Every lifter has tried an exercise once and known, mid-set, that they would never do it again. We wondered which exercises collect the most of those verdicts, and how fast the verdicts arrive.
So we followed 7 million “first tries”: every time one of our users logged an exercise for the first time between September 2025 and February 2026, provided they kept training afterward. Then we watched what happened, did the exercise become part of their routine, or disappear? And when it disappeared, was that a real rejection, or did the app simply stop serving it? The result is a clear picture of exercise abandonment data.
The answer has a shape, and the shape is abrupt.
Key takeaways
- Abandonment is front-loaded. When an exercise disappears from someone’s logs for good, 70% of the time it happens within the first two tries. The risk of quitting an exercise falls nearly sixfold between the first performance and the tenth: survive the audition, and you’re in the program.
- The fastest-abandoned exercises share a family resemblance: balance-trainer moves, medicine-ball accessories, combo lifts, and pause variations lose about 9 in 10 first-timers within two tries.
- Disappearing isn’t always rejecting. When the app kept re-offering an exercise someone had tried, they refused it for good only 9.5% of the time, but that ranges from 1.3% (lat pulldown) to 29% (cable hip adduction). We rank the truly vetoed separately.
- One skip is not a breakup: half the people who decline an exercise the next time it’s offered eventually come back to it. For the lat pulldown it’s 84%. For single-leg glute bridges, under 30%.
Join our email newsletter for more fitness insights!
The verdict comes in the first two sessions
Start with the universal pattern. For every first try in our cohort, we asked: given that you’ve now done this exercise n times, what’s the chance that was the last time?
| Performances so far | Chance that was the last one |
|---|---|
| 1 | 38.6% |
| 2 | 25.7% |
| 3 | 19.2% |
| 4 | 15.8% |
| 5 | 12.8% |
| 6 | 11.0% |
| 7 | 9.5% |
| 8 | 8.4% |
| 9 | 7.5% |
| 10 | 6.7% |
The first performance is the most dangerous moment in an exercise’s life. Nearly four in ten first tries are also last tries, and the risk falls with every repetition after that, by the tenth session it’s down nearly sixfold. Add it up and 70% of all abandonments happen within the first two performances.
In other words: lifters don’t slowly drift away from exercises. They audition them, and the casting decision comes almost immediately. An exercise that survives your first three sessions has essentially made the roster.
The curve’s level varies enormously by exercise, though. The lat pulldown’s quit risk is 7–8% flat at every stage, it barely has an audition phase. The Turkish Get Up loses 75% of its first-timers on try one.
The exercises that die on the first date
Here are the fastest-abandoned exercises: among people who tried them (and kept working out), the share gone within two performances. The typical exercise in our catalog loses 76% of its new adopters this way, but the top of the list approaches total mortality:
| Exercise | First-timers | Gone within 2 tries |
|---|---|---|
| Wall Knee Drive | 1,179 | 92.7% |
| Medicine Ball Hip Thrusts | 1,628 | 91.7% |
| Single Leg Glute Bridge and Squeeze | 1,827 | 91.5% |
| Balance Trainer Push Up | 1,762 | 90.9% |
| Overhead Dumbbell Lunge | 1,546 | 90.5% |
| Windmill | 1,131 | 90.5% |
| Box Front Squat | 1,011 | 90.5% |
| Balance Trainer Dome Push Ups | 1,415 | 90.5% |
| Medicine Ball Glute Bridge | 1,438 | 90.4% |
| Pause Barbell Decline Bench Press | 1,811 | 90.0% |
| Medicine Ball Side Lunge | 1,457 | 89.9% |
| Archer Pull Up | 1,032 | 89.7% |
Just below the cutoff, at 89–90%: pause squats, the Romanian-deadlift-to-bent-over-row, the dumbbell clean, and the lunge-with-a-twist family.
Read the list and you can almost feel why. Balance-trainer anything. Medicine-ball accessories. Two-lifts-stapled-together combos. Pause variations of already-hard lifts. These are exercises that demand coordination, setup, or patience before they pay anything back, and the payback window, we now know, is about two sessions long.
At the other end, the survivors’ club is familiar: the lat pulldown loses only 14% of first-timers within two tries, the dumbbell bicep curl 18%, leg extension and cable row about 20%. Simple, loaded, immediately rewarding.
Vanished or vetoed?
Here’s the nuance that keeps the list above honest: Fitbod’s algorithm rotates exercises. When a move disappears from your logs, that’s sometimes your verdict, and sometimes the app just stopped bringing it up. Disappearance alone can’t tell you which.
