Pilates ball exercises: the small ball and the gym ball, what each adds, and a routine

Two different balls

A small, soft ball you squeeze or rest under your back, and a large gym ball (also called a Swiss, exercise, stability or yoga ball) that you sit, lie or balance on.

Sources 1, 2

Get the size right

Sitting on a gym ball, your hips and knees should be at right angles with your feet flat. An NHS trust's leaflet matches the ball's size to your height.

Source 2

Your trunk works a little harder

Across 86 studies, unstable surfaces raised trunk muscle activity by a small to moderate amount, but not leg muscle activity, and some ball moves lowered it.

Sources 3, 4

The move matters more than the ball

Roll-outs and pikes on a gym ball drew the most tummy muscle activity and a seated march the least. A ball under the lower back doubled crunch activity compared with a ball under the shoulder blades, and Pilates back extensions worked the back muscles less on a ball than on the floor.

Sources 5, 6, 7

More activity isn't more strength

Unstable surfaces reduce the force you can produce, so the Canadian Society for Exercise Physiology (CSEP) says they shouldn't be athletes' main training, though they have a place in rehabilitation and for people who don't lift weights. Squats and deadlifts work the trunk as hard or harder.

Sources 8, 9

Sitting on a ball isn't core training

In lab studies, sitting on a ball changed little or made the lower back work harder, it was less comfortable, and getting used to it didn't improve posture.

Sources 10, 11, 12, 13

The small ball

Teachers call the small, squashy ball one of the most versatile props in a Pilates studio. Between the inner thighs, ankles or palms, it "invites the engagement of the surrounding muscles"; under the spine or hips, it gives support; and pregnant women can use it behind the mid-back to avoid lying flat on their back for long.

Sizes on sale vary and no health source gives one; our adductor squeeze uses a soft ball about the size of a football. A cushion, a clean soft football or a yoga block can stand in.

Sources 1, 14

The gym ball

An NHS physiotherapy leaflet describes it as "a large ball made of plastic vinyl and elastic and filled with air", also called a Swiss ball, exercise ball, yoga ball or stability ball. NHS physiotherapists use it for seated exercises, bridges, crunches and back lifts.

Sources 2, 15

How a wobbly surface changes an exercise

CSEP's position stand says core and limb muscles work harder on unstable surfaces than on stable ones during similar exercises, but that free-weight lifts such as squats and deadlifts can match or beat that activity. Because unstable surfaces cut force, power, speed and range of movement, they aren't recommended as athletes' main training, but they can play a part in rehabilitation, in planned training cycles and for people who prefer not to lift free weights.

It also says bracing your tummy (tightening it all round) has been shown to keep the spine more stable than hollowing (drawing it in).

Sources 8, 16

Where researchers differ

A later review by the same lead author is warmer, finding training on unstable surfaces gives similar results to stable training in recreational exercisers and recommending it for young people, older adults and regular exercisers. A lab study in trained men found squats and deadlifts worked the back muscles harder than three ball exercises and questioned the ball's role in strength training. Both can be true: the ball adds a balance challenge at a low load, but it doesn't replace getting stronger.

Sources 17, 9

A wobble isn't only a ball

On a reformer, a lighter spring makes the carriage less stable: in 18 women, a low-resistance spring raised core muscle activity in basic exercises compared with a fixed platform. Our reformer guide has more.

Source 18

Gym ball size by height

An NHS trust's physiotherapy leaflet gives these sizes. The bands overlap, so if you're between two, sit on the ball: your hips and knees should be at right angles with your feet flat on the floor.

• Under 150 cm tall: a 45 cm ball
• 150 to 165 cm: 55 cm
• 162 to 183 cm: 65 cm
• 180 to 200 cm: 75 cm
• Over 198 cm: 85 cm

Source 2

How firm

"The ball should be burst resistant", and "The firmer the ball the more difficult the exercise will be". A slightly softer ball is easier while you learn (our suggestion).

Source 2

Safety

Use a ball that holds your weight without rolling, on a floor where it can't slide, with space around you (the advice on our ball move pages). NHS physiotherapists' ball leaflet adds: "If you feel dizzy, experience pain or any new symptom during exercise, please stop immediately and seek advice from your Doctor".

Source 15

Unstable surfaces raise trunk muscle activityStrong evidence

A 2024 meta-analysis of 86 studies (1,783 people) found higher activity on unstable surfaces in the straight tummy muscle (standardised mean difference 0.51), the outer and inner obliques (0.44 and 1.04) and two back muscles (0.37 and 0.35), but not in the leg muscles. A 2025 meta-analysis of 28 studies (579 people) agreed. These are small to moderate effects on muscle activity during an exercise, not results of training.

