Massage for Sore Muscles: It Eases the Ache. A Strength Boost Is Unproven

Somebody finally lined them all up.

In 2018 a French research group pooled 99 studies of post-exercise recovery methods — active recovery, water immersion (cold, cool or warm), contrast baths, cryotherapy, compression garments, electrical stimulation, stretching, massage and hyperbaric (pressurised-chamber) therapy — analysing each method’s studies on their own and then setting the results side by side. One question: which of these actually reduces soreness, fatigue and the blood markers of muscle damage? [1]

Massage came out on top. Their sentence, not ours: “Massage was found to be the most powerful technique for recovering from DOMS and fatigue.” DOMS is delayed onset muscle soreness — the ache that shows up a day or two after a hard session, not during it. Massage also came top, alongside cold, for reducing inflammation. [1]

So in that line-up the rubdown came out ahead of cold water and compression tights. Compression boots, the inflatable sleeves, were not in it, and nor were laser light or vibration. Case closed?

Not quite. What massage wins is a specific thing, and it is not the thing on the poster.

And even the win is softer than it sounds. The ranking came from setting each method’s pooled result beside the others’, not from trials that put the methods against each other, and massage’s own soreness figure rested on 158 people in 14 study groups [1]. The trials that do test massage directly against another method are small, and in the ones we read neither side came out clearly ahead on soreness. In 72 ultramarathon finishers, a 20-minute massage eased pain straight after the race better than lying down, massage and compression boots both eased fatigue, and from the next day to the seventh no group differed from the others [7]. In 46 runners after a half-marathon, massage and a cold-water bath both left legs less sore the next day than resting did [8]. In 60 amateur athletes, massage and a cold-water bath each eased pain over the following days compared with doing nothing [9]. And in the 2025 review of reviews further down, heat and light therapy showed bigger average effects than massage at 24 and 48 hours, on far fewer trials [5].

What it definitely does: the soreness goes down

A separate 2017 meta-analysis went at soreness directly — 11 trials, 239 people, all of them put through deliberately hard exercise and then massaged or not. (The review’s own total, 504, counts the same people again at each of up to three time points.) [2]

Before the numbers, the scale, because a number like “1.51” means nothing on its own. Researchers express these as effect sizes. Roughly: 0.2 is barely there, 0.5 is moderate, 0.8 and above is what they call large. A minus sign just means the thing went down.

Soreness in the massaged groups fell by 0.61 at 24 hours, 1.51 at 48 hours and 1.46 at 72 hours. The 48-hour figure is large on that scale, though three of the eleven trials carry much of it: without them, the other eight average about 0.8, which is our rough arithmetic from the review’s own chart, not a published figure. This kind of soreness usually peaks somewhere between 24 and 72 hours. Creatine kinase — an enzyme that leaks into the blood when muscle is damaged, used as a rough proxy for how beaten up you are — fell by 0.64. [2]

That is a real, repeated, decent-sized effect on how bad you feel. Three separate reviews agree on it, though they are less independent than that sounds: the 2025 review’s massage figures are built partly from the 2017 review’s trials, and all of them rest on small trials in which people knew whether they had been massaged. If your goal is to be less sore on Thursday, this is one of the better-supported things you can do.

What it does not reliably do: make you perform better

Here is the biggest massage-only review on this page, and it is the one that complicates everything above.

In 2020 a team at the University of Sheffield pulled together 29 randomised studies covering 1,012 participants and ran separate analyses for each outcome rather than lumping them together — strength, jump, sprint, endurance, flexibility, fatigue and soreness, each on its own. [3]

Their finding, verbatim: “We found no evidence that massage improves measures of strength, jump, sprint, endurance or fatigue, but massage was associated with small but statistically significant improvements in flexibility and DOMS.” [3]

Read that twice, because it is the whole verdict. Soreness: yes. Flexibility: a little. Everything you would actually measure in a gym or on a track: no clear benefit. No evidence is not the same as no effect, and the authors say so: “it is impossible to conclude that massage cannot improve performance if the correct timing and indication could be defined.” [3] The 2016 review below found small gains when people were retested within minutes.

One caution on fatigue. The review’s pooled result for fatigue leaned toward massage without reaching significance, and four of its five trials pointed that way [3]. The fifth, a trial of eight amateur boxers, is plotted against massage; its own paper reports that the boxers rated themselves more recovered after massage than after rest, 19.0 against 12.1 on a 28-point scale [10]. Turned the right way round, the five trials favour massage clearly, by our rough arithmetic from the review’s chart rather than any published figure. The 2018 review, too, put massage first for perceived fatigue [1].

The two reviews disagree, and we are not going to hide which side is bigger

The 2017 soreness meta-analysis did report improvements in muscle performance — maximal isometric force up 0.56, peak torque up 0.38, both small-to-moderate. [2] The 2020 review found no clear effect on strength. [3]

So which is right? We cannot tell you for certain, and anyone who says otherwise is guessing. What we can tell you is the shape of the disagreement: the review finding no performance benefit had roughly three times the studies and four times the people, and it counted each person once, at the measurement soonest after the massage, where the 2017 review pooled up to three time points into a single total effect. Those are reasons to weight it more heavily. They are not proof, and one point runs the other way: the 2025 review of reviews further down rated the 2020 review’s methods critically low, and the 2017 review’s low.

