Providing an Advanced Wall Street Tool to Individual Investors

Factor Modeling compares two or more assets to determine if their price trajectories are statistically linked with one another. These complex calculations reveal the effects of market-moving events between assets, sectors, and entire markets.

Due to the resources required, Factor Modeling is primarily used by hedge funds and institutional investors with teams allocated to the task. The MarketReader team is putting our decades of institutional-side experience into building a similar tool accessible to individual investors at a fraction of the cost.

Mapping the Market

MarketReader answers the question, “what assets are moving together, right now, in real time?”

For each significant price move detected, our systems look at the entire market and calculate how many times in history other assets were at the same points in their contextual curves. We then surface only the most significant relationships that are relevant in real time.

Pinpointing these relationships helps identify causality, e.g. if poor earnings results for one tech company causes other tech stocks to drop in price. Overlaying our News, Social Media, and Calendar models with this data allows us to determine the most plausible root cause of the original price move.

Knowing Where to Look

Because MarketReader systems are continually monitoring all price moves across the market, we are able to filter by price action to surface only the most relevant news and social media content that is actually moving markets. 

Rather than searching endless content feeds hoping to find valuable information, we focus first on unusual price activity. Once we flag a significant price movement, we conduct an informed search to identify the most plausible explanation for the price move in near real time.

Adding Valuable Context

The rapid pace of social media serves as a valuable tool for our systems to determine what headlines were actually interesting to market participants. Of the myriad topics being discussed in a given moment, which were relevant to price movement in a specific asset or sector? 

By applying Natural Language Processing (NLP) only in specific, targeted areas of our analysis, we are able to simplify the problem and use real-time, real world signaling to identify what topics actually had an effect on the market.

High-Impact Macro Events

The MarketReader Calendar Model is designed to search for macro events with the potential for high impact on asset prices and pinpoint those that had a causal effect.

We look at wide-scale events including, but not limited to, FOMC announcements, CPI prints, Jobs reports and more, as well as events like dividends, stock splits, earnings announcements, filings, equities offerings from individual companies. These disparate sources of calendar information serve to paint a clear picture of what was happening during the time a significant price movement occurred.

Calendar Causality

When a large proportion of stocks are moving unusually across the entire market, our system is able to identify the macro calendar events that are truly “market moving.” When we flag unusual price or volatility movements in a disproportionate number of assets, the system scans all available data sources and attempts to match those price movements to a logical root cause.

The output of this model alerts users to movements in individual stocks and larger groups of stocks and then contextualizes those market movements around a vast array of possible calendar events occurring at the time.