Learning the lifts: technique, light weights and video

Learn each lift with a weight you can control. End every set while your reps still look the same. And use video: a demonstration before you start, and footage of yourself after. The American College of Sports Medicine's position is that lifting weights is safe for healthy adults of all ages.^[1]

So why doesn't everyone just say that? Because the loudest advice runs ahead of the evidence. "Focus on the bar, not your body" sounds like settled science. It isn't.^[2] And nobody has measured well how often gym beginners get hurt. Some of what follows is coaching practice, and we say so each time.

You've already warmed up (see how to warm up for lifting). This article covers the next part: the first working sets of a lift you've never done before.

Why it matters

Technique, or form, means how you move the weight: the path it travels, where your joints go, and how fast you lower it. A beginner can get it wrong in two opposite ways.

The first is never starting. ACSM notes that many adults, older adults in particular, avoid lifting because they think it's riskier than it is.^[1] That costs them every benefit of training, for a risk that turns out to be small.

The second is starting too heavy. A beginner sees someone with ten years of training squat 140 kg and loads the bar to match. A weight you can't control is harder to lower, harder to rack and harder to learn with. You also can't tell what your knees are doing when every rep is a fight.

What the research says

A few terms first. A position stand is an official summary of the evidence by a group of experts. Here that's the American College of Sports Medicine (ACSM). A meta-analysis pools the results of many studies on the same question into one estimate. A systematic review collects every study on a question in a set way, with or without pooling them.

Lifting is low-risk by sports standards

ACSM's 2026 position stand says resistance training (lifting against a load, such as weights) is safe for healthy adults of all ages.^[1] Expert consensus. This is the panel's view, not a tested finding.

It points to an analysis of trials that used exercise as a treatment, with more than 38,000 people, over 6,700 of whom lifted weights.^[1] The people in these trials were patients, not healthy gym-goers.^[3] Exercise did not raise the risk of serious harm.^[1][3] It did make minor problems such as pain or fatigue more common than not exercising, by about a fifth.^[3] Those minor problems were no more common with lifting than with aerobic exercise (steady exercise such as walking or cycling).^[1] Moderate evidence. More than 11,000 of those people were older adults, so this isn't only about the young and fit.^[1]

Limited evidence on the exact numbers. The direction is consistent across both reviews.

What the emergency-room data show

One large US study counted weight-training injuries treated in emergency departments from 1990 to 2007: 25,335 cases. The most common cause was a weight dropping on the person, at 65.5%. The most common diagnosis was a sprain or strain, at 46.1%.^[6]

Limited evidence. This counts injuries. It doesn't say how often they happen per hour of lifting, and it includes children. It also can't tell you whether form played a part. It does tell you where to pay attention: the moments you pick a weight up, put it down and move it around.

Light weights still build muscle

A meta-analysis of 21 studies compared light loads with heavy ones. "Light" meant 60% or less of the person's one-rep max (1RM: the most weight you can lift once). Every set went to failure: the point where you can't complete another rep. Muscle growth was about the same with light and heavy loads. Heavy loads built more maximal strength: a bigger one-rep max.^[7]

ACSM's 2026 review of 137 systematic reviews, covering more than 30,000 people, reached the same place. Muscle growth wasn't affected by load. Strength went up more with heavier loads, around 80% of your max or above.^[1] Strong evidence. Much of the evidence ACSM pooled came from people new to training, like you.^[1]

There's a catch: the light sets have to be hard. In the light-load studies above, every set went to failure.^[7] ACSM found that training to failure isn't needed in general. Its panel proposes that stopping a few reps short is enough effort, for example 2–3 reps in reserve.^[1] RIR means how many more reps you could have done before you had to stop. The panel also says the evidence can't yet pin down an exact number.^[1] Expert consensus. More on that in how hard a set should be.

ACSM adds a warning for some people, such as older adults. For them, lifting to fatigue may be unwise, partly because of "an increased risk of injury resulting from poor form".^[1] Expert consensus, and ACSM hedges it itself.

Starting light is accepted, but untested

ACSM notes that people without experience may need to start with low loads, for safety among other reasons, and then progress.^[1] Expert consensus.

Does starting light prevent injuries? Nobody has studied that well yet. Starting light is coaching practice, not a tested finding.

ACSM's review also found strength gains were about the same on machines and free weights.^[1] Moderate evidence. So a machine version of a lift is a fair place to start while you learn the barbell one.

Watching first helps you copy the movement

A meta-analysis of motor-learning studies looked at people who watched someone demonstrate a movement before practising it. Compared with people who only practised, they copied the shape of the movement much better. The gain in the end result (such as hitting a target) was smaller.^[8] Moderate evidence. These studies were about learning movements in general, not lifting.

Video feedback may help too. In one trial, 30 healthy people practised the squat. After each set, one group was shown video of their own squats, along with feedback on what needed correcting. They improved their knee position and how far they leaned forward. When tested again, they still showed the gains: three days later for knee position, and a week later for forward lean. A group that got hands-on (touch) feedback from a physical therapist also improved, but didn't keep it.^[9] Limited evidence. It's one small trial, and someone pointed out the fixes, so whether checking your own clips works as well hasn't been tested.

