Products
REACTION
Post-event predictions
JUMP
Pre-event predictions
SENTIMENT
Predicting returns with news sentiment
REACTION
Markets react every time a company reports their latest financial results. This event provides a lucrative trading opportunity given the volatility surrounding these events. The Leapday REACTION model uses information about historical reactions to identify predictable patterns of behavior. This model provides an event-based signal that investors and traders can use to predict short-term price moves for stocks after financial reporting events.
COVERAGE
Events - quarterly and annual financial statement releases
Markets - all companies listed on the NYSE, NASDAQ, and NYSE American
CUSTOMIZABLE
Signal Threshold - set the trade-off between the number of actionable trades and the accuracy of the signal
Signal Timing - receive signals shortly after the open (to incorporate them into an intra-day process), on the close, or prior to the next open
Holding Periods - signals for durations up to 15 days are derived from independent models that are optimized for each specific holding period
ACTIONABLE SIGNALS
The trade signal is delivered as a numerical value between -100 and 100. A positive value indicates a buy opportunity while a negative value indicates a short opportunity. The magnitude of the value represents the degree of confidence behind the signal.
VERIFIED ALPHA
High hit rates, positive return skews, and significant average returns per trade using realistic assumptions about liquidity and trading costs. Request our white paper to learn more about these results.
SENTIMENT
SENTIMENT is a state of the art natural language processing model that provides traders and investors with meaningful predictions of short-term price direction following news. It goes beyond traditional sentiment analysis which identifies feeling or emotion in news without linking it to market behavior. Instead, we train each sentiment model on historical price reactions to news so that the model can learn the behavioral biases of investors as they react to news. The result is a sentiment score that forecasts short-term market price reactions.
COVERAGE
News Sources - press releases, premium news sources, original web sources
Markets - all companies listed on the NYSE, NASDAQ, and NYSE American
PREDICTION-DRIVEN SENTIMENT
The SENTIMENT model is trained to identify “sentiment charged terms” in press releases based on historical market price reactions. It analyzes the text of each press release using a non-parametric approach to learn the sentiment of a document. The result is a sentiment score between -100 (negative sentiment) and +100 (positive sentiment) that acts as a forecast for short-term market price reaction.
VERIFIED ALPHA
Large annual returns, high Sharpe ratios, and uncorrelated returns in a standalone portfolio using Leapday's SENTIMENT scores (inclusive of conservative trading cost assumptions). Request our white paper to learn more about these results.
DELIVERY OPTIONS
Real-Time - Integrate SENTIMENT directly into your trading and risk-management systems
Daily - Use SENTIMENT to drive your portfolio construction process
