DISCLAIMER: This note is intended for US recipients only and, in particular, is not directed at, nor intended to be relied upon by any UK recipients. Any information or analysis in this note is not an offer to sell or the solicitation of an offer to buy any securities. Nothing in this note is intended to be investment advice and nor should it be relied upon to make investment decisions. Read our full disclaimer, here.

The SignalFlow AI Algo Suite

by Alex King, CEO, Cestrian Capital Research, Inc

As you know, everyone has a trading algorithm now. All you need is a $20/month LLM subscription, some ideas about parameters and goals, and you can create one. There is zero differentiation in just having an algo.

What is not generally available is an algorithm family built on top of decades of quantitative experience.

The differentiation in algorithmic trading today is not access to the mathematics or the computing power, each of which is now commoditized. The differentiation is the expertise in creating models based on a deep understanding of distribution patterns in securities pricing. This is rare indeed.

And with that in mind, a word from my colleague Jay on the background to this algorithm family.

Why I Created SignalFlow

by Prof. Jay Urbain, Ph.D

I've always been interested in applying my quantitative background to developing financial models. The original idea for SignalFlow was to increase returns by avoiding significant drawdowns. And given a particular market state, to determine what is the best trade or investment.

The baseline goal was to exceed benchmark results and maintain a high Sharpe ratio. I, like many of us suffered steep losses in 2008, 2018, and 2022. Thanks in part to SignalFlow, my losses during market selloffs in 2025 and 2026 have been manageable and as a result my gains have been more substantial.

The basis of SignalFlow is a risk model and a ranking model. The risk model, I believe, is state-of-the-art.

The SignalFlow Algo Family, Explained

Using the same underlying risk and ranking model methodology, we offer the following SignalFlow algo services.

SignalFlow AI Growth is designed to press on the gas in bull markets. The algo rotates through a set of high-beta ETFs such as QQQ, SMH, XLK, IGV and others with the intent of outperforming buy and hold by (i) avoiding overall market drawdowns and (ii) owning the strongest three ETFs at any one time. The algo falls back to SPY and then to cash in the event that the high-growth ETFs lose their luster. You can read more about it here.

SignalFlow Sector Rotation is designed to outperform buy & hold of the S&P500 by rotating through the S&P500’s constituent sector ETFs. It is a market-neutral algorithm ie. it is designed to try to achieve long-only gains regardless of the market environment - since if tech ($XLK) is in decline there is a good chance that energy ($XLE), metals ($XME) or other sectors are on the up, and vice versa. It falls back to cash in the event of cross-sector weakness. You can read more about it here.

SignalFlow Global gives you access to international equity markets by way of simple US-listed ETFs. It rotates through a long list of country-specific ETFs; it has enjoyed great success lately with $EWY, South Korea, and $EWZ, Brazil. In the event that US markets are stronger than international markets, it falls back to $SPY; if all markets in its universe are weak, it falls back to cash. You can read more about it here.

SignalFlow Long-Only SPY is designed to outperform buy and hold by moving to cash during S&P500 selloffs. You can read more about it here.

SignalFlow Long/Short is for investors who wish to take advantage of market volatility, holding light- or heavily-weighted long positions in the S&P500 during bull markets, or short positions in the S&P during bear phases. The algo falls back to cash if it believes the market is directionless. You can read more about it here.

SignalFlow Bonds is designed to help investors be long bond ETFs during bond bull markets, and in cash during bond bear markets. You can read more about it here.

SignalFlow Core

Our premium SignalFlow service is SignalFlow Core. It is designed for professional clients to achieve an optimized balance of performance and volatility. It is disarmingly simple in its output, rotating through $SPY, $SCHD, and cash. The high returns, Sharpe and Sortino ratios achieved by the Core service make it suitable for (i) managers running risk-averse client funds on an unlevered basis and (ii) managers seeking a safer way to deploy levered funds. Contact us to discuss pricing of SignalFlow Core.

Bespoke SignalFlow Strategies

For our professional clients, be that funds, family offices or RIAs - we offer custom SignalFlow strategies. We can develop a rotation algorithm based on the ETFs or single-name stocks that you wish to run in a particular strategy or theme. In most cases we can produce a rotation output that can outpeform simple buy and hold; the SignalFlow model at its core was built to minimize portfolio drawdowns. Examples of such themes might include sector focus (e.g., semiconductor, software, defense, financials, healthcare, industrials and so forth), ethical considerations (faith-based, environmental and so forth), or others. We need a reasonable number of stocks and/or ETFs in your strategy and we need them to be reasonably liquid and long-lived. The model trains on the price history of each individual security, so very recent IPOs aren’t suitable.

Contact us to discuss your approach and will we work with you to develop an algorithm that you can use in your strategy. Email Alex personally at alex.king@cestriancapital.com, or use this contact form.

About Prof. Jay Urbain, Ph.D


Jay Urbain, Chicago native. Deep interest in analyzing the fractal and temporal relationships in market data.

  • PhD, Computer Science, Illinois Institute of Technology
  • MBA, University of Wisconsin, Madison
  • MS Computer Engineering,  Illinois Institute of Technology
  • BS Electrical Engineering and Computer Science, University of Illinois, Chicago

I have 35+ years of experience in private sector and academic research and development. Research and projects have included algorithms for radar electronic countermeasures, 3D robotic guidance and calibration, algorithms and leadership for three generations of critical care patient monitoring systems, data mining system to prospectively identify patients at high medical risk, machine learning and natural language processing models for search and construction of temporal graph entity relationships for intelligence documents, development of search systems for medical college medical datawarehouse, brain cancer MRI segmentation and image synthesis of pseudo contrast images, brain cancer data mining system, graphical machine learning models for matching similar products, and development of machine learning pipelines for risk and ranking of financial assets over time.

20 years undergraduate and graduate computer science instruction.  Retired as full professor. Currently teaching part-time.

Cestrian Capital Research, Inc - May 2026