Ranked by evidence

Strength training exercises, graded by evidence

The lifts people actually do, with what the research says about what each one builds, how to perform it, sets and reps, who should be careful, and alternatives.

Back squat Strong evidence
The barbell back squat reliably builds the quadriceps, glutes and adductors and is one of the best-studied lower-body exercises in training research.
Barbell bench press Strong evidence
The barbell bench press is one of the best-studied ways to build the chest and triceps, and the standard test of upper-body strength.
Biceps curl Strong evidence
Curls reliably build the elbow flexors, and in a direct comparison they grew them about twice as much as rows did.
Leg extension Strong evidence
The leg extension isolates the quadriceps and is the best-supported way to grow the rectus femoris, the one quad head that squats and leg presses barely touch.
Leg press Strong evidence
The leg press builds the quadriceps about as well as free-weight squats and also grows the glutes and adductor magnus, with no balance or spinal-loading limit.
Barbell hip thrust Moderate evidence
The barbell hip thrust grows the glutes, and it does so with little growth of the quads or adductors, which makes it a good targeted addition.
Barbell row (bent-over row) Moderate evidence
The bent-over barbell row activates the whole back, from the lats and mid-back down to the spinal erectors, and builds pulling strength.
Bulgarian split squat Moderate evidence
The Bulgarian (rear-foot-elevated) split squat trains one leg at a time and builds the quads and glutes with far less load on the bar than a back squat.
Cable fly Moderate evidence
Flyes isolate the chest without the triceps.
Conventional deadlift Moderate evidence
The conventional deadlift trains the whole posterior chain and grip and is a core strength lift, but research on its growth effects is thin: trials either use other deadlift variants or measure strength rather than muscle size.
Dumbbell bench press Moderate evidence
The dumbbell bench press works the chest about as hard as the barbell version, with roughly 15% less weight and a freer path for the shoulders.
Front squat Moderate evidence
The front squat holds the bar on the front of the shoulders, which keeps the torso more upright and loads the knees less than the back squat while recruiting the thigh muscles about as much.
Goblet squat Moderate evidence
The goblet squat (one dumbbell or kettlebell held at the chest) is an easy-to-learn deep squat that works the quadriceps hard.
Hack squat (machine) Moderate evidence
The machine hack squat lets you squat deep on a fixed track with your back supported, which makes it easy to push the quads close to failure.
Incline bench press Moderate evidence
The incline bench press builds the chest and pressing strength about as well as the flat bench, and the little evidence available suggests it adds a bit more growth to the upper chest.
Incline dumbbell curl Moderate evidence
The incline curl puts the arm behind the body, stretching the long head of the biceps.
Lat pulldown Moderate evidence
The lat pulldown trains the same vertical pull as the pull-up but lets you pick any load, which makes it ideal for beginners and for high-volume back work.
Lateral raise Moderate evidence
Lateral raises grow the side of the shoulders: in a randomized within-person trial, trained lifters added about 3–5% lateral deltoid thickness in 8 weeks, with dumbbells and cables working equally well.
Lunges Moderate evidence
Lunges (forward, reverse or walking) train the quads and glutes one leg at a time and highly activate the glutes.
Lying leg curl Moderate evidence
The lying (prone) leg curl builds the hamstrings: in a 12-week MRI trial it grew them by about 9%.
Nordic hamstring curl Moderate evidence
The Nordic curl is the most studied exercise for preventing hamstring strains: programs that include it had about half the injuries in a large meta-analysis, though a stricter reanalysis of randomized trials found the effect uncertain.
One-arm dumbbell row Moderate evidence
The one-arm dumbbell row trains the lats and mid-back one side at a time, with a long range of motion and a supported torso.
Overhead triceps extension Moderate evidence
Training the triceps with the arm overhead stretches the long head, and in a 12-week MRI trial it grew the triceps about 1.
Pull-ups (and chin-ups) Moderate evidence
Pull-ups load the lats and biceps heavily with nothing but a bar, and they are one of the best tests of relative upper-body strength.
Push-ups Moderate evidence
Push-ups build the chest and pressing strength about as well as the bench press when they are made hard enough: in small randomized trials, push-ups matched to bench press load or effort gave similar strength and muscle gains.
Romanian deadlift Moderate evidence
The Romanian deadlift is a hip hinge that loads the hamstrings and glutes at long muscle lengths, and trials of stiff-leg deadlifts and similar hip-extension exercises show real hamstring growth, mainly in the semimembranosus and biceps femoris long head.
Seated cable row Moderate evidence
The seated cable row trains the lats and mid-back with steady tension and very little demand on the lower back, which makes it easy to push close to failure.
Seated leg curl Moderate evidence
The seated leg curl trains the hamstrings with the hip flexed, which keeps them long, and in the only head-to-head trial it grew them more than the lying curl (+14% vs +9% in 12 weeks).
Skull crusher (lying triceps extension) Moderate evidence
The skull crusher is a lying elbow extension with the arm at about shoulder height, a position between the pushdown and the overhead extension.
Standing calf raise Moderate evidence
The standing (straight-knee) calf raise is the best-supported calf builder: with the knee straight the gastrocnemius works at long lengths, and in a 12-week MRI trial it grew far more than with seated raises.
Step-ups Moderate evidence
Step-ups train one leg at a time and, in a randomized trial with rugby players, built strength about as well as back squats, with gains carrying over between the two.
Sumo deadlift Moderate evidence
The sumo deadlift uses a wide stance and an upright torso, shifting work toward the quadriceps and adductors and reducing the load on the lower back compared with the conventional deadlift.
Triceps pushdown Moderate evidence
The cable pushdown builds the triceps and lets you handle more weight than overhead work.
Back extension (Roman chair) Limited evidence
The 45° or horizontal back extension is a decent accessory for the glutes, hamstrings and back-muscle endurance, but it is a weak builder of lower-back strength: in two trials, 10–12 weeks of Roman chair training did not raise back-extension strength.
Dips (parallel bars) Limited evidence
Dips almost certainly train the chest and triceps hard, but no training trial has measured what they build, so the evidence is limited to muscle-activation and movement studies.
Face pull Limited evidence
Face pulls are popular for the rear delts and for 'shoulder health', but no trial has measured what they build or whether they prevent shoulder problems.
Hip abduction machine Limited evidence
The seated hip abduction machine targets the gluteus medius well in EMG studies, but evidence that it builds muscle is thin: in the one trial, 10 weeks of cable hip abduction grew the gluteus medius and minimus by about 4% and mostly the upper gluteus maximus.
Overhead press (shoulder press) Limited evidence
The overhead press is the classic shoulder builder and the standard test of overhead strength, but direct evidence is thin: we found no randomized trial measuring how much shoulder muscle it builds, and the data are mostly activation studies.
Rear-delt fly (reverse fly) Limited evidence
Rear-delt flyes, on the reverse-fly machine or bent over with dumbbells, load the back of the shoulder more than rows do, but the evidence is limited to activation and modelling studies: we found no trial measuring rear-delt growth from them.
Seated calf raise Limited evidence
The seated calf raise bends the knee, which slackens the gastrocnemius, so it is a poor calf builder: in the one long-term trial the gastrocnemius barely grew (about 1%), versus 9–12% with standing raises, and the soleus grew no more than with standing raises.