Starting Strength Review: Does Novice Linear Progression Still Hold Up?

TL;DR: Starting Strength remains the best on-ramp in lifting. Three sets of five on a handful of barbell lifts, more weight every session, no decisions to make. The program's flaws are real but narrow: not enough volume to be a good muscle-building plan, and a glaring lack of horizontal pulling. The bigger failure is user error. Almost nobody exits on time, and the months spent grinding a dead linear progression are the most wasted training months in the sport.

---

Mark Rippetoe published Starting Strength in 2005, and two decades later it is still the first program most new lifters are handed. That's unusual. Training fashions turn over every few years, and the fact that a book older than the smartphone is still the default answer says something about how well it solved its particular problem.

The short verdict: for an untrained adult who wants to get strong, it's still close to optimal. For anyone else, including the version of you that exists four months from now, it's the wrong tool.

What the program actually is

Two alternating full-body workouts, run three days a week on non-consecutive days.

| | Workout A | Workout B | |---|---|---| | 1 | Squat 3x5 | Squat 3x5 | | 2 | Bench press 3x5 | Overhead press 3x5 | | 3 | Deadlift 1x5 | Power clean 5x3 |

That's the whole thing. You start light enough that the first few weeks feel almost pointless, and you add weight every single session: roughly 5 pounds on the upper-body lifts, 10 on the squat, 10 to 15 on the deadlift early on. When you miss reps, you repeat the weight. When you miss three times, you reset to 90 percent and climb back up.

No percentages. No RPE. No weekly undulation. The genius of it is that a nervous 22-year-old in a commercial gym has exactly one decision to make each session, and that decision is "put more weight on the bar."

Why it works so well at the start

Untrained lifters adapt to almost anything, and they adapt fast. Neural efficiency improves session to session in the first few months, which means strength can climb faster than muscle can actually be built. A linear progression harvests that window better than any complicated template can, because it pushes load up at exactly the rate a novice can absorb.

There's also a dose argument. Beginners need shockingly little volume to add strength. A systematic review of the minimum effective dose for 1RM strength in trained men found meaningful strength gains from as little as a single working set per lift per session (Androulakis-Korakakis et al., 2020), and novices sit even further down that curve. Three sets of five, three times a week, is comfortably enough stimulus while you're new.

And the frequency is right. Squatting three times a week means roughly 150 practice sessions a year on the movement pattern, which matters more than anything else when technique is the actual limiting factor. If you want the mechanics behind why adding load beats adding effort at this stage, we broke that down in how to actually apply progressive overload.

Where it falls short

The hypertrophy volume is thin. Three sets of five on the squat is three hard sets for your quads in a session, nine per week. That's fine for strength and mediocre for size. Muscle growth scales with weekly hard sets well into the double digits per muscle group, with roughly 10 or more sets per week producing clearly more hypertrophy than fewer (Schoenfeld et al., 2017). Starting Strength gets you strong on a squat while leaving a lot of quad on the table, and it does essentially nothing for arms, side delts, or hamstrings in isolation. If your actual goal is a physique rather than a total, know that going in.

There is almost no horizontal pulling. Power cleans are a pull, but they're explosive, low-rep, and technically demanding enough that most self-coached lifters butcher them or skip them entirely. What's missing is rows and chin-ups, which means a beginner can spend six months benching three times a week with nothing balancing it. Adding two or three sets of rows and chin-ups per week fixes this and costs you nothing. Do it from day one.

The nutrition advice aged badly. The book's mass-gain guidance, GOMAD included, was written for skinny teenage boys and has been applied for twenty years to 38-year-old office workers who did not need a gallon of whole milk a day. Eat in a modest surplus if you're underweight and roughly at maintenance if you're not. Our lean bulk calculator will give you a saner number in thirty seconds.

The part almost everyone gets wrong

Linear progression has an expiry date, and it is earlier than you want it to be.

A realistic run is three to six months. Some people get eight. Nobody gets two years. The failure mode is not that the program stops working, it's that lifters refuse to accept that it stopped: they reset the squat, climb back to the same weight, stall again, reset again, and burn four months oscillating around a number they already hit in month three.

Two resets on the same lift with no new ground gained is your signal. That's not a technique problem or a discipline problem, it's the program telling you the novice window has closed and you now need something with more volume, more variation, and planned recovery. Where you go next depends on where you actually sit, and time in the gym is a poor proxy for that. Our breakdown of beginner, intermediate, and advanced lifters is a better yardstick. StrongLifts 5x5 is the sideways move most people make, and it buys a bit more volume but the same expiry problem eventually.

One practical note on the exit. Most intermediate programs want a one-rep max to build percentages from, and you should not test a true single the week your linear progression dies. Estimate it from your best set of five instead, and understand the error bars on that estimate before you build a program on it: we looked at how accurate those formulas really are in how accurate are 1RM calculators.

Who should run it

Good fit: genuine beginners with barbell access, anyone returning after years away, and lifters whose main problem is that they can't stick to anything because the program has too many moving parts.

Poor fit: anyone already past a 1.5x bodyweight squat, people training primarily for physique, and anyone without a rack and a coach or a camera. Squatting to failure alone in a garage with no safeties is how linear progressions end in an emergency room rather than a plateau.

How to run it well

1. Start lighter than your ego wants. The empty bar for two weeks costs you nothing and buys you a longer runway. Starting 40 pounds too heavy costs you a month.
2. Add rows and chin-ups. Two to three sets of each, twice a week. This is the single highest-value modification to the program.
3. Film your squat once a week. Not every set, just one. Technique breakdown is what usually ends a linear progression, and you can't feel it happening.
4. Log every session and watch the trend, not the day. Two failed resets on the same lift means you're done. Move on that week, not two months later.
5. Eat like an adult. Modest surplus if you're underweight, roughly maintenance if you're not. Calculate a real number instead of guessing.

The verdict

Starting Strength is a very good answer to a very specific question: how does an untrained person get strong as fast as possible with minimal complexity? Twenty years of newer programs haven't improved much on it for that job. It just isn't a general-purpose training program, and it was never sold as one.

The hard part was never the first three months. It's knowing when the window has closed and what to run next, which is exactly the judgment call a beginner is least equipped to make. Protokl handles that transition: it tracks the weekly volume each muscle group is getting, watches your logged performance for the stall pattern instead of waiting for you to notice it, and moves you onto a program with real volume when linear progression is genuinely done. If you're running a linear progression right now, keep running it. Just don't be the person still resetting the same squat in February.