{"paper_id":"2507.08584","title":"To Trade or Not to Trade: An Agentic Approach to Estimating Market Risk Improves Trading Decisions","authors":["Dimitrios Emmanoulopoulos","Ollie Olby","Justin Lyon","Namid R. Stillman"],"paper_date":"2025-07-11","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2507.08584v1","page_url":"https://thequant.space/flowcharts/to-trade-or-not-to-trade-an-agentic-approach-to-estimating/","scores":{"empirical_rigor":6,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes an agentic system that uses large language models to iteratively discover stochastic differential equations for financial time series, generating risk metrics to inform daily trading decisions.","quote":"We address this gap by developing an agentic system that uses LLMs to iteratively discover stochastic differential equations for financial time series.","summary":"Large language models (LLMs) are increasingly deployed in agentic frameworks, in which prompts trigger complex tool-based analysis in pursuit of a goal. While these frameworks have shown promise across multiple domains including in finance, they typically lack a principled model-building step…","strategy":"model-informed LLM trading strategy"},"assumptions":{"status":"available","items":[{"assumption":"The Brownian motion increment dW_t is sampled from a normal distribution with mean zero and standard deviation proportional to the time step.","type":"model","stated":"explicit","quote":"the stochastic term includes composition with Brownian motion, ◦dWt, which we assume to be sampled from the normal distribution with mean zero and standard deviation proportional to the time step, dWt∼N(0,dt)."},{"assumption":"The price process S_t is a continuous R^n‑valued process representing asset price dynamics.","type":"model","stated":"explicit","quote":"we assume that St is a continuous Rn-valued process representing the evolution of price dynamics of an asset."},{"assumption":"Only a single independent variable (price) is modeled; volatility and other factors are not included.","type":"model","stated":"explicit","quote":"we restrict the space of SDEs to a single independent variable, considering only the evolution of the price, St, and not, for example stochastic models of both volatility and price paths."},{"assumption":"Transaction costs are a fixed proportion κ of the trade value, with κ = 8.8 bps.","type":"execution","stated":"explicit","quote":"for an assumed transaction charge, κ, given in basis points (bps). NYSE stocks are estimated to have an average charge of 8.8 bps of the total value of the trade, such that κ = 8.8 × 10−4."},{"assumption":"Calibration loss is measured as mean absolute error between synthetic and historical price paths.","type":"statistical","stated":"explicit","quote":"we calculate the loss, L, as the MAE between synthetic and historical price paths, L(θ)=MAE(St;θ,S∗t;θ)= ΣNsims i |St,i,θ−S∗t,i,θ| / N sims"},{"assumption":"All LLMs are pre‑trained on a large portion of the internet up to their cut‑off date, which may introduce bias.","type":"other","stated":"explicit","quote":"Assuming that all LLMs have been trained on a significant proportion of the internet up to their cut‑off date, we cannot preclude the possibility of a biasing factor from their pre‑training data affecting their trading performance."}]},"datasets":{"status":"available","items":[{"name":"Simudyne Horizon synthetic market and news generator","provider":"Simudyne Horizon","kind":"synthetic","access":"unknown","period":"2023-01-01 to 2023-07-01","universe":"S&P 500","quote":"generate an alternative price path and synthetic news cycle for the S&P 500 covering the 6 month trading period from 1st January 2023 to 1st July 2023."},{"name":"Historical price data S∗t","provider":null,"kind":"market-prices","access":"unknown","period":null,"universe":null,"quote":"given some historical price data, S∗t , and model-specific parameters, θ."}],"known":[],"sample":{"start":"2024-09","end":"2025-04","market":"US equities (e.g., AAPL)","asset_class":"equities","frequency":"daily"},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":true,"detail":"Transaction cost TCt = κ Pt St with κ = 8.8×10⁻⁴ (8.8 bps) per trade.","quote":"for an assumed transaction charge, κ, given in basis points (bps). NYSE stocks are estimated to have an average charge of 8.8 bps of the total value of the trade, such that κ = 8.8 × 10−4.","source":"front_matter","rule":"R3","candidates":{"front_matter":"yes","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"Transaction cost TCt = κ Pt St with κ = 8.8×10⁻⁴ (8.8 bps) per trade.","card_quote":"for an assumed transaction charge, κ, given in basis points (bps). NYSE stocks are estimated to have an average charge of 8.8 bps of the total value of the trade, such that κ = 8.8 × 10−4.","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"Model‑informed trading strategies outperform standard LLM‑based agents, yielding higher Sharpe ratios across multiple equities.","kind":"performance","quote":"We find that model‑informed trading strategies outperform standard LLM‑based agents, improving Sharpe ratios across multiple equities."},{"finding":"Adding the model‑discovery loop increases Sharpe ratio by an average of 22% compared with using news‑only context.","kind":"performance","quote":"We see an average 22% increase in SR when comparing LLMs using contextN to contextN M."},{"finding":"Most LLMs outperform a simple buy‑and‑hold strategy when using news‑only signals, and performance further improves with model discovery.","kind":"performance","quote":"We find that most LLMs are able to outperform the simple buy-and-hold strategy by using news stories only, contextN. The trading results are further improved when combined with a model discovery loop."},{"finding":"Sonnet 3.5v2, Sonnet 3.7, and 4o‑mini achieve profitable trading decisions in the synthetic backtest, while R1 and o1‑mini perform worse.","kind":"performance","quote":"Sonnet 3.5v2, and Sonnet 3.7 and 4o-mini resulted in profitable trading decisions across the synthetic backtest."},{"finding":"No LLM outperforms the buy‑and‑hold baseline for NVIDIA due to its large 47% return over the period.","kind":"negative","quote":"for NVIDIA, no LLM is able to outperform the buy-and-hold baseline. This is likely due to the drastic increase in value of the NVIDIA stock over this period (roughly 47% returns)."}],"reported_performance":{"strategy":"model-informed LLM trading strategy","sharpe":4.56,"sharpe_benchmark":null,"annual_return_pct":null,"total_return_pct":null,"max_drawdown_pct":7},"out_of_sample":{"method":"holdout","headline_out_of_sample":true}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":44,"citations":{"cited_by_in_corpus":1,"by_independent_authors":1,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":1,"unclassified":0},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":1,"extracted_at":"2026-11-14T16:38:00Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}