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 · 132 shown
Portfolio optimization techniques often use historical data to estimate parameters for Markowitz-type or ML models. Effective allocation requires a way to align the available data with conditions expected over the investment horizon.
PAPER REPORTS · Markowitz, lambda=0 (historical mean only, no future information in the view): net Sharpe 0.019 (gross 0.050), net MDD… · LSTM mean of 100 seeds, lambda=0: net Sharpe -0.021 (gross 0.037), net MDD -24.7%, final wealth 0.978.
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…
Worst-case risk evaluation is central to finance and operations research, yet existing analytical results remain fragmented and largely confined to ambiguity sets with specific structures.
OUR BACKTEST · Sharpe 0.67 · Return +44.6% · Max DD -28.9%
We propose a regime-aware tactical asset allocation (TAA) framework integrating market regime detection with Nested Clustered Optimization (NCO) for multi-asset portfolio construction.
PAPER REPORTS · Regime-NCO (proposed), OOS Jan 2021 to Aug 2025, monthly, 10bp proportional costs, rf 0.479%/month: cumulative return… · Regime-NCO, same period and settings: Sharpe -0.9522, Sortino -1.2271.
This paper investigates the ability of random feature models to learn more complex patterns than the simple negative autocorrelation of financial returns.
PAPER REPORTS · Market timing, RFRR strategy, Jan 1990–Dec 2020, annualized certainty-equivalent returns, weights in [-1,2]: 8.84%… · Same strategy and period, weights in [-0.5,1.5]: 8.81% (a=2, p=0.04%), 7.26% (a=4, p=0.01%), 5.55% (a=6, p=0.02%).
This paper studies continuous-time portfolio selection under Range Value-at-Risk (RVaR), with both portfolio performance and risk evaluated in terms of logarithmic returns.
OUR BACKTEST · Sharpe 0.18 · Return +15.4% · Max DD -38.2%
PAPER REPORTS · DCC-ePOET-1, out-of-sample global MVP on 331 S&P 500 stocks: Var 0.445, Annual-SD 10.840, mean Sharpe ratio 2.281,… · DCC-ePOET-2 (latent factors only, no observable factors), same setup: Var 0.464, Annual-SD 11.049, mean Sharpe ratio…
PAPER REPORTS · Recommended allocation (20% equities + 100 tons gold), historical simulation Nov 2002–Oct 2022, rolling 1-year returns:… · Same recommended allocation, rolling annualized 5-year returns, Nov 2002–Oct 2022: CZK average 3.48%, volatility 2.79%,…
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%
The Oracle Approximating Shrinkage (OAS) estimator provides a closed-form, data-driven shrinkage intensity for regularizing sample covariance matrices toward a scaled identity target, but its derivation assumes Gaussian data.
PAPER REPORTS · OAS-Extended GMV portfolio, p=50 US large caps, Jan 2010–Dec 2023, 155 monthly-stepped windows: mean annualized… · OAS-Extended vs OAS realized-variance improvement: t=3.83, p<0.001; bootstrap 95% CI for mean variance reduction [2.0,…
OUR BACKTEST · Sharpe 0.53 · Return +40.1% · Max DD -29.9%
This paper studies decision functionals under ambiguity when pure-risk evaluation is determined by a general, possibly non-affine statistic.
OUR BACKTEST · Sharpe 0.54 · Return +30.9% · Max DD -25.2%
We study liquidity provision in automated market makers when trading in the external market incurs proportional transaction costs. We consider both separately collected fees and fees retained in reserves.
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.
Abstract Implied temperature rise (ITR) metrics translate corporate emissions pathways into a temperature-based measure of climate alignment at the company and portfolio levels.
We study dynamic portfolio decarbonization under incomplete information about firms' carbon intensities and the underlying transition state.
In stochastic control problems involving optimal portfolio choice, one often seeks an optimizer of some local criterion (dynamic programming value function) generated by some global criterion (e.g., expected utility).
OUR BACKTEST · Sharpe 0.80 · Return +34.1% · Max DD -15.2%
To extend conventional model-level feature importance to economically meaningful portfolio outcomes and decisions, we propose a local first-order attribution-propagation framework that maps allocation-function feature exposures to local sensitivities of…
PAPER REPORTS · XGB baseline, OOS 2006-2024: net Sharpe 0.6175 (annualised, 5bp per unit turnover); gross Sharpe 1.0207; net mean 2.55%… · FFN baseline, OOS 2006-2024: net Sharpe 0.5750 (5bp); gross Sharpe 1.3051; net mean 2.65% p.a.; net FF5 alpha 1.12%…
OUR BACKTEST · Sharpe -0.82 · Return -24.2% · Max DD -31.5%
Portfolio decisions are often made under uncertainty about expected returns, and investors revise their beliefs as new information arrives.
