Daily research update
we backtested it — Sharpe 0.26 · quant relevance 0.98
Sparse-autoencoder reconstructions produce contrarian portfolios with lower measured volatility than matched-dimension PCA, but do not establish superior risk-adjusted returns or stable factor identities. The return-based method is implementable on the secondary US-equity market, although a practical test must replace the study's explicitly non-executable same-close entries with subsequent-price execution.
we backtested it — Sharpe 0.65 · quant relevance 0.95
The study explains leveraged ETF compounding through convexity, volatility drag, and costs, and advocates realized terminal wealth and expected log growth over compounded arithmetic averages. Its framework directly supports leverage selection and investment-performance evaluation using available daily ETF prices, although it does not establish a predictive trading signal.
we backtested it — Sharpe 0.64 · quant relevance 0.94
The study derives an optimal portfolio allocation under logarithmic utility in an Itô-Liu market, recovering the Merton excess-drift-over-variance rule. A daily implementation is backtestable, but it would test the allocation rule rather than validate the continuous-time uncertainty framework. The supplied material describes theoretical results and numerical examples, not measured trading performance.
The rest of the batch
The study introduces hourly Pendle yield-derivative data and an asymmetric Tail-MSE forecasting loss used in a systematic yield-trading strategy.
The paper develops an optimal-execution mean field game in which aggregate market orders change stochastic limit-order-book densities, liquidity depth, and prices.
The study develops and empirically calibrates a queue-sensitive market-impact model for passive limit orders using E-mini S&P 500 futures order-book data.
The study evaluates monthly US healthcare-equity rankings combining price-volume models, fundamentals, attention measures and language-model text representations.
The study tests Chronos forecasts and conventional risk indicators as triggers for defensive stock–bond allocation during retirement withdrawals.
Analyst upgrades predict higher subsequent returns than downgrades in Chinese equities, with stronger effects among smaller and less-covered firms.
The study tests whether prospectively identified coordinated wallets on pump.fun provide actionable token-entry signals, using time-consistent discovery and reserve-based execution accounting.
The study examines rapid automated hedging of market-maker inventory shocks in Italian sovereign bonds using MTS trade-level data.
The study models synchronized liquidation by AI wealth-management agents and tests portfolio-risk and execution mitigations.
The study develops a risk-neutral valuation framework for discretely monitored variance swaps and absolute-return-based volatility swaps, incorporating regime switching and nonlinear liquidity feedback.
The study compares daily Bitcoin volatility forecasts from GARCH, LSTM, and hybrid models and finds that forecast rankings depend on market regimes and predictor sets.
The study identifies rights-issue arbitrage conditions by comparing outstanding share prices with the cost of acquiring shares through subscription rights.
The study uses Twitter and Telegram sentiment to forecast Bitcoin, Ether, and Binance cryptocurrency returns with ARIMA, SARIMA, LSTM, and Bi-LSTM models, alongside correlation and VaR analysis.
The study decomposes fixed-rate pass-through MBS prices, duration, and convexity to characterize when prepayment behavior produces negative convexity.
and 12 more · see the full radar
Backtests are hypothetical research results; forward-portfolio figures are a live record. Neither is a guarantee of future performance. Research and education only: not individualised investment advice, not an offer to sell or a solicitation of an offer to buy any security.