This paper proposes PALM (Point-in-time Adaptation for financial Language Models) as an alternative to annual pre-training of financial language models. It achieves adaptation by fitting low-rank adapters on permitted financial text without modifying any pre-trained weights.
Read originalThis arXiv paper defines the recovery of tick-level queue history from aggregated L2 limit order book snapshots as a conditional partial identification problem. It constructs a path-preserving compiler that only varies among three cancellation allocation rules—front-of-queue, quantity-weighted random, and back-of-queue—while fixing the aggregated path, reconciled market removals, latent order partitioning, and new order arrivals.
Read originalThis arXiv paper evaluates the reliability and investment consequences of TimeGAN-generated scenarios in Rachev ratio portfolio allocation. Using 12 ETFs, it compares historical, substitution, and hybrid strategies under delayed execution and transaction costs. The study finds that generated scenarios underestimate volatility and tail losses and distort dependencies between assets; extending training only reduces reconstruction loss without improving cross-asset correlation accuracy. Substitution strategies increase turnover and are sensitive to the lower bound of the objective function denominator, and adjusting this lower bound does not consistently improve realized downside risk.
Read originalQuantConnect Lean commit number 18197 adds the ShareClassMarketCap field, with MarketCap still retained as the company's total market capitalization. This commit is verified by GitHub signature and co-authored by Claude Fable 5.1.
Read originalvectorbt v1.2.0 has been officially released. This version introduces the ulcer index and Martin ratio indicators, adds the nbformat>=4.2.0 dependency, and fixes multiple issues including OBV calculation, EWM standard deviation ddof handling, portfolio P&L expectation statistics, and drawdown logging.
Read originalA former equity strategy partner and independent investment firm operator used AI in retirement to backtest his long-standing personal strategy, which combined Peter Lynch's growth stock selection, Hilary Kramer's undervalued obscure stock selection, and earnings revision discipline. He corrected for survivorship bias, used point-in-time data, calculated pre-fee returns, and compared it against random stock selection from the same investment pool. The results showed that the strategy did not outperform random selection, with ordinary index funds performing better than the strategy.
Read originalQuantocracy has compiled the latest quant research-related links as of October 18, 2026, covering topics including trend following, futures continuous contract construction, in-sample to out-of-sample Sharpe ratio decay, the impact of liquidity on stock returns and option pricing, and cryptocurrency cross-sectional trend factors. One study based on 12 commodity futures found that the expected retention ratio of in-sample Sharpe ratios from a 10-year backtest on new data is 30%.
Read originalThis arXiv paper proposes a NoLBERT encoder pre-trained only on texts from 1976 to 1995, comparing it with LDA, MiniLM, and FinBERT on 1.26 million news articles from 2012 to 2024 to predict returns of the 1000 largest U.S. stocks.
Read originalZipline Reloaded 3.2.0 is a maintenance release. It supports Python 3.11–3.14, removes support for Python 3.10, and is not yet compatible with Pandas 3 or Python 3.15. The pandas version remains constrained to below 3.0.
Read originalCode change commit number 18193 in QuantConnect Lean migrates BaseExtendedDictionary, ReadOnlyExtendedDictionary.
Read originalQuantConnect Lean Release 18192 migrates BaseExtendedDictionary, ReadOnlyExtendedDictionary, and SessionConsolidator.
Read originalQuantConnect Lean released commit 18189 (corresponding hash 1a91415), which has a verified signature on GitHub. This update hides FundamentalProperty and the Get methods that accept this parameter from the code stubs.
Read originalQuantocracy compiled recent quant research links on October 15, 2026. This roundup includes Allocate Smartly's backtest of the OptimumV strategy from Financial Mentor (backtest period starting in 1987).
Read originalNautilusTrader 2.0.0rc7 was released on October 15 (UTC), removing Python 3.12 support and adding compatibility with Python 3.13 and 3.14. This update adds features such as negative price support for futures, local calendar-based scheduled dispatch, batch order book incremental replay, and CME Globex option candidate IDs, while also including multiple breaking changes to interfaces and configurations.
Read originalQuantocracy released its latest roundup of quantitative research links on October 14, 2026, featuring multiple research pieces and content related to the quantitative field. These include an out-of-sample evaluation study of Hidden Markov Model (HMM) market state identification, an analysis of the impact of complexity on strategy performance, and research on cross-sectional momentum ranking methods. The related research covers backtest results across different sample periods and discussions of strategy logic.
