Data-Driven Decision Making

The Foundation of Systematic Investing

Welcome to the core of systematic value investing, where data is the foundation of every decision. Unlike discretionary investors who might follow a narrative, we use a quantitative framework to identify undervalued opportunities objectively.

Mandatory Educational Disclaimer

Before we proceed, it is critical to acknowledge that this course is for educational purposes only. We do not provide financial advice or specific stock recommendations. Investing involves significant risk, and you should always conduct independent research or consult a professional.

The Data Pyramid

To build a screener, you must manage three layers of data. At the base, we have Market Data for prices and volume. Above that is Fundamental Data from financial statements. Finally, Metadata provides the context of sectors and corporate actions.

Point-in-Time (PIT) Data

Point-in-Time data is the most critical concept for backtesting. It ensures your simulation only uses information that was actually available to the market on that date, accounting for reporting lags and avoiding the use of future restatements.

Recap: Data Foundations

Let's pause and review the foundational layers of our data framework before we look at quality metrics.

Identify the Backtest Error

Look at this backtest setup. A company's Q3 earnings ended in September, but were filed in November. The backtest 'buys' the stock in October based on those Q3 results. What is the error? Not quite. Consider the timing of the filing versus the timing of the trade. Correct! That is Look-Ahead Bias. You are using data in October that wasn't public until November.

Data Quality Metrics

High-quality data rests on three pillars: Accuracy, Timeliness, and Completeness. Without these, even the best algorithm will succumb to 'Garbage In, Garbage Out'.

The Survivorship Bias Trap

Imagine testing a strategy from 2010 using only today's S&P 500 list. You'd ignore every company that went bankrupt in the interim. This is Survivorship Bias, and it's a silent killer of value strategies.

Recap: Quality and Bias

Before we look at how to vet your sources, remember: clean data is often a bigger competitive edge than a complex model.

Vetting Your Data Source

When choosing a data provider, use this checklist. You must ensure they handle delisted securities and corporate actions correctly, or your value strategy will be built on sand.

Case Study: Vetting 'AlphaData'

You ask a new data provider if they are suitable for backtesting. They say: 'We provide the most accurate, restated financial data for all current S&P 500 companies.' Type your critique of this provider.