Training science, no guesswork

Articles on readiness, neural demand and programming — written by a sports scientist, with the data behind every claim.

Programming

A deload is a budget, not a percentage

A typical heavy session costs 43 units of neural cost. Halving the sets brings it to 21.5. Training the same muscles on cheaper movements brings it to 17 — and keeps more stimulus than the heavy day had.

17 October 2026 · 6 min read
Training science

The leg curl paradox

Both leg curls return 4.50 per unit of neural cost — nothing in 272 movements returns more. The Romanian deadlift returns 0.94 and the barbell deadlift 0.45. Why the least fashionable machine in the gym scores highest.

13 October 2026 · 5 min read
Data analysis

Every expensive movement in the library returns less than it costs

59 of 272 movements give back less stimulus than they charge in neural cost — and 58 of those are simply every movement priced at 8 or above. The full list, and the four that still belong in most programmes.

11 October 2026 · 6 min read
Training science

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

Handstand Walk returns 0.44 per unit of neural cost. Ring Muscle-Up returns 0.44. They are the only two of 272 movements below 0.45 — and neither number is a programming error.

3 October 2026 · 5 min read
Training science

The biggest tier in the library is also the one nobody programmes

Sort all 272 movements by stimulus-to-fatigue ratio and the median is 1.38. The band around it holds 81 movements — more than any other tier in the library, and the one least written about.

29 September 2026 · 5 min read
Training science

The real cost of training on less than eight hours of sleep

Adolescent athletes who averaged under eight hours of sleep a night carried 1.7 times the injury risk across a season. Why that number belongs next to the neural-cost ledger, not instead of it.

25 September 2026 · 5 min read
Training science

The exercises that cost you least aren't there to build muscle

77 of 272 movements sit at the lowest neural cost there is. Sixteen build muscle efficiently. The other sixty-one — mobility, prehab, activation, breathing — are insurance for the expensive work.

22 September 2026 · 5 min read
Training science

What readiness actually means, and why it isn't HRV

A systematic review set subjective wellbeing checks against HRV, resting heart rate and biochemical markers to see which one actually tracked training response. The chest strap lost — consistently.

19 September 2026 · 5 min read
Training science

The 18 exercises that build muscle almost for free

Out of 272 movements, only eighteen reach 9.0 stimulus per unit of fatigue — and sixteen of them have a neural demand of 2. The complete list, grouped by muscle, for the days your body isn't there.

17 August 2026 · 5 min read
Data analysis

The exercises that cost you most are the ones that give you least

272 movements analysed. The deadlift returns 0.45 for every unit of fatigue it charges. A leg curl returns 4.50. That is a ten-fold difference — and almost nobody builds a week around it.

14 August 2026 · 6 min read
Next up: Cable Lateral Raise at 4.50 against Barbell Deadlift at 0.45, and the programming mistake a ten-fold gap exposes · why 34 movements share the maximum cost of 10, and what that ceiling means · training age against neural cost, and why the same session costs a 19-year-old and a 34-year-old different amounts.