Calendar tap any day
UAlift
LRlong run
Qrun
Srest
deload & test
This week tap a day
When the week falls apart
1 · Sat long runreschedule
Move to Sunday, drop Upper A. Never lose it.
2 · Wed temporeschedule
Run it that evening — run first, then a short lift.
3 · Mon quality runmove
Push to Tuesday AM. Don't double it with the lower lift.
4 · Mon Lower (light)skip
The flexible one — Thursday's heavy squat still holds most of the leg volume.
5 · Upper A / Bskip
Zero cost to the race.
6 · Thu Lower (heavy)protect
If it must go, drop to squat only, 2 sets.
7 · Easy runsskip
That's what they're for.
Doubling rules
Monday's run and lift are both on the lighter side of the week (the AM run is hard, but the PM lift is the flexible posterior-chain session) — if you must double, that's the day to do it. Thursday has no run at all, which is exactly why the heavy squat lives there now: nothing to double against.
The long arc
Aug — Dec 13
Race build
Running leads, lifting protects it. Legs stay heavy and low-rep. The goal is arriving healthy, not peaking.
Dec 14 — 27
Recovery
No hard running for two weeks. Lifting returns first, running last. Resist signing up for anything.
Jan — Mar
Strength block
Running drops to three easy plus one quality. Leg volume climbs into hypertrophy ranges, quads come off the cap, and the hip gets resolved rather than managed around.
Apr onward
Next goal
Spring race off a real base, or keep building. December is a checkpoint on a much longer curve.
Week by week whole plan at a glance
Miles are derived from the selected athlete's easy pace — the plan is prescribed in minutes on purpose, so the same 120-minute long run becomes more miles as you get fitter and the load self-corrects. D marks a deload, T a taper week, and · a test. The current week is highlighted.
Big Apple Half — 13 December, 8:00 AM, Central Park.
The course 2.5 loops · Harlem ×2 · Cat ×3
Schematic, not surveyed. Hill order and count come from the official course description and race reports; exact mile placement isn't published.
StartWest Drive · 96th St
Shape2.5 loops via 72nd transverse
Finish102nd St transverse
Harlem Hill~100 ft · ×2
Cat Hill~50 ft · ×3
Total gain~470–590 ft
The trap
Cat Hill comes a third time after both Harlem Hills are behind you — a 50-foot climb somewhere past mile 11 on empty legs. Race reports repeatedly describe running out of gas on the flats, not the climbs. Running the first Harlem Hill slower than feels right is the single decision that protects your last 5K.
Hill strategy
Start → Harlem #1≤ 180 bpm
Deliberately easy. It should feel too slow.
Harlem Hill ×2to 188
Effort constant, pace drops 30–60 s/mi. Let it.
Descentsdrops on its own
Relax, don't brake, don't attack.
Cat Hill ×1, ×2183–188
Steady, no surging.
Flats after Harlem #2185–188
This is where the race is won or lost.
Cat Hill ×3188+
Whatever's left.
Your paces set from the last 2-mile trial
This setting drives every pace shown on the daily cards. Riegel's power law from 2 miles to 13.1 is a long extrapolation — treat the projection as a range of several minutes, not a number. Week 12's test switches to a 10k for exactly this reason.
Progression rule
Paces change at a test, never mid-block. Hitting every rep means you earned faster paces at the next test — not that today's tempo should be quicker. If you can't hold pace on the final rep, the session ends there. Don't set a goal time until week 12.
Blocks and taper
| Block | Wks | Mon AM | Wed AM | Long |
|---|---|---|---|---|
| Base | 1-4 | Easy + strides | Easy | 35→41 |
| Strength | 5-8 | Easy + strides | Hills | 45→53 |
| Speed | 9-12 | 5×3 min | Tempo 2×10 | 58→69 |
| Race-specific | 13-15 | 6×800m | 3×10 at HMP | 75→90 |
| Taper | 16-17 | 4×3 min | 2×10 at HMP | 45→32 |
| Race week | 18 | 3×3 min | Shakeout | 13.1 mi |
Taper, built to the evidence
Weeks 16-17 cut total running volume 43% then 56% from the week-15 peak — inside the 41-60% band that meta-analyses identify as optimal. Every run gets shorter, but you keep the same number of runs and keep doing half-marathon-pace work, because holding frequency and intensity constant is precisely what makes a taper work.
