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 · 47 shown
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%).
Accurate prediction of stock price movements is of substantial importance to investors, portfolio managers, and financial regulators, and although machine learning models are increasingly deployed in quantitative finance, the limited transparency of many such…
PAPER REPORTS · XGBoost long/flat, hold-out period, as-is configuration: annualized return −11.59%, Sharpe −0.241 at 10 bps one-way… · XGBoost long/short, hold-out period, as-is configuration: annualized return −34.29%, Sharpe −1.014 at 10 bps one-way…
OUR BACKTEST · Sharpe 0.09 · Return +2.4% · Max DD -14.9%
Bitcoin's extreme price volatility poses significant challenges for investors and traders in making informed investment decisions. Conventional technical analysis approaches are often subjective and unable to capture complex market dynamics accurately.
PAPER REPORTS · Trailing-stop strategy, threshold 0.37, test period 1 Jan 2025-1 Mar 2026: net profit +8,491 USD on 10,000 USD (total… · No-trailing-stop strategy, threshold 0.35, same test period: net profit +8,978 USD (total return +89.78%), win rate…
In its traditional form, the smart order router only cares about price when deciding where to send a trade; send the order to the exchange currently showing the tightest spread.
PAPER REPORTS · Mean effective execution cost 6.561 bp (latency-aware router) vs 6.548 bp (price-only baseline). · Hit rate vs baseline: 2.2% of 14,400 held-out rows (latency-aware cheaper than price-only).
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%
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%
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%,…
Global financial markets are increasingly exposed to shocks that propagate rapidly across countries, sectors, and asset classes, generating synchronized volatility that reflects the systemic dimension of financial risk.
Finding which variables cause which others in multivariate time series, and at what lags, is central to science and policy, yet existing methods force a choice between flexible confounder adjustment, data-driven lag selection, and inference that controls the…
PAPER REPORTS · ACLE-VAR (the paper's adaptive-lag procedure, no confounders): weighted long-short, SPY dollar-hedged, open-to-close. · ACLE-VAR by strategy (Ann. Return % / Sharpe), 2004–2020, no costs: Naive 4.64/1.15; Top 50% 8.54/1.31; Top 25%…
OUR BACKTEST · Sharpe -0.33 · Return -2.2% · Max DD -4.6%
Forecasting stock index prices remains a challenging task because financial markets are volatile, non-stationary, and influenced by several economic and behavioral factors.This study evaluates 12 supervised machine learning models against a Naïve Persistence…
Power systems with increasing variable renewable generation face greater uncertainty in scheduling and balancing. Intraday and balancing electricity markets facilitate position adjustments and real-time balancing close to delivery.
In quantitative finance, standard regression losses are misaligned with the economics of return prediction.
PAPER REPORTS · BTC 15-min, CZAR tuned (alpha about 0.054). Test window 2,000 candles (about 21 days); costs, slippage and sizing… · BTC 15-min, CZAR fixed alpha=0.005 to 1: naive annualized Sharpe of 5.62, -0.95, 3.95, 1.15, 4.76 and 5.08…
Monetary-policy announcements and central-bank communications play a central role in foreign exchange markets, yet their qualitative, unstructured form makes their forecasting value difficult to quantify.
PAPER REPORTS · Directional hit rate, out-of-sample rolling windows, 2024-06-13 to 2026-06-12, no transaction costs considered: 66.08%… · Directional AUC 0.744 at 1h, 0.729 at 3h, 0.707 at 8h, 0.634 at 24h. Balanced accuracy 0.721, 0.698, 0.705, 0.626.
Abstract Binary direction classifiers can appear stable on one chronological split while offering little predictive or economic value out of sample.
PAPER REPORTS · Full AND-gate (LightGBM + trend + Hurst), 3 expanding walk-forward folds, Mar-May 2025, 9 configurations: precision… · Full AND-gate unlevered next-bar accepted-signal mean return: -0.16 bps at 0 cost, -10.16 bps at 10 bp round-trip cost,…
Abstract This research utilizes explainable artificial intelligence (XAI) techniques to explore the relationship between size-sorted environmental, social, and governance (ESG) portfolio indices (large-cap, mid-cap, and small-cap ESG portfolios) and their…
Abstract We develop a modular forecast-to-construction framework for static option-portfolio design with piecewise-affine contingent claims.
