References · 421 verified sources

[1] J.P. Morgan Markets — e-Trading Survey Report 2026 (GenAI 43%, ML/NLP 18%, API/EMS 15%)[EB/OL]. Survey Jan 12–27, 2026 (accessed 2026-07-29). https://markets.jpmorgan.com/discover-more/e-trading-survey-report [2] oneZero — Liquidity and AI are top priorities in JPMorgan's annual e-Trading survey (series 25%→53%→65%)[EB/OL]. 2024-03-05. https://www.onezero.com/in-the-news/liquidity-and-ai-are-top-priorities-in-jpmorgans-annual-e-trading-survey/ [3] Generative AI and Asset Management (Rutgers Business School / CoreData Research; 70% hedge funds; 21%→33%→63%)[EB/OL]. Aug-2025. https://www.business.rutgers.edu/sites/default/files/documents/generative-ai-and-asset-management.pdf [4] EY — GenAI in Wealth & Asset Management Survey 2025 (95% scaled, 78% agentic, ~25% substantial impact)[EB/OL]. 2025-09-19. https://www.ey.com/en_us/insights/wealth-asset-management/gen-ai-in-wealth-asset-management-survey [5] Kensho — How Kensho built a multi-agent framework with LangGraph to solve trusted financial data retrieval (Grounding; router + DRAs; exact-match eval)[EB/OL]. 2026-03-26. https://kensho.com/news/how-kensho-built-a-multi-agent-framework-with-langgraph-to-solve-trusted-financial-data-retrieval [6] The TRADE — Bloomberg embeds agentic AI into the Terminal (ASKB, beta Feb 23, 2026)[EB/OL]. 2026-02-23. https://www.thetradenews.com/bloomberg-embeds-agentic-ai-into-the-terminal/ [7] Finance Director Europe (citing FT) — JPMorgan rolls out AI-based chatbot LLM Suite (~50,000 employees)[EB/OL]. 2024-07-29. https://www.financedirectoreurope.com/news/jpmorgan-rolls-out-ai-based-chatbot/ [8] Microsoft Cloud Blog — BlackRock Aladdin Copilot (no investment advice; anti-hallucination filters)[EB/OL]. 2024-09-30. https://www.microsoft.com/en-us/microsoft-cloud/blog/financial-services/2024/09/30/elevating-investment-management-tech-ai-powered-leadership-from-blackrock-and-microsoft/ [9] Man Group — What AI Can (and Can't Yet) Do for Alpha (AlphaGPT, 3 roles, thresholds identical to human research)[EB/OL]. 2025-11-13. https://www.man.com/insights/what-ai-can-do-for-alpha [10] Longterm Wiki — Bridgewater AIA Labs (2,000M USD macro fund Jul-2024; 11.9% 2025; guardrails 8%→1.6%; kill switch)[EB/OL]. compilation 2026-02-01. https://www.longtermwiki.com/wiki/bridgewater-aia-labs [11] Bloomberg via fa-mag — AQR bets on machine learning (ML ~1/5 of Apex signals; Asness)[EB/OL]. 2025-04-23. https://www.fa-mag.com/news/aqr-bets-on-machine-learning-as-cliff-asness-becomes-ai-believer-82224.html [12] OpenAI — Balyasny Asset Management (95% of investment teams; Central Bank Speech Analyst 2 days → 30 min)[EB/OL]. 2026-03-06. https://openai.com/index/balyasny-asset-management/ [13] HyperAI — Point72 Turion (Eric Sanchez's AI strategy, outperforms the flagship in 2025)[EB/OL]. 2026-07-28. https://hyper.ai/en/stories/7986a83e5e06a56baada10b139e6ad6a [14] S&P Global — press release: app for ChatGPT via Kensho MCP connector (Capital IQ, transcripts, no training on licensed data)[EB/OL]. 2026-02-09. https://www.spglobal.com/en/press/press-release/sp-global-delivers-trusted-financial-data-and-insights-to-customers-through-app-for-chatgpt [15] Alpha FMC — AI transformation across middle and back office operations (91% using/planning; 7% scaled; measurable reconciliation)[EB/OL]. 2026-05-28. https://alphafmc.com/blog/2026/05/28/ai-transformation-across-middle-and-back-office-operations/ [16] WatersTechnology / MD Market Insights — AI reconciliation >99% matches, ~10 high-risk exceptions[EB/OL]. 2025-02-26. https://www.waterstechnology.com/topics/artificial-intelligence [17] FinBen analysis (beancount.io): forecasting ~0.54 accuracy (barely above chance); FinQA EM 0.63[EB/OL]. 2026-04-15. https://beancount.io/bean-labs/research-logs/2026/04/15/finben-financial-llm-benchmark [18] Advisor Perspectives / fa-mag — Asness (AQR): from ML skeptic (FT 2017) to admitting a 1–2 year lag[EB/OL]. 2024-12-06. https://www.advisorperspectives.com/articles/2024/12/06/asness-ai-end-human-fund-managers [19] Fortune — MIT report: 95 percent of generative AI pilots at companies failing (NANDA, The GenAI Divide)[EB/OL]. 2025-08-18. https://fortune.com/2025/08/18/mit-report-95-percent-generative-ai-pilots-at-companies-failing-cfo/ [20] Timspark — Why AI projects fail (S&P Global 42% of initiatives abandoned in 2025 vs 17%; RAND 80%)[EB/OL]. 2025-09-25. https://timspark.com/pl/blog/why-ai-projects-fail-artificial-intelligence-failures/ [21] Resonanz Capital — How Hedge Funds Are Really Using Generative AI (anonymized abandoned pilots)[EB/OL]. 2025-04-20. https://resonanzcapital.com/insights/how-hedge-funds-are-really-using-generative-ai-and-why-it-matters-for-manager-selection [22] SEC — Press Release 2024-36: AI washing, Delphia USD 225,000 and Global Predictions USD 175,000[EB/OL]. 2024-03-18. https://www.sec.gov/news/press-release/2024-36 [23] Yu et al. — FinCon (arXiv:2407.06567; NeurIPS 2024): CR 113.84%, SR 3.269; CVaR gating; manager-analyst hierarchy[EB/OL]. 2024-07. https://arxiv.org/html/2407.06567v3 [24] ThesisAgent (GitHub fafawlf/thesis-agent): separation of LLM reasoning / deterministic math decision (Decision Hub)[EB/OL]. 2026-04-04. https://github.com/fafawlf/thesis-agent [25] Islam et al. — FinanceBench (arXiv:2311.11944): GPT-4-Turbo+RAG fails/refuses 81%[EB/OL]. 2023-11-20. https://arxiv.org/abs/2311.11944 [26] Macfarlanes — ESMA supervisory briefing (Feb 26, 2026): AI in algorithmic trading integrated into RTS 6 self-assessment; non-delegable responsibility[EB/OL]. Feb-2026. https://www.macfarlanes.com/insights/102mpep/algorithmic-trading-and-artificial-intelligence-esma-supervisory-briefing [27] Xiao et al. — TradingAgents (arXiv:2412.20138), Table 1: CR 26.62% / Sharpe 8.21 on AAPL[EB/OL]. 2024-12-28 (v1). https://arxiv.org/pdf/2412.20138 [28] TradingAgents, footnote 1: self-critique of the 8.21 Sharpe outside the empirical range (3-month window)[EB/OL]. 2025. https://arxiv.org/html/2412.20138v5 [29] Li et al. — HedgeAgents (arXiv:2502.13165; WWW 2025 Companion): 70% annualized, 400% over 3 years[EB/OL]. 2025-02. https://arxiv.org/html/2502.13165 [30] Li et al. — Profit Mirage / FinLake-Bench (arXiv:2510.07920): returns evaporate after the knowledge window[EB/OL]. 2025-10-09. https://arxiv.org/abs/2510.07920 [31] Lopez-Lira & Tang — Can ChatGPT Forecast Stock Price Movements? (arXiv:2304.07619): Sharpe 3.8 (v4) → 3.28 (v5); >500% before costs[EB/OL]. 2023-04-15 (v1); v5 2024-09-11. https://arxiv.org/pdf/2304.07619 [32] Lopez-Lira, Tang & Zhu — The Memorization Problem (arXiv:2504.14765)[EB/OL]. 2025-04. https://arxiv.org/abs/2504.14765 [33] Li, Kim, Cucuringu & Ma — FINSABER (arXiv:2505.07078): 20 years × 100+ symbols, erosion of LLM edges[EB/OL]. 2025-05-11 (v1). https://arxiv.org/abs/2505.07078 [34] Li et al. — DeepFund, Time Travel is Cheating (arXiv:2505.11065; NeurIPS 2025 D&B): net losses in live trading post-cutoff[EB/OL]. 2025-05-16 (v1). https://arxiv.org/abs/2505.11065 [35] A Verifiable Correctness Property for Backtesting and Agentic Trading Pipelines (arXiv:2607.04958): parametric look-ahead undetectable by code inspection[EB/OL]. 2026-07-06. https://arxiv.org/html/2607.04958 [36] McLean & Pontiff (JoF 2016) and Bailey & López de Prado (DSR 2014), synthesis: −26% OOS, −58% post-publication, ~3 trials[EB/OL]. primary sources 2014–2016; synthesis 2026. https://www.turbinefi.com/blog/why-backtests-lie-prediction-market-overfitting-2026 [37] The Alpha Illusion (arXiv:2605.16895): positive short backtest ≠ deployable alpha[EB/OL]. 2026-05-16. https://arxiv.org/html/2605.16895v1 [38] EY-Parthenon — Generative AI in wealth and asset management, survey highlights (52% of hedge funds deployed)[EB/OL]. Mar-2024. https://www.ey.com/content/dam/ey-unified-site/ey-com/en-gl/industries/wealth-asset-management/documents/ey-gl-genai-wam-survey-highlights-03-2024.pdf [39] LangChain Blog — "LangChain and LangGraph Agent Frameworks Reach v1.0 Milestones"[EB/OL]. 2025-10-22. https://www.langchain.com/blog/langchain-langgraph-1dot0 [40] PyPI JSON API — langchain (1.0.0 release timestamps)[EB/OL]. accessed 2026-07-29. https://pypi.org/pypi/langchain/json [41] LangChain Docs — Release policy (LTS, minor/patch cadence)[EB/OL]. 2026-07-28. https://docs.langchain.com/oss/python/release-policy [42] LangChain Docs — LangChain v1 migration guide (breaking changes, mirror imports)[EB/OL]. 2026-07-28. https://docs.langchain.com/oss/python/migrate/langchain-v1 [43] LangChain Docs — "What's new in LangChain v1" (reduced namespace, langchain-classic content)[EB/OL]. 2026-07-28. https://docs.langchain.com/oss/python/releases/langchain-v1 [44] LangChain Docs — Models (init_chat_model, 6x backoff retries, InMemoryRateLimiter, configurable_fields, bind_tools on configurable)[EB/OL]. 2026-07-28. https://docs.langchain.com/oss/python/langchain/models [45] LangChain API Reference — init_chat_model (signature, configurable_fields, config_prefix)[EB/OL]. 2026-07-14. https://reference.langchain.com/python/langchain-classic/chat_models/base/init_chat_model [46] LangChain Blog — "Standard message content" (standard content blocks)[EB/OL]. 2025-09-03. https://www.langchain.com/blog/standard-message-content [47] LangChain Docs — Messages (content block reference, AIMessage attributes)[EB/OL]. 