Testing weeks 4 · 8 · 12 · 16
Weeks 4 & 82-mile trial
Week 1210k trial — sets your goal
Week 165 mi at goal pace
WhenMonday AM, deload week
Effort~95% — not a race
Abort ifany sharp or unusual pain
Race morning
Arrival6:00 AM · start 8:00
Wake~4:45, eat by 5:15
Rehearsalwks 16-17 long runs on Sunday
Fuelling practicefrom week 9, every long run
Nothing newshoes, food, kit — all tested
Cold start ~35-45°F at the gun, 2 hours outside before it
Typical Dec 13 8AM tempupper 30s–low 40s°F
Time outside pre-race~2 hrs (6:00 arrival → 8:00 start)
Corral waitthrowaway layers — sweats, old gloves, a trash-bag poncho
Warm-uplonger than usual — cold muscle takes more strides to feel normal
First milestay conservative — cold air makes easy pace feel harder than it is
Hands / extremitiesthrowaway gloves; they're the first thing you'll regret skipping
NYC Central Park December mornings average a 42-49°F daily high (NWS climate normals); 8 AM is well before the daily peak, so upper-30s to low-40s°F is the realistic range, with real year-to-year variance — check the actual forecast in race week rather than planning off an average. Layer to shed, not to arrive comfortable: you should feel slightly cold standing in the corral if you're dressed right for mile 3.
Worth repeating once
Peak weeks put roughly 26 miles a week through legs that are also climbing 480 feet of hills and lifting four times a week. Strength training is the single best-evidenced way to stay injury-free while running — that's the whole reason the lifting stays in through a race build instead of being cut for "recovery."
The part that outlasts any race.
Hybrid principle
Upper body carries the volume; legs carry the load. Running doesn't interfere with upper-body hypertrophy, so that's where set volume lives. Legs go heavy and low-rep instead — loads at or above 80% of 1RM are what improve running economy, while high-rep leg work stacks fatigue onto five days of running without the same return.
Sun — Upper A push emphasis · 45 min
Incline DB press4×8-10 ▶
Chest-supported row4×8-10 ▶
Standing overhead press3×6-8 ▶
Dip or machine press3×10 ▶
Lateral raise3×15 ▶
Triceps pushdown3×12 ▶
Mon — Lower posterior chain · 40 min
The lighter leg day — glute and hamstring focus, and the flexible one to drop in a tight week. Seated calf trains the soleus directly; the standing/heavy variant on Thursday already covers the gastrocnemius, so between the two sessions both calf heads get dedicated work.
Wed — Upper B pull emphasis · 45 min
Weighted pull-up / pulldown4×6-8 ▶
Flat DB press4×8-10 ▶
Cable row4×10 ▶
Face pull3×15 ▶
Barbell or DB curl3×10 ▶
Hammer curl2×12 ▶
Thu — Lower heavy squat · 45 min
Before this sessionFive minutes of calf and hamstring prep, every time. No 1RM has ever been tested here — RPE 7-8 means the last rep of each heavy set should feel like you had 2-3 more in you, never a grind. The only day with no run before it, and now the only day with a run scheduled after it (Wednesday's tempo) more than 4 days out.
Weekly sets per muscle
Back12 sets
Optimal band. Rows twice, vertical pull once.
Chest11 sets
Optimal band. Incline, flat, and a dip or machine press.
Glutes13 sets
Highest in the plan — squat, RDL, thrust, split squat. Glute and hip strength is one of the best-evidenced ways to run injury-free (Lauersen 2014, 2018).