PAPER REPORTS · Out-of-sample predictive-region coverage on 200 purged test origins (20 Jun 2025 – 7 Apr 2026): 200/200 = 1.000000 at… · Deterministic hedge certificate for the single deployed portfolio: Profit(x) ≥ 0 on [4.0, 8.5] (e.g.
Early warning of market turbulence is usually treated as a classification problem, yet a usable system must also produce trustworthy probabilities, stay parsimonious and convert forecasts into defensible positions.
PAPER REPORTS · Optimized policies, mean of the eleven metaheuristics on the untouched holdout (25 Apr 2019-21 Feb 2020, 5 bp per unit… · Best optimized method by mean return (CS), 30 runs: balanced accuracy 0.693 +/- 0.026, AUC 0.763 +/- 0.019, return…
OUR BACKTEST · Sharpe 0.89 · Return +22.5% · Max DD -8.6%
Forecasting financial markets remains challenging due to nonlinear dynamics, structural changes, volatility clustering, and uncertainty transmission across asset classes.
PAPER REPORTS · No trading strategy, Sharpe ratio, return, alpha or drawdown is reported anywhere in the paper; performance is reported… · Best-model out-of-sample R2 by target (one-step-ahead monthly, price levels, final 20% of sample): VIX 0.999982…
Abstract Whether richer information systematically improves portfolio allocation remains an open question in extreme-risk digital asset markets.
PAPER REPORTS · Gross of costs (2021-01-01 to 2024-03-31, out-of-sample rolling windows): GBRT+MV cumulative return 4.978, annual SD… · Gross: RF+MV return 4.806, SD 1.195, Sharpe 4.020; RF+MMV return 4.944, SD 1.386, Sharpe 3.567
MASTER's inter-stock multi-head attention -- the module responsible for modeling cross-sectional stock relationships -- accounts for 42.5% of model parameters and 25% of predictive value.
PAPER REPORTS · Long-short portfolio Sharpe at N~3,500 (full A-share, test 2019-2020, 4 seeds, frictionless): PureLSTM 5.6 +/- 0.5;… · IC / Rank IC as the primary reported performance: CSI300 Original 0.058 +/- 0.007 / 0.066 +/- 0.003 (5 seeds); CSI800…
OUR BACKTEST · Sharpe -0.58 · Return -50.8% · Max DD -63.3%
Frequent market instability and the lack of rigorously validated forecasting frameworks pose significant challenges for predicting market stress in the Dhaka Stock Exchange (DSE).
PAPER REPORTS · Random Forest crash-gated risk-off strategy, pooled equal-weighted, 2019-2022 test period: total return 67.26% (15.01%… · Annualized volatility 16.39% (strategy) vs. 18.11% (buy-and-hold); maximum drawdown -32.21% vs.
OUR BACKTEST · Sharpe 0.47 · Return +25.2% · Max DD -21.6%
This study constructs a climate risk attention indicator for Chinese funds by applying Word2Vec-based text analysis to annual fund reports.
PAPER REPORTS · Carhart four-factor alpha regression coefficients (2013-2023, annual fund-year panel, fund and year fixed effects):… · Authors' summary claim: "an annualized alpha of approximately 2.7% associated with acute risk focus" (conclusion…
Artificial intelligence (AI) now supports investment workflows from data and prediction through research, portfolios, execution, and tool use. Technical capability, however, is not evidence of investment profitability.
Let a finite population of n labelled examples carry a class-weighted loss, with pi*n in a rare positive class weighted by N0/N1. We study estimation of total risk from a subsample K << n under designs allocating K0 and K1 draws to the two strata.
This study investigates the forecasting performance of machine learning models and traditional econometric volatility models in predicting daily stock price volatility across selected Southern African Development Community (SADC) markets from 02 January 2015…
This study comparatively examined the forecasting performance of machine learning and traditional volatility models in predicting daily exchange rate volatility across selected economies from 01 January 2015 to 08 May 2026.
Foundation models promise accurate forecasts with little or no task-specific training, but whether they can replace models designed specifically for electricity price forecasting remains unclear.
