© 2019 Boosted.ai. All rights reserved.

Boosted Insights augments the investment process by empowering investment professionals to source new ideas and create models based on their unique financial domain expertise.

Dynamic Stock Screening

Use Case

Solution

Output

Create a more dynamic tool to screen for ideas. Traditional screening tools can be backward looking and do not adapt to new regimes. 

Boosted Insights continuously learns from your inputs to predictively rank stocks. It discovers and explains which combination of features are important in different periods.

Our Machine Learning provides dynamic and adaptive stock rankings. It compares each stock in a universe to each other (250,000 comparisons in S&P 500).

Rankings are on a 5 star scale - 5 stars are strong buys/longs, 0 stars are strong sells/shorts

The graph below shows the performance of the highest rated Boosted Insights picks (5 stars = green line) versus the lowest (0 stars = red line) - a 200% outperformance over the benchmark (SPY) since January 2008

 
 
 
 

Added Conviction For Your Picks

Use Case

Solution

Output

A portfolio manager looks for validation of their fundamental work by investing in the intersection where their stock picks and Boosted Insights agree.

The portfolio manager can increase conviction and position sizing when their picks and Boosted Insights agree. Where they disagree, it serves as a flag to possibly sell or re-examine their position. 

The portfolio manager can cross reference Boosted's top rankings with their own to invest in positions that have the most overlap.

A portfolio manager invested in the intersection of their own fundamental analysis and the machine's predictions. The portfolio outperformed the manager and machine alone. 

Targeted Baskets

Use Case

Solution

Output

A portfolio manager can create a more targeted basket of longs or shorts within distinct sectors or with specific attributes.

Combine your own fundamental views and select your own variables to create a more targeted basket of stocks that the machine predicts will outperform/underperform the market.

A highly customized basket of stocks is selected for a specific purpose, allowing the portfolio manager to create additional alpha over using a broad market index. 

A portfolio manager created a short basket based on these lowest ranked stocks in the S&P 500 on May 1, 2019. As of September 3, 2019, it underperformed the benchmark by 6.15%, adding significant alpha to the client's risk mitigation strategy. 

Sector Allocation

Use Case

Solution

Output

A portfolio manager can use Boosted Insights to determine which sectors are most attractive for long and short exposures to allocate their portfolio most efficiently.

The portfolio manager inputs their fundamental, technical or economic variables from a top down perspective. Boosted Insights learns from the data to provide a detailed sector allocation recommendation. 

The machine returns predictive sector weightings, along with which securities within these sectors provide the best long and short opportunities.

Below are the Boosted Insights sector weightings versus the S&P 500 and the relative outperformance. 

A full sector allocation recommendation is provided by Boosted Insights. Boosted's sector weightings outperformed the broader market by 1.9%.

 
 
 

Custom Data Uploads

Use Case

Solution

Output

A fund believes that their custom data is predictive and wants to generate maximum value from it.

The fund uploads their custom data (alternative data, like credit card, geolocation, insider transactions) to Boosted Insights and creates a model.

Boosted Insights determines if the custom data is in fact predictive and generates trade ideas, rankings and a full analysis on the strength of this data.

Rankings are on a 5 star scale - 5 stars are strong buys/longs, 0 stars are strong sells/shorts

Portfolio Optimization

Use Case

Solution

Output

Once a portfolio manager has their positions, they want to optimize the sizing in their portfolio.

The portfolio manager uploads their positions and runs it through the Boosted Insights optimizer. They can choose to optimize the portfolio to maximize returns, alpha, or minimize volatility across markets or sectors. 

Boosted Insights dynamically returns position weightings for the chosen optimization. 

Below are the results from the same model with the volatility optimizer turned off and on. With the optimizer on, volatility is reduced by 4% with similar returns (the Sharpe ratio increases 0.22 points). 

Actively Traded Models

Use Case

Solution

Output

A portfolio manager wants to operate on a fully quantitative basis using machine learning.

The portfolio manager creates a model (using their own unique variables), sets portfolio constraints and then has the model “go live” with current data. They then execute the model’s dynamic trading strategy. 

A fully quantitative portfolio runs using technical variables and/or fundamental data.

Since going live in October 2018, this quantitative model returned 36.8%, with gross exposure of 150% long and 50% short, compared to the benchmark return of 4.9%.