QuantDesk® Machine Learning Forecast

for the Week of January 22nd, 2018

Betting On US Defense Stocks

Erez Katz writes about Betting On US Defense Stocks

by Erez Katz, CEO and Co-founder of Lucena Research.

We continue to engage traditional deep value managers who are looking to extend their offerings with AI and predictive analytics. On that topic, I wanted to present another example of how Lucena’s technology can be effective in advancing a researcher’s deep value analysis, specifically in the context of the defense sector which represents military, aerospace and defense contractors.

Why Defense Contractors?

From a macro perspective, it’s not a secret that the Trump administration is keen on strengthening our military. Escalating tensions with North Korea, Iran, Russia and China along with America’s attempts to strengthen its influence in the Middle East are strong enough reasons to rebuild an aging and deteriorating military. This is one area in which the Trump administration will likely face less resistance politically, as voting for a weaker military could come at a high political cost. Indeed, 2016-2017 have been excellent years to invest in aerospace, defense, and military stocks. Over the past year alone, the five biggest defense stocks have averaged returns of more than 57%. Unfortunately, as evidenced by a much higher average PE ratio of 34 relative to the SP’s PE ratio of 26, these stocks have become a whole lot more expensive.

Image 1: Top five defense contractors trailing twelve months (TTM returns, PE ratio and dividend yield).

As the Trump administration continues to authorize billions of dollars’ worth of arms sales to foreign allies, the outlook for the industry remains extremely bullish. However, the opportunities may now lie with the mid-size to smaller defense players who have yet to fully realize the benefits of additional defense spending.

Today, I want to showcase how with QuantDesk®, we can apply science to constructing a defense contractor portfolio geared to outperform the defense sector and maximize its risk adjusted return.

Step 1 – QuantDesk® Portfolio Replication Engine

Using QuantDesk® portfolio replication technology, build a 20-position portfolio from the Russell 1000 that holds similar characteristics to those of the defense sector. We chose to represent the defense sector with PPA, PowerShares Aerospace & Defense ETF. The portfolio replication engine utilizes a unique pattern-matching technology that identifies a set of stocks with distinguishing characteristics that together perform as close as possible to a given time series. In our case, we are looking to find a collection of securities that move in tandem with PPA. Research that would traditionally take at a minimum a few days, and in some cases weeks, can be accomplished with uncanny accuracy in less than a minute!

Image 2: Replication Wizard – takes you step-by-step through the replication engine. You are now ready to replicate.

After running through a set of Wizard screens, we are now ready to replicate. The results appear a few short seconds later.

As you can see, the time series represented by the replication engine tracks almost identically the price history of PPA.

Image 3: Visual representation of our target portfolio (in orange) tracking our target time series, PPA (in green).

What’s even more interesting is that the set of constituents is mostly defense companies. Remember, I have asked to identify securities from the Russell 1000, with no specific instructions to concentrate on defense stocks.

Image 4: Replication Wizard output – List of constituents and their respective allocation to represent PPA.

Step 2 QuantDesk® Portfolio Optimizer

Now that we have a portfolio with 20 qualified constituents, we can optimize their allocation in order to maximize the portfolio’s risk adjusted return. QuantDesk Portfolio Optimizer utilizes Markowitz’s Nobel winning approach to portfolio optimization. Mean Variance Optimization (MVO) identifies scientifically how to divide the allocation among the constituents of a portfolio in order to maximize the returns for a predetermined risk profile. With QuantDesk it is nothing more than following a set of screens. The system utilizes Lucena’s machine learning Price Forecaster in order to optimize the portfolio allocations for a future price target. If the price target assumes a downward pressure, the portfolio automatically adjusts to a more defensive posture and, conversely, if the target price has a strong conviction higher, the portfolio will adjust to maximize its return potential.

Image 5: Optimizing the defense portfolio. The blue line and cone represent the current portfolio before optimization and the orange line represents the target portfolio after optimization. You can see on the right side the improvement in higher Sharpe, higher returns and lower volatility.

Step 3 QuantDesk® Backtester – Assessing How Our Portfolio Performs Over Time

Rolling back time and assessing how the portfolio performs against PPA will give us a good idea if the science indeed adds value in different market regimes. A simple Wizard-like set of screens will launch a backtest and generate a comprehensive performance report.

