Understanding the FVEr Trading Strategy (Part 1)
Welcome to the FVEr Invest Blog: Your Guide to Data-Driven Market Analysis
Hello FVEr Invest Subscribers!
Thank you to our new and continuing subscribers for your support of our platform. Our goal is to bring you frequent market insights, using our proprietary FVE algorithm to help you gain a clearer understanding of market dynamics.
Updates
Our Models page went down for some time over the weekend, and we apologize for the inconvenience. The page is functional again, and you can access it here.
https://fver.shinyapps.io/shinyapplication/
Recap
Last week, we discussed the concept of the FVE residual, which is the percentage difference between the price of the ETF and the fair value estimate (FVE) from our algorithm. We outlined the importance of the residual distribution (which buckets the residuals into different ranges), and how it can be used for predictive capabilities about the likely direction of the price of the ETF in the future. In developing this algorithm, we intentionally engineered a significant price divergence between the ETF and its fair value estimate (5-15% standard deviation depending on the ETF). This divergence is designed to offer compelling buying opportunities for investors focused on the long term. One drawback of this is that the model admits notable autocorrelation between nearby residuals, which limits the statistical underpinnings of the model.
One of the ways that we have validated the model is by the empirical success of backtests of the trading strategy that we will start discussing below. Ultimately, the success of the model will depend on how the trading strategy performs forward-looking (not just backtested), and we have been rolling out the strategy to a number of ETFs since mid-April. Updates on performance will be shared every month.
Importantly, we have strived to avoid ‘over-fitting’ our trading strategy to various ETFs – we apply the same statistical parameters across every ETF.
For further research, readers might explore the ARIMA forecasting model, which stands for Auto-Regressive Integrated Moving Average, and is intended to minimize autocorrelation. This is a popular algorithm in quantitative finance which is used best for forecasting short term price movements for individual stocks, or ETFs, and is suited for short term trading.
https://www.investopedia.com/terms/a/autoregressive-integrated-moving-average-arima.asp
Understanding the FVEr Trading Strategy (Part 1)
With this review, we will begin introducing the FVEr Trading Strategy which is built upon the fair value estimator algorithm.
In simple terms, the idea of the strategy is to take a higher level of risk when an ETF is flashing undervalued criteria, take a neutral level of risk when the ETF is fairly valued, and take a lower level of risk when the ETF is flashing overvalued criteria. This aligns with Warren Buffett’s famous quote “to be fearful when others are greedy and to be greedy when others are fearful”. Our goal at FVEr has been to formalize this quote in a sophisticated and data driven way, at the broad market and sector level (we don’t model individual stocks). Indeed, there needs to be a robust valuation measure that can signal when an ETF is overvalued, undervalued, or neutrally valued.
Historically, one of the most common valuation tools for broad market index valuation is the aggregate price-to-earnings ratio (P/E), which is found by dividing the total market capitalization of the constituent companies in an index by the total earnings of those companies. We certainly believe that this valuation metric is an important one to watch, but there are some limitations:
The ratio is only easily available for large indices such as S&P 500, Russell 2000, S&P Small Cap 600, and S&P Mid Cap 400.
The ratio is subject to the discrete timing of corporate earnings (four times a year per company). The denominator of the P/E depends on the earnings of the constituent companies which are either outdated (previous earnings), or predicted future earnings.
The ratio is a blunt valuation tool and can go haywire if there is an extreme earnings collapse such as what happened during the financial crisis. In early 2009, the S&P 500 Aggregate P/E rose to around 100 (which would normally signal an extremely expensive market, but it turned out to be a great time to buy)
When developing our new valuation tool, we decided to make it only price-driven (independent of earnings) for two main reasons: to avoid the complexities of future/past earnings discrepancies, and to ensure easy application across various ETFs by solely relying on readily available historical price data. In theory the price of an ETF should dynamically take into account all of the information about earnings performance of the constituent companies, because earnings are what support the price of equities.
The question remains how determinative the FVE algorithm is in predicting price returns, and thus representing a “good” measure of market valuation. As an example, if we take the SPY ETF (SPDR S&P 500 ETF) over the last decade, we can calculate the 6 week forward returns when SPY was undervalued (having a residual less than 0), versus when the SPY was overvalued (having a residual greater than 0).
When SPY was undervalued, the percentage return during the ensuing six weeks had an average of 1.82%, and a median of 2.23%. When SPY was overvalued, the average return was 0.44%, and the median return was 1.51%. Looking at the distribution of these returns in both cases (see charts below), we see a left skew when the FVE is overvalued (meaning there were a number of 6 week returns with sizable sell-offs, which pulled the average return notably lower than the median).
This analysis indicates a tendency for 6-week performance to be modestly higher when the FVE is undervalued. In the realm of investing, where perfection is unattainable, the key to outperformance lies in identifying and exploiting such subtle edges. Next week, we will discuss how we can use the FVE as a signal to modulate risk levels by tactically allocating to leveraged ETFs under certain conditions.
FVEr Weekly Market Update: July 28, 2025
Current Allocation Status: As of our model updates on Friday, here are the major updates:
Since last week, there have been no model changes for the ETFs covered on the allocation page. We are working on expanding that list to more sectors soon.
XLV (SPDR S&P 500 Healthcare) remains undervalued in leveraged status. There was a modest rally over the last week.
All other ETFs are unleveraged.
See you next week. In the meantime, please don't hesitate to reach out if you have any questions.
The FVEr Team
Unlock Deeper Insights: Schedule a Learning Session
As a valued member, we encourage you to take advantage of a personalized 30-minute learning session with one of our co-founders. This is your opportunity to get tailored guidance on how to interpret our data and effectively implement our strategies in your own investment approach.
To schedule your session, simply email us at info@fverinvest.com with the subject line: "Learning Session". (Please note: We do not provide specific investment advice.)