2026-07-27. https://docs.langchain.com/oss/python/langchain/messages [48] LangChain API Reference — Runnable (interface, methods, composition, with_retry, with_fallbacks, config, callbacks)[EB/OL]. accessed 2026-07-29. https://reference.langchain.com/python/langchain_core/runnables/base/Runnable/ [49] LangChain API Reference — RunnablePassthrough (assign)[EB/OL]. accessed 2026-07-29. https://reference.langchain.com/python/langchain_core/runnables/passthrough/RunnablePassthrough/ [50] GitHub — langchain-playground Runnables.md (RunnableLambda as a callable wrapper)[EB/OL]. 2026-01-31. https://github.com/himanshu231204/langchain-playground--for-llms-/blob/main/Runnables/Runnables.md [51] qiankunli.github.io — LangGraph (RunnableParallel 3 equivalent forms, RunnableBranch routing)[EB/OL]. 2026-03-05. http://qiankunli.github.io/2024/05/16/langchain_graph.html [52] LangChain Docs — Streaming (stream modes, event streaming v1.3, tool_call_chunk, tool_choice="any")[EB/OL]. accessed 2026-07-29. https://docs.langchain.com/oss/python/langchain/streaming [53] 51CTO blog — config example with tags/metadata (traceability)[EB/OL]. 2026-07-17. https://blog.51cto.com/u_16213580/14771671 [54] LangChain API Reference — DynamicRunnable (inherited methods: bind, pick, assign, as_tool)[EB/OL]. 2026-06-20. https://reference.langchain.com/python/langchain-core/runnables/configurable/DynamicRunnable [55] LangChain Docs — LangSmith Model fallbacks (LLM Gateway, triggers HTTP 429/5xx)[EB/OL]. 2026-07-24. https://docs.langchain.com/langsmith/llm-gateway-fallbacks [56] LangChain API Reference — with_structured_output (Pydantic/TypedDict/dict schema, include_raw raw/parsed/parsing_error)[EB/OL]. 2026-07-24. https://reference.langchain.com/python/langchain-core/language_models/chat_models/BaseChatModel/with_structured_output [57] LangChain Forum — how Pydantic descriptions are used in with_structured_output (schema as prompt; tool vs prompt+parser strategy)[EB/OL]. 2025-09-23. https://forum.langchain.com/t/clarification-on-how-pydantic-schema-descriptions-are-used-in-with-structured-output/1612 [58] LangChain Forum — with_structured_output with method="json_schema", strict=True[EB/OL]. 2026-02-22. https://forum.langchain.com/t/parsing-error-with-structured-output-with-model-output/2985 [59] GitHub — langchain issue #38223 (Responses API + streaming + nested Pydantic)[EB/OL]. 2026-06-17. https://github.com/langchain-ai/langchain/issues/38223 [60] OpenAI structured outputs JSON schema: a practical guide (CodeWords) — 100% vs ~86% schema compliance[EB/OL]. 2026-06-09. https://www.codewords.ai/blog/openai-structured-outputs-json-schema [61] Structured outputs engineering (datarekha) — "confidently wrong" failure, null as part of the schema[EB/OL]. 2026-04-12. https://datarekha.com/blog/structured-outputs-engineering/ [62] Mind Your Step (by Step): CoT can Reduce Performance (arXiv:2410.21333) — +331% iterations on classification with exceptions[EB/OL]. 2025-06-13. https://arxiv.org/html/2410.21333v4 [63] Gao et al., PAL: Program-aided Language Models (ICML 2023) — GSM-Hard 61.2% vs ~20% CoT; dominant failure = arithmetic[EB/OL]. 2023. https://www.cs.cmu.edu/~callan/Papers/icml23-Luyu-Gao.pdf [64] LangChain API Reference — bind_tools (signature, tool_choice)[EB/OL]. accessed 2026-07-29. https://reference.langchain.com/python/langchain-core/language_models/chat_models/BaseChatModel/bind_tools/ [65] zenvanriel — "Why Senior Engineers Are Ditching LangChain for Plain Python" (the LLM proposes, the code executes)[EB/OL]. 2026-07-15. https://zenvanriel.com/ai-engineer-blog/ditching-langchain-for-plain-python/ [66] Ry Walker Research — Octomind (Jun-2024 post, 480 HN points, CEO response)[EB/OL]. 2026-06-11. https://rywalker.com/research/octomind [67] Octomind Blog — "Why we no longer use LangChain for building our AI agents"[EB/OL]. 2024-06. https://octomind.dev/blog/why-we-no-longer-use-langchain-for-building-our-ai-agents [68] Skywork — "Octomind and the Great Migration" (comparison of 1 vs 4 abstractions)[EB/OL]. 2025-10-14. https://skywork.ai/skypage/en/octomind-great-migration-teams-langchain/1976832104900653056 [69] Anthropic — "Building effective agents" (simple, composable patterns; reduce layers in production)[EB/OL]. 2024-12-19. https://www.anthropic.com/engineering/building-effective-agents [70] Theodo — "Don't use langchain anymore" (hidden costs, framework constraint)[EB/OL]. 2025-01-31. https://www.theodo.com/blog/dont-use-langchain-anymore-atomic-agents-is-the-new-paradigm [71] GitHub community Discussion #182015 (simple RAG → vanilla Python + Pydantic; 5+ layer stack traces)[EB/OL]. 2025-12-16. https://github.com/orgs/community/discussions/182015 [72] GroovyWeb — "LangChain vs LlamaIndex in 2026" (hybrid retrieval + orchestration pattern)[EB/OL]. 2026-06-23. https://www.groovyweb.co/blog/langchain-vs-llamaindex-comparison [73] Priorise — "RAG Framework Comparison 2026" (LlamaIndex indexes as LangChain tools)[EB/OL]. 2026-07-15. https://priorise.co/blog/rag-framework-comparison-langchain-vs-llamaindex-in-2026/ [74] arXiv — "ADK Arena: Evaluating Agent Development Kits" (top-5 >93% of downloads; LangChain ~233M/month)[EB/OL]. 2026-06-04. https://arxiv.org/html/2606.05548v1 [75] the-agent-report — "State of Agent Engineering 2026" (57.3% agents in production; 67% large enterprises; Datadog telemetry 9%→18%)[EB/OL]. 2026-05-23. https://the-agent-report.com/2026/05/state-of-agent-engineering-2026-langchain-datadog/ [76] alatirok — "State of AI Agent Adoption 2026" (1,340 respondents, fieldwork Nov 18–Dec 2, 2025; 57.3%)[EB/OL]. 2026-05-31. https://alatirok.com/ai-agent-adoption-2026/ [77] Stackademic — LangChain in Chains #54: Create Agents (create_agent signature)[EB/OL]. 2026-03-02. https://blog.stackademic.com/langchain-in-chains-54-create-agents-32362eaca10f [78] LangChain Docs — Middleware overview[EB/OL]. 2026-06-13. https://docs.langchain.com/oss/python/langchain/middleware/overview [79] LangChain Docs — Context engineering in agents[EB/OL]. 2026-07-28. https://docs.langchain.com/oss/python/langchain/context-engineering [80] LangChain Docs — Human-in-the-loop[EB/OL]. 2026-07-28. https://docs.langchain.com/oss/python/langchain/human-in-the-loop [81] LangChain Docs — Use the graph API (reducers, add_messages, Overwrite)[EB/OL]. 2026-07-27. https://docs.langchain.com/oss/python/langgraph/use-graph-api [82] LangChain Docs — Graph API overview (Command)[EB/OL]. 2026-07-28. https://docs.langchain.com/oss/python/langgraph/graph-api [83] LangChain Docs — Use time-travel (interrupt, replay, fork)[EB/OL]. 2026-06-11. https://docs.langchain.com/oss/python/langgraph/use-time-travel [84] LangChain Docs — Multi-agent (subagents/handoffs/skills/router; 40–50% savings)[EB/OL]. 2026-07-28. https://docs.langchain.com/oss/python/langchain/multi-agent [85] Reference — langgraph-supervisor create_supervisor[EB/OL]. 2026-07-27. https://reference.langchain.com/python/langgraph-supervisor/supervisor/create_supervisor [86] Reference — langgraph-swarm create_handoff_tool (+ internals active_agent, CSDN)[EB/OL]. 2026-07-12. https://reference.langchain.com/python/langgraph-swarm/handoff/create_handoff_tool [87] Anthropic — Building effective agents (workflows vs agents; simplest solution)[EB/OL]. 2024-12-19 (archived 2026-06-13). https://www.anthropic.com/research/building-effective-agents [88] zylos.ai — Finite State Machines and Statecharts for AI Agent Orchestration[EB/OL]. 2026-04-02. https://zylos.ai/research/2026-04-02-finite-state-machines-statecharts-ai-agent-orchestration/ [89] LangChain Docs — What's new in LangGraph v1 ("create_agent runs on LangGraph")[EB/OL]. 2026-07-26. https://docs.langchain.com/oss/python/releases/langgraph-v1 [90] LangChain Docs — Streaming (v2, subgraphs=True)[EB/OL]. 2026-07-28. https://docs.langchain.com/oss/python/langgraph/streaming [91] LangChain Docs — LangGraph v1 migration guide[EB/OL]. 2026-07-27. https://docs.langchain.com/oss/python/migrate/langgraph-v1 [92] LangChain Docs — Changelog (langchain v1.1.0: model retry / content moderation middleware)[EB/OL]. 2026-07-24. https://docs.langchain.com/oss/python/releases/changelog [93] arXiv 2603.27299 — Appendix B: LangGraph Implementation Notes (StateGraph, channels, Pregel, v1.1.3)[EB/OL]. 2026-03-28. https://arxiv.org/html/2603.27299v1 [94] ActiveWizards — LangGraph State Management: Checkpointing & Recovery (persistence decision table)[EB/OL]. 2026-05-06. https://activewizards.com/blog/langgraph-state-management-checkpointing-recovery-and-the-persistence-layer-decision/ [95] GitHub dbrowneup/Linus repo-notes/TradingAgents.md + juejin (StateGraph, InvestDebateState, SQLite checkpoints, deep/quick tiers; 30–50 calls via A-share fork)[EB/OL]. 2026-04-22 / 2026-06-30. https://github.com/dbrowneup/Linus/blob/main/docs/repo-notes/TradingAgents.md; https://juejin.cn/post/7656801430955704358 [96] Vectorize/Hindsight — LangGraph Short-Term State vs Long-Term Memory (checkpointer vs Store)[EB/OL]. 2026-07-17. https://hindsight.vectorize.io/guides/2026/07/17/guide-langgraph-state-vs-long-term-memory [97] Towards AI — LangGraph HITL: Pausing, Reviewing, Rewinding (time travel; "checkpoint between decision and consequence")[EB/OL]. 2026-06-15. https://pub.towardsai.net/langgraph-human-in-the-loop-pausing-reviewing-and-rewinding-your-agent-4028bd05b049 [98] ctaio.dev — Monolith, Handoff, or Swarm? (5-agent swarm ≈3× tokens) + skywork.ai (reliability 90%^10 = 34.9%)[EB/OL]. 