · FinSMART (periodically retrained every 6 months): cumulative return 406.2%, annualized return 125.7%, Sharpe 2.41,…
OUR BACKTEST · Sharpe -0.46 · Return -48.2% · Max DD -75.2%