PAPER REPORTS · Total BESS arbitrage profit over 2021-2025 (1,826 days, 1 MWh battery, 25 EUR round-trip operating cost deducted per… · Total profit Poland: unlimited-bid best 106,685 EUR (TabPFN-3) vs Oracle 123,439 EUR, 86.4% of perfect foresight; Spain…
Algorithmic trading now represents a market exceeding $20 billion, where even marginal gains in signal robustness can translate into economically significant returns.
PAPER REPORTS · Hybrid ensemble, OOS 2025 (252 trading days), 2.2bp per trade per leg: total return 51.26%, Sharpe 2.44, Sortino 5.35,… · Hybrid ensemble, OOS 2025: annualised CAPM alpha 0.423 (p = 0.011), beta 0.048, Probabilistic Sharpe Ratio 0.960
OUR BACKTEST · Sharpe 0.75 · Return +696.5% · Max DD -365.1%
Lead-lag relationships are widely used in financial time series, and many clustering algorithms based on them have been developed. The traditional DTW-KMedoids algorithm performs well both on the synthetic dataset and the real financial dataset.
PAPER REPORTS · Sharpe 0.866, annual return 6.21%, annual volatility 7.17%, max drawdown -63.908, hit rate 0.520, profit-loss ratio… · Sharpe 0.808 / 0.790 (KShape mod / med), lead strategy, 679 assets, same period; drawdowns -67.604 / -69.418
OUR BACKTEST · Sharpe 0.39 · Return +27.5% · Max DD -39.3%
This paper develops the first end-to-end application of cross-sectional learning-to-rank to the S&P 500 weekly options (SPXW) zero-day-to-expiration surface, integrated with margin-aware position sizing, an abstention rule driven by model uncertainty, and a…
PAPER REPORTS · Out-of-time 2025 annualized Sharpe 4.308 to 5.761 across seven sizing methods, net of Reg-T margin, tiered IBKR fees,… · Headline Edge Allocation OOT 2025: Sharpe 5.7612, Sortino 7.0291, annualized return 10.48% (excess of risk-free),…
KellyBoost is a single multi-output XGBoost model whose softmax output is the portfolio: with y the vector of per-asset holding-period returns, the training loss is - log(1 + w y), the negative log growth rate, so the fitted model is the growth-optimal…
PAPER REPORTS · KellyBoost (searched, gross of costs), 2013-01 to 2026-07, 163 monthly decisions: mean log growth 0.47 (x100 per 20-day… · KellyBoost hand-built feature pipeline, same period: logG 0.39, annualized return 5.7%, vol 22.9%, Sharpe 0.28, max DD…
PAPER REPORTS · Optimized ESG portfolio: average conditional volatility 1.29% (2020) declining to 1.18% (2022) and 1.23% (2023-2024),… · Optimized ESG portfolio breach rates: 0.3526 volatility violations and 0.4261 drawdown violations on average 2020-2024…
PAPER REPORTS · Out-of-sample RMSE (70-30 split, 2021–2024, no transaction costs): LLF BTC 0.541, ETH 0.256, USDT 0.244, BNB 0.676, BCH… · Out-of-sample MAE: LLF BTC 0.377, ETH 0.199, BNB 0.446, BCH 0.567, LTC 0.458, ICP 0.519, MATIC 0.612, USDT 0.137 (RF…
Investors interpret social disclosures from a risk perspective, yet relevant information can reach them through channels that differ sharply in regulatory enforcement and materiality: SEC filings, sustainability reports, or financial reports.
Using neural networks for stock return prediction typically requires choices about depth and hidden-layer width that are difficult to connect to financial interpretation.
PAPER REPORTS · HNN (marginal): pooled out-of-sample R2 0.509% (vs zero forecast), 1987-2016 · HNN (marginal): gross equal-weighted decile long-short spread 3.95% per month, 1987-2016
As more investors contemplate private markets and contend with limited transparency, sparse disclosures, and infrequent transactions, identifying economically meaningful peer companies for comparison is a fundamental challenge for valuation, due diligence,…
PAPER REPORTS · Out-of-sample log-valuation MAE 1.08, RMSE 1.44, R^2 0.46, MAPE 0.06, MdAPE 0.05 (CatBoost, 20% held-out test set… · Relative improvement over OLS baseline: MAE +8%, RMSE +8%, R^2 +28%, MAPE +14%, MdAPE 0%
Foundation models for time series forecasting demonstrate impressive zero-shot generalization but often underperform on specialized domains such as high-frequency finance.