2026-04-23 / 2026-04-15. https://ctaio.dev/en/labs/agentic-orchestration/topology-patterns/; https://skywork.ai/skypage/en/ai-agent-skills-2025-2026/2064636941351976960 [99] mer.vin — When Not to Build AI Agents (workflow-vs-agent playbook)[EB/OL]. 2026-05-22. https://mer.vin/2026/05/when-not-to-build-ai-agents-anthropics-workflow-vs-agent-playbook/ [100] arXiv 2505.00753v5 — LLM-Based Human-Agent Collaboration Survey (accountability gap in finance)[EB/OL]. 2026-05-06. https://arxiv.org/html/2505.00753v5 [101] LangChain Docs — Deep Agents overview (FilesystemMiddleware allowlist)[EB/OL]. 2026-07-28. https://docs.langchain.com/oss/python/deepagents/overview [102] LangChain Docs — Deep Agents human-in-the-loop (interrupt_on per subagent)[EB/OL]. 2026-07-03. https://docs.langchain.com/oss/python/deepagents/human-in-the-loop [103] LangChain Docs — Model Context Protocol (MCP)[EB/OL]. 2026-07-28. https://docs.langchain.com/oss/python/langchain/mcp [104] Reference — MultiServerMCPClient (stateless, handle_tool_errors)[EB/OL]. 2026-07-23. https://reference.langchain.com/python/langchain-mcp-adapters/client/MultiServerMCPClient [105] GitHub MeyerThorsten/TradingAgents (operational fork: 65-ticker watchlist, LLM cost vs P/L dashboard, FSI MCP)[EB/OL]. 2026-05-07. https://github.com/MeyerThorsten/TradingAgents [106] arXiv — TradingAgents: Multi-Agents LLM Financial Trading Framework[EB/OL]. v1 2024-12-28; v7 2025-06-03. https://arxiv.org/abs/2412.20138 [107] arXiv 2606.08283 — critique of controlled tests of financial multi-agent reasoning[EB/OL]. 2026-06-06. https://arxiv.org/html/2606.08283v1 [108] Investopedia — Bloomberg Terminal; GodelDiscount — Bloomberg Terminal Cost 2026[EB/OL]. 2025 / 2026-03-05. https://www.investopedia.com/terms/b/bloomberg_terminal.asp; https://godeldiscount.com/blog/bloomberg-terminal-cost-2026 [109] GlobalDatabase — Bloomberg vs Refinitiv vs S&P; UF Business Library FAQ[EB/OL]. 2026-02-16 / 2025-08-02. https://www.globaldatabase.com/bloomberg-vs-refinitiv-vs-sp-capital-iq-which-financial-terminal-is-worth-it; https://answers.businesslibrary.uflib.ufl.edu/genai/faq/401428 [110] WallStreetPrep — terminal comparison; Vendr — FactSet (real-world contracts ~$45K)[EB/OL]. 2025-04-07 / 2026-05-20. https://www.wallstreetprep.com/knowledge/bloomberg-vs-capital-iq-vs-factset-vs-thomson-reuters-eikon/; https://www.vendr.com/marketplace/factset [111] GodelDiscount — Bloomberg Terminal Cost 2026 (B-PIPE, Data License, enterprise feeds)[EB/OL]. 2026-03-05. https://godeldiscount.com/blog/bloomberg-terminal-cost-2026 [112] Official Massive blog — Polygon is now Massive; EINPresswire[EB/OL]. 2025-10-30 / 2025-11-03. https://massive.com/blog/polygon-is-now-massive/; https://www.einpresswire.com/article/863823068/polygon-io-is-now-massive [113] BrightData — Top 5 Stock Data Providers of 2026 (Massive plans)[EB/OL]. 2026-07-07. https://brightdata.com/blog/web-data/best-stock-data-providers [114] QVeris — Alpha Vantage pricing guide; alphavantage.co/premium[EB/OL]. 2026-07-24. https://qveris.ai/guides/alpha-vantage-pricing-alternative/; https://www.alphavantage.co/premium/ [115] apilayer — Best Financial Market APIs 2026; DailyTickers — Price Data APIs[EB/OL]. 2026-07-17. https://blog.apilayer.com/12-best-financial-market-apis-for-real-time-data-in-2026/; https://articles.dailytickers.com/series/finance-apis/part1-price-data/ [116] QuantVPS — Databento review; Databento blog — pricing update[EB/OL]. 2026-03-31 / 2026-05-29. https://www.quantvps.com/blog/databento-review; https://databento.com/blog/updates-to-subscription-pricing [117] DataResearchTools — data marketplaces; FXMacroData — vs Nasdaq Data Link[EB/OL]. 2026-04. https://dataresearchtools.com/best-data-marketplaces-dataset-websites/; https://fxmacrodata.com/articles/fxmacrodata-vs-quandl-nasdaq-data-link [118] Scrapfly — Yahoo Finance API guide; MarketXLS; ML4T provider audit[EB/OL]. 2026-07-24 / 2025-12-27. https://scrapfly.io/blog/posts/guide-to-yahoo-finance-api; https://marketxls.com/blog/yahoo-finance-api-ultimate-guide; https://www.ml4trading.io/docs/data/providers/PROVIDER_AUDIT/ [119] NYSE Proprietary Market Data Pricing Guide; Nasdaq Equity 7; Exegy — non-display fees[EB/OL]. 2026-05-14. https://www.nyse.com/publicdocs/nyse/data/NYSE_Market_Data_Pricing.pdf; https://listingcenter.nasdaq.com/rulebook/nasdaq/rules/nasdaq-equity-7; https://www.exegy.com/market-data-fees-planning-infrastructure/ [120] TheSmartInvestor — Benzinga Pro pricing; Benzinga APIs blog (MCP)[EB/OL]. 2025-03-14 / 2025-11-21. https://thesmartinvestor.com/investing/benzinga-free-plan-review/; https://www.benzinga.com/apis/blog/ [121] Apify — SEC EDGAR filings scraper docs; Fundamentalshub[EB/OL]. 2026-06-01 / 2026-05-15. https://apify.com/devilscrapes/sec-edgar-filings-scraper; https://fundamentalshub.com/blog/data-sec-gov-submissions-json [122] tldrfiling — SEC EDGAR XBRL API Python tutorial; apis.io OpenAPI spec[EB/OL]. 2026-04-10. https://tldrfiling.com/blog/sec-edgar-xbrl-api-python-tutorial; https://apis.io/apis/sec-edgar/sec-edgar-xbrl-api/ [123] tldrfiling — Free SEC EDGAR API guide; Fundamentalshub — submissions JSON[EB/OL]. 2026-07-24 / 2026-05-15. https://tldrfiling.com/blog/free-sec-edgar-api-guide/; https://fundamentalshub.com/blog/data-sec-gov-submissions-json [124] tldrfiling — EDGAR Full-Text Search API; explainx.ai[EB/OL]. 2026-04-09 / 2026-05-21. https://tldrfiling.com/blog/sec-edgar-full-text-search-api; https://explainx.ai/skills/sec.gov/search-edgar-fulltext-dpk6r2/search-edgar-fulltext [125] LangChain official docs — tools; Latenode — LangChain tools[EB/OL]. 2026-07-28 / 2026-06-11. https://docs.langchain.com/oss/python/langchain/tools; https://latenode.com/blog/langchain-tools [126] Thunderbit — 8 Best News APIs in 2026 (NewsAPI cliff)[EB/OL]. 2026-07-01. https://thunderbit.com/blog/best-news-apis-compared [127] Apify — GDELT scraper; gdelt-doc-api (GitHub); official GDELT blog[EB/OL]. 2026-07-18. https://apify.com/logiover/gdelt-news-scraper; https://github.com/alex9smith/gdelt-doc-api; https://blog.gdeltproject.org/announcing-the-gdelt-context-2-0-api/ [128] Massive × Benzinga (ratings, earnings, guidance, news $99/month)[EB/OL]. 2025-04-11 (verified 2026-07). https://massive.com/partners/benzinga [129] NexusFi — RavenPack; MediaWatcher — RavenPack alternatives[EB/OL]. 2026-04-17 / 2026-05-20. https://nexusfi.com/d/data-providers/ravenpack/; https://mediawatcher.ai/ravenpack-alternative/ [130] Apify — SEC Form 4 scraper; sec-api-python (GitHub)[EB/OL]. 2026-07-26 / 2026-03-31. https://apify.com/parseforge/sec-form4-scraper; https://github.com/janlukasschroeder/sec-api-python [131] edgarscout — 13F guide; sec-api.io — 13F holdings API[EB/OL]. 2026-05-28 / 2026-07-02. https://edgarscout.com/13f-filings/; https://sec-api.io/docs/form-13-f-filings-institutional-holdings-api [132] ForTraders — backtesting biases; StockFit — point-in-time (Plug Power case); HedgeFundAlpha[EB/OL]. 2026-07-25 / 2026-06-16 / 2026-04-13. https://fortraders.com/blog/how-to-avoid-bias-in-backtesting; https://developer.stockfit.io/blog/point-in-time-data-backtesting; https://hedgefundalpha.com/education/backtesting-mistakes-kill-quant-strategies-guide/ [133] PromptHub — RAG for Beginners (vector store vs SQL partitioning)[EB/OL]. accessed 2026-07-29. https://www.prompthub.us/blog/retrieval-augmented-generation-for-beginners [134] Streamkap — Flink financial market data; Mage — Kafka crypto pipeline[EB/OL]. 2026-02-25 / 2025-06-25. https://streamkap.com/resources-and-guides/flink-financial-market-data; https://www.mage.ai/blog/building-real-time-crypto-trading-pipelines-with-kafka-and-mage-pro [135] TraderMade — TimescaleDB tick→OHLC tutorial; DriTe Studio[EB/OL]. 2025-12-09 / 2026-02-17. https://tradermade.com/tutorials/6-steps-fx-stock-ticks-ohlc-timescaledb; https://dritestudio.co.th/article/timescaledb/ [136] Official Feast docs; GUVI — Feast feature store tutorial[EB/OL]. 2026-06-15 / 2026-07-10. https://docs.feast.dev/; https://www.guvi.in/blog/feast-feature-store-tutorial/ [137] QuantifiedStrategies — survivorship bias; StockFit — PIT backtesting[EB/OL]. 2026-04-01 / 2026-06-16. https://www.quantifiedstrategies.com/survivorship-bias-backtesting/; https://developer.stockfit.io/blog/point-in-time-data-backtesting [138] FinanceBench — arXiv 2311.11944 (Patronus AI; closed-book 9%, shared-store 19%, per-doc 50%, long-context 76%, oracle 85%; Llama2 70%/54% incorrect)[EB/OL]. 2023-11-20. https://arxiv.org/pdf/2311.11944 [139] meetdewey/financebench-eval — agentic RAG (Dewey+Claude Opus 4.6) 83.7%; full-context 76.0%[EB/OL]. 2026-04-05. https://github.com/meetdewey/financebench-eval [140] AlphaCreek — How to Build an SEC Filing Agent Without Naive RAG[EB/OL]. 2026-07-07. https://www.alphacreek.ai/blog/how-to-build-sec-filing-agent-without-naive-rag [141] Red Hat Developer — Stop chunking tables: agentic GraphRAG for financial disclosures with Docling[EB/OL]. 2026-07-22. https://developers.redhat.com/articles/2026/07/22/how-we-built-agentic-graphrag-financial-disclosures [142] arXiv 2505.20650 (citing XBRL-Agent, Han et al. 2024): +17% taxonomy / +42% numeric reasoning with RAG + symbolic calculators[EB/OL]. 2025. https://arxiv.org/html/2505.20650v3 [143] FinQA — arXiv 2109.00122 (8,281 QA; 11 professionals; S&P 500 earnings reports)[EB/OL]. 