Shoulders9 sets
Overhead press, laterals, face pulls.
Quads9 sets, heavy
Fewer sets than upper, but at higher intensity. Running supplies the volume; the gym supplies the load.
Hamstrings6 sets
RDL (3) plus leg curl (3). At the low end of the landmark band — running supplies real hamstring load too, but this is the one muscle group worth an extra set if you have room.
Arms11 sets
Direct work plus what the compounds supply.
Core4 sets
Loaded carry (anti-lateral-flexion) plus Pallof press (anti-rotation) — the two patterns nothing else in the plan trains directly. Light, on purpose; running and the compounds already load the trunk isometrically.
Progression
Upper — hit top of rangenext dumbbell up, rebuild reps
Upper effort1-2 reps in reserve
Lower — squat3×5 @ RPE 7-8 + 1×8 back-off
Squat effortRPE 7-8, never grind
Squat progresses+load next week once all 3×5 hit RPE ≤8
Add setsupper only, 1/muscle/month
Deload weeks (4, 8, 12)lift sets cut ~35%
Taper (16, 17)lift sets cut 25-40%
Leg volume risesJanuary, not now
A meta-regression across 55 hypertrophy studies found growth improves as sets approach failure but the curve flattens past about 2 reps in reserve — and proximity to failure did not help strength at all. Grinding out the last rep costs recovery you need for running and buys almost nothing.
Every rule in this plan, and what it actually rests on.
How to read the grades
A — multiple meta-analyses or large trials agree. B — direction is clear, magnitude or population uncertain. C — contested, inferred, or the evidence cuts against what this plan does. Where a C sits under something you're doing anyway, the reason is stated.
BWhy the Base block still looks this easy when your legs remember running 1,000 miles last year
Strength and endurance detrain and retrain on genuinely different clocks, and both cut against starting hard right now.
On the lifting side, "muscle memory" is real and mechanistic, not just a phrase: myonuclei and satellite cells added during earlier training appear to persist through a layoff, giving retraining a head start over a true first exposure. The clearest controlled data on this comes from older adults (12 weeks training, 12 weeks detraining, 12 weeks retraining) rather than a population that matches you exactly, so treat the mechanism as well-supported and the exact timeline as a reasonable guess, not a number to plan around.
On the running side, the effect is closer to the opposite. A 24-month study in competitive runners found injury risk rose specifically when the acute:chronic workload ratio increased quickly — i.e. when training resumed faster than the body's recent average could absorb. Tendon and bone don't get a "memory" discount the way muscle does. Coming back from on-and-off running with a nonexistent current base is exactly the scenario that statistic describes, regardless of last year's mileage or how strong the lifts still feel.
Practical result: the Base block's easy volume isn't paced for a beginner and it isn't paced for someone at full fitness — it's paced to let running catch up to what the muscles already remember, not the other way around. If the lifts start feeling too easy before week 4, that is expected and not a sign the Base block needs to change.
On the lifting side, "muscle memory" is real and mechanistic, not just a phrase: myonuclei and satellite cells added during earlier training appear to persist through a layoff, giving retraining a head start over a true first exposure. The clearest controlled data on this comes from older adults (12 weeks training, 12 weeks detraining, 12 weeks retraining) rather than a population that matches you exactly, so treat the mechanism as well-supported and the exact timeline as a reasonable guess, not a number to plan around.
On the running side, the effect is closer to the opposite. A 24-month study in competitive runners found injury risk rose specifically when the acute:chronic workload ratio increased quickly — i.e. when training resumed faster than the body's recent average could absorb. Tendon and bone don't get a "memory" discount the way muscle does. Coming back from on-and-off running with a nonexistent current base is exactly the scenario that statistic describes, regardless of last year's mileage or how strong the lifts still feel.
Practical result: the Base block's easy volume isn't paced for a beginner and it isn't paced for someone at full fitness — it's paced to let running catch up to what the muscles already remember, not the other way around. If the lifts start feeling too easy before week 4, that is expected and not a sign the Base block needs to change.
