Predictive Media Content Analytics: 24/7 Information Has Forever Changed Financial Market Strategies

Executive Summary:
TrendPointers, a provider of predictive analytics, describes how to use RavenPack’s Equity Indicators to enhance a set of macro core signals and intervention metrics that are applicable to both long and short portfolios over time frames from one to eight weeks.

The performance of the indicators is tested against a buy-and-hold strategy on the S&P 500. Specifically, they find that from 2010 to mid 2014:

• The bi-weekly signal directional accuracy is high, averaging 57%
• A simple binary entry/exit signal execution of the signal outperforms the benchmark by up to 11,87% on an annualized basis

The New Methodology of MacroSentiment Analytics© (MSA)

The application of causal or anticipatory MacroSentiment Analytics is shown to significantly outperform the benchmark S&P 500 over a recent nearly five-year period. The complex-text analytics process, applied to the 24/7 flow of an adaptive lexicon of financial-relevant information, is a new design to extract the net meaning of content from the continuous and cumulative body of decision-relevant “news”. The results provide new leading measures of macro-influences on market behavior which are used for biweekly models of market direction.

Applications

There are many styles of investment management, from long-term asset allocation to active trading strategies. But all styles are subject to the same continuous flow of economic reports, measures of consumer activity, and many analyses of the influences on forward expectations and economic behaviors. Large-scale portfolio management can apply the continuous MacroSentiment Analytics trends and signals to their asset management strategies. Traders and funds that plan around shorter time frames and persistent volatility can use the biweekly signals and interventions. The TrendPointers/RavenPack analytics provide a superior set of core signals and intervention metrics that are applicable to long and short portfolios, and over time frames from one to eight weeks.

Predictive Media Content Analytics - White Paper's Guideline

• Introduction
• The MacroSentiment Concept: Why Better Indicators Are Now Possible
• MacroSentiment Analytics Design Objectives
• Financial Market Application Objectives
• Architecture of the MacroSentiment Content Analytics
• MacroSentiment Signal Methodology
• MacroSentiment Model Results
• Summary and Conclusions

Introduction

How The Internet Has Provided Us With Predictive Content Analytics

The 18th-century philosopher/economist David Ricardo is credited with popularizing the investing concept of "buy low, sell high." But as one 21st century writer noted, Ricardo failed to explain how to identify the “high” and the “low”. We've now spent more than two centuries trying to reliably identify and quantify those decision points. Ricardo had the right instincts, but lacked the proverbial 30,000-foot view of the markets to observe and anticipate the speed and shifts of lows and highs along continuous trajectories. The necessary analytics Ricardo could not have dreamed of are now available.

The rapid development of 24/7 Internet information has put the entire realm of the potential knowledge of the buy low/sell high quest at our literal fingertips --if we know exactly what data to look for, how to extract it from all the noise, and how to correlate the data with our ultimate investment targets and risk parameters.

In the late 1990s the founders of TrendPointers began working on Internet-based analytics. It was evident that information available via the net would soon replace survey, consensus and audit systems with more accurate and timely research methods. We could now capture nearly real-time activity to better isolate the decision-relevant variables, eliminate many data collection limitations, and continuously test and update for statistical correlation to the targets.

Hence, TrendPointers developed a set of indicators that model the antecedents of macroinfluences on economic attitudes and behavior that eventually manifest in consumer spending and market performance.

But with the “Big Data” explosion of investment-relevant information, there is even less time to acquire, analyze and act upon the available information. The increasingly complex, volatile and uncertain corporate, macroeconomic and geopolitical influences have driven the interest to find analytics that provide measurable anticipatory benefits over conventional measures.

To capture the most contemporaneous market relevant influences, TrendPointers has partnered with RavenPack to provide timely, forward-looking, market-specific sentiment. The respective data sets are combined and used as both initial inputs and feedback within the nonlinear modeling process.

This paper describes how TrendPointers and RavenPack data produce new leading indicators that:

• Accurately capture the variable and often unstructured influences on overall economic behavior and specifically financial markets.
• Are able to continuously adapt to the changing configuration of the influences.
• Create effective leading indicators and models of the shifts in market influences to provide more time to plan, modify or take preemptive asset management actions.