2021. https://sites.cs.ucsb.edu/~william/papers/FinQA.pdf [144] arXiv 2305.05862 — Are ChatGPT and GPT-4 General-Purpose Solvers for Financial Text Analytics? (GPT-4: 62.87 FinQA / 76.48 ConvFinQA; human 91.16 / 89.44)[EB/OL]. 2023-05. https://arxiv.org/pdf/2305.05862v1.pdf [145] ConvFinQA — arXiv 2210.03849 (3,892 conversations; human 89.44%)[EB/OL]. 2022-10. https://ar5iv.labs.arxiv.org/abs/2210.03849 [146] TAT-QA — ACL 2021.acl-long.254 (16,552 QA; TAGOP 58.0 F1 vs human 90.8 F1)[EB/OL]. 2021. https://aclanthology.org/2021.acl-long.254/ [147] FailSafeQA — 2024 study: hallucination in up to 41% of financial queries[EB/OL]. 2025-02-15. https://ajithp.com/2025/02/15/failsafeqa-evaluating-ai-hallucinations [148] mixpeek — Multimodal financial benchmarks leaderboard (15% scale errors; 10% figure hallucination; CoT 60.0→76.9%)[EB/OL]. 2025-12-05. https://github.com/mixpeek/multimodal-benchmarks/blob/main/finance/LEADERBOARD.md [149] LangChain docs — Split markdown (MarkdownHeaderTextSplitter + RecursiveCharacterTextSplitter)[EB/OL]. 2026-04-23. https://docs.langchain.com/oss/python/integrations/splitters/markdown_header_metadata_splitter [150] FinTech Studios — RAG for Finance: Beyond the Tutorial (+34% accuracy with speaker-turn chunking; vendor figure)[EB/OL]. 2026-02-28. https://www.fintechstudios.com/blog/rag-pipelines-financial-intelligence-best-practices [151] arXiv 1906.02868 — Modeling financial analysts' decision making via earnings calls (speaker turns as the unit; filter out operator and <10 tokens)[EB/OL]. 2019-06. https://arxiv.org/pdf/1906.02868.pdf [152] FutureAGI — What is LangChain? A 2026 Production Engineer's Guide[EB/OL]. 2026-05-14. https://futureagi.com/blog/what-is-langchain/ [153] official langchain-skills — RAG skill (pre-search metadata filtering)[EB/OL]. 2026-07 (accessed). https://github.com/langchain-ai/langchain-skills/blob/main/config/skills/langchain-rag/SKILL.md [154] supermemory.ai — Hybrid Search Guide (RRF: 91% vs 78%/65% recall@10; ~6ms overhead)[EB/OL]. 2026-04-23. https://supermemory.ai/blog/hybrid-search-guide [155] oneuptime.com — Hybrid Retrieval with LangChain (EnsembleRetriever RRF, 0.6/0.4 weights)[EB/OL]. 2026-02-17. https://oneuptime.com/blog/post/2026-02-17-how-to-implement-hybrid-retrieval-with-langchain-using-vertex-ai-vector-search/view [156] arXiv 2504.06293 — Generative AI Enhanced Financial Risk Management Information Retrieval (domain embeddings 88% HR@5)[EB/OL]. 2025. https://arxiv.org/pdf/2504.06293 [157] arXiv 2411.07142 — Greenback Bears and Fiscal Hawks (BAM embeddings +8 pts on FinanceBench)[EB/OL]. 2024-11. https://arxiv.org/pdf/2411.07142v1 [158] Elastic Search Labs — Cohere Rerank 3 (4k-token context; nDCG@10)[EB/OL]. 2025-09-12. https://www.elastic.co/search-labs/blog/elasticsearch-cohere-rerank [159] LangChain API Reference — ParentDocumentRetriever (children in vectorstore, parents in docstore)[EB/OL]. 2026-07-27. https://reference.langchain.com/python/langchain-classic/retrievers/parent_document_retriever/ParentDocumentRetriever [160] LangChain docs (Mintlify mirror) — Retrieval (RAG): search types similarity/mmr/threshold[EB/OL]. 2026-03-04. https://mintlify.com/langchain-ai/langchain/guides/retrieval [161] reference.langchain.com — MaxMarginalRelevanceSearchOptions (MMR lambda_mult / fetch_k)[EB/OL]. 2026-07-18. https://reference.langchain.com/javascript/langchain-core/vectorstores/MaxMarginalRelevanceSearchOptions [162] GitHub anthony-chaudhary/fak — Agent optimization methods survey (multi-query = production standard)[EB/OL]. 2026-06-23. https://github.com/anthony-chaudhary/fak/blob/main/docs/notes/RESEARCH-agent-optimization-methods-survey-2026-06-23.md [163] abstractalgorithms.dev — LangChain RAG in Practice (contextual compression ~2x latency)[EB/OL]. 2026-03-28. https://abstractalgorithms.dev/langchain-rag-retrieval-augmented-generation-in-practice [164] LangChain — Customers: Kensho (S&P Global); multi-agent framework on LangGraph with multi-stage exact-match evaluation of routing and tool calling[EB/OL]. 2026 (accessed). https://www.langchain.com/blog/customers-kensho [165] descope.com — DeepEval vs. RAGAS vs. LangSmith (evaluate retrieval and generation separately, then end-to-end)[EB/OL]. 2026-04-17. https://www.descope.com/blog/post/deepeval-vs-ragas-vs-langsmith [166] arXiv 2409.09046 — HyPA-RAG §5.2 (formal definitions of the 4 RAGAS metrics)[EB/OL]. 2024-09. https://arxiv.org/pdf/2409.09046 [167] qaskills.sh — Ragas Faithfulness & Answer Relevancy: The 2026 Guide (retrieval vs generation diagnosis; CI thresholds 0.85/0.80/0.75/0.80)[EB/OL]. 2026-06-27. https://qaskills.sh/blog/ragas-faithfulness-answer-relevancy-guide [168] LangChain/LangSmith docs — Evaluate a RAG application (structured correctness grader)[EB/OL]. 2026-07-28. https://docs.langchain.com/langsmith/evaluate-rag-tutorial [169] Financial-RAG-System (ICAIF-24 Finance RAG Challenge) — faithfulness ~86% in a financial pipeline on FinanceBench[EB/OL]. 2025-12-20. https://github.com/shivam1423/Financial-RAG-System [170] OpenAI Prompt Engineering Guide (official)[EB/OL]. accessed 2026-07-29. https://platform.openai.com/docs/guides/prompt-engineering [171] Fine-tuning of lightweight LLMs for sentiment classification on heterogeneous financial textual data (arXiv:2512.00946)[EB/OL]. 2025-11-30. https://arxiv.org/html/2512.00946v1 [172] Prompt Engineering in Finance: An LLM-Based Multi-Agent Architecture for Decision Support (ersj.eu)[EB/OL]. 2025. https://ersj.eu/journal/4220/download/ [173] Look-Ahead-Bench (arXiv:2601.13770; HAL 2026)[EB/OL]. 2026-01-20. https://arxiv.org/pdf/2601.13770; https://inria.hal.science/hal-05466549v1 [174] cat-llm: Best Practices for Classification (GitHub, empirical testing)[EB/OL]. 2026-06-03. https://github.com/chrissoria/cat-llm [175] EMNLP 2025 findings: prompting methods on FinQA/TAT-QA[EB/OL]. 2025. https://aclanthology.org/2025.findings-emnlp.1108.pdf [176] Production GenAI — LLM reasoning failures (technical compilation)[EB/OL]. 2026-02-25. https://github.com/h9-tec/Production_GenAI_interview/blob/main/sections/15-llm-reasoning-failures.md [177] Can Small Models Reason About Legal Documents? (arXiv:2603.25944)[EB/OL]. 2026-03-26. https://arxiv.org/html/2603.25944v1 [178] When Thinking Fails (arXiv:2505.11423)[EB/OL]. 2025-05-16. https://arxiv.org/abs/2505.11423 [179] Prompt repetition / inference-time techniques in stochastic forecasting (preprint) + FutureAGI glossary (self-consistency)[EB/OL]. 2026. https://rxiv.org/pdf/2602.0135v1.pdf; https://futureagi.com/glossary/self-consistency-prompting/ [180] Zhou et al., Least-to-Most Prompting (arXiv:2205.10625)[EB/OL]. 2022-2023. https://arxiv.org/pdf/2205.10625 [181] Test-time Scaling of LLMs: A Survey from a Subproblem Structure Perspective (arXiv:2511.14772)[EB/OL]. 2025. https://arxiv.org/html/2511.14772v1 [182] NumericBench: Exposing Numeracy Gaps (arXiv:2502.11075)[EB/OL]. 2025. https://arxiv.org/html/2502.11075v2 [183] Why AI Thinks 9.11 Is Bigger Than 9.9 (unrote)[EB/OL]. 2026-06-27. https://unrote.com/ai/nine-eleven-vs-nine-nine/ [184] Comprehension Without Competence (TMLR)[EB/OL]. 2025-10. https://openreview.net/pdf?id=Gz5HMiJLqv [185] AI Finance & Financial LLMs 2026 (youngju.dev)[EB/OL]. 2026-05-16. https://www.youngju.dev/blog/culture/2026-05-16-ai-finance-financial-llms-2026... [186] InvestPhilBench (arXiv:2606.25984)[EB/OL]. 2026. https://arxiv.org/html/2606.25984v2 [187] Toolformer (Failure-First daily paper)[EB/OL]. 2026-01-13. https://failurefirst.org/daily-paper/230204761/ [188] Structured Outputs Across LLM Providers (NiteAgent)[EB/OL]. 2026-06-04. https://niteagent.com/blog/2026-06-04-structured-outputs-across-providers/ [189] OpenAI JSON Mode vs Structured Outputs (AI JSONMedic)[EB/OL]. 2026-05-13. https://aijsonmedic.com/blog/openai-json-mode-structured-output [190] OpenAI Structured Outputs: Strict JSON Schema Mode (AI/TLDR)[EB/OL]. 2026-06-14. https://ai-tldr.dev/learn/llm-apis/function-calling/openai-json-schema-mode/ [191] LangChain langchain-openai: with_structured_output (official reference)[EB/OL]. 2026-07-08. https://reference.langchain.com/python/langchain-openai/chat_models/base/ChatOpenAI/with_structured_output [192] LangChain Structured output (official OSS docs)[EB/OL]. 2026-07-28. https://docs.langchain.com/oss/python/langchain/structured-output [193] GitHub langchain issue #34098: OutputFixingParser[EB/OL]. 2025-11-25. https://github.com/langchain-ai/langchain/issues/34098 [194] Agent开发面经: reasoning_effort / Thinking Budget (cnblogs)[EB/OL]. 2026-05-07. https://www.cnblogs.com/cwp0/p/19990412 [195] FinTradeBench (arXiv:2603.19225)[EB/OL]. 2026-06-03. https://arxiv.org/html/2603.19225v3 [196] Extended Thinking in Claude: production guide (developersdigest)[EB/OL]. 2026-04-29. https://www.developersdigest.tech/blog/extended-thinking-claude-production-guide [197] GitHub qwen-code issue #2508 (max_tokens > budget_tokens)[EB/OL]. 2026-03-19. https://github.com/QwenLM/qwen-code/issues/2508 [198] Fine-Tuning vs Prompting: When to Use Each (productgrowth)[EB/OL]. 2026-02-01. https://productgrowth.in/insights/ai-ml/fine-tuning-vs-prompting/ [199] Prompt Engineering vs Fine Tuning (Codecademy)[EB/OL]. 2025-07-27. https://www.codecademy.com/article/prompt-engineering-vs-fine-tuning [200] MetricGate — RiskMetrics EWMA Volatility Forecast[EB/OL]. 2026-05-20. https://metricgate.com/docs/ewma-volatility-riskmetrics/ [201] Ryan O'Connell, CFA — Portfolio VaR & Risk Decomposition[EB/OL]. 