It's worth reading even more about the origin story of BioAge's "flexibility" value before you read this post. 

Please note that the final formulas are proprietary to EGYM. This article describes the general rationale behind the calculations.

Review - The Calculation of the BioAge

In a detailed article, we have already described development from idea to feature of the " flexibility " value of the BioAge. 

In short, if I am as flexible as the average of my peer group, my BioAge should be equal to my chronological age. So, underlying this is a norm curve that describes the decline in flexibility over chronological age. This norm curve was derived from scientific literature, internal studies, and the help of test studios. 

To calculate a person's BioAge, the flexibility achieved is determined using this norm curve. The more flexible a person is than the norm, the lower the BioAge, and vice versa. 

One of the primary objectives of BioAge is to represent physical reality as accurately as possible. To achieve this, of course, we first have to know what that reality looks like. The first version of the "flexibility" value of the BioAge, released at the time of the Fitness Hub launch, was based on scientific studies, internal testing, and selected pilot gyms that were allowed to pre-test an early version of the Fitness Hub. 

First, let's look at the first version of the Norm Flexibility of the Neck test, which was used to release the Fitness Hub in June 2021. The orange line ("v0.1") describes the formula of the target flexibility, i.e., the assumed average of this age group.

The graph should be read as follows: A 40-year-old athlete should reach approximately 35° on the neck flexibility test for his BioAge to match his chronological age. The formula was determined with the help of scientific literature, internal tests, and through the help of test studios. 

However, opportunities to collect real-world data were less than optimal because of the closures of gyms during the Corona lockdowns. Fundamentally, however, it was clear that we will be able to build one of the world's largest flexibility databases over time. Thus, we knew that we would continually expand the meaningfulness of the "flexibility" value and its relationship to age, further optimizing its usability.
 

A few weeks after release - we are on our way to becoming the most comprehensive database for flexibility worldwide!

However, at the time of writing, gyms have reopened, and the Fitness Hub is already in heavy use in over 100 gyms, just a few months after the start of series delivery, meaning that tens of thousands of flexibility measurements have already been taken from real customers. This data can now be used as planned to improve the BioAge calculation.

The data is DSGVO compliant and only in anonymized form. For each test performed, we know the following data points:

• User ID in encrypted form
• Date of the test
• Chronological age of the client
• Type of test (neck, lat flex, hip)
• Test results (left and right):
  • Degree
  • Normalized
  • BioAge

Real-world test data will keep improving BioAge accuracy!

We have now collected a lot of real-world data about flexibility with Fitness Hub for our customers. One month after release, we have already collected enough real data points to make first improvements. Let's now plot these individual measurements on the graph as well. In addition to our initial estimate, we now see the actual degrees achieved plotted against the chronological age of the customers categorized into female and male. The lines drawn in are the respective fitted polynomial functions, including confidence interval (gray area), which best describe the course of the data. 

What we see is that the original formula was designed to be a bit too "aggressive." The performance of the young users was overestimated, resulting in the fact that it was challenging for them to get a low BioAge. For the user group around 60 years of age, the formula seems to have worked well. 

• Furthermore, we can draw the following conclusions from the actual distribution of results:
• Flexibility remains relatively constant up to the age of 30.
• From the age of 30, flexibility declines.
• The loss of flexibility accelerates with age.
• The results of the age groups scatter relatively strongly upwards and downwards.
• The overlapping confidence intervals (gray areas) indicate that there is no significant difference between females and males.

In the next step, in order to create a motivating BioAge that matches reality as closely as possible, we adapt our target flexibility formula to reflect the actual data. The new formula corresponds to the line "v0.2".

The new formula now estimates the target flexibility of young users to be significantly lower. However, we see that the actual average performance of young users is still slightly below what our formula predicts – this is the way it should be! For a motivating system, it is an advantage if the majority of the under-30s do not have a BioAge of "21" from the get-go. Achieving this best value is and will continue to be a challenge!

Additionally, the decrease in flexibility over age is a bit less estimated. If you look closely, you can see that the new target flexibility formula is slightly above the fitted polynomial functions. This results in these users having a slightly higher average BioAge relative to their chronological age. The majority of the data comes from customers taking this measurement for the first time, and these users have not yet performed muscle length training on EGYM Smart Flex. Therefore, we can assume that follow-up measurements will be significantly better and that most users with appropriate training will progress to a better BioAge. 

Please note that the calculation of target flexibility is only one part of the BioAge calculation. In the further steps, the deviation of the achieved degree from the target norm is calculated, normalized, and converted into the BioAge. Accordingly, this article presents the data basis, including the first step of the calculation. 
 

Improved accuracy = increased motivation!

A complex feature like the EGYM BioAge always goes through several phases of development. First comes the idea, which we consider and critically scrutinize based on the latest scientific findings. A combination of scientific literature, internal experiments, and testing in selected pilot studios then results in a first version.

But the true advantage of a data-driven feature can only fully unfold when the data is available in sufficient quantity. Thanks to EGYM's extensive ecosystem, we were able to release a vastly improved version of Flexibility BioAge within a short period of time. This further development is only possible through the usage of the product in a natural gym environment by as many customers as possible. The data is always provided anonymously with the utmost care, and traceability to individuals is not possible. 

We wish you all a lot of fun and success with the new version of Flexibility BioAge and look forward to the following months, in which more and more tests will be conducted, and our database will continue to grow.