Training science

The two worst-value movements in the library are both skills, not lifts

Sort all 272 movements by what they return per unit of neural cost and the bottom two are Handstand Walk at 0.44 and Ring Muscle-Up at 0.44. Neither is a programming error. Both are the clearest case in the library for when a bad ratio is the right answer.

Alfredo Silva · Sports Scientist MSc Physical Education · MSc High Performance Training 2 of 272 movements

Every movement in the Forta library carries two numbers: neural cost (1–10, what it charges your nervous system) and stimulus (1–10, what it gives back). Divide the second by the first and you get the return on the transaction.

The best number in the library is 4.50 — the leg curl, the cable fly, the back extension. The median is 1.38. And right at the floor, below everything else, sit exactly two movements.

Handstand Walk is the only movement in the whole library scoring 3.5 on stimulus — the lowest value any of the 272 returns. Ring Muscle-Up is the only one at 4.0. Below 0.45 there is nothing else.

Compare that to the top of the ledger and the gap is roughly ten-fold: a lying leg curl returns 4.50 for every unit of fatigue it charges, a handstand walk returns 0.44. Ten leg curls' worth of return for one handstand walk's worth of cost.

What the bottom thirty are made of

Thirty movements in the library score under 5.0 on stimulus, and all thirty of them cost 8 or more. There is nothing in the library that is both cheap and low-return — if a movement gives little back, it is because it is expensive, not because it is weak.

Twenty-seven of those thirty are barbell or plyometric: snatch, clean, jerk, thruster, depth jump, broad jump. That list has already had its article. The remaining three are the ones above — and they are a different animal entirely.

A deadlift has a poor ratio because it is enormously expensive. A handstand walk has a poor ratio because it was never trying to produce stimulus in the first place.

The ratio prices one outcome, and only one

This is the limit of the number, stated plainly. Stimulus in this model means stimulus to adapt — hypertrophy, strength, the tissue-level return. It does not price coordination. It does not price the motor pattern you are trying to acquire. A handstand walk charges 8 for balance, shoulder stability upside down, and a pattern your nervous system has to build from nothing. The muscular return on that is genuinely small. The number is correct.

Impellizzeri, Marcora and Coutts made the general version of this argument in 2019: a training-load measure is only meaningful in relation to the specific outcome it is meant to explain, and the work an athlete does is not the same thing as the response it produces. One quantity cannot stand in for every goal you might have. A stimulus-to-cost ratio is a hypertrophy-and-strength instrument. Hand it a skill and it will tell you, accurately, that the skill is a bad way to grow muscle.

Which it is. That was never the reason to do it.

The ladder, and where it falls off

Here is what makes the handstand case readable. The library holds four points on the same progression, and the cost climbs while the stimulus collapses only at the last step.

The wall walk returns 1.38 — exactly the library median. The handstand hold and the handstand push-up both return 1.17, which is unremarkable but honest work. Then the moment the shoulders have to produce force and move the body across the floor, cost goes to 8 and the muscular return drops to the lowest figure in the library.

The rings tell the same story more sharply. A Bar Muscle-Up costs 6 and returns 7.0 on stimulus — a ratio of 1.17. Move the same movement onto rings and the cost goes to 9 while the stimulus falls to 4.0. The bar version returns about two and a half times what the ring version does, for a muscular outcome that is nearly identical.

So when is a bad ratio fine

When the thing you want back is not stimulus. There are three cases, and they are narrow.

The skill is the goal. If you compete in a sport that contains ring muscle-ups, you have to do ring muscle-ups. No amount of cable work substitutes for the pattern. The 0.44 is not a warning, it is a price tag.

The skill is early. Pattern acquisition happens fastest when you are fresh and the nervous system is clean. That is an argument for putting the expensive skill at the front of a good day, not for skipping it.

The volume is small. A handstand walk is practised in metres, not sets to failure. Ten minutes of skill work at cost 8 is not the same weekly bill as five sets of deadlifts at cost 10, and treating them as equal is the mistake the ratio invites if you read it carelessly.

What is not fine is reaching for a 0.44 movement on a day when the account is empty. A bad check-in — poor sleep, residual soreness, low mental energy — is the worst possible day to practise a skill that demands precision under load. You will pay the 8, you will collect the 3.5, and you will groove a pattern badly while you do it.