Bloom et al., Exp Gerontol 2020;133:110860 (muscle memory mechanism; population: older men) · Dijkhuis et al., Int J Sports Med 2020 (n=23 competitive runners, 24 months)
Hybrid structure final audit — the newest part of the plan
AUpper body carrying the volume is correct — interference is body-part specific
This was the central bet of the hybrid restructure and it holds. Wilson's meta-analysis found the interference effect to be body-part specific: upper-body strength was not affected by concurrent endurance training. Sabag's 2018 meta-analysis on concurrent strength and HIIT reported the same pattern — lower-body strength attenuated versus strength training alone, upper-body strength unaffected. Endurance volume was a moderating variable: longer and more frequent endurance training correlated with greater reduction in lower-body strength and power.
So loading chest, back, shoulders and arms to 11-12 sets a week while holding legs at 9 heavy sets isn't a compromise — it's the structure the evidence points at.
So loading chest, back, shoulders and arms to 11-12 sets a week while holding legs at 9 heavy sets isn't a compromise — it's the structure the evidence points at.
Wilson et al., J Strength Cond Res 2012 · Sabag et al., 2018
ADJUSTEDThe squat had one more set than the evidence calls for
The ACSM's 2026 Position Stand — synthesising 137 systematic reviews and over 30,000 participants — found voluntary strength enhanced by lifting ≥80% of 1RM, through a complete range of motion, for 2-3 sets, at the beginning of the session, ≥2 sessions per week. Runner-specific reviews converge on the same place: 2-3 sets of 4-5 reps at about 85% 1RM.
My prescription was 4×5. Changed to 3×5 plus one back-off set of 8 — three heavy sets for the strength adaptation, one lighter set for the hypertrophy stimulus, and less leg fatigue heading into a race build. Load is prescribed by RPE (7-8 on the heavy sets) rather than a %1RM, since no 1RM has actually been tested — RPE is what the weekly log can act on. The same Position Stand also found that training to momentary muscle fatigue did not consistently improve outcomes, which is why nothing in this plan asks you to grind.
My prescription was 4×5. Changed to 3×5 plus one back-off set of 8 — three heavy sets for the strength adaptation, one lighter set for the hypertrophy stimulus, and less leg fatigue heading into a race build. Load is prescribed by RPE (7-8 on the heavy sets) rather than a %1RM, since no 1RM has actually been tested — RPE is what the weekly log can act on. The same Position Stand also found that training to momentary muscle fatigue did not consistently improve outcomes, which is why nothing in this plan asks you to grind.
ACSM Position Stand 2026 (137 reviews, >30,000 participants) · Llanos-Lagos et al., Sports Med 2024
CORRECTEDThe Eddens citation doesn't cover the actual scheduling question here
Eddens's meta-analysis found resistance-before-endurance favoured lower-body dynamic strength over ≥5 weeks — but it only included studies with 15 minutes or less between modalities, and the authors scope the finding to people who can't separate their sessions. Nothing here runs a same-session double; every day is AM run, PM lift, hours apart. That paper doesn't actually speak to which day the heavy lift should fall on.
The real scheduling question — and the one that matters — is the between-day gap: this build used to put the heavy squat on Monday, ~36 hours before Wednesday's tempo, inside the window resistance training measurably depresses running economy (see below). Fixed by moving the heavy squat to Thursday — the only day with no run at all — and the light posterior lift to Monday. Wednesday's tempo now sits ~5 days post-heavy-squat instead of 36 hours.
The real scheduling question — and the one that matters — is the between-day gap: this build used to put the heavy squat on Monday, ~36 hours before Wednesday's tempo, inside the window resistance training measurably depresses running economy (see below). Fixed by moving the heavy squat to Thursday — the only day with no run at all — and the light posterior lift to Monday. Wednesday's tempo now sits ~5 days post-heavy-squat instead of 36 hours.
