Machines vs Free Weights for Muscle Growth: What the Research Actually Shows

TL;DR: Machines and free weights grow muscle about equally. The best available meta-analysis puts the hypertrophy difference at essentially zero. Strength is where specificity bites: you get better at the exact thing you train. So choose per exercise based on which tool lets you load the target muscle hard and take the set close to failure without the setup becoming the limiting factor.

---

Walk into any gym and you'll find someone happy to tell you that machines are for people who aren't serious. Free weights recruit stabilizers, build "functional" strength, hit the core. Machines put you on rails and do the work for you.

It's a great story. The trouble is that researchers have run this comparison directly, more than a dozen times, and the muscle does not appear to care.

The Pooled Evidence

Haugen et al. (2023), published in BMC Sports Science, Medicine and Rehabilitation, is the largest synthesis on this question. Thirteen studies, 1,016 participants, 789 men and 219 women, comparing free-weight training against machine-based training head to head.

For hypertrophy, the pooled standardized mean difference was -0.055 with a p-value of 0.751. In plain terms: no difference, and not even a hint of one. That result is about as flat as findings get in exercise science.

Heidel, Novak and Dankel (2022) reached the same conclusion in the Journal of Sports Medicine and Physical Fitness, reporting comparable outcomes for both muscle size and power across modalities.

The individual trials line up. Schwanbeck et al. (2020) randomized 46 young adults to eight weeks of free weights or machines, three to four sets of four to ten reps, each muscle group two to three times per week. Muscle thickness and strength improved similarly in both groups. The free-weight group did show larger acute spikes in salivary testosterone and cortisol, which is exactly the kind of finding that gets clipped into a caption. It didn't translate into more muscle, because acute hormone bumps within a session have repeatedly failed to predict growth.

Three independent lines of evidence, same answer. If your goal is size, the equipment is not your bottleneck.

Where Specificity Genuinely Matters

The strength picture is different, and this is the part people gloss over in both directions.

Both reviews found a clear specificity effect. When strength was tested on a free-weight exercise, the free-weight groups won (SMD -0.210, p = 0.023). When it was tested on a machine, the machine groups won (SMD 0.291, p = 0.064). Head to head on neutral tests, including isometric strength, the difference vanished.

That isn't a point about muscle. It's a point about skill. A barbell squat is a motor pattern with a bar path, a bracing strategy, and a balance requirement you have to rehearse. A leg press removes most of that. If your goal includes a number on a specific lift, whether that's a powerlifting total or just wanting a confident 315 squat, you have to practice that lift. No amount of leg pressing will make you good at squatting.

So the honest framing is:

• Want more muscle? Either works. Choose on other grounds.
• Want a bigger squat, bench or deadlift? Train the squat, bench and deadlift. Everything else is accessory work.
• Want general strength and health? The reviews explicitly conclude it's down to preference and adherence.

If you're estimating maxes from your working sets, remember that a machine-based estimate does not transfer to a barbell version of the same movement. A 1RM calculator is only accurate for the exact lift you fed it.

What Actually Decides the Outcome

Once you accept that the equipment is close to neutral, the useful question becomes: which tool lets you do the things that do drive growth?

Those things are well established. Enough hard sets per muscle each week. Sets taken close enough to failure. Load and reps that climb over time. Everything else is detail.

Refalo et al. (2023) found hypertrophy improves as sets are taken nearer to failure, with the returns flattening in the last rep or two. That's the criterion I'd use to pick equipment, because it's the one where the two options genuinely differ in practice.

Machines usually win when the setup is the limiting factor. A set of Bulgarian split squats often ends because your balance gives out or your front foot is screaming, not because the quad is done. On a leg press or a hack squat, the quad is the thing that fails. Same for lateral raises versus a machine lateral raise late in a session, or a barbell row versus a chest-supported row when your lower back is fried from deadlifts. If you can't reach the muscle's true failure point, you're leaving stimulus on the table.

