Newsletter, January 2026

Platform updates

🔧 Health-check Robot

We deployed a monitoring system that performs routine checks on the availability of key InvestLens features. Powered by AWS Lambda, it runs independently from the platform itself, making it a reliable detector of service disruptions. At regular intervals, it verifies that core tools are functioning as expected and alerts us immediately if any disruption occurs. This has already reduced out-of-service time by many hours, and in some cases potentially by days, as we continue weekly deployments.

📚 Help Center

As part of our commitment to high standards and transparency, we launched the InvestLens documentation platform. Built with MkDocs, it is publicly accessible and provides a clear overview of what InvestLens is: professional-grade tools and resources for portfolio managers and sophisticated retail investors. The Help Center also supports our Finance in Education initiative, with content written in an encyclopedia-style format and a strong emphasis on rigor. Over time, we expect it to become a key point of contact for retail investors, practitioners, and finance enthusiasts.

🎨 General UI

The user interface across the platform now features a 3D, elevated look, creating a sleeker, more dashboard-style experience. A subtle light-blue glow is applied across the UI to improve visual comfort during extended use.

🔬 Research and Development

December was especially busy on the R&D side. We focused on the second version of our price prediction model, upgrading it to use a richer feature set now available from our data vendor. This includes trading volume signals and an expanded set of fundamental statistics. As testing continues, we will share more details and performance results.

🌌 Universe Screener

As a direct response to feedback from our design testers, we have moved the Universe Screener to the center of the InvestLens workflow. It is now easier to access, with users routed to the Screener after login. We also refreshed the UI with a more futuristic, dashboard-style design and improved consistency in slice selection across stocks and ETFs. Both instrument types now support selecting ALL assets without requiring a benchmark, giving users more flexibility and control when customizing their universe.

Upcoming

We are expanding into offerings that include derivative products with an options-based downside protection strategy designed to help investors build a portfolio-specific set of instruments that can provide protection against severe drawdowns.

To refine volatility estimates, we are developing a VIX-inspired method for estimating implied volatility, a forward-looking indicator of risk. We plan to use implied volatility in reporting, since many portfolio managers view it as more informative than purely history-based measures. We will also incorporate these forward-looking estimates into our asset allocation logic to support more risk-aware capital allocation.

Analytics in Practice

Our goal is to build a focused US Financials basket for broad sector coverage that is not riskier than the overall equity market. Using the Universe Screener (Custom tab), we define “slices” of the market by combining country, benchmark, and sector filters, as shown in the table below (US Financial Services is drawn primarily from IXF, with a small complement from S&P 100 Financials). We constrain the basket's beta to 1.0, ensuring the idea delivers broad Financials exposure without taking on more systematic risk than the overall market.

Country

Benchmark

Sector

Position

IXF

Financial Services

S&P 100 (SP100)

Financial Services

We transfer Idea A from the Universe Screener into a new portfolio and apply allocation guardrails to preserve diversification. Position weights are constrained to 2%–12% (0.02–0.12), preventing holdings from becoming immaterial or overly concentrated contributors to risk. The portfolio is titled US IXF SP100.

We then run the portfolio through the Asset Allocator to optimize weights under our selected constraints. For risk, we use a constant-correlation historical covariance matrix, and for expected returns we use Fama-French Five-Factor return estimates (defaults).

The chart compares cumulative performance over the last 252 trading days for the US IXF S&P100 portfolio under four weighting approaches: the baseline basket (blue), identical to the equal-weight portfolio in this case, Global Minimum Variance (GMV), and Maximum Sharpe Ratio (MSR). All series start at 1.0 and track through the same market swings, but diverge as weighting choices change the path of returns. Over this window, the MSR portfolio ends with the highest cumulative return, while GMV shows the most conservative profile, illustrating that portfolio construction can materially affect outcomes even when the underlying holdings are identical.

In conclusion, this example shows how InvestLens turns a broad investment objective into a repeatable, decision-ready workflow: we define transparent universe slices, identify an idea that respects a market-risk beta constraint, and then translate it into an implementable portfolio using optimized weights under clear allocation bounds. The backtest illustrates that weighting choices meaningfully impact performance, reinforcing why disciplined portfolio construction matters as much as asset selection.

NEWSLETTER

Platform updates

Analytics in Practice