2026-03-30. https://ryanoconnellfinance.com/portfolio-var-risk-decomposition/ [202] Ryan O'Connell, CFA — VaR Methods Compared[EB/OL]. 2026-03-30. https://ryanoconnellfinance.com/var-methods-comparison/ [203] arXiv — Comparative Evaluation of VaR Models: HS, GARCH-Based Monte Carlo, and FHS[EB/OL]. 2025-05-08. https://arxiv.org/html/2505.05646v1 [204] arXiv 2203.02599 — A reverse ES(CVaR) optimization formula[EB/OL]. 2022. https://arxiv.org/pdf/2203.02599 [205] AL Capital Advisory — CVaR framework (Artzner et al. 1999 primary citations)[EB/OL]. 2026-07-09. https://alcapitaladvisory.com/research/frameworks/cvar.html [206] arXiv 2606.00320 — Adversarially Robust Control of CVaR via Rockafellar-Uryasev[EB/OL]. 2026-05-29. https://arxiv.org/html/2606.00320v1 [207] pfolio academy — Backtesting VaR: Kupiec and Christoffersen[EB/OL]. 2026-05-08. https://www.pfolio.io/academy/var-backtesting [208] MetricGate — VaR Backtesting (Christoffersen Test)[EB/OL]. 2026-05-05. https://metricgate.com/docs/var-backtesting-christoffersen/ [209] SAS/UPenn (Diebold) — Global Equity Market Volatility Spillovers[EB/OL]. n.d. https://www.sas.upenn.edu/~fdiebold/papers/misc/Gisler2015.pdf [210] MSCI (primary document) — RiskMetrics Technical Document, Fourth Edition[EB/OL]. 1996-12-17. https://www.msci.com/documents/10199/5915b101-4206-4ba0-aee2-3449d5c7e95a [211] Learnsignal — GARCH Model Explained[EB/OL]. 2026-06-24. https://www.learnsignal.com/blog/garch-model/ [212] Ryan O'Connell, CFA — Maximum Drawdown Calculator[EB/OL]. 2026-03-30. https://ryanoconnellfinance.com/calculators/maximum-drawdown-calculator/ [213] arXiv survey 2408.06361 — Large Language Model Agent in Financial Trading[EB/OL]. 2024-07-10. https://ar5iv.labs.arxiv.org/html/2408.06361 [214] Ernie's Trading Blog — Sharpe Ratio (Lo 2002, 65% overestimation)[EB/OL]. 2026-06-21. https://ernie55ernie.github.io/trading/2026/06/21/sharpe-ratio.html [215] Calcoi — Sortino Ratio Calculator[EB/OL]. 2026-04-30. https://calcoi.com/en/calculator/sortino-ratio-downside-risk-calculator/ [216] arXiv 1911.10254 — Omega and Sharpe ratio[EB/OL]. 2019. https://ar5iv.labs.arxiv.org/html/1911.10254 [217] MetricGate — Omega Ratio Calculator[EB/OL]. 2025-10-02. https://metricgate.com/docs/omega-ratio/ [218] QuantDecoded — Maximum Drawdown[EB/OL]. 2026-02-25. https://quantdecoded.com/en/maximum-drawdown-the-risk-metric-investors-fear-most [219] MyPivots Dictionary — Sterling Ratio[EB/OL]. n.d. https://www.mypivots.com/dictionary/definition/187/sterling-ratio [220] Marketopia — Sterling Ratio Guide[EB/OL]. 2026-06-13. https://www.marketopia.org/blog/sterling-ratio/ [221] CRAN — PerformanceAnalytics (UlcerIndex)[EB/OL]. 2026-04-09. https://cran.r-project.org/web/packages/PerformanceAnalytics/refman/PerformanceAnalytics.html [222] Schwab — What's the Information Ratio?[EB/OL]. 2026-05-18. https://www.schwab.com/learn/story/hows-that-fund-doing-check-information-ratio [223] Blank Capital Research — Grinold (1989): Fundamental Law of Active Management[EB/OL]. 2026-02-01. https://blankcapitalresearch.com/learn/grinold-fundamental-law-active-management [224] Ryan O'Connell, CFA — Skewness and Kurtosis[EB/OL]. 2026-05-09. https://ryanoconnellfinance.com/skewness-kurtosis-returns/ [225] MetricGate — VaR via Cornish-Fisher Expansion[EB/OL]. 2026-05-20. https://metricgate.com/docs/value-at-risk-cornish-fisher/ [226] ADVISORI — FRTB Expected Shortfall FAQ (traffic light, PLAT)[EB/OL]. 2026-04-08. https://www.advisori.de/services/regulatory-compliance-management/frtb/frtb-expected-shortfall-en [227] MetricGate — Component & Marginal VaR Calculator[EB/OL]. 2026-06-09. https://metricgate.com/docs/component-marginal-var-decomposition/ [228] Figy — Mathematics behind Diversification (Markowitz)[EB/OL]. 2025-11-07. https://figy.app/en/content/the-mathematics-behind-diversification [229] Santander (via HKMA mirror) — Scenario analysis (historical 2008/COVID, worst case ±3σ-±6σ)[EB/OL]. 2021. https://vpr.hkma.gov.hk/statics/assets/doc/100289/ar_21/ar_21_pt02_eng.pdf [230] Emergent Financial Group — Portfolio Risk During Market Stress[EB/OL]. 2026-06-05. https://emergentfingrp.com/knowledge-base/best-financial-planning-articles/how-portfolio-risk-changes-during-market-stress-lessons-from-1929-2008-2020-2022-and-todays-markets/ [231] arXiv 2601.03983 — Reverse Stress Testing Geopolitical Risk (cites BCBS 2009, p.14)[EB/OL]. 2025-09-05. https://arxiv.org/html/2601.03983v1 [232] S&P Global Market Intelligence — Reverse Stress Testing[EB/OL]. 2021-08-10. https://www.spglobal.com/market-intelligence/en/news-insights/research/reverse-stress-testing-assessment-tool-risk-managers-regulators [233] BIS — Stress testing principles (BCBS d450)[EB/OL]. 2018-10-17. https://www.bis.org/bcbs/publ/d450.htm [234] BIS FSI — Stress testing: Executive summary[EB/OL]. n.d. https://www.bis.org/fsi/fsisummaries/stress_testing.pdf [235] Risk Publishing — Quantitative Risk Management: Concepts, Techniques (three VaR failures)[EB/OL]. 2026-03-25. https://riskpublishing.com/quantitative-risk-management-concepts-and-tools/ [236] Risk Hub — Expected Shortfall Under FRTB (2.326σ vs 2.338σ)[EB/OL]. 2026-07-21. https://riskhub.org/blogs/expected-shortfall-under-frtb [237] AnalystPrep — FRTB and Market Risk Capital Framework[EB/OL]. 2026-04-17. https://analystprep.com/study-notes/frm/part-2/operational-and-integrated-risk-management/fundamental-review-of-the-trading-book-frtb/ [238] arXiv 2605.19337 — Agentic Trading: When LLM Agents Meet Financial Markets[EB/OL]. 2026-05-19. https://arxiv.org/html/2605.19337v1 [239] PyPI/libraries.io — wraquant (MCP server, 218 tools)[EB/OL]. 2026-04-02. https://libraries.io/pypi/wraquant [240] GitHub langchain-samples — risk-assessment-agent[EB/OL]. 2026-04-06. https://github.com/langchain-samples/risk-assessment-agent [241] GitHub — DART (Deep Adaptive Reinforcement Trader)[EB/OL]. n.d. https://github.com/ItzSwapnil/DART [242] New York Times Magazine (via Shareholder Forum) — "Risk Mismanagement", Joe Nocera (Taleb "a fraud", Einhorn "airbag")[EB/OL]. 2009-01-04. https://shareholderforum.com/sop/Library/20090104_NYT.htm [243] Picture Perfect Portfolios — Black Swan Events (correlations to 1.0)[EB/OL]. 2026-05-11. https://pictureperfectportfolios.com/black-swan-events-building-portfolios-that-survive-the-unexpected/ [244] ACM Digital Library — FAITH: Assessing Intrinsic Tabular Hallucinations in Finance[EB/OL]. 2025-11-14. https://dl.acm.org/doi/10.1145/3768292.3770433 [245] arXiv 2512.01123 — Hybrid Architecture for Options Wheel Strategy (LLM stochasticity)[EB/OL]. 2025. https://arxiv.org/html/2512.01123v1 [246] GitHub — VectorQuant (deterministic reasoning engine, verify_numeric)[EB/OL]. 2026-03-16. https://github.com/Sahilgitlab/VectorQuant [247] Risk Publishing — Model Risk Management: SR 11-7 (London Whale \(6.2B)[EB/OL]. 2026-07-23. https://riskpublishing.com/model-risk-management-sr-11-7-guidance/ [248] Signzy — SR 11-7: Federal Reserve Model Risk Management Guidance[EB/OL]. 2026-05-06. https://www.signzy.com/regulation-glossary/model-risk-management-SR-11-7 [249] Federal Reserve — SR 26-2 (rescinds SR 11-7; excludes GenAI/agentic AI; non-binding)[EB/OL]. 2026-04-17. https://www.federalreserve.gov/supervisionreg/srletters/SR2602.htm [250] Libertify — GenAI Model Risk Management in Finance[EB/OL]. 2026-04-18. https://www.libertify.com/interactive-library/genai-model-risk-management-financial-institutions-sr11-7/ [251] Michaud (1989): MVO as an "error maximization" device, FAJ 45(1):31–42 — quantdecoded.com[EB/OL]. 2026-01-17. https://quantdecoded.com/en/the-science-of-diversification-from-markowitz-to-modern-portfolios [252] Markowitz, "Portfolio Selection", Journal of Finance 7(1), 77–91 (1952) — via sesen.ai, MPT guide[EB/OL]. 2026-03-20. https://sesen.ai/blog/modern-portfolio-theory-markowitz-efficient-frontier [253] Merton (1980): Σ estimable, μ not; Σ⁻¹ amplifies errors; 2 p.p. example → weight 0.50→0.18 (Best & Grauer 1991) — ResearchGate[EB/OL]. 2025-08-06. https://www.researchgate.net/publication/5025643 [254] Closed-form GMV formula w = Σ⁻¹1/(1ᵀΣ⁻¹1) — mbrenndoerfer.com, MVO guide[EB/OL]. 2025-12-13. https://mbrenndoerfer.com/writing/modern-portfolio-theory-mean-variance-optimization [255] Tobin, "Liquidity Preference as Behavior Toward Risk", Review of Economic Studies 25(2), 65–86 (1958) — via Cambridge Core[EB/OL]. 1958. https://www.cambridge.org/core/journals/journal-of-financial-and-quantitative-analysis/article/E6A093D218636E28E03FD1FB8EC169C3 [256] Tangency = maximum Sharpe; CML; Tobin separation — pfolio.io academy[EB/OL]. 2026-03-28. https://www.pfolio.io/academy/tangency-portfolio [257] Sharpe, "Capital Asset Prices", Journal of Finance 19(3), 425–442 (1964); CAPM formula verified — MDPI JRFM 14(6):136[EB/OL]. 2026-05-29. https://www.mdpi.com/2227-7072/14/6/136 [258] Fama-French 3/5 factors: regression form, ~90% vs ~70% of variation explained, Ken French Data Library — ryanoconnellfinance.com[EB/OL]. 2026-03-04. https://ryanoconnellfinance.com/calculators/fama-french-calculator/ [259] Carhart 4 factors (1997), WML momentum factor — WallStreetMojo[EB/OL]. 2026-07-03. https://www.wallstreetmojo.com/carhart-four-factor-model/ [260] Fama & French, "A five-factor asset pricing model", JFE 116(1), 1–22 (2015) — arXiv 2605.30393 references[EB/OL]. 