Company: elb_research
Categories: Finance, Macro Economics, Quantitative
Analysis
We believe that the biggest problem with current quantitative research—and finance research in general—is the significant lack of objectivity in the analysis, with far too much subjectivity and emotion involved. We have never seen a single analysis for retail investors that includes a valid machine learning approach to eliminate any emotional bias in trading decisions. The goal of our work is for you to receive the latest report from us every morning, providing a clear answer to the central question for every investor or trader: What is more likely: that a market will go up or down? Below, you can find our approach to provide a quantitative answer to this question.
| Metric | Value |
|---|---|
| Sharpe Ratio | 1.1245 |
| Sortino Ratio | 0.1026 |
| Max Drawdown | 0.1637 |
| Annualized Return | 0.1058 |
| Win Rate | 0.2075 |
| Loss Rate | 0.1459 |
| Profit Factor | 1.3858 |
If you had followed the trading signals of the index, you would have achieved a remarkable return of 4x without any significant drawdowns. From our perspective, you should once again follow Jim Simons’ advice:
“Past performance is the best predictor of success.” – Jim Simons
The interpretation of every metric in this section is the same as above. As you can see in the figures below, the index provides a robust estimation of the current macroeconomic environment in any market context, protects our portfolio against significant drawdowns, and delivers excellent buying opportunities.
| Metric | Value |
|---|---|
| Sharpe Ratio | 0.7805 |
| Sortino Ratio | 0.0729 |
| Max Drawdown | 0.2344 |
| Annualized Return | 0.0832 |
| Win Rate | 0.2007 |
| Loss Rate | 0.1499 |
| Profit Factor | 1.2711 |
“The best decisions are often made with data and evidence, not just gut feelings” – Jim Simons
This section remains the same every day. If we change or add additional information or introduce a new index, we will note this at the top of the daily paper. We calculated this index using multiple machine learning methods and simulate it on a daily basis with out-of-sample data in a so-called “online simulation.” We were very careful to avoid any look-ahead bias, randomness, or overfitting, and we produce reproducible and deterministic results. We create this paper with R Studio in R Markdown. Please understand that we do not share the code with the public nor delve deeply into our methodology. The potential for more advanced analysis is huge. Currently, we are working on an add-on called the “Short Macro Index,” which is designed to generate more consistently significant outperformance during bear markets when used in combination with the Kaya Regime Index. In addition, we are developing a Stock Regime Index to identify price movements that are not driven by macroeconomic factors.
The index contains a sample of every major US macroeconomic data series from the FRED database. In this section, we show you which data is used for the calculation and simulation of the index. You will recognize most of these data series.
All data is from the FRED database.