Short answer: both, and the interesting question is the ratio. The strength-versus-cardio framing is a false choice — the two produce different adaptations that barely overlap, and public health guidance has recommended both for years. What is genuinely personal is how you split limited time between them, and that depends on your goal, your age, your training history, and what you're currently worst at.
Here is how to decide, and where personalised analysis actually adds something beyond a generic template.
What each one actually does
These are not two routes to the same destination. They train different systems.
Cardiovascular training
- Increases stroke volume and cardiac efficiency
- Improves insulin sensitivity and glucose handling
- Raises VO₂ max — one of the strongest single predictors of all-cause mortality yet measured
- Supports mood and sleep quality, with reasonably robust evidence
- Contributes to energy expenditure for weight management
Resistance training
- Builds and preserves muscle mass, which declines steadily from roughly the fourth decade onward
- Increases bone mineral density — the main non-pharmacological defence against osteoporosis
- Raises resting metabolic rate modestly, by adding metabolically active tissue
- Improves joint stability, balance and fall resistance, which becomes the dominant benefit later in life
- Preserves lean mass during weight loss, which is what separates losing fat from losing weight
Notice how little the two lists overlap. Cardio will not meaningfully protect your bone density. Lifting will not meaningfully raise your VO₂ max. Choosing one means accepting the absence of what the other provides.
The baseline everyone should start from
WHO physical activity guidance for adults is 150–300 minutes of moderate-intensity aerobic activity per week (or 75–150 minutes vigorous), plus muscle-strengthening activity on two or more days per week working all major muscle groups.
Most people fall short on the second part. Aerobic activity happens somewhat incidentally — walking, commuting, sport. Resistance training almost never does. If you're currently doing neither, the strength side is usually where the larger unclaimed benefit sits.
How to split your time by goal
Fat loss
Diet drives the deficit; training determines what you lose. Losing weight without resistance training means losing a meaningful fraction as muscle, which lowers your metabolic rate and tends to produce the "skinny but soft" outcome people are unhappy with. Suggested split: 2–3 strength sessions, 2 cardio sessions, with adequate protein — this is the single highest-leverage combination for body composition, and it depends heavily on getting protein intake right.
Building muscle
3–4 strength sessions, 1–2 easy cardio sessions. Keep cardio low-intensity and separated from lifting where possible. High-volume endurance work performed close to strength sessions can blunt strength adaptations — the "interference effect" — though this is often overstated by people using it as an excuse to skip cardio entirely. At recreational volumes it is a minor consideration.
Cardiovascular health and longevity
3–4 cardio sessions, 2 strength sessions. Include some higher-intensity work; VO₂ max responds to intensity more than duration. But keep the two strength sessions — muscle mass is independently associated with better outcomes, and the last decades of life are shaped more by whether you can stand up unassisted than by your treadmill time.
Over 50
Strength becomes the priority, not the supplement. Sarcopenia and bone loss accelerate with age, and both are substantially modifiable. Two to three resistance sessions weekly is a floor rather than a ceiling here, with cardio built around walking and low-impact work.
Managing blood glucose
Both help through different mechanisms — cardio improves insulin sensitivity acutely, while added muscle increases the body's capacity to store glucose. Roughly even split, with a preference for shorter, more frequent sessions. Anyone with diagnosed diabetes should set this up with their doctor rather than from an article.
Where AI-driven analysis genuinely helps
Most of what gets marketed as AI personalisation is a decision tree with better graphics. Three places where it does earn its keep:
Detecting what you actually do, not what you plan
Almost everyone believes their training is more balanced than it is. Logged data reliably shows the opposite — a "balanced" programme is usually four cardio sessions and one half-hearted lifting session. Being shown your real 90-day ratio is often the most useful output of any fitness app.
Adjusting for recovery, not just the calendar
A fixed weekly template ignores whether you slept five hours, are travelling, or are three weeks into a stressful period. Systems that read sleep and resting heart rate can shift a heavy session to an easy one before you dig a hole. This is where the cross-domain view stops being a slogan and starts being useful.
Progressive overload without arithmetic
Strength progress requires gradually increasing demand. Doing this well by hand means tracking sets, reps, loads and perceived effort across dozens of exercises. Doing it badly — same weights for months — is the most common reason people plateau and quit. Automating the increment is unglamorous and genuinely valuable.
Practical scheduling
- Separate hard cardio and hard lifting by several hours where you can. If you must combine, do the one that matters most for your goal first.
- Don't schedule heavy leg work the day before a long run, or vice versa. Most "overtraining" at recreational level is just bad ordering.
- Two full-body strength sessions beat a five-day body-part split for anyone training under four hours a week — more frequent stimulus per muscle, fewer sessions to miss.
- Walking is not your cardio session if you're training for cardiovascular fitness — though it remains valuable for overall activity and metabolic health. Different jobs.
- Take a genuinely easy week every 6–8 weeks. Adaptation happens during recovery, and most people under-recover rather than under-train.
The bottom line
Nobody who has looked at the evidence recommends choosing. The split that works is the one that covers your weakest domain, fits the time you actually have, and that you'll still be doing in six months. For most people reading this, that means adding resistance training — because that's the half that almost never happens by accident.
BodyFastLane's AI coach builds adaptive plans across both modalities, adjusts for sleep and recovery data, and handles progressive overload automatically — free on iOS. Check with a doctor before starting a new training programme if you have any cardiovascular or musculoskeletal condition.