2026-05-28. https://arxiv.org/html/2605.30393v1 [261] "The Alpha Illusion" (2026): end-to-end LLM agent alpha is not deployment evidence; linguistic confidence ≠ tradable probability; priors = implicit factor exposures — arXiv 2605.16895[EB/OL]. 2026-05-16. https://arxiv.org/abs/2605.16895 [262] Black & Litterman, "Global Portfolio Optimization", FAJ 48(5), 28–43 (1992) — alcapitaladvisory.com[EB/OL]. 2026-07-09. https://alcapitaladvisory.com/research/frameworks/black-litterman.html [263] Alternative form of the BL posterior and π = λΣw_m — arXiv 2601.18811[EB/OL]. 2026. https://arxiv.org/html/2601.18811v1 [264] BL precision form; τ ∈ [0.01, 0.05] or τ=1/T; Idzorek ω_k = τ p_k Σ p_k′ — pictureperfectportfolios.com[EB/OL]. 2026-07-09. https://pictureperfectportfolios.com/breaking-down-the-black-litterman-model-for-optimal-asset-allocation/ [265] Kelly: G(f), f* = [bp−(1−p)]/b, partial losses f* = p/a − (1−p)/b — arXiv 2508.16598[EB/OL]. 2025. https://arxiv.org/html/2508.16598v1 [266] Fractional Kelly: Thorp ~half-Kelly; ~75% of the growth with ~50% of the volatility — quantmemo.com[EB/OL]. 2026-07-13. https://www.quantmemo.com/writing/paper-thorp-kelly-criterion-markets [267] ERC: Maillard, Roncalli & Teïletche (2010), JPM 36(4), 60–70; equalized risk contributions — arXiv 2508.11856[EB/OL]. 2025-08-16. https://arxiv.org/pdf/2508.11856 [268] Ledoit & Wolf 2003 (JEF 10(5):603–621) and 2004 (JMA 88(2):365–411) — arXiv 2605.30464 references[EB/OL]. 2026. https://arxiv.org/html/2605.30464 [269] LW estimator Σ̂ = ρT + (1−ρ)Σ; closed-form δ* — arXiv 2503.01581[EB/OL]. 2025. https://arxiv.org/pdf/2503.01581v2 [270] Numerical shrinkage example: 0.45×0.42 + 0.55×0.72 = 0.585 — alcapitaladvisory.com[EB/OL]. 2026-04-05. https://alcapitaladvisory.com/research/frameworks/ledoit-wolf.html [271] Choueifaty & Coignard, "Toward Maximum Diversification", JPM 34(4), 40–51 (2008) — Springer Empirical Economics[EB/OL]. 2026-03-10. https://link.springer.com/article/10.1007/s00181-026-02900-x [272] DR(ω) = Σᵢ ωᵢσᵢ/√(ω′Σω) — RDocumentation FRAPO[EB/OL]. 2026-01-25. https://rdrr.io/cran/FRAPO/man/DivRatios.html [273] No-short MDP solution: normalized w ∝ Σ⁻¹σ — MetricGate[EB/OL]. 2026-06-09. https://metricgate.com/docs/maximum-diversification-portfolio/ [274] DR² = number of independent risk factors (Choueifaty, Froidure & Reynier 2012) — Advisor Perspectives[EB/OL]. 2018/2025. https://www.advisorperspectives.com/articles/2018/11/12/portfolio-optimization-and-the-sharpe-multiplier-a-case-study-on-managed-futures [275] López de Prado, "Building Diversified Portfolios that Outperform Out of Sample", JPM 42(4), 59–69 (2016), DOI 10.3905/jpm.2016.42.4.059 — Revista Brasileira de Finanças[EB/OL]. 2026-05-16. https://periodicos.fgv.br/rbfin/article/view/97825 [276] skfolio: HRP Ward linkage; SchurComplementary (Cotton 2024) var(Schur) ≤ var(HRP); NCO (López de Prado 2019) — skfolio.org[EB/OL]. 2026-04-21. https://skfolio.org/generated/skfolio.optimization.SchurComplementary.html [277] PyPortfolioOpt: mean_historical_return → sample_cov → EfficientFrontier → max_sharpe flow; black_litterman and hierarchical_portfolio modules — CSDN[EB/OL]. 2026-04-15. https://blog.csdn.net/gitblog_00650/article/details/154044676 [278] Green & Hollifield (1992); Jobson & Korkie (1980, JASA 75:544–554); Britten-Jones (1999, JF 54:655–671) — Cambridge Core / Scientific Portfolio[EB/OL]. 2026-03-07. https://www.cambridge.org/core/journals/probability-in-the-engineering-and-informational-sciences/article/5778993FF88178B28275907D18E7AA81 [279] DeMiguel-Garlappi-Uppal 2009 (RFS 22:1915–1953): 1/N dominates out of sample; Kirby-Ostdiek (JFQA 2012): low-turnover MVO does beat 1/N net of costs — Cambridge Core JFQA[EB/OL]. 2026-07-22. https://www.cambridge.org/core/services/aop-cambridge-core/content/view/05D18E04B226C6E9B8441482CC92F943/S0022109012000117a.pdf [280] Jagannathan-Ma (2003): constraints = covariance shrinkage; DRO limit = 1/N (Pflug et al. 2012) — arXiv 2606.12612[EB/OL]. 2026-06-10. https://arxiv.org/html/2606.12612v1 [281] Abstract JM2003: with no-shorting, sample Σ performs like factor/shrinkage models — Scientific Portfolio[EB/OL]. 2025-11-16. https://scientificportfolio.com/external-research-anthology/ravi-jagannathan-tongshu-ma-2003/ [282] DeMiguel et al. (2009, Mgmt Sci): norm constraints nest Ledoit-Wolf, Jagannathan-Ma and 1/N — Scientific Portfolio[EB/OL]. —. https://scientificportfolio.com/anthology/generalized-approach/ [283] Net return with costs R = wᵀr − φ(Δw) — Portfolio Optimization Book (Palomar 2025), ch. 6.1[EB/OL]. 2025-05-01. https://portfoliooptimizationbook.com/book/6.1-fundamentals.html [284] Deflated Sharpe Ratio (Bailey & López de Prado 2014): full formula and E[max SR]; DSR > 0.95 rule — lessons.alejandrofernandezcamello.me[EB/OL]. 2026-06-17. https://lessons.alejandrofernandezcamello.me/machine-learning-for-alpha/backtest-overfitting-and-deflated-sharpe/ [285] Anti-overfitting toolkit: PSR, DSR, PBO/CSCV, Purged/Embargoed CV, Haircut Sharpe (Harvey-Liu 2015) — GitHub OutOfSampleLab/oos-lab[EB/OL]. 2026-06-24. https://github.com/OutOfSampleLab/oos-lab [286] LLM-Enhanced Black-Litterman (Kim, Lee et al. 2025): 100 queries/stock; top LLMs beat baselines; regime-dependent style — arXiv 2504.14345[EB/OL]. 2025-04-19 (v2: 2025-10-19). https://arxiv.org/abs/2504.14345 [287] Mechanics of LLM views in BL: q = mean of N=100; diagonal Ω = variance; P = I; GLS form — arXiv 2504.14345 (HTML)[EB/OL]. 2025-10-19. https://arxiv.org/html/2504.14345v2 [288] Lopez-Lira & Tang: GPT-4 post-cutoff, ~90% initial-reaction hit rate (not tradable), subsequent drift; returns decay with adoption — arXiv 2304.07619[EB/OL]. 2023-04-15 (v6: 2025-10-28). https://arxiv.org/abs/2304.07619 [289] Glasserman & Lin: bias = look-ahead + distraction; anonymized headlines perform better in-sample — arXiv 2309.17322[EB/OL]. 2023-09-29. https://arxiv.org/abs/2309.17322 [290] "Agentic Trading" (2026): 0 of 77 studies with full reproducibility; Profit Mirage (arXiv 2510.07920): information leakage in LLM financial agents — GitHub NyxFoundation[EB/OL]. 2026-07-18. https://github.com/NyxFoundation/interests/issues/20 [291] Rahimikia & Drinkall: FinText (50× smaller than LLaMA) beats FinBERT/LLaMA; Sharpe 3.45; robust to CAPM/FF/costs — Rebellion Research[EB/OL]. 2025-06-02. https://www.rebellionresearch.com/revisiting-large-language-models-in-finance-a-review [292] FinMem: layered memory, SR>2.0 and CR>0.35 (arXiv:2311.13743)[EB/OL]. 2023-11 / AAAI Symp. 2024. https://ar5iv.labs.arxiv.org/html/2311.13743 [293] FinAgent: multimodal, 92.27% return on 1 of 6 datasets (arXiv:2402.18485, KDD 2024)[EB/OL]. 2024-02-28. https://arxiv.org/abs/2402.18485 [294] StockAgent: anti-leakage simulation (arXiv:2407.18957, ACM TIST)[EB/OL]. 2024-07-15. https://arxiv.org/abs/2407.18957 [295] FinRobot: 4-layer platform + Smart Scheduler (arXiv:2405.14767)[EB/OL]. 2024-05-27. https://arxiv.org/pdf/2405.14767 [296] MarketSenseAI: +13% vs S&P 100, up to 40% (arXiv:2401.03737, ESWA)[EB/OL]. 2024-01-08. https://arxiv.org/abs/2401.03737 [297] MarketSenseAI: commercial conflict-of-interest statement[EB/OL]. 2024. https://arxiv.org/html/2401.03737v1 [298] arXiv 2510.07920, Profit Mirage: LLM agents fall to random baseline post-cutoff; best agent −50%; FinMem PC 0.8213[EB/OL]. 2025-10-09. https://arxiv.org/html/2510.07920v1 [299] arXiv 2505.07078v5: bias taxonomy, 0.1–0.9%/year distortions[EB/OL]. 2026-02-12. https://arxiv.org/html/2505.07078v5 [300] StockBench (arXiv:2510.02209): live benchmark Mar–Jul 2025, post-cutoff[EB/OL]. 2025-09-14. https://arxiv.org/html/2510.02209v1 [301] Chung & Tanaka-Ishii (ICAIF '23): ECC contextual features, +53–354 bps/trade PEAD OOS[EB/OL]. 2023-11. https://dl.acm.org/doi/10.1145/3604237.3626861 [302] UC Berkeley iSchool MIMS: ECC transcripts do not beat baseline (58.0% vs 61.2%)[EB/OL]. 2024-04-16. https://www.ischool.berkeley.edu/projects/2024/assessing-predictive-power-earnings-call-transcripts-next-day-stock-price-movement [303] Official PyPI nautilus_trader (Rust/PyO3 core, backtest-live parity)[EB/OL]. 2024-12-15. https://pypi.org/project/nautilus_trader/1.208.0/ [304] python.financial, The Python Backtesting Landscape (2026)[EB/OL]. 2026-03-02. https://python.financial/ [305] BullAlert, Best Python Backtest Engines in 2026 (zipline "glacial")[EB/OL]. 2026-05-18. https://bullalert.ai/blog/best-python-backtest-engines-2026 [306] AutoTradeLab, Backtrader vs NautilusTrader vs VectorBT vs Zipline ("execute it elsewhere")[EB/OL]. 2025-09-02. https://autotradelab.com/blog/backtrader-vs-nautilusttrader-vs-vectorbt-vs-zipline-reloaded [307] NYCServers, QuantConnect Review 2026 (LEAN: 10 years ≈ 33 s, T+3, fees/slippage/spread)[EB/OL]. 2026-02-15. https://newyorkcityservers.com/blog/quantconnect-review [308] Official GitHub QuantConnect/Lean (event-driven, backtest/live parity)[EB/OL]. accessed 2026-07-29. https://github.com/quantconnect/lean [309] DeepWiki nautilus_trader: three contexts, same Strategy class[EB/OL]. 2025-08-06. https://deepwiki.com/nautechsystems/nautilus_trader/1.1-getting-started [310] Quanthedge AI: DSR primary citation, Bailey & López de Prado JPM 40(5) 94–107 (2014)[EB/OL]. 2026-06-29. https://www.quanthedgeai.com/blog/the-deflated-sharpe-ratio-honestly-implemented/ [311] interactiveml.org, The Deflated Sharpe Ratio read closely (0.95 threshold)[EB/OL]. 