In your 2022 Year In Review, you’ll notice a new metric: Muscle Strength. It will be in the Fitbod app experience soon, but we wanted to give you all a sneak peek of it in your 2022 Year In Review. Read on to learn more!
See how your strength changes over time
People love resistance training because it’s easy to gauge your personal progress. Put on another plate, do another rep, hold a position longer: Your numbers tell you where you stand. But that’s harder to nail down if your training regimen is varied.
With Fitbod’s new Muscle Strength score, you’ll see exactly how your strength changes over time. That means you’ll get the feedback you need to know where your hard work is paying off—and where you might want to put in a little more focus.
With your Muscle Strength score, you’ll find out to answers to these questions:
- What are the relatively strongest and weakest parts of my body?
- Which muscle group is holding me back from progressing on my bench press?
- I’m trying some new leg exercises—how are they impacting my quadricep and calf strength?
- Am I making any progress toward my goal to increase my back strength?
- And, settle it once and for all: Which of my friends has the strongest quads?
By compiling billions of data points and feeding them through our personal trainer-approved machine learning algorithm, we created the Muscle Strength score. Use it as a read out on just how strong your muscle groups are.
Great. So how does it work?
Every Muscle Strength score is on a scale of 0-100+ where a score of 50 is what an average Fitbod user can lift (see FAQ below for more info). Lower is weaker, higher is stronger. The numbers are based on the relative strength of each muscle group, so having a higher score in biceps than quadriceps does not mean your biceps can lift more weight than your quads.
Once Muscle Strength launches in the Fitbod app, you’ll see how your workouts have affected each muscle group’s strength every week. Did you switch up your routine? Did you add new exercises? Are you training with a friend? Our Muscle Strength scores will show you how those changes have impacted your strength progression.
How did Fitbod develop this metric?
Buckle up because we’re handing the keys over to our data science team! (Don’t have the stomach for stats? Jump to the FAQ for additional user tips.)
Imagine you have a bench press one rep max of 235 lbs and a pushup one set max of 41 reps. What are your tricep and chest strengths?
This is exactly what we set out to answer with Muscle Strength scores.
Our first step was to use a static representation (i.e., a snapshot) of Fitbod population statistics to put every exercise effort on the same scale. Once we had all the exercises on the same scale, we had some questions to answer:
- How important are different types of strength in performing each exercise?
- How do we quantify the importance of different strength types within an exercise?
- Is chest or tricep strength more important in performing barbell bench press?
- Which exercises are the most and least indicative of each type of strength?
- How do we quantify the importance of different exercises within strength types?
- Are lat pulldowns or bicep curls more indicative of bicep strength?
We could have enlisted our personal trainers to manually estimate the contributions of each muscle group to each exercise, but that would have taken months and would be an approximation at best. So when you have billions of data points, why not use them to get an unbiased and data-based assessment of muscle group usage in each exercise? To do this, we created a novel machine-learning approach aimed at answering the exact questions posed above.
The machine learning model followed the flow depicted below. For each exercise a user performed, the model took the importance-weighted average of all other normalized exercise strengths (excluding the chosen exercise) to generate muscle group strength scores. The model then took the importance-weighted average of those muscle group strength scores to predict the strength of that chosen exercise. The gap between the actual strength of the exercise and the predicted strength was then used to adjust the importance weights. The algorithm ran through our data until it had a sufficient understanding of the mapping between exercise and muscle group importance.
The algorithm ran through data until it had a sufficient understanding of the mapping between exercise and muscle group importance.
The result? We now know the importance of all of those arrows in the flow shown above. They show us just how important each muscle group is to other muscle group.
For example, we found out that the Barbell Bench Press has more to do with shoulder strength than triceps strength. We also have now figured out that when you do Hip Thrusts, most of you out there are putting your heels too close to your butt, so you’re working your calves more than your glutes.
Most importantly though, we now can provide you with Muscle Strength scores that are reflective of just how strong you are in each muscle group according to all your lifts.
So, get out there! And get ready to dive into your Muscle Strength scores.
Q: Why am I compared to the average Fitbodder? Why not compare me to people who are the same gender, age, weight, etc?
A: We wanted to start everyone on the same line. That means the same level of strength on the same lifts will get you the same strength score and gaining 10 lbs in a particular lift will mean the same amount of progress, no matter who you are.
We plan to roll out rankings and percentiles of your strength score based on cross-sections of age, gender, height, and weight. That means that if you’re a 35-year-old 5’4” woman who weighs 125 lbs, you would be able to see that you are in the 80th percentile of users like you with your quadriceps strength score of 66.
Q: Is this score a percentile?
A: No, this score is not a percentile. While this score and percentiles are both calculated using population statistics, percentiles skew the data based on the curve of a normal distribution. The skew of percentiles causes the same 5kg 1 rep max (1RM) gain to be a large jump for users near the average and to make almost no difference for users farther away from the average. Our score will increase by the same amount for any given 5kg change in 1RM.
Q: Will my scores change as the Fitbod population changes?
A: No. We have created a snapshot of the Fitbod population that will be fixed for the future, barring any unforeseen issues. This means that, if you do the same sets, reps, and weight of the same exercise for eternity, your associated strength scores will not change.
Q: Can my score go above 100?
A: There is always more progress to be made, so we didn’t want to limit the strength scores you can achieve. 1.67% of all weekly strength score data points are above 100 in our current historical data, so know that you are in rarefied air if you get above 100.
Q: What if I cheat?
A: You won’t reap the benefits of this metric if you try to fake it. Plain and simple: You might see higher scores but those numbers won’t translate to more strength. Your experience with the app will also likely suffer, as we plan on using these progress metrics to optimize your workout recommendations and fit your personal goals. If you enter fake values, your recommendations will get worse. In short, don’t do it.
Q: What is “mSTRENGTH”?
A: We needed to give Muscle Strength scores a unit to help you all understand that the number you’re seeing isn’t a percentile, nor is it pounds or kilograms. So we chose “mSTRENGTH” to convey Muscle Strength—Muscle Strength and mSTRENGTH are exclusive to Fitbod.
Q: How can I see my Muscle Strength scores?
A: As of December 2022, you can see some of your Muscle Strength scores in your 2022 Year in Review. We’re planning to add Muscle Strength scores into the Fitbod app soon so you can access scores for all of your muscle groups at any time. Stay tuned!