Sources 3, 4

Some ball moves lower itEmerging evidence

The 2025 meta-analysis found less activity in the back muscles during a shoulder press on a Swiss ball, and in the muscle that holds the shoulder blade against the ribs when working on a Swiss ball. A ball made no difference to trunk activity in a face-up bridge, and in 15 women with Pilates experience, back extension exercises worked the back muscles less on the ball than on the floor.

Sources 4, 19, 7

Which move, and where the ball sitsEmerging evidence

In 18 adults, the Swiss ball roll-out and pike drew the most tummy muscle activity and a seated march the least, and the lower back muscles stayed below 10% of their maximum in every ball tummy exercise. In 41 adults, a crunch with the ball under the lower back beat a floor crunch, which beat a crunch with the ball under the shoulder blades. The ball raised tummy muscle activity in a face-down bridge, and in a single-leg hold and the top of a press-up in 8 adults, but not in a face-up bridge.

Sources 5, 6, 20, 19

A ball crunch with a bandEmerging evidence

In 42 untrained adults, a crunch on a Swiss ball against an elastic band worked the straight tummy muscle harder than a crunch machine (104% against 84% of its maximum) with much less work from the hip flexors.

Source 21

Pilates moves with a ballEmerging evidence

In 40 adults, a roll-up holding a ball worked the tummy muscles as much as a standard curl-up. In 28 men doing the Hundred, a small soft ball raised core activity above no prop, though less than a Pilates ring. Both were single-session studies.

Sources 22, 23

Activity isn't strengthStrong evidence

CSEP says unstable surfaces reduce force, power, speed and range of movement, and its review says they aren't recommended as the main exercises for building strength, muscle size or power. They suit rehabilitation, and people who prefer not to lift free weights.

Sources 8, 16

Squats and deadlifts work the trunk tooEmerging evidence

In 9 trained men, squats and deadlifts near their maximum worked the back muscles harder than three ball exercises. In 16 highly trained people, standing on a Swiss ball raised muscle activity, but a BOSU (a half ball) and an inflatable disc didn't. Both were small lab studies.

Sources 9, 24

Training on a ballMixed evidence

Short trials in untrained people found benefits: 5 weeks of ball curl-ups and back extensions improved balance more than the same moves on the floor in 30 women, though strength didn't differ, and 10 weeks twice a week raised back endurance from 149 to 195 seconds and side-bridge time from 45 to 71 seconds in 20 sedentary adults, with no change in those who didn't train. In 18 young athletes, 6 weeks improved a core stability test but not running economy. Two other positive studies had no comparison group.

Sources 25, 26, 27, 28, 29

Ball programmes against other exerciseMixed evidence

In 60 people with fibromyalgia, strengthening on a Swiss ball eased pain and improved quality of life more than stretching. After childbirth, ball training and floor exercises helped back pain equally in 30 women, and a BOSU and a Swiss ball did equally well on the main outcomes in 36 students with back pain. For period pain, Pilates beat gym ball exercises in 30 young women, and Pilates ball exercises improved balance in 62 preschool children. All were small single trials.

Sources 30, 31, 32, 33, 34

Older adults and balanceEmerging evidence

In 42 older women, step aerobics improved two balance and mobility tests, while a stability ball programme didn't, though both improved how healthy the women felt. So don't count on ball work for balance; our balance exercises train it directly.

Source 35

AthletesEmerging evidence

In 12 studies of athletes, core training on unstable surfaces improved trunk strength and sprinting more than core training on stable ground. An earlier review by the CSEP lead author advised using both: stable exercises for force and unstable ones for balance.

Sources 36, 37

Pilates with small props or on apparatusEmerging evidence

In 24 women after the menopause, 8 weeks of Pilates with an elastic band and a gym ball, and of Pilates on a reformer and Cadillac (Pilates apparatus), both improved fitness and mood scores; blood pressure and grip improved only on the apparatus. It was one small trial with no group that didn't exercise.

Source 38

As an office chairEmerging evidence

In 10 women typing for an hour, sitting on a ball meant more trunk movement but 78% higher average lower-back muscle activity and more spinal shrinkage (the spine losing a little height under load). In 14 adults it brought slightly more upper-back muscle activity and more discomfort, and in 8 men it made no difference to muscle use or modelled spine load. In 12 adults, 9 days of getting used to the ball eased discomfort in women, but the researchers found no evidence that using a ball as a chair improves trunk strength or posture. Our sitting guide covers chairs.