OUR BACKTEST · Sharpe 1.07 · Return +152.8% · Max DD -57.4%
We study a two-sided singular stochastic control problem in which changes in consumption carry utility costs and the risk premium follows the mean-reverting Kim-Omberg model.
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 study examines whether Shannon entropy regularization can mitigate the concentration associated with classical minimum-variance portfolio optimization while preserving risk-adjusted performance.
PAPER REPORTS · ERMV (λ=0.01), full sample (in-sample Σ), 2015–Jul 2026, daily simple returns, no transaction costs: annualized mean… · ERMV (λ=0.01), full sample, no costs: Sharpe 0.2966 at 4.5% rf, 0.5140 at 2% rf, 0.6879 at 0% rf.
This paper studies entropy regularization for the discrete-time dynamic mean--variance portfolio problem with no-shorting constraints.
PAPER REPORTS · Rolling Fama-French 10 industries, 72 non-overlapping six-month out-of-sample episodes, first beginning Jan 1990 and… · Simulated data (lognormal mixture, 10,000 test paths, no costs), deterministic mode policy: - tau=10^-4: terminal mean…
Abstract Time series forecasting models play a crucial role as decision support tools in various real-world domains.
PAPER REPORTS · Multi-stock swing trading (17 stocks, equal budget, 3-day confirmation), uptrend Jul 1–Dec 31 2021: average gain 7.64%… · Multi-stock swing trading, downtrend Jan 1–Sep 20 2022: average gain -0.34% across Max Shares 1/5/10 (-7.16%, 3.61%,…
The rapid growth of crypto markets is increasing investment opportunities, but also volatility risks. There is a need to automate portfolio construction to rapidly adapt investment decisions to rapidly changing markets.
PAPER REPORTS · Crew B (30-day rolling Sharpe max), out-of-sample per Table 6: expected return 10%, volatility 10%, Sharpe 0.72,… · Crew B (30-day window), out-of-sample per the window-sensitivity table: expected return -20.58%, volatility 39.88%,…
We represent (weighted-)selection-elliptical distributions as an affine combination of the $q$ selection variables plus an elliptical term whose direction alone is independent.
OUR BACKTEST · Sharpe 0.75 · Return +27.1% · Max DD -21.2%
Expectiles are the only law-invariant risk measures that are both coherent and elicitable. Unlike Conditional Value-at-Risk (CVaR), however, they do not admit a Rockafellar--Uryasev representation that admits tractable Wasserstein reformulations.
PAPER REPORTS · Proposed model, (rho, lambda)=(0.005, 0.10), FTSE 90 stocks, 3,341 OOS days 2012-2026, gross of transaction costs: ann. · Proposed model, (0.001, 0.10), same sample, gross: ann. return 20.62%, ann. vol 20.52%, Sharpe 1.005, MaxDD 44.41%,…
We develop a framework for an investor who trades until she either reaches a financial goal or an exogenous deadline arrives. Analogous to utility functions over wealth, we measure satisfaction with the timing of reaching a goal by a discount function.
OUR BACKTEST · Sharpe 0.63 · Return +44.3% · Max DD -26.3%
Risk evaluation under distributional ambiguity is central to decision making in finance, economics, and operations research. Wasserstein balls provide a natural way to describe uncertainty around a reference distribution.
OUR BACKTEST · Sharpe 0.58 · Return +43.5% · Max DD -29.9%
In this paper, we present robust portfolio selection models by incorporating a reward and penalty mechanism into portfolio management.
OUR BACKTEST · Sharpe 0.97 · Return +77.6% · Max DD -27.8%
This article studies a retirement planning problem from a new perspective in which a retiree delegates an initial lump sum to a professional fund manager.
Factor mining seeks to discover signals from financial data that predict future asset returns and guide portfolio construction. Automated factor mining now spans genetic programming, reinforcement learning, generative models, and large language model agents.
Taiwan’s central role in global semiconductor manufacturing exposes Taiwan-related ETFs to technology concentration, geopolitical uncertainty, and supply-chain disruptions, resulting in return distributions characterized by heavy tails, volatility clustering,…
OUR BACKTEST · Sharpe 0.37 · Return +28.9% · Max DD -34.3%
We develop a price-based framework for stochastic portfolio theory in which trading strategies are generated from nominal price weights and evaluated relative to a price-weighted benchmark.
PAPER REPORTS · Cap-based p-diversity-weighted portfolios, terminal log relative wealth vs cap-weighted index on U^S (2002–2021,… · Cap-based p-diversity-weighted portfolios, terminal log relative wealth vs cap-weighted index on U^P (2002–2021, no…
OUR BACKTEST · Sharpe 0.38 · Return +39.6% · Max DD -47.5%
The enhanced index tracking (EIT) portfolio selection problem aims to construct a portfolio that is expected to outperform a benchmark index.