2026-07-19. https://interactiveml.org/blog/deflated-sharpe-ratio [312] arXiv 1905.05023: PBO definition and CSCV procedure[EB/OL]. 2019. https://arxiv.org/pdf/1905.05023v1.pdf [313] MDPI Mathematics 14(12):2182: PBO→1 as N grows[EB/OL]. 2026-06-17. https://www.mdpi.com/2227-7390/14/12/2182 [314] Harvey & Liu, Backtesting (JPM 42(1), 13–28): non-linear haircut[EB/OL]. 2015. http://www.followingthetrend.com/?mdocs-file=2597 [315] C. Harvey, Jacobs Levy Center slides (Wharton): Sharpe 1.0 → 0.438 BHY / 0.232 Bonferroni[EB/OL]. 2015. https://jacobslevycenter.wharton.upenn.edu/wp-content/uploads/2015/05/3-Campbell-Harvey.pdf [316] Divitae Assets: implementation shortfall as the canonical measure; Almgren-Chriss with λ[EB/OL]. 2026-05-04. https://divitaeassets.com/insights/algorithmic-execution-strategies.html [317] curvedtrading.com: SMH borrow fee 0.7%/year, 1.4M availability (IBKR feed)[EB/OL]. 2026-07-22. https://curvedtrading.com/stocks/smh/ [318] Almgren & Chriss, Optimal Execution of Portfolio Transactions (primary paper PDF, eqs. 17–20)[EB/OL]. 1999/2001. https://quantitativebrokers.com/s/Optimal-Execution-of-Portfolio-Transaction-_-AlmgrenChriss-1999.pdf [319] SimTrade: differentiable efficient frontier at minimum cost; naive suboptimal[EB/OL]. 2025-12-28. https://www.simtrade.fr/blog_simtrade/understanding-almgren-chriss-model-for-optimal-trade-execution/ [320] arXiv 1701.03941: square-root law I=Yσ(Q/V)^δ; "1 day of vol per 1 day of volume" rule[EB/OL]. 2017. https://arxiv.org/pdf/1701.03941 [321] arXiv 2311.18283 (Durin, Rosenbaum, Szymanski): two square-root laws; MI∝√participation[EB/OL]. 2023-12-01. https://arxiv.org/pdf/2311.18283 [322] QuestDB Glossary, Market Access Rule 15c3-5 (mandatory kill switch in DMA)[EB/OL]. 2025-01-29. https://questdb.com/glossary/market-access-rule/ [323] pomegra.io Wiki: 15c3-5 non-delegable; Flash Crash origin[EB/OL]. 2026-06-08. https://pomegra.io/wiki/rule-15c3-5-market-access-rule/ [324] Alpaca Docs, Trading API (4x/2x buying power, shorting, advanced orders)[EB/OL]. 2025-09-24. https://docs.alpaca.markets/us/docs/trading-api [325] AlgoTrading101, Alpaca Trading API Guide (idempotent client_order_id, bracket, 200 req/min)[EB/OL]. 2021-02-28 (API still current). https://algotrading101.com/learn/alpaca-trading-api-guide/ [326] GitHub ib-api-reloaded/ib_async (ib_insync successor; paper port 7497; bracket orders)[EB/OL]. accessed 2026-07-29. https://github.com/ib-api-reloaded/ib_async [327] Horizon Trading: OMS/EMS distinction[EB/OL]. 2026-05-04. https://www.horizontrading.io/what-is-an-order-management-system-oms/ [328] Quod Financial: OMS as system of record with pre-trade controls; OEMS[EB/OL]. 2026-06-29. https://www.quodfinancial.com/what-is-an-order-management-system-oms-a-complete-guide-for-institutional-trading-desks/ [329] lobehub skills (integration-patterns): FIX message structure, key MsgTypes[EB/OL]. 2026-05-15. https://lobehub.com/zh-TW/skills/joellewis-finance_skills-integration-patterns [330] FIXopaedia B2BITS: FIX 4.2 dictionary tag 35 (D/8/F/G)[EB/OL]. accessed 2026-07-29. https://www.b2bits.com/fixopaedia/fixdic42/tag_35_MsgType.html [331] sanj.dev, FIX Protocol Production Patterns (gapless sequence numbers)[EB/OL]. 2026-04-03. https://sanj.dev/post/fix-protocol-production-patterns/ [332] ESMA statement May 30, 2024 (ESMA35-335435667-5924) — AI in investment services; para. 24 records[EB/OL]. May 30, 2024. https://veritaschain.org/blog/posts/2026-01-27-verification-imperative-eu-regulatory-convergence/ [333] PYMNTS — Moffatt v. Air Canada (2024 BCCRT 149): CA\)812.02; "remarkable submission"[EB/OL]. Feb 14, 2024 (decision). https://www.pymnts.com/news/artificial-intelligence/chatbot-tracker/2026/courts-tell-companies-they-own-what-their-chatbot-says/ [334] Sullivan & Cromwell — SR 26-2 / OCC 2026-13: rescinds SR 11-7; GenAI/agentic AI exclusion; $30B threshold[EB/OL]. Apr 29, 2026. https://www.sullcrom.com/insights/memo/2026/April/OCC-Fed-FDIC-Issue-Revised-Guidance-Model-Risk-Management [335] Containment.ai / Trepp — the GenAI/agentic AI governance gap after SR 26-2[EB/OL]. Apr–Jun 2026. https://containment.ai/blog/sr-26-02-genai-governance-gap.html; https://www.trepp.com/trepptalk/the-agencies-replace-sr-11-7 [336] Trepp — downstream consequences of GenAI components remain in scope[EB/OL]. 2026. https://www.trepp.com/trepptalk/the-agencies-replace-sr-11-7 [337] Fluxforce — SR 11-7: incomplete inventories and shadow models as exam findings[EB/OL]. n.d. https://www.fluxforce.ai/regulations/us-occ-sr-11-7-model-risk-management/ [338] Bank of England — PRA SS1/23 Model risk management principles for banks (includes AI; SMF)[EB/OL]. May-2023. https://www.bankofengland.co.uk/prudential-regulation/publication/2023/may/model-risk-management-principles-for-banks-ss [339] Regulations.ai — PDA conflicts rule (S7-12-23): proposed Jul 26, 2023, withdrawn Jun 17, 2025[EB/OL]. updated Jul-2026. https://regulations.ai/regulations/RAI-US-NA-S712230-2023 [340] Debevoise — SEC 2026 Examination Priorities (AI: representations, controls, disclosures)[EB/OL]. Nov 21, 2025. https://www.debevoise.com/-/media/files/insights/publications/2025/11/2026-sec-division-of-examinations-priorities.pdf [341] SEC PR 2024-31 — Delphia: AI washing, $225,000 (Mar 18, 2024)[EB/OL]. Mar 18, 2024. https://www.sec.gov/newsroom/press-releases/2024-31 [342] Debevoise / Mayer Brown — FINRA Regulatory Notice 24-09 (technology neutrality, Rule 3110)[EB/OL]. Jul-2024. https://www.debevoisedatablog.com/2024/07/01/finra-issues-brief-reminder-on-managing-generative-ai-risk-in-supervisory-tools/ [343] Smarsh / Cohasset — SEC Rule 17a-4: WORM or audit-trail (amended Oct-2022)[EB/OL]. n.d. (amended 2022). https://www.smarsh.com/regulations/sec-rule-17a-4-records-preservation/ [344] AI Act Blog / Licentium — post-Digital Omnibus timeline (high-risk postponed, Art. 50 intact)[EB/OL]. Jun–Jul 2026. https://www.aiactblog.nl/en/posts/article-50-transparency-deadline-2-august-2026; https://www.licentium.io/post/eu-ai-act-article-50-transparency-obligations-august-2026 [345] Comma Compliance — DORA (Regulation (EU) 2022/2554): ICT third parties, Arts. 28–30[EB/OL]. in force Jan 17, 2025. https://commacompliance.com/regulations/dora [346] Pertama Partners — MAS Consultation Paper on Guidelines on AI Risk Management (Nov 13, 2025)[EB/OL]. Feb 12, 2026. https://www.pertamapartners.com/insights/singapore-mas-ai-risk-management-guidelines-financial-services [347] Zyphe — Art. 50 AI Act and sanctions regime (EUR 15M / 3%)[EB/OL]. Jul-2026. https://www.zyphe.com/resources/news/eu-ai-act-transparency-obligations-deepfake-august-2026 [348] SEC PR 2024-167 — Rimar Capital: $310k fines, \(213,611 disgorgement, 5-year bar (Oct 10, 2024)[EB/OL]. Oct 10, 2024. https://www.sec.gov/newsroom/press-releases/2024-167 [349] SEC (AP File 3-22413) — Presto Automation: AI washing at a listed company, no fine due to cooperation (Jan 14, 2025)[EB/OL]. Jan 14, 2025. https://vorplabs.com/ai-regulatory-updates/federal-enforcement [350] LeapXpert — record-keeping fines timeline (>\)2,000M; JPMorgan $200M; 16 firms $1,100M, Sep 27, 2022)[EB/OL]. n.d. https://www.leapxpert.com/electronic-messaging-compliance-investigation-and-regulatory-fines-summary/ [351] Bank of England / PRA — Citigroup Global Markets Ltd fine: £33,880,000 PRA + £27,766,200 FCA (May-2024)[EB/OL]. May-2024. http://bankofengland.co.uk/news/2024/may/pra-fines-citygroup-global-markets-limited [352] Springer (European Business Organization Law Review) — AI and MiFID II: outputs explainable, auditable, consistent[EB/OL]. Feb-2026. https://link.springer.com/article/10.1007/s12027-026-00871-1 [353] Judiciary of Singapore — speech by Justice Philip Jeyaretnam (APAC Legal Congress 2026): open questions on deployer defense[EB/OL]. May 22, 2026. https://www.judiciary.gov.sg/news-and-resources/news/news-details/justice-philip-jeyaretnam--speech-at-the-asia-pacific-(apac)-legal-congress-2026 [354] Interviewbaba (citing Reuters) — Mata v. Avianca (SDNY, Jun 22, 2023): $5,000 sanction, 6 fabricated citations[EB/OL]. Jun 22, 2023. https://interviewbaba.com/ai-engineer-interview-questions/ [355] Blacksight — ChatGPT restrictions at JPMorgan, Goldman, Citi, BofA, Deutsche, Wells Fargo (2023)[EB/OL]. Apr 28, 2026. https://blacksight.ai/blog/jpmorgan-chatgpt-ban [356] Articsledge (citing Bloomberg 2023) — Samsung data leak via ChatGPT (3 incidents in 20 days)[EB/OL]. 2023. https://www.articsledge.com/post/ai-security [357] Warrant — mapping agent traces to MRM pillars (industry practice, not a standard)[EB/OL]. n.d. https://warrant.build/regulators/sr-11-7 [358] Moody's — "From SR 11-7 to SR 26-2: Managing model risk when models don't stand still"[EB/OL]. Jun 22, 2026. https://www.moodys.com/web/en/us/insights/banking/from-sr117-to-sr262-managing-model-risk-when-models-dont-stand-still.html [359] Brilo.ai (citing Xu et al. 2024) — hallucination as a non-eliminable property of LLMs[EB/OL]. 2024. https://www.brilo.ai/resources/ai-hallucination-cost-businesses-2024-statistics [360] European Parliament — written question E-001210/2026: AI Act cost €320,000–600,000 (high-risk)[EB/OL]. Mar 24, 2026. https://www.europarl.europa.eu/doceo/document/E-10-2026-001210_EN.html [361] Wavect — AI Act implementation estimates (transparency €15–31k; recurring 30–50%)[EB/OL]. n.d. https://wavect.io/blog/eu-ai-act-compliance-cost-startup/ [362] BIS FSI — FSB, "The Financial Stability Implications of Artificial Intelligence" (Nov 14, 2024)[EB/OL]. Jun 26, 2025. https://www.bis.org/fsi/fsisummaries/exsum_23904.htm [363] Kosmoy — Best LLM Observability & AI FinOps Platforms 2026[EB/OL]. 