Fitness Benchmaxxing: The Fallacies of Exercise Scores
Written by Kyle Ligon - MovementLink.FIT Head Coach
The fitness industry operates on a broken assumption: if you improve your back squat, drop your marathon time, or crush a CrossFit benchmark, you are getting closer to your real goals.
There is a systemic issue in the fitness industry where programs target certain performance benchmarks, causing everyday people to be happy with their programs as they make progress on their benchmarks while unknowingly falling well short of their actual goals.
In the tech industry, this exact phenomenon is currently holding back AI models. Engineers call it "Benchmaxxing."
Benchmaxxing occurs when developers tailor an AI to score well on standardized benchmark tests, only to discover that the strategies used to obtain the high test scores have degraded the real-world potential of the model.
Replace "AI model" with "human body," and the fitness industry is guilty of the exact same error. When you work out in a way that exclusively targets gym scores, you overfit to the test and sacrifice potential real-world performance, body composition, tissue & joint health, and overall health & wellness—what we call our Non-Negotiable Goals.
This can mean overfitting to exercise sports like weightlifting, powerlifting, CrossFit, and kettlebell sport, but also to metrics like heart rate and VO2 Max, or single-modality programs like yoga, running, and cycling. When you over-focus on simplified metrics, results typically do not align with comprehensive progress toward your core goals.
Here are the psychological traps and logical fallacies driving this error:
Goodhart's Law “When a measure becomes a target, it ceases to be a good measure.” Benchmarks are useful for identifying the effectiveness and flaws of your training, but only if you pretend like you don’t know what your benchmarks are. As you target specific benchmarks, a positive feedback loop takes over. You stop using the benchmark to measure fitness and start training explicitly to beat the benchmark. Benchmarks are inherently more narrow in focus than the actual results we are after. If we overfit to benchmarks, the gaps between the benchmarks don’t get the focus they need.
The Streetlight Effect An observational bias where people only look for answers where it is easiest to measure. Gym-goers obsess over barbell metrics because the "light is better" there. It is much harder to measure how well you play with your kids or your physical readiness for a random weekend activity. We dismiss what matters simply because it defies easy measurement.
Even within barbell exercises, difficulty (intensity—reps, time, load) is easy to measure, while execution (intent and technique) is hard. Because execution defies easy measurement, it is routinely ignored as long as basic competition standards are met. The only measurement that goes into a workout journal is difficulty.
The McNamara Fallacy Because the Streetlight Effect biases what we measure, the McNamara Fallacy takes it one step further: our decisions shift solely based on easily quantifiable data while ignoring all other factors. Since it is easy to quantify a max back squat, our workouts start pointing more and more toward what improves that number. We employ techniques, gear, and training and forget that some strategies that may produce higher back squat numbers on paper may be at the expense of our Non-Negotiable Goals. It’s hard to imagine that heavier weights moved wearing weight belts, squat suits, and ultra-specific techniques improve our Non-Negotiable Goals more than a raw squat done with functional movement patterns.
The Availability Heuristic Substituting a complex question with an easily recalled cultural symbol. When someone has a new goal, they focus on society's most visible proxy for that goal. Those wanting flexibility look to yoga; those wanting strength look to powerlifting; those wanting cardio look to marathon runners. They adopt the culturally available symbol rather than analyzing if extreme distance running or squatting like a powerlifter will actually produce the systemic results they need within the context of their life.
Construct Validity Failure (The Proxy Fallacy) In statistics and psychometrics, construct validity is the degree to which a test actually measures what it claims to be measuring. A construct validity failure occurs when you choose a proxy that does not represent the overarching trait. I have written about this many times: the heavy barbell snatch is common in CrossFit as one measure of functional fitness, yet it is hard to find a real-world example that resembles anything like the hyper-specific demands of the lift at high loads (Should Your Workouts Include Heavy Barbell Snatches). I have yet to see athletes outside of Olympic weightlifting and CrossFit train the heavy barbell snatch regularly and outperform others in their sport. To me, it has not yet been proven worth the time, effort, and risk, and is likely a very poor benchmark and a very poor exercise choice for those not competing in those two sports.