Eddens et al., Sports Med 2018 · Robineau et al., J Strength Cond Res 2016
AThe AM/PM split is doing real work — not just fitting your schedule
Separating the two modalities by at least six hours meaningfully reduces acute interference: in 58 athletes over 7 weeks, a 0-hour gap between strength and endurance work blunted strength gains, while 6-hour and 24-hour gaps did not differ from strength-training-alone. The mechanism is measurable too: AMPK rises sharply after high-intensity endurance work and takes at least three hours to return to baseline, while mTORC1 activity from strength training persists around eighteen hours.
Your structure puts roughly ten to twelve hours between morning runs and evening lifts — comfortably past the 6-hour threshold. Where the real caution sits is the next day's run, not the same day's: submaximal running performance can stay measurably depressed for up to 48 hours after heavy resistance training. That's the actual reason the heavy squat now falls on Thursday (see above) rather than Monday — it puts ~5 days, not 36 hours, between the heavy squat and the next demanding run.
Your structure puts roughly ten to twelve hours between morning runs and evening lifts — comfortably past the 6-hour threshold. Where the real caution sits is the next day's run, not the same day's: submaximal running performance can stay measurably depressed for up to 48 hours after heavy resistance training. That's the actual reason the heavy squat now falls on Thursday (see above) rather than Monday — it puts ~5 days, not 36 hours, between the heavy squat and the next demanding run.
Robineau et al., J Strength Cond Res 2016 · Doma et al., Sports Med Open 2019
AHeavy and low-rep beats light and high-rep for runners — clearly
A comparison of heavy/low-rep, light/high-rep and complex protocols found the heavy and complex interventions produced superior improvements, while the light/high-rep version produced no distinguishable improvements despite being load-matched. Heavier, lower-rep work also generates less residual fatigue than high-rep sets, which matters when you're running five days a week. This is the direct justification for legs going 5-rep and upper body staying in the 8-12 range.
Runner-specific strength protocol comparisons · Rønnestad & Mujika, Scand J Med Sci Sports 2014
BTwo lifting sessions a week is the demonstrated minimum, and you're above it
Several studies show benefits for runners from twice-weekly strength training, and the ACSM stand specifies ≥2 sessions per week for strength adaptation. You run four — two upper, two lower. The lower frequency is what the running evidence asks for; the upper frequency is free, because that's the half that doesn't interfere.
ACSM Position Stand 2026 · runner strength-training reviews
Race design validated against the endurance literature
Two things this audit changed
The taper was too shallow and has been rebuilt. And what I've been calling "blocks" is traditional periodization, not block periodization — both are supported, but the label was wrong. Details below.
FIXEDThe taper cut too little volume
Bosquet's meta-analysis of 27 studies found the most effective strategy is a two-week taper with training volume reduced 41-60%, without altering intensity or frequency. A later meta-analysis of 14 endurance studies confirmed it: a ≤21-day taper reducing volume 41-60% while maintaining intensity and frequency significantly improved time-trial performance, with the largest effect at 8-14 days. It also found tapering after a pre-taper overload period beat conventional tapering.
My original taper cut only 18% then 25% — well outside the effective range. Rebuilt: week 16 is now −43%, week 17 is −56%, achieved by shortening every run rather than dropping sessions. You still run five to six days and still do half-marathon-pace work, because frequency and intensity are exactly what the evidence says to hold constant. Week 15 stays as the overload peak that makes the taper work.
My original taper cut only 18% then 25% — well outside the effective range. Rebuilt: week 16 is now −43%, week 17 is −56%, achieved by shortening every run rather than dropping sessions. You still run five to six days and still do half-marathon-pace work, because frequency and intensity are exactly what the evidence says to hold constant. Week 15 stays as the overload peak that makes the taper work.
Bosquet et al., Med Sci Sports Exerc 2007 (27 studies) · Wang et al., PLOS ONE 2023 (14 studies)
FIXED"Blocks" was the wrong word for what this plan does
Block periodization means concentrated blocks of a single intensity — typically high-intensity work clustered together. What this plan actually runs is traditional periodization: a gradual progression from high volume toward higher intensity and greater specificity as the race approaches. That's a real distinction, not semantics.