Sources 10, 11, 12, 13

As exercise, it's different

The Hypermobility Syndromes Association lists sitting on a gym ball among ways to exercise while seated, and NHS physiotherapists use short seated ball exercises. Short bouts of movement on the ball are a different use from sitting on it all day.

Sources 39, 15

Lying flat in later pregnancyStrong evidence

The NHS says to try to avoid lying flat on your back for more than a few minutes in later pregnancy. Teachers use a small ball behind the mid-back so you can do Pilates without lying flat for long.

Sources 40, 1

Birth balls in labourEmerging evidence

In 7 trials of 533 women, using a birth ball in labour lowered pain scores by 1.7 points on average, with no difference in the type of birth. Our Pilates in pregnancy guide covers pregnancy and birth.

Source 41

Balance on the ball

The NHS says not to do sports or activities where there's a risk of falling or bumping your tummy. Our suggestion: in pregnancy, sit on the ball rather than balancing on it, with a wall or chair within reach.

Source 40

An NHS physiotherapists' gym ball programme

Kent and Medway NHS physiotherapists give these exercises and numbers. There are no trials of this programme, so treat the numbers as a guide.

• Sitting tall: roll the ball forward and back 20 times, side to side 20 times, and in circles 10 times each way
• Sitting balance: feet together and eyes shut for 10 to 20 seconds, 5 times; then one foot lifted for 10 to 20 seconds, 5 times each leg
• Bridges with your upper back on the ball, and with your feet on the ball: 20 each
• Crunches with your shoulder blades on the ball: 15
• Face down over the ball, lifting the opposite arm and leg and holding for 5 seconds: 10 each side
• Back lifts over the ball: 15; side leg lifts over the ball: 15 each side; wall squats with the ball behind your lower back: 15

Source 15

Our starting doses

The stability ball hip bridge and leg curl move pages' dose: 2 to 3 sets of 8 to 12, stopping with 2 or 3 reps in reserve (their editorial starting point). For crunches and back extensions over the ball, 8 to 15 slow repetitions is our editorial starting point, up to the NHS leaflet's 15.

How often

The ball training trials used 2 or 3 sessions a week. Using the ball in 2 or 3 of your sessions a week is our suggestion.

Sources 26, 27, 38

Stop if

NHS physiotherapists' ball leaflet: stop if you feel dizzy, get pain or any new symptom, and seek advice.

Source 15

Call 999 (911 in the US)

If someone faints while exercising, call 999. Call 999 or go to A&E too if back pain comes with pain, tingling, weakness or numbness in both legs, a loss of feeling around your genitals or anus, or changes in your bladder or bowels.

Sources 42, 43

Who should change what

Our suggestions, from the sources here:

• Poor balance, or older and new to exercise: sit on the ball next to a wall or use a chair, and leave out the push-up with your feet on the ball, the knee tuck and the downward balance
• Osteoporosis: leave out crunches over the ball; bridges and back extensions are fine, as the UK consensus recommends spinal extension exercise
• Pregnancy: a small ball behind your mid-back instead of lying flat; sit on the gym ball rather than balancing on it
• A shoulder that has dislocated: skip the kneeling chest stretch on the ball
• Hypermobility: Swiss balls are used to retrain balance; keep drapes over the ball small

Sources 35, 44, 40, 45

"A stability ball makes any exercise better for your core"Mixed evidence

Unstable surfaces raise trunk activity a little on average, but some ball moves lower it, the ball made no difference in a face-up bridge, and Pilates back extensions worked the back less on a ball. More activity also isn't more strength: force and range drop, and squats and deadlifts work the trunk as hard.

Sources 3, 4, 19, 7, 8, 9

"The wobblier, the better"Emerging evidence

In highly trained people, a BOSU and an inflatable disc didn't raise muscle activity, and in older women a ball programme didn't improve the balance tests that step aerobics did.

Sources 24, 35

"Sitting on an exercise ball at your desk strengthens your core"Emerging evidence

In lab studies it changed little or raised the load on the lower back, it was less comfortable, and getting used to it didn't improve trunk strength or posture.

Sources 10, 11, 12, 13

"A small ball under your lower back takes pressure off your abdomen"

A TV fitness page says so, but we found no study. Teachers use the ball under the back as a support, which can make some moves more comfortable.

Source 1

"Ball crunches burn belly fat"Strong evidence

Training one area doesn't remove fat from it: across 13 studies there was no spot reduction, and 6 weeks of abdominal exercises didn't change belly fat in 24 adults.

Sources 46, 47