The Synecdoche Fallacy (Part-to-Whole Bias) Mistaking a single component of a subject for the entirety of the subject. People consider that they are "exercising" when they participate in yoga, powerlifting, or running, without realizing that participating exclusively in these single-modality programs is actually causing them to be detrained in vital areas of fitness. Additionally, even within an area of fitness like strength, back squats, bench presses, and deadlifts executed to powerlifting standards are often considered the gold standards. But these specific exercises are not comprehensive. They ignore odd objects, rotational power, and dynamic sports strength. Prioritizing the standards of powerlifting changes the gear and techniques used to benchmaxx, rather than building real-world strength.
Surrogation vs. Means-End Inversion These two traps work in tandem to destroy a program:
Surrogation: The psychological trap of replacing the overarching goal (functional performance, body composition, tissue & joint health, and overall health) with the proxy metric (CrossFit benchmark scores). I fell for this for years, assuming that improving my CrossFit performance would equate to reaching my goals. It wasn’t until I directly targeted my goals that I started making extreme progress.
Means-End Inversion: The behavioral result. The original goal is forgotten entirely, and the proxy metrics become the obsession. People employ tactics at the expense of their actual goals to target the proxies. They prioritize their score that day over long-term progress and use specialized gear as they game the workout. The score is higher that day, but the future potential is squashed.
Because none of these fallacies are immediately clear, people get boiled like frogs. Because untrained individuals see initial gains in both benchmark scores and real-world capability when they first start, these traps go unnoticed. But over time, the two diverge more and more.
We see the exact same thing happen with nutrition all the time. Why does every nutrition fad have people who swear by it? Because shifting from a terrible diet to almost anything else yields results. But "better" is not the same as "optimal." Doing yoga classes is better than sitting on the couch all day, but yoga alone is far from enough to achieve the Non-Negotiable Goals required to optimize our bodies for a long, active life.
Why are these problems so common? In marketing, the saying is, "Built to show, built to grow." Just like AI models that blow benchmarks out of the water get all the attention, people gravitate toward fitness programs that show massive deadlift PRs as opposed to those that struggle to quantify a decade of being active with pain-free joints.
At MovementLink and CrossFit City Limits, we utilize the base principles of CrossFit but make the real world the target, not the whiteboard. We don't throw out benchmarks; we anchor them with subjective reality checks to ensure the program is actually moving you in the right direction.
Functional Performance
Objective: Fitness Benchmarks that test speed, power, strength, muscle, VO2 Max, anaerobic thresholds, endurance, stamina, mobility, and balance.
Subjective: Do you feel capable of saying “yes” to physical opportunities outside the gym?
Body Composition
Objective: DEXA Scan for body fat, muscle mass, and bone density.
Subjective: Are you happy with how you look and feel?
Tissue & Joint Health
Objective: Movement and Mobility Tests
Subjective: Do you have nagging pains? Are you restricted? How long does it take you to warm up?
Overall Health and Wellness
Objective: Blood Work
Subjective: How do you feel on a daily basis?
At CrossFit City Limits, instead of over-prioritizing the easy-to-measure reps, load, and times of workouts, we pretend we are being scored by gymnastics judges who are also grading us on how well we are adhering to the intent of the workout and the exercises, along with the quality of our technique. For this, we have coined the acronym F.I.I.T. (For Intent, Intensity, & Technique) to represent the balance of priorities we keep during high-intensity efforts.
We do not workout For Time, we workout F.I.I.T.
Our focus is to maximally improve peoples’ lives in the real world. The benchmarks we use are our current best attempt at quantifying the program’s effectiveness, but we are also hyper-aware of not falling for the benchmaxxing traps that are holding back the rest of the industry.
MAKE THE RESULTS YOU WANT A REALITY
I have been running clients through my MOMENTUM transformation program for 8 years with tremendous success. But, with the amount of 1-on-1 attention I provide in the program, it creates a cost that has been prohibitive for some. I have finally consolidated the MOMENTUM program into one, easy to navigate EBOOK, so that more people can finally access the information that has changed my life and so many others.
When you finally “Get It”, the results you’ve been after are inevitable.