PAPER REPORTS · M-ER, known mean/covariance, short selling allowed, OOS Jan 2020-Dec 2024, monthly rebalancing, transaction costs not… · M-TSV, known mean/covariance, short selling allowed, same period, costs not stated: mean return 0.048939, SD 0.150532,…
OUR BACKTEST · Sharpe 0.70 · Return +62.4% · Max DD -32.5%
Abstract The fact that the solution of the Markowitz model is numerically unstable is recognized as one of the principal reasons for its poor out-of-sample results when applied in practice.
PAPER REPORTS · Out-of-sample Sharpe ratio of the μp=β (δ-stable) portfolio, daily data 2000–2020, rolling window, net of 10 bp ×… · Same setup, DAX: 0.0099 (w=120), 0.013 (w=180), 0.0206 (w=240), 0.0312 (w=480).
OUR BACKTEST · Sharpe 0.66 · Return +49.0% · Max DD -31.1%
We propose an interactive robo-advising framework that learns personalized risk preferences from scores provided by clients.
In private wealth management, a manager delegating to artificial intelligence (AI) acts as the client's agent and the system's principal.
When developers change one component of an agent, such as its controller, a learned model or its verifier, they usually judge the change by an aggregate task score.
PAPER REPORTS · ARC-AGENT (full): Sharpe 0.56±0.29 (across-seed s.d.), MaxDD 10.5%, mean per-rebalance turnover 18.1%. · ARC-AGENT Sharpe excess vs static EW/MinVar frontier at matched drawdown: −0.04 (synthetic U3, 8 bps costs).
We study finite-horizon portfolio optimization with proportional transaction costs and trading opportunities arriving at the jump times of a Cox process.
OUR BACKTEST · Sharpe 0.52 · Return +0.0% · Max DD -0.0%
Abstract This paper addresses the issue of time inconsistency in recursive stochastic control problems, in which the forward state process evolves under the influence of an additional recursive utility system.
We develop a spouse-protected tontine in which a first death changes the household state but generates no pool transfer.
PAPER REPORTS · Out-of-sample learned policy, payment-seeking (λ=0.3): expected cumulative real payments 1860.4 (thousand AUD over 35… · Out-of-sample learned policy, intermediate (λ=1): expected cumulative real payments 1704.8, CVaR_0.95 terminal…
Decision-focused learning (DFL) trains predictors through downstream objectives, but a different loss need not provide an independent parameter-update direction.
PAPER REPORTS · Neural DFL vs MSE, N=100, K=20, 9 annual test folds 2016–2025: annualized TE 0.0276 vs 0.0331 (−16.5%), one-sided… · Same neural folds, K=20: worst-fold TE 0.0360 vs 0.0564 (−36.3%), max tracking drawdown 0.0306 vs 0.0362 (−15.5%),…
OUR BACKTEST · Sharpe 0.71 · Return +69.7% · Max DD -35.0%
Bank treasury portfolios must balance yield, liquidity, and interest-rate risk across bonds of different maturities.
PAPER REPORTS · DP mean terminal wealth over 6 months, starting from 1 (simulated DNS-VAR chain, unconditional-mean initial curve):… · DP mean terminal wealth over 12 months: 1.1801 at c=0; 1.1166 at c=0.01.
Global shocks - ranging from climate-related disasters and geopolitical conáicts to pandemics - transmit rapidly across borders and shape Önancial market co-movements.
The article examines the investment universe of the Russian financial market under conditions of restricted access to global capital markets.
In a recent paper, Schmelzer and Hastie argue that Markowitz's Critical Line Algorithm and the LASSO path trace the same curve. Here we use that identity to compute efficient frontiers with a stock LASSO solver, \texttt{lars\_path} from \texttt{scikit-learn}.
OUR BACKTEST · Sharpe 0.00 · Return +0.0% · Max DD 0.0%
In this work, we study portfolio optimization under the stochastic discount factor (SDF) framework by learning market state representations that capture the underlying risk structures of financial data.
PAPER REPORTS · Sharpe ratio 1.062 (out-of-sample rolling 1-year test windows within June 2011–June 2020, 7-day rebalancing;… · Sortino ratio 1.352; Calmar ratio 0.924 (same setting; costs not stated)
News may reveal systematic risk, but whether its context enhances the construction of systematic risk factors is still unclear.
PAPER REPORTS · Spherical ST + sparse-soft c-TF-IDF + multi-horizon IPCA, L=60: annualized OOS excess-return Sharpe 1.0334, Sep 2020 to… · Same raw spherical portfolio: annualized volatility 81.68%, maximum drawdown 62.23%; not risk-managed, not…