Read originalQuantpedia released a study comparing four four-state Gaussian HMMs that only differ in training start year (1970, 1980, 1990, 2000), evaluating market state recognition performance on the same 2010-2023 test set.
Read originalAuthor Kris Longmore shares content from his 72-minute interview on The Algorithmic Advantage podcast, looking back on an early experience at a Sydney traders' meetup where he was unable to answer a proprietary trader's question about his own trading edge. He later joined that proprietary firm as a researcher and partner, and summarized relevant insights after leaving to trade independently.
Read originalQuantocracy compiled recent quantitative content on October 12, 2026, covering an introduction to Quant Observer, a quantitative tools directory platform; a historical effectiveness analysis of excess returns from insider buying; a backtest of volatility clustering effects in S&P 500 constituents from 2006 to 2025; a guide to building a multi-asset, multi-timeframe trading pipeline; and AI answers on the replicability of empirical finance research.
Read originalAlpha Architect interprets Dmitriy Muravyev's 2026 working paper, which uses an in-house LLM replication pipeline to reconstruct the core 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 were eligible for subsequent out-of-sample testing.
Read originalvectorbt v1.1.2 has been released. This update fixes multiple functional defects, including the isclose helper function returning True for any finite number, NaN consistency issues in is_sorted_nb, portfolio NaN trade plotting marker handling, ignored NaN returns in cond_value_at_risk, and intraday time retention during NumPy datetime64 conversion.
Read originalIn September 2026, Quantpedia added two new fields, Trading Costs and Strategy Robustness, to its Screener and API. Access to the related API requires a Quantpedia Pro subscription and a separate application.
Read originalQlib v0.9.8 has been released, including security hardening completed at Microsoft's request, merging changes from #2339 and #2340. This version restricts the loading of executable recorder artifacts; loading artifacts from trusted storage requires explicitly setting trusted=True, and configuration-driven local .py imports also require explicitly enabling trusted: true. Before upgrading production workflows, please review the two migration guides. Source archives are available on GitHub, and PyPI packages can be queried separately.
Read originalQuantpedia analyst Cyril Dujava released a study based on a sample of 13 US federal elections from 2000 to 2024, covering XLK, SPY, IYR, IEF, EEM, and GLD.
Read originalvn.py version 4.5.0 has been released. This update includes feature adjustments, performance optimizations, and fixes for multiple issues. Key changes include: vnpy_ctabacktester backtest configurations are now saved grouped by strategy name; vnpy.alpha has been refactored for factor and signal performance analysis with a new vwap matching mode; vnpy_polygon has been renamed to vnpy_massive; and several trading interfaces have updated API versions. The update also fixes issues related to UI sorting, cross-contract indicator calculation, market data processing, and connection timeouts.
Read originalQuantocracy has compiled recent quantitative research-related links as of October 5, 2026, covering 5 quantitative research pieces on different topics. The content includes a study on the century-long history of meme stocks, methods for constructing statistically valid backtests, market reactions following analyst downgrades, a comparison of 200-day moving average and volatility targeting strategies, and an introduction to the Orthogonal Clustering Risk Parity (OCRP) strategy.
Read originalNautilusTrader 2.0.0rc6 was officially released on October 4, 2026 (UTC). This version introduces several features, including configurable WebSocket inbound size limits, expanded outputs for multiple technical indicators, a Parquet directory migration tool, and enhanced adapters for venues such as Architect AX, Binance, and Bybit. It also removes the adapter for the closed BitMEX exchange and adjusts certain fee configurations and serialization-related interface definitions.
Read originalQuantocracy compiled recent links related to quantitative research on October 3, 2026, covering five topics: a backtest of Warren Buffett's contrarian investment advice, the performance of a Nasdaq-100 polynomial regression band strategy, the causes of the asset growth anomaly, a test of the U.S. stock turn-of-the-month effect, and a method for calculating transaction costs in macro trading strategies.
Read originalThe Quantpedia Awards 2027 is now open, with a total prize pool of USD 25,000, for quantitative research papers published on open-access platforms such as SSRN and arXiv between May 1 and December 31, 2026.
Read originalOpenBB has released the V5 major version line. The openbb meta-package is at 5.0.0, while the core, CLI, API, MCP server, router, and most providers are at 2.0.0. openbb-devtools is at 2.0.1, openbb-charting is at 4.0.0, and openbb-imf is at 3.0.0.
