The goal was a program Taylor — returning to training after time off — could trust claim-by-claim: drop body fat without losing muscle, rebuild an aerobic base with better running form, keep skill work (handstands, splits) alive, and protect an achy hip and a recently-healed wrist along the way. The findings below are the ones a captured source actually supported. Where the evidence is thin, small-sample, or drawn from an analogous group, it says so.
Running: form, return, and the dog
The headline question was a personal one: Taylor liked a cue — lean forward and push the ground out behind — that turned out not to be supported, and that likely loaded an already-taxed ankle. The evidence pointed somewhere else: build form around a quicker cadence and a whole-body lean from the ankle. A systematic review and meta-analysis of changing running step rate put it directly:
“Increasing running step rate will broadly reduce kinematic and kinetic variables at the ankle, knee and hip.” “There is insufficient evidence to determine the effects of changing running step rate on injury or performance.” — Step-rate systematic review & meta-analysis [1]
That second sentence is the honest part: a quicker cadence reliably redistributes joint load, but the direct injury-and-performance payoff isn't proven — so the program treats cadence as a load-management tool, not a magic fix. The cues to drop had clearer evidence against them: a four-week trial switching runners to a forefoot/“lean and pull” style produced no significant change in impact loading,[2] forefoot striking in minimal shoes raised Achilles loading,[3] and a hard forward trunk lean traded knee load for hip-extension load rather than removing it.[4]
Returning without re-injuring
For a returner, the protective rule is to cap how fast volume jumps: injury rates rose when a single run exceeded roughly 10% of the longest run in the prior 30 days.[5] And for someone rebuilding, fancy intensity distribution isn't the lever — an eight-week trial found polarized 80/20 training no better than focused/threshold work for recreational runners;[6] consistent easy-dominant volume is what moves the needle.
The dog runs
One of the easy-run partners is a dog, which makes heat a real safety variable, not a footnote: a large veterinary-records study found canine heat-related illness can be fatal, with worse odds for heavier, older, and flat-faced dogs.[7] The dog's runs are kept to the easy sessions, built up gradually, and gated on weather.
Nutrition: a muscle-sparing cut, not a crash
Taylor's instinct was a flat low calorie number. The evidence pointed to anchoring on a rate instead. For resistance-trained people, the muscle-sparing zone is a loss of about 0.5–1.0% of body weight per week with protein around 2.2–3.0 g/kg/day, with creatine and caffeine the supplements that actually carry evidence.[8] Resistance training itself does most of the muscle-protecting work in a deficit,[9] and under-eating too hard carries its own cost — low energy availability harms health and performance in male athletes too, not only female athletes.[11]
Two levers people skip turned out to matter. Sleep is one of them: at equal calories, cutting sleep from 8.5 to 5.5 hours shifted the body toward losing muscle instead of fat.[10] And the supplement list is short and honest — an expert consensus puts good evidence behind only a handful (caffeine, creatine, and a couple of others),[12] with creatine well-supported for safety and effect in healthy people.[13]
Lifting: machines, preservation, and not chasing failure
The plan leans on machines, and that's well-supported: across thirteen studies, machines built muscle as well as free weights.[14] Training to failure isn't required either — leaving reps in reserve produced no significant hypertrophy disadvantage,[15] which spares recovery for the skill and cardio work. And strength training earns its place twice over: it's the best-evidenced injury-prevention tool there is.
A meta-analysis of 25 trials found that strength training reduced sports injuries to less than a third, and overuse injuries by almost a half — while stretching showed no protective effect. — Strength-training injury-prevention meta-analysis [16]
Swimming: offload the joint, gate the breaststroke
Water genuinely offloads the hip and knee — in-vivo measurements showed aquatic exercise cut joint force substantially, though faster, more forceful movement loaded it back up.[17] So easy swimming is restorative; hard kicking is not automatically safe. That matters most for the breaststroke, which Taylor wanted to build. The classic study of breaststroker's knee is specific about the risk and the lever:
“An extremely high incidence of knee pain was documented both among breaststroke specialists (73%) and nonbreaststrokers (48%)… dramatic differences in the injury rate were noted when hip abduction angles at kick initiation were less than 37 degrees or greater than 42 degrees.” — Breaststroker's knee: epidemiological & biomechanical analysis [18]
The injury rises at the extremes of the hip-abduction window, which makes hip-led kick mechanics the protective lever — supporting Taylor's own hunch that adductor and glute strength were the thing to build first.
Mobility & skill: the hip and the wrist
The deep hip ache fits an end-range-loading pattern: studies of dancers and gymnasts with extreme hip range show that the range itself can produce impingement and instability, even with normal bony anatomy.[19] That points toward building range with loaded, active end-range control rather than forcing passive stretches — and backing off deep-compression positions during a flare. The recently-healed wrist gets the same careful treatment: handstand wrist pain is common and tracks with intensity and biomechanics, not just hours,[20] so the on-ramp is graded and technique-focused rather than max volume from day one.
What this writeup does not claim
- It does not offer medical or dietary advice. The hip and wrist are program-around-and-monitor items, not diagnoses; the aggressive cut is framework-level and an RD is a reasonable second opinion.
- It does not claim cadence prevents injury. The evidence is that it redistributes joint load; the direct injury-and-performance benefit is explicitly unproven in the source.
- It does not treat every finding as equally strong. Several rest on small or analogous-population studies (the sleep and aquatic-load samples are small; the hip-impingement work is on dancers and gymnasts) — treat those magnitudes as directional.
- It does not claim full-text depth on every source. Some claims were grounded against published abstracts where the full text was access-blocked; the findings are present, the method detail sometimes isn't.
- It does not claim completeness. Nineteen load-bearing claims were verified verbatim; the program also draws on practitioner doctrine (warm-up drills, end-range mobility methods) that has thin formal evidence, and that is flagged rather than dressed up.
How this was verified
Candidate sources were gathered across all six domains, retrieved, and captured as verified content. A finding entered the program only when a captured passage actually supported it — bound to that passage by a content fingerprint, not by a model's recollection. Nineteen load-bearing findings cleared that bar, backed by twenty cited sources (the cadence recommendation draws on its own meta-analysis). Every source is listed on the citations page with its link, its evidence tier, and the date it was retrieved, so you can open any one and check it yourself.
See how it was built
This research is the backbone of a 10-week program delivered through a logging app. Read the case study for the end-to-end build, or open the live demo to try the app yourself.
Read the Case Study Try the Demo