2026-07-15. https://www.kosmoy.com/resources/blog/best-llm-observability-ai-finops-platforms-2026/ [364] LangChain (official) — LangSmith Observability[EB/OL]. Accessed 2026-07-29. https://www.langchain.com/langsmith/observability [365] NextFuture — Langfuse vs Helicone hands-on 2026[EB/OL]. 2026-07-15. https://nextfuture.io.vn/blog/langfuse-vs-helicone-i-tested-both-for-llm-observability-2026 [366] LangChain Docs — LangSmith Evaluation Concepts[EB/OL]. Accessed 2026-07-29. https://docs.langchain.com/langsmith/evaluation-concepts [367] LangChain Docs — LangSmith Dashboards[EB/OL]. 2026-07-27. https://docs.langchain.com/langsmith/dashboards [368] LangChain Docs — LangSmith Alerts[EB/OL]. Accessed 2026-07-29. https://docs.langchain.com/langsmith/alerts [369] TenTrillionTriangles — The Open-Source Lie: 2026 Pricing (LangChain prices verified)[EB/OL]. 2026-06-15. https://tentrilliontriangles.com/blog/2026-06-15-the-open-source-lie-2026-pricing [370] UsagePricing — LangChain Blueprint (LCU/LSU restructuring Jul 21, 2026)[EB/OL]. Snapshots 2026-06-10 and 2026-07-21. https://www.usagepricing.com/blueprint/langchain [371] Langfuse (official) — Langfuse vs LangSmith[EB/OL]. 2026-07-24. https://langfuse.com/resources/engineering/langsmith-alternative [372] AgenticWire — Langfuse vs LangSmith self-hosted (ClickHouse acquisition)[EB/OL]. 2026-07-24. https://www.agenticwire.news/article/langfuse-vs-langsmith-self-hosted [373] genai.qa — AI Agent Trajectory Testing 2026 (Phoenix OTel/OpenInference)[EB/OL]. 2026-07-02. https://genai.qa/ai-agent-trajectory-testing-2026/ [374] Laminar — Braintrust Alternatives 2026 (Phoenix OSS / Arize AX pricing)[EB/OL]. 2026-06-29. https://laminar.sh/article/braintrust-alternatives-2026 [375] xseek.io — Best AI Observability Platforms 2026 (Galileo pricing)[EB/OL]. 2026-07-09. https://www.xseek.io/blogs/articles/best-ai-observability-platforms-in-2026-galileo-langsmith-more [376] Braintrust (official) — Best AI Evals Tools CI/CD[EB/OL]. 2026-05-21. https://www.braintrust.dev/articles/best-ai-evals-tools-cicd-2025 [377] Braintrust (official) — Braintrust vs promptfoo[EB/OL]. 2026-04-29. https://www.braintrust.dev/articles/braintrust-vs-promptfoo [378] Kosmoy — Best AI Agents Management Platforms 2026[EB/OL]. 2026-07-16. https://www.kosmoy.com/resources/blog/best-ai-agents-management-platforms-2026/ [379] FinanceBench paper (Patronus AI), arXiv 2311.11944[EB/OL]. 2023 (accessed 2026-07-29). https://ar5iv.labs.arxiv.org/html/2311.11944 [380] FinAgent-RAG, arXiv 2605.05409 (FinQA/ConvFinQA/TAT-QA sizes)[EB/OL]. 2026. https://arxiv.org/pdf/2605.05409 [381] Morph — AI Agent Evaluation (three layers; LLM-as-judge limits)[EB/OL]. 2026-06-20. https://www.morphllm.com/ai-agent-evaluation [382] AIFinHub — Hallucination Detector: Numeric Source Grounding (EPS $0.81 vs \(0.78 case)[EB/OL]. 2026-05-25. https://aifinhub.io/articles/hallucination-detector-numeric-source-grounding/ [383] World Bank — Proof-Carrying Numbers (official GitHub)[EB/OL]. 2025-12-15. https://github.com/worldbank/pcn [384] qaskills.sh — Promptfoo LLM Testing Guide (CI gates)[EB/OL]. 2026-07-01. https://qaskills.sh/blog/promptfoo-llm-testing-guide [385] SECQUE, arXiv 2504.04596 (financial benchmarks landscape, LLM-as-judge)[EB/OL]. 2025. https://arxiv.org/pdf/2504.04596 [386] is4.ai — Guardrails AI vs NeMo Guardrails Comparison 2026[EB/OL]. 2026-03-26. https://is4.ai/blog/our-blog-1/guardrails-ai-vs-nemo-guardrails-comparison-2026-352 [387] Rajath J. Bosco — LLM Financial Agent Hallucination (15%→8%→2%)[EB/OL]. 2024-02-21. https://www.rajathjohn.com/writing/llm-financial-agent-hallucination [388] ai-tldr.dev — What is Microsoft Presidio[EB/OL]. 2026-06-14. https://ai-tldr.dev/learn/evaluation-safety/alignment-safety/what-is-microsoft-presidio/ [389] Grepture — Grepture vs Presidio (OSS MIT, self-host)[EB/OL]. 2026-07-05. https://grepture.com/compare/grepture-vs-presidio [390] GitHub shivshankar20 — PII Anonymization with Presidio (risk-based strategy)[EB/OL]. 2026-03-09. https://github.com/shivshankar20/pii-anonymization-presidio [391] OneUptime — LLM Rate Limiting[EB/OL]. 2026-01-30. https://oneuptime.com/blog/post/2026-01-30-llm-rate-limiting/view [392] Reintech — LLM Rate Limiting & Quota Management: Production Best Practices[EB/OL]. 2025-12-31. https://reintech.io/blog/llm-rate-limiting-quota-management-production-best-practices [393] TrueFoundry — Rate Limiting AI Agents: Preventing LLM API Exhaustion[EB/OL]. 2026-05-12. https://www.truefoundry.com/blog/rate-limiting-ai-agents-preventing-llm-api-exhaustion [394] GitHub kagenti — Issue #1274 (\)47k/11-day incident, ZenML analysis)[EB/OL]. 2026-04-17. https://github.com/kagenti/kagenti/issues/1274 [395] Fountain City — AI Agent Cost Circuit Breaker (incidents, 2.5×/5× thresholds)[EB/OL]. 2026-04-06. https://fountaincity.tech/resources/blog/ai-agent-cost-circuit-breaker/ [396] SupraWall — AI Agent Runaway Costs[EB/OL]. 2026-01-01. https://www.supra-wall.com/learn/ai-agent-runaway-costs [397] arunbaby.com — AI Agents: Error Handling & Recovery (circuit breaker types, requestId)[EB/OL]. 2025-01-24. https://www.arunbaby.com/ai-agents/0033-error-handling-recovery/ [398] Alice Labs — LangGraph Guide 2026 (enterprise checklist, interrupt_before)[EB/OL]. 2026-05-23. https://alicelabs.ai/en/insights/langgraph-guide-2026 [399] PlanTwin, arXiv 2603.18377 (isolation: Firecracker, gVisor, Kata)[EB/OL]. 2026-03-01. https://arxiv.org/html/2603.18377v1 [400] Beam — Best Stateful Sandbox Code Execution 2026 (E2B/Modal/Northflank costs)[EB/OL]. 2026-06-05. https://www.beam.cloud/blog/best-stateful-sandbox-code-execution-2026 [401] BenchLM — OpenAI API Pricing (synced with official page)[EB/OL]. Sync 2026-07-28. https://benchlm.ai/openai/api-pricing [402] BenchLM — Anthropic API Pricing (synced with official pricing)[EB/OL]. Sync 2026-07-28. https://benchlm.ai/anthropic/api-pricing [403] BenchLM — Google API Pricing (synced with Gemini Developer API pricing)[EB/OL]. Sync 2026-07-28. https://benchlm.ai/google/api-pricing [404] Klymentiev — LLM Router (difficulty distribution 60–75/15–25/5–15)[EB/OL]. 2026-05-10. https://klymentiev.com/blog/llm-router [405] Coursiv — Claude Pricing 2026 (cache write, fast mode, tokenizer)[EB/OL]. 2026-07-25. https://coursiv.io/blog/claude-pricing-2026 [406] VibeEngines — LLM API Pricing Handbook (batch −50%, cache −90% across the board)[EB/OL]. 2026-07-21. https://vibeengines.com/handbook/llm-api-pricing [407] SumatoSoft — AI Cost Reduction Playbook (Batch API 50% off)[EB/OL]. 2026-07-01. https://sumatosoft.com/blog/ai-cost-reduction-playbook [408] Pinnacle J. (JSISI) — Comparison of agentic strategies in finance[EB/OL]. 2026-05-13. https://pinnaclepubs.com/index.php/JSISI/article/view/733 [409] G. Avodagbe — LLM Inference Costs Exploding: Cut 60% (Thomson Reuters; Singapore bank \(180k→\)71k)[EB/OL]. 2026-07-22. https://godwinavodagbe.com/llm-inference-costs-exploding-cut-60-percent/ [410] Parse — Semantic Caching Solutions (40–70% hit rates, 0.85–0.90 threshold)[EB/OL]. Sources 2026-07-08. https://parse.gl/prompts/p/my-goal-is-to-cache-llm-responses-to-reduce-latency-and-cost-whats-the-best-semantic-caching-solution--521a09eb-9569-4300-9cd9-c3b785f4603e [411] Parse — LLM Caching & Semantic Caching Tools (RedisSemanticCache, LangCache −73%)[EB/OL]. Sources 2026-04/07. https://parse.gl/prompts/p/i-am-planning-to-optimize-our-token-usage-costs-who-offers-llm-caching-and-semantic-caching-tools--e3288045-7775-48e0-b580-096cb4de552d [412] NeuralTrust — LLM Model Routing (RouteLLM, FrugalGPT, gateways)[EB/OL]. 2026-07-23. https://neuraltrust.ai/blog/llm-model-routing [413] TechnoLynx — RouteLLM Explained (per-segment calibration)[EB/OL]. 2026-07-11. https://www.technolynx.com/post/routellm-explained-how-model-routing-cuts-llm-inference-cost/ [414] GitHub simonlin1212 — TradingAgents-astock (~30–50 LLM calls per analysis)[EB/OL]. 2026-05-12. https://github.com/simonlin1212/TradingAgents-astock [415] skywork.ai — Most Critical Skills for AI Agents in 2025 & 2026 (reliability 90%^10 ≈ 34.9%)[EB/OL]. 2026-04-15. https://skywork.ai/skypage/en/ai-agent-skills-2025-2026/2064636941351976960 [416] Kunal Ganglani — AI Agent Latency Optimization Budget (2–5 hops, P99, SLO tiers)[EB/OL]. 2026-07-06. https://www.kunalganglani.com/blog/ai-agent-latency-optimization-budget [417] McNeil, A. J., Frey, R. & Embrechts, P. — Quantitative Risk Management: Concepts, Techniques and Tools (Princeton University Press; EVT/POT/GPD, chs. 5 and 7). 2005. [418] Cherubini, U., Luciano, E. & Vecchiato, W. — Copula Methods in Finance (Wiley; non-Gaussian tail dependence). 2004. [419] J.P. Morgan — CreditMetrics Technical Document (Gupton, Finger & Bhatia; rating migrations and credit VaR). 1997-04-02. [420] agentmarketcap.ai — AI Agent Framework Lock-In (breaking-change cycles; version pinning)[EB/OL]. 2026-04-06. https://agentmarketcap.ai/blog/2026/04/06/ai-agent-framework-lock-in-langchain-crewai-autogen-migration-costs [421] tessl.io — Riskfolio-Lib 7.0.0 PyPI record (24 convex risk measures, HRP/HERC/NCO, cvxpy engine)[EB/OL]. 2026-01-29. https://tessl.io/registry/tessl/pypi-riskfolio-lib/7.0.0