PAPER REPORTS · Mean per-day correlation 0.3730 (GatedLinear+RF, 40 test days per stock, Dec 2024-Jan 2026 sample, transaction costs… · Pooled correlation 0.5972 and cross-day correlation 0.5631 (GatedLinear+RF, same 40-day test window)
OUR BACKTEST · Sharpe -0.42 · Return -0.6% · Max DD -0.8%
Cryptocurrency exchange-traded products (ETPs) listed on European exchanges provide a regulated environment for studying intraday market anomalies.
PAPER REPORTS · AUC-ROC up to 0.823 for one-bar-ahead DPOT prediction (LR, cumulative features, VIRBTC.ST), out-of-sample last 20% of… · AUC-ROC 0.821 for DPOT, LR, cumulative, VBTC.XE; 0.812 session-based
The paper considers the problem of variable selection for forecasting electricity spot prices.
PAPER REPORTS · BMT hourly rMAE averaged across six areas: 0.501 (2022-2025 out-of-sample, no transaction-cost concept; rMAE < 1 means… · BMT daily baseload rMAE_t averaged across six areas: 0.408 (2022-2025)
Neural and numerical policy solvers can produce feasible controls even when the optimal rule and its binding constraints are unavailable.
Financial sentiment classifiers are commonly evaluated against human labels, but strong linguistic performance does not necessarily imply economically useful return predictability. This study separates these questions through two experiments.
PAPER REPORTS · One-day mean rank IC, 2019 Benzinga S&P 100 sample: FinBERT 0.0143 (largest), Financial-RoBERTa 0.0141, Naive Bayes… · FinBERT one-day long–short: gross total return 12.96%, annualized Sharpe 1.11, max drawdown -6.37% (2019, gross — no…
OUR BACKTEST · Sharpe -1.16 · Return -61.4% · Max DD -64.2%
Informed traders are supposed to need anonymity: they profit by hiding among the uninformed. A decentralized exchange now publishes the counterparty. Every committed order, cancellation, rejection, and fill carries a persistent pseudonymous wallet address.
PAPER REPORTS · One-second out-of-sample R2: 10.88% anonymous vs 12.31% with identity, +13.2% relative (t=9.2), ridge, evaluation July… · Gradient-boosted trees, one second: 19.48% -> 20.65%, +6.0% (t=5.0), same evaluation window, no costs
Factor structures are central to empirical work in economics and finance, and are usually used to model time-varying unobserved heterogeneity through interactive fixed effects (IFE).
We introduce data-driven measures of high-frequency trading (HFT) that distinguish between liquidity-supplying and liquidity-demanding strategies.
Production forecasting systems retrain models regularly, but a retrained candidate does not necessarily outperform a continuously maintained incumbent that has continued to learn.
PAPER REPORTS · Metric is negative log-likelihood of a 3-class 300s direction forecast, not trading P&L; no Sharpe, return, alpha or… · Pooled 48 weeks (4 Aug 2025 - 5 Jul 2026, Binance USD-M + COIN-M, 8 underlyings, 3 seeds): SBS relative NLL reduction…
We audit whether candle-based machine-learning models can turn predictions of cryptocurrency extrema or short-horizon outcomes into positive Binance Spot paper policies after assumed costs.
PAPER REPORTS · Frozen mandatory-daily selector: -6.72% compounded, 2026-07-01 to 07-19, 19 cycles, 3 wins/16 losses, 31 bps… · Frozen mandatory-daily selector cost stress, same period: -4.74% at 20 bps, -6.72% at 31 bps, -10.21% at 51 bps…
OUR BACKTEST · Sharpe 0.00 · Return +0.0% · Max DD 0.0%
Using the outbreak of COVID-19 in Singapore as a quasi-natural experiment, we investigate tenants' changing responses to road traffic noise in the rental housing market, using 46,980 transaction records between 2006 and 2022.