Free weights usually win on loadability and range. Dumbbells let you find a groove that fits your joints. A barbell lets you keep adding weight for years without running out of plates or hitting the top of a stack. And free-weight versions of many movements load the muscle better in the stretched position, which matters given what we know about training at long muscle lengths.

Machines win on safety near failure. Failing a rep on a pec deck costs you nothing. Failing a rep under a loaded barbell without a spotter or safeties costs you a lot more. That difference alone can be worth more than any theoretical modality advantage, because it changes how hard you're willing to push.

The Stabilizer Argument, Examined

The most common defense of free weights is that they recruit stabilizers and therefore build more total muscle. There is a real kernel here: free-weight variations do produce more activation in stabilizing musculature.

The problem is the leap from activation to growth. EMG amplitude during a set is a measure of electrical activity in a moment, not a measure of the mechanical tension a fiber experiences over a full set, and it has a poor track record predicting long-term hypertrophy. If extra stabilizer recruitment produced meaningfully more muscle, the free-weight groups in these trials would have grown more. Across 1,016 people, they didn't.

The stabilizers do get trained, and that's genuinely useful for balance, coordination and confidence under load. It just isn't a hypertrophy argument.

How to Choose, Exercise by Exercise

Stop treating this as a gym-wide philosophy. Decide movement by movement:

1. Anchor your program with a few free-weight compounds you intend to get strong at. Squat or a variation, a hinge, a press, a row. These carry the specificity benefit and the long runway for progressive overload.
2. Use machines and cables for the isolation and high-fatigue slots. Late-session work, single-joint work, and anything where stability fails before the muscle does.
3. Pick whichever version you can add weight or reps to most consistently. That's the real test.
4. Fill the week to 10 to 20 hard sets per muscle regardless of what equipment those sets came from. Weekly set volume is the variable with the clearest dose response in the literature, and it doesn't ask where the load came from.
5. Take the sets seriously. Within one to three reps of failure on your working sets, per the proximity-to-failure evidence.

A crowded gym is a perfectly good reason to swap a barbell bench for a machine press that day. So is a cranky shoulder. Neither costs you growth, as long as the set is hard and it goes in the log.

The Part That Actually Costs People Muscle

Here's what I've noticed watching training data: almost nobody stalls because they chose the wrong equipment. They stall because their working weights haven't moved in two months and they never noticed, or because their weekly set count for a muscle quietly dropped from 14 to 6 while they were focused on something else.

Those are accounting problems, not equipment problems. And they're invisible unless something is keeping score.

Protokl tracks hard sets per muscle per week and watches whether load and reps are climbing on each movement, machine or barbell. When a lift flattens, it shows up in the data well before it shows up in the mirror. If you've been arguing with yourself about equipment for months, that's almost certainly not the variable holding you back.

References

• Haugen, M.E., Vårvik, F.T., Larsen, S., Haugen, A.S., van den Tillaar, R., & Bjørnsen, T. (2023). Effect of free-weight vs. machine-based strength training on maximal strength, hypertrophy and jump performance: a systematic review and meta-analysis. BMC Sports Science, Medicine and Rehabilitation, 15(1), 103.
• Heidel, K.A., Novak, Z.J., & Dankel, S.J. (2022). Machines and free weight exercises: a systematic review and meta-analysis comparing changes in muscle size, strength, and power. Journal of Sports Medicine and Physical Fitness, 62(8), 1061-1070.
• Schwanbeck, S.R., Cornish, S.M., Barss, T., & Chilibeck, P.D. (2020). Effects of training with free weights versus machines on muscle mass, strength, free testosterone, and free cortisol levels. Journal of Strength and Conditioning Research, 34(7), 1851-1859.
• Refalo, M.C., Helms, E.R., Trexler, E.T., Hamilton, D.L., & Fyfe, J.J. (2023). Influence of resistance training proximity-to-failure on skeletal muscle hypertrophy: a systematic review with meta-analysis. Sports Medicine, 53(3), 649-665.