Every day, we express ourselves in 500 million tweets and 64 billion WhatsApp messages. On Facebook, 864 million of us log in to post status updates, comment on news stories, and share videos. We enable our clients to quickly extract value and insights from large amounts of information. Explore News and Job Analytics with a knowledge graph across 12 million business-relevant entities. By keeping institutional investors and traders apprised of critical earnings date revisions, they
can take advantage of – or avoid – short-term volatility in a given security.

1. Peter Hafez, Chief Data Scientist at RavenPack, analyzes the impact and limitations of LLMs in stock price predictions.
2. We will handle your personal information in compliance with our Privacy Statement.
3. By providing your personal information and submitting your details, you acknowledge that you have read, understood, and agreed to our Privacy Statement and you accept our Terms and Conditions.
4. We attempt to rationalise whether a stock is „cheap‟ or „expensive‟ for a reason, or „mispriced‟.

By providing your personal information and submitting your details, you acknowledge that you have read, understood, and agreed to our Privacy Statement and you accept our Terms and Conditions. We will handle your personal information in compliance with our Privacy Statement. You can exercise your rights of access, rectification, erasure, restriction of processing, data portability, and objection by emailing us at in accordance with the GDPRs. You also are agreeing to receive occasional https://g-markets.net/ updates and communications from RavenPack about resources, events, products, or services that may be of interest to you. We value your privacy and assure you that we will handle your personal information with the utmost care and in compliance with our Privacy Statement. You can exercise your rights of access, rectification, erasure, restriction of processing, data portability, and objection by emailing us at in accordance with the General Data Protection Rules (GDPR).

At the 1-day mark, annualized returns reach 8.0% for the mid-/large-caps and 19.7% for small-caps, with information ratios of 0.8 and 1.2, respectively. Investors and researchers have suspected for decades that text could be used to predict markets, some trying and failing. As other studies have demonstrated and the RavenPack study confirms, there is immense value in staying on top of corporate events like quarterly earnings reports and their changes. Anyone familiar with the Wall Street Horizon DateBreaks Factor or Late Earnings Report Index (LERI) knows that academic research supports the idea that companies that advance their earnings date tend to share good news on their earnings calls, while those that delay tend to share bad news. The RavenPack Earnings Dates dataset consists of Wall Street Horizon earnings calendar
change records for over 8,000 stocks globally since 2006.

Identifying the Drivers of Stock Returns

The company’s products allow clients to enhance returns, reduce risk and increase efficiency by systematically incorporating the effects of public information in their models or workflows. RavenPack’s clients include the most successful hedge funds, banks, and asset managers in the world. Financial professionals rely on RavenPack for its speed and accuracy in analyzing large amounts of unstructured content. The company’s products allow clients to enhance returns, reduce risk or increase efficiency by incorporating the effects of public information in their models or workflows.

Again, small-cap companies were more sensitive to the changes and therefore exhibited greater price reactions between advance and delay events than the mid-/large-cap companies. Ke, Kelly, and Xiu created a model that essentially automatically generates a dictionary of relevant words and allows for contextually specific sentiment scores. Using supervised machine learning and a method that required only a laptop and basic statistical capabilities, the researchers analyzed more than 22 million articles published from 1989 to 2017 by Dow Jones Newswires.

Insights and Limitations in Stock Price Prediction Using LLMs

Prior to joining the firm, he was Director of Product & Strategy at Third Point, a $15 billion Event-Driven Hedge Fund, where he co-founded and led their data science team. Aakarsh was previously Senior Product Manager at FactSet, responsible for re-architecting & launching a web-based analytics platform servicing over 100,000 enterprise clients. Figure 8 shows that both mid-/large- and small-cap long-only strategies decay more slowly than short-only; additionally, advance events have higher momentum than delay events.

Schedule a personalized trial to see how our actionable insights can boost your investment strategies. As mentioned in the previous two points, results were most profound for small-caps, those companies with revenues below $250M. This is not surprising considering these companies have lower liquidity and are therefore more volatile in nature. It’s important to note that this analysis doesn’t diminish the potential value of LLMs in systematic investing. While achieving Sharpe Ratios above 3 may pose challenges, the RavenPack Data Science team remains optimistic about the applications of LLMs in finance based on internal research and we anticipate sharing more of our findings on this topic throughout 2024. Peter Hafez, Chief Data Scientist at RavenPack, analyzes the impact and limitations of LLMs in stock price predictions.