So we tightened the definition. Take only the cases where someone tried an exercise once or twice, and the app then demonstrably put it back in front of them, at least twice, in workouts they went on to perform, and they still never touched it again. No ambiguity: they saw it, they passed, repeatedly. Call it a veto.
By that strict standard, vetoes are rarer than raw disappearance suggests, 9.5% overall, but the exercise-to-exercise spread is wide and telling:
| Most vetoed | Rate | Least vetoed | Rate |
|---|---|---|---|
| Dumbbell Fire Hydrant Circle | 29% | Lat Pulldown | 1.3% |
| Cable Hip Adduction | 28% | Dumbbell Bicep Curl | 1.9% |
| Step Up | 28% | Hammer Curls | 2.1% |
| Dumbbell Superman | 26% | Dumbbell Shoulder Press | 2.5% |
| Cable Hip Abduction | 26% | Cable Rope Tricep Extension | 2.5% |
| Cable Wood Chop (Low to High) | 26% | Leg Extension | 2.9% |
| Rack Pulls | 25% | Dumbbell Row | 3.1% |
The two rankings agree only partially, and the disagreements are the interesting part. The standing dumbbell tricep extension vanishes from 65% of its adopters’ logs within two tries, but when it’s actually re-offered, only 4% refuse it. It isn’t disliked; it’s just easily displaced. The cable hip adduction, meanwhile, is high on both lists: it disappears and it’s refused to its face.
That’s a genuine no.
One skip isn’t a breakup
A related surprise: declining an exercise once doesn’t mean it’s over. Among people who passed on an exercise the very next time it was offered after their first try, half, 51.9%, eventually did it again anyway.
But the range is the story. Skip the lat pulldown once and there’s an 84% chance you’ll be back; the cable row and leg extension run close behind. Skip a single-leg glute bridge and the return odds drop below 30%. For the staples, a skip is a scheduling decision. For the awkward accessories, it’s usually the beginning of the end.
Same fate, different tempo
One more cut, because it’s the cleanest illustration of what “fastest” means. Some exercise pairs end up equally abandoned, but get there at completely different speeds:
| Exercise | Eventually abandoned | Quit within 2 tries | Median tries before quitting |
|---|---|---|---|
| Dead Bug | 76% | 71% | 1 |
| Deadlift | 74% | 47% | 3 |
| Kettlebell Sumo Squat | 72% | 74% | 1 |
| Back Squat | 68% | 41% | 3 |
The dead bug and the deadlift lose almost the same share of their adopters in the long run. But the dead bug is dismissed on sight, median one try, while the deadlift gets a real trial: people work with it for several sessions before moving on. Big, familiar lifts earn deliberation. Accessories get a single date and a quiet unfollow.
What to do with this
Two practical readings.
If you’re adding a new exercise to your routine, know that you’ll be tempted to judge it inside one session, everyone is, and that the first session is the worst possible evidence: you’re uncoordinated at it, the weight is a guess, and none of the payoff has arrived. The data says exercises that survive to a third session tend to stay for good. Give a new movement three honest tries before the verdict.
And if there’s an exercise you’ve now vetoed repeatedly, you’re in good company, and it’s not a character flaw. The moves people actually refuse are overwhelmingly the awkward, low-feedback accessories, not the core lifts. Swap in something that trains the same muscle from the survivors’ club, and spend your discipline where it compounds.
How we measured this
We followed every (person, exercise) first try from September 2025 through February 2026, 7 million of them, from people who logged at least eight workouts in the twelve weeks after the try, so ordinary gym-quitting doesn’t masquerade as exercise-quitting. Performances are distinct training days, tracked through June 2026; an exercise counts as abandoned only when its last appearance left at least two months of silence before our cutoff. “Quit risk after the nth performance” is computed among pairs that reached n. Vetoes use Fitbod’s recommendation logs: the exercise had to reappear at least twice in workouts the person actually performed after their last try, with no return. Standard catalog only, custom, cardio, and mobility exercises excluded.
Final thoughts
Seven million first tries point to the same conclusion: exercise adherence isn’t a slow fade, it’s a snap decision. The audition happens in the first two sessions, and everything after that is mostly loyalty. Knowing that changes how you should treat a new movement, and how you should read your own history with old ones. The exercises you’ve quietly dropped weren’t a failure of willpower; they were, more often than not, just poor first dates. The ones that survived did so because they gave something back early, and that’s the real design lesson buried in this data: the exercises worth keeping are the ones that pay off before the patience runs out.