Key Takeaways:OpenBB V5 changes the license to Apache-2.0 and migrates to hatchling/uv packaging, a layered openbb.toml configuration, and a CLI 2.0 spec file mechanism. These changes can be used to evaluate upgrade paths and available data sources.
OpenBB has released ODP Desktop v2.0.0, corresponding to the V5 iteration. This update completes the initial refactoring of openbb-core, achieves 100% test coverage for testable code paths, and introduces a backward-compatible layered configuration system, openbb.toml.
Read originalvectorbt v1.1.1 is officially released. This version optimizes the numerical stability of rolling standard deviation for the Numba and Rust engines, corrects the calculation logic for deflated Sharpe ratio, and adjusts the alignment rules for cross-engine expanding mean/standard deviation. This version is adapted to the built-in plotting themes of Plotly 7, while the documentation site has been migrated to Zensical.
Read originalNautilusTrader released version 2.0.0rc5 on September 15, 2026 (UTC). This release introduces a rolling ZScore indicator, HTTP streaming reads, and Python message publishing and subscription.
Read originalRob Carver has completed the final edits on his fifth trading book, "The Art And Science of Trading," scheduled for official publication on December 1 and now available for pre-order. Aimed at the mass market, this book serves as the unleveraged counterpart to his third book, "Leveraged Trading." It covers an introduction to trading delta-one instruments such as stocks, ETFs, and cryptocurrencies, explains trading science concepts like backtesting and the Kelly Criterion, and explores the integration of art and science within a semi-automated trading framework.
Read originalbacktesting.py v0.6.6 has been released, tagged by kernc on July 22 at 15:22, corresponding to commit hash 9c2b8e1, which carries a verified committer signature. This version includes only one documentation change: hiding backtesting.Pool in the API documentation generated by pdoc3.
Read originalvectorbt v1.1.0 has been officially released, adding support for Python 3.14, pandas 3, NumPy 2.4+, Numba 0.66+, and Telegram Bot API v20+. This release optimizes Rust engine packaging and consistency coverage, refreshes documentation, examples, resources, and feature previews, and fixes multiple issues related to data processing, indicator calculation, and plotting.
Read originalIn the 10th installment of his 2026 portfolio optimization series, Rob Carver conducts rolling out-of-sample backtests to compare four return pooling schemes. The test sample consists of 50 instruments randomly selected from 214 available instruments, with in-sample lengths of 5, 10, and 20 years and out-of-sample lengths of 1 and 5 years. The optimization uses shrinkage methods with an SR shrinkage of 0.5 and a correlation shrinkage of 0.75, while also incorporating the reverse rule of each trading rule, with only positive SR rules retained for optimization.
Read originalThe authors use random subsample backtests to examine the effectiveness of exponentially weighted window estimates of the Sharpe ratio for portfolio optimization. They compare the out-of-sample Sharpe ratios of EWM spans of 5, 10, 20, and 30 years against full-sample estimates across various in-sample and out-of-sample periods.
Read originalThis article is the eighth installment in the Portfolio Optimization Series. The author proposes using the Euclidean distance of SR vectors across multiple trading rules for each instrument to determine whether to combine cross-instrument returns, calibrates 95% critical values for different historical lengths using random data, and then obtains pseudo-instruments through hierarchical merging. The author compares three approaches—full merging, algorithmic merging, and no merging—using in-sample backtests across 214 futures instruments. The results show inconsistent out-of-sample performance, with longer sample data favoring no merging. The author ultimately concludes that the algorithm is highly complex and the grouping results lack intuitive logic, so it will not be adopted for now, and future research will explore simpler merging rules.
Read originalThis article introduces a method for automatically identifying structural breakpoints in the Sharpe ratio of trading rules, using independent t-tests to compare mean differences across subsamples, recursively splitting the sample while requiring each subsample to retain at least 5 years of data. The authors provide complete Python implementation code, using the momentum4 rule for CORN futures as an example, with the algorithm automatically identifying August 26, 1982 as the breakpoint.
Read originalThis is the sixth article in a portfolio optimization series. The author conducts hundreds of random backtests to examine the impact of different numbers of clusters, correlation shrinkage, SR shrinkage, and in-sample vs. out-of-sample lengths on out-of-sample Sharpe ratios across asset sizes of 20, 30, and 40. The results show that most parameter combinations have no significant differences, with only weak evidence that moderate clustering outperforms no clustering, and that the number of clusters should not be too large. The author ultimately recommends using an empirical value of 6 clusters when clear statistical evidence is lacking, while noting that clustering significantly increases optimization time.
Read original