To complement the slew of new reports, Wall Street Horizon looks at recent findings from RavenPack that continue to highlight the important information that can be culled from the timing of earnings dates. Internal research within RavenPack suggests that the outcomes can be sensitive not only to the version of the GPT model they used but also to the strategy implementation. The robust performance depicted in the paper relies heavily on the assumption of attaining the open-price, a scenario proven impractical in real-world contexts.

Financial firms have long understood that actionable information is increasingly found in the oceans of news and digital content available. In the nearly 20 years since the technology firm was founded, RavenPack has built a sterling reputation on Wall Street for the unparalleled breadth and quality of its low-latency text processing and data products. With Edge, RavenPack sets a new standard for its traditional user base, and further extends its reach by helping non-financial firms better mitigate risk exposures in investments, supply chain, client compliance, reputation management, competitive analysis, and sustainability. Capable of understanding content in 13 different languages, Edge can extract insights from all types of documents —from short news articles to complex legal filings. As a result of the observable differences in how advance and delay events affect stock performance, the first strategy pursued by RavenPack trades stocks just ahead of the confirmed earnings announcement date, going long on stocks that advance their earnings date and shorting those that delay. As Figure 6 shows, companies that advanced an earnings date outperformed those that delayed after the change was made.

General Business Overview

Chicago Booth Review
Research driven insights on business, policy, and markets. Some funds have likely been using natural language processing to trade for several years, with dubious success. A 2016 article in MIT Technology Review called analyzing language data to predict markets “one of the most promising uses of new AI techniques,” but one of the handful of funds it mentioned, Sentient, liquidated in 2018. The research by Ke, Kelly, and Xiu provides an academic framework for applying such processing to markets. 80+ fields describe every entity detected including over 20 sentiment indicators. Contextual analytics on temporal, territory, segment and similarity aspects are also delivered.

Insights

Figure 3 below shows how mid-/large-cap companies have a greater reaction to delayed dates and experience more momentum on the negative leg, while the small-caps react more to advanced dates resulting in greater momentum on the positive leg. The signal, although strong, decays ravenpack pricing relatively quickly, with the difference between the average advance and delay reactions reaching a peak in just a few days. For every entity and event detected in a story, RavenPack provides advanced analytics including relevance scoring, novelty tracking, and impact analysis.

Since 2003, RavenPack has pioneered investment-grade sentiment analysis in financial services. We do not believe in “one size fits all” and have developed multiple sentiment techniques where some leverage millions of rule sets while others use sophisticated machine learning algorithms. Another Wall Street Horizon proprietary metric that considers these important earnings announcement changes and therefore offers a view on corporate confidence is the Late Earnings Report Indicator (LERI).

Five Research Findings on Earnings Date Timing That Affect Trading

Her research has been widely featured in financial news outlets including regular appearances on networks such as CNBC and Fox Business to talk about corporate earnings and the economy. Ms. Short earned a BA in International Relations and English from Fairfield University. Both mid-/large- and small-caps long-only strategies decay more slowly than short-only, and advance events have higher momentum than delay events. This is consistent with findings from Figure 6 where the short-only strategy does not perform well, as delayed events do not deviate from zero significantly. RavenPack also found that stocks not only reacted to what was being shared in quarterly earnings reports as predicted by the confirmed earnings date, but that stocks also reacted to the earnings date changes themselves (prior to the actual report being released). The second strategy focused on the price reactions around these sequential changes in earnings announcement dates.

New RavenPack Earnings Date Dataset Powered by Wall Street Horizon Provides Institutional Investors with Alpha Boost

Classifying words as either positive or negative, the researchers generated article-level sentiment scores—to highlight how news likely to be perceived as positive or negative would impact stock prices. Traditionally finance researchers and market practitioners have relied on accounting data and fundamentals to predict where the market is headed. But quarterly reports arrive slowly for a market moving at warp speed, which led researchers and traders to look for other sources of predictive information, including news. To find out if news reports could be used to predict stock prices, Ke, Kelly, and Xiu borrowed machine-learning techniques used by computer scientists, who are increasingly training machines to understand text.