Image 6: Backtesting bi-weekly optimization of the defense portfolio from January 1, 2010 to present. The blue line represents PPA and orange line represents our dynamically optimized portfolio. On the right side you can see the improvement in higher Sharpe, higher returns with slightly higher volatility. Transaction costs and slippage are also taken into account.
Past performance is no guarantee of future returns.

Lastly, in order to assess the portfolio on a roll forward basis, we can click one button (see Paper Trade), after which the portfolio follows the same set of rules of the backtest but on a roll forward basis.


The defense sector remains attractive for investment. Traditional advisors are now looking for mid-size and small cap alternatives to the large caps which may have advanced disproportionately ahead of the defense sector. With QuantDesk, even a non-technical user can take advantage of alternative data and machine learning in order to overlay science on top of their traditional expertise. We will continue to follow and report on our defense portfolio in the weeks to come and see if the portfolio indeed lives up to its expectations in both absolute and PPA relative performance.

Strategies Update

As in the past, we will provide weekly updates on how the model portfolios and the theme-based strategies we cover in this newsletter are performing.

Tiebreaker – Lucena’s Long/Short Equity Strategy - YTD return of 3.03% vs. benchmark of 2.76%
Image 1: Tiebreaker YTD– benchmark is VMNIX (Vanguard Market Neutral Fund Institutional Shares)
Past performance is no guarantee of future returns.

Tiebreaker has been forward traded since 2014 and to date it has enjoyed remarkably low volatility and boasts an impressive return of 55.98%, low volatility as expressed by its max-drawdown of only 6.16%, and a Sharpe of 1.99! (You can see a more detailed view of Tiebreaker’s performance below in this newsletter.)

BlackDog – Lucena’s Risk Parity - YTD return of 4.40 % vs. benchmark of 2.34%

We have recently developed a sophisticated multi-sleeve optimization engine set to provide the most suitable asset allocation for a given risk profile, while respecting multi-level allocation restriction rules.

Essentially, we strive to obtain an optimal decision while taking into consideration the trade-offs between two or more conflicting objectives. For example, if you consider a wide universe of constituents, we can find a subset selection and their respective allocations to satisfy the following:

  • Maximizing Sharpe
  • Widely diversified portfolio with certain allocation restrictions across certain asset classes, market sectors and growth/value classifications
  • Restricting volatility
  • Minimizing turnover

We can also determine the proper rebalance frequency and validate the recommended methodology with a comprehensive backtest.

Image 2: BlackDog YTD– benchmark is AQR’s Risk Parity Fund Class B
Past performance is no guarantee of future returns.

Forecasting the Top 10 Positions in the S&P

Lucena’s Forecaster uses a predetermined set of 10 factors that are selected from a large set of over 500. Self-adjusting to the most recent data, we apply a genetic algorithm (GA) process that runs over the weekend to identify the most predictive set of factors based on which our price forecasts are assessed. These factors (together called a “model”) are used to forecast the price and its corresponding confidence score of every stock in the S&P. Our machine-learning algorithm travels back in time over a look-back period (or a training period) and searches for historical states in which the underlying equities were similar to their current state. By assessing how prices moved forward in the past, we anticipate their projected price change and forecast their volatility.

The charts below represent the new model and the top 10 positions assessed by Lucena’s Price Forecaster.

Image 3: Default model for the coming week.

The top 10 forecast chart below delineates the ten positions in the S&P with the highest projected market-relative return combined with their highest confidence score.

Image 4: Forecasting the top 10 position in the S&P 500 for the coming week. The yellow stars (0 stars meaning poorest and 5 stars meaning strongest) represent the confidence score based on the forecasted volatility, while the blue stars represent backtest scoring as to how successful the machine was in forecasting the underlying asset over the lookback period -- in our case, the last 3 months.

To view a brief video of all the major functions of QuantDesk, please click on the following link:
QuantDesk Overview


The table below presents the trailing 12-month performance and a YTD comparison between the two model strategies we cover in this newsletter (BlackDog and Tiebreaker), as well as the two ETFs representing the major US indexes (the DOW and the S&P).

12 Month Performance BlackDog and Tiebreaker
Image 5: Last week’s changes, trailing 12 months, and year-to-date gains/losses.
Past performance is no guarantee of future returns.