It would have made a better headline to quote only the 2017 numbers. That is precisely why we are showing you both.

The dose finding is the useful part, and it runs backwards

A 2016 meta-analysis of 22 randomised trials asked a question nobody selling massage asks: under what conditions does it actually work? [4]

Two answers, both counter-intuitive:

Shorter tended to beat longer. Five to twelve minutes did roughly six times as much as anything over twelve on the performance tests, a gap the authors call a tendency. That is about performance, not soreness: the soreness trials further up used massages of 6 to 30 minutes.

And it shows up soon after the effort, or barely at all. The benefit was clearest when people were retested within ten minutes of finishing an exhausting bout, with the massage filling most of that break, and small when the retest came later, anywhere from more than 20 minutes to two weeks after. That clock starts at the end of the exercise, not at the massage. Our reading, not the review’s: whatever massage does for performance looks like a short lift between bouts rather than faster repair.

The reviewers’ own conclusion is blunter than anything we would write: “it remains questionable if the limited effects justify the widespread use of massage as a recovery intervention in competitive athletes.” [4]

Note who that sentence is aimed at: competitive athletes. The review also found a tendency for untrained people to gain a little more than trained athletes did: +6.5% against +2.3%, or effect sizes of 0.23 against 0.17, both small.

So what about the gun?

Percussion massage devices sell the same promise for use at home, so it is worth asking whether the machine reproduces the hands.

A 2023 systematic review screened 281 records and found 11 studies worth including — ten at moderate risk of bias, one at high. Massage guns improved flexibility and range of motion, and helped stiffness and strength after a fatigue protocol. But for strength, balance, acceleration, agility and explosive work, the review reports the device “either did not have improvements or it even showed a decrease in performance”. [6]

Two other analyses point different ways. A 2023 review of 13 device studies reported short-term gains in strength, explosive strength and flexibility after a single session, from studies its authors say all had weaknesses [11]. A 2026 meta-analysis of 12 randomised trials after muscle-damaging exercise found better recovery of jump height (at low certainty) and no clear effect on soreness (very low certainty) or on maximum strength (moderate certainty); it was funded through a joint laboratory of its authors’ university and a fitness company that sells massage guns [12]. So the device evidence is thinner than the hands’, and on soreness its one pooled estimate is unclear. If you use one immediately before lifting something heavy, the evidence does not obviously support that and in places argues against it.

The caveat that applies to this whole page

In 2025 a Polish team did the unglamorous job: they collected 29 systematic reviews covering 863 unique randomised trials across 24 different recovery treatments, and graded the reviews themselves. [5]

Seventeen of the twenty-nine were rated critically low methodological quality. Only two were rated high. [5]

That includes reviews cited on this page: the 2020 review this page leans on for performance was rated critically low, and the 2017 and 2018 reviews low. It does not make the soreness finding disappear — massage still showed significant effects at 24, 48 and 72 hours in their evidence map. At 24 hours its evidence earned the best grade any treatment got (Class III on a five-step scale, where Class V means no significant effect); at 48 hours it dropped a grade, below compression, contrast baths, kinesiotape and cryostimulation. The authors list it among the treatments with some strong evidence behind them. But it means the literature underneath every recovery claim you have ever read, including the ones we are reporting, is thinner and messier than the confident tone of the average article suggests. [5]

We would rather hand you that than pretend our sources are cleaner than they are.

The verdict

Supported. Massage reliably reduces how sore you feel after hard exercise, with the biggest effect around 48 hours, and a 2018 review that set the common recovery methods side by side put it first for soreness. That is a genuine finding and it is not a small one. It is not a head-to-head win: in the small trials we read that pitted massage directly against cold water or compression boots, neither side came out clearly ahead on soreness.

There is no good evidence it makes you stronger or faster in the hours and days after. The biggest massage-only review on this page found no clear effect on strength in 12 studies of 346 people, or on jump, sprint or endurance in smaller analyses, which is not the same as showing it does nothing, and its authors say so. Feeling less fatigued is another matter: the 2018 review put massage first for perceived fatigue, and the 2020 review’s fatigue trials lean the same way, clearly so once the boxing trial plotted the wrong way round is turned (our arithmetic, above). Rated on its own, the performance question would be Preliminary: real trials, reviews that disagree, nothing good enough to settle it.

Which is a perfectly good reason to get one anyway. Feeling less wrecked is worth something on its own, and it is a more honest thing to sell than a performance gain the trials do not clearly show. If you want the version the evidence actually supports: the soreness trials in the 2017 review gave massages of 6 to 30 minutes after hard exercise, some repeated daily, and what they reliably changed was how sore people felt over the next three days; whether you will test better is the unsettled part. The five-to-twelve-minute figure further up comes from performance tests, not from soreness.

The rest of the recovery aisle, checked the same way: ice baths, which ease soreness and, used straight after lifting, may cost a little of the muscle growth you trained for; compression boots, which in pooled trials ease soreness a little, with no significant effect on strength or power; and tart cherry juice. A different massage claim entirely — lymphatic drainage for bloating and detox — did not survive contact with the evidence, and it is worth knowing the two are not the same product.