Where to put your attention: the studies disagree

A cue is a short instruction you think about during a rep. An external cue points your attention outside your body: "push the floor away". An internal cue points it at a body part: "straighten your knees".

Mixed evidence. The studies disagree. Here's what each side found.

• For external cues. A 2021 meta-analysis of 73 studies, with 1,824 people, found a small advantage for an external focus on performance. The advantage was bigger when people were tested again later.^[10] These covered many kinds of movement tasks, not lifting specifically.
• Against. A 2024 re-analysis of the same data found moderate to strong signs of publication bias. That means studies with positive results were more likely to be published than studies that found nothing. After correcting for that, the average benefit of an external focus was close to zero. The authors say the effect seems to depend on things nobody has pinned down yet.^[2]
• In lifting. A meta-analysis of 10 studies found an external cue gave a small boost to strength on the day. Over weeks of training, the studies couldn't tell it apart from no difference, apart from a possible edge in lower-body lifts.^[11] Only 3 of those studies ran for weeks, and they all used men.^[11]
• For muscle growth. One trial put 30 untrained young men on 8 weeks of training. Those told to "squeeze the muscle" grew the front of the upper arm more (12.4%) than those told to "get the weight up" (6.9%). Thigh growth was about the same either way.^[12] One small trial, so it shows that "it depends", not a new rule.

The practical upshot: a cue is a tool, not a rule. Use the one that makes the rep look right. Coaching practice: the research hasn't settled it.

How to apply it

Before your first set: watch it

Watch a demonstration of the lift before you try it. Watch it twice. The first time, follow the weight from start to finish. The second time, follow the joints: hips, knees, elbows.

Exercises in the Gymbers app come with a video you can watch before your first set. You can save the videos to your phone, so they're on hand at the gym.

In the session

These steps are coaching practice, not tested findings.

1. Pick a weight you can control on every rep. For a barbell lift, that may be the empty bar. A machine or dumbbell version is fine too.
2. Ramp up first. Do a few lighter sets before your working sets, as in the warm-up article.
3. End the set while the reps still match. If a rep looks different from the first one (hips rise first, back rounds, knees cave in), that was the last rep. Aim to finish with about 2–3 reps left in the tank.
4. Film one set from the side, at about hip height. Prop your phone on a bench or a plate. Watch the clip before your next set. In the video trial above, someone else pointed out the fixes. Checking your own clips, and the best camera angle, haven't been tested.
5. Use one cue at a time. Try "push the floor away" on squats and leg presses. If it makes the rep look better, keep it. If it doesn't, swap it for a different cue, internal or external.
6. Add weight once every rep looks the same. Then add a little and check again. That's progressive overload, the topic of stage 2.
7. Take care between sets. Lower the weight under control, rack it gently and put dumbbells down instead of dropping them. Use collars on the bar and the safety arms in a rack.

A worked example

Here's a 55 kg beginner learning the goblet squat (a squat holding one dumbbell against your chest). Three sets of 8.

• Session 1: 8 kg × 8, 8, 8. On the clip of set 2, her heels lift on the last rep. Next time she'll think "whole foot down".
• Session 2: 8 kg × 8, 8, 8. Heels stay down. Every rep on the clip looks like the first.
• Session 3: 10 kg × 8, 8, 7. The last rep of set 3 leans forward more. She stops there.
• Session 4: 10 kg × 8, 8, 8. Clean on the clip. Next session, 12 kg.

The numbers are an example, not a schedule. Nobody has studied well how long a learning phase should last.

What's practical versus optimal. Your first sets will be easier than the hard sets the muscle-growth studies used. That's a trade: a few easy sessions in exchange for a movement you can repeat under load. Coaching practice.

Common mistakes

Copying someone else's weight. The person on the next rack looks like a fair benchmark. They may have had years to learn the lift. Pick the weight you can control today and add to it from there.

Forcing out the ugly reps. It feels like the last rep is the one that counts. But a rep that looks different from the first is a different movement. End the set there and log what you did.

Stacking five cues at once. Videos and coaches give plenty of them, and it's tempting to use them all. Five things is too many to think about during one rep. Pick one cue per set, and change it only if it isn't helping.

Never checking. You can't see your own knees in the middle of a squat, and a mirror shows you only one angle. Film a set every session or two while you learn a new lift. Coaching practice.

Getting careless between sets. Most attention goes to the lift itself. Yet in the US emergency-department data, the most common cause of injury was a weight dropping on the person.^[6] Rack the bar with both hands, and lower dumbbells to the floor instead of letting go.

How Gymbers fits

Learning a lift is a loop: watch it, try it light, check it, repeat. Exercises in the Gymbers app come with a video, and you can save the videos to your phone. The app's log records the weight and reps for every set, and last session's numbers sit beside each set. Programs in the app set a target RIR for each working set, so you know how many reps to leave in the tank.