On the merits, block periodization has a small edge: 20 studies were screened for eligibility, but only 6 were actually pooled in the meta-analysis, which found small favourable effects for VO2max and maximal power versus traditional. The authors flagged low methodological quality across the included studies, with a mean PEDro score of 3.7 out of 10 — a small edge from a thin evidence base. Traditional periodization also remains dominant in high-performance endurance sport. Progressive volume-then-intensity is the sound default regardless — the naming is what needed fixing.
On the merits, block periodization has a small edge: 20 studies were screened for eligibility, but only 6 were actually pooled in the meta-analysis, which found small favourable effects for VO2max and maximal power versus traditional. The authors flagged low methodological quality across the included studies, with a mean PEDro score of 3.7 out of 10 — a small edge from a thin evidence base. Traditional periodization also remains dominant in high-performance endurance sport. Progressive volume-then-intensity is the sound default regardless — the naming is what needed fixing.
Mølmen, Øfsteng & Rønnestad, Open Access J Sports Med 2019 (20 studies)
CI was too confident that you don't need a long run near race distance
A prospective study of 556 half-marathon runners found that a longest endurance run over 21 km was associated with a faster finish time and less pace decline during the race, as was weekly volume above 32 km — and found no association between training characteristics and new injuries in that cohort. Your peak long run is 90 minutes, roughly 8.4 miles (13.5 km) — well under 21 km.
So capping there is a deliberate trade, not a free lunch: less demonstrated race-day benefit than a longer peak run would give, in exchange for a lower total training-stress ceiling heading into the race. If the build stays completely clean through week 12, extending the peak long run toward 100-105 minutes is a legitimate option.
So capping there is a deliberate trade, not a free lunch: less demonstrated race-day benefit than a longer peak run would give, in exchange for a lower total training-stress ceiling heading into the race. If the build stays completely clean through week 12, extending the peak long run toward 100-105 minutes is a legitimate option.
Fokkema et al., Scand J Med Sci Sports 2020 (n=556 half-marathoners)
BThe long run sits marginally above the conventional ceiling
The widely used guideline, attributed to Daniels and Noakes, is that the long run should not exceed 25-30% of weekly volume. At peak, yours is about 9 miles of a 28-mile week — roughly 32%. Marginally over, and worth knowing that the fix is adding weekday easy volume rather than shortening the long run. This is a coaching convention with long practical standing rather than a trial result.
Daniels' Running Formula · convention, widely cited
BYour weekly volume is below the point where returns flatten
Analyses of recreational endurance performance consistently find weekly mileage and mean training pace explain most of the variance in race outcomes, and the return on additional volume, while diminishing, has not been shown to flatten out anywhere near this plan's range. Peaking at ~26 mi/week means more volume would still plausibly buy you performance — it's being held back deliberately, to keep the build sustainable through a full race cycle, not because a hard ceiling has been hit. The January strength block is where base volume can grow without a race deadline forcing the pace.
General endurance training-volume literature — direction well-supported, no single number verified for a plateau point
FIXEDThe Riegel exponent was 1.06 — corrected to 1.12
Riegel's power law needs an exponent, and 1.06 (the textbook default) is calibrated on elite/high-volume runners. It does not fit this build: Blythe & Kiraly's analysis of 164,746 runners and 1.4M performances found individual exponents average 1.12 (5th-95th percentile 1.10-1.15) for the general competitive-club population, and that a lower exponent tracks with higher ability — so 1.12 is a reasonable floor, not a midpoint, for a runner at this volume. Vickers & Vertosick (the previous citation here) fixes its own exponent at 1.07 and never validates an input shorter than 5K, so it couldn't actually license a 2-mile-trial projection in the first place.
At a 2-mile trial, this pace calculator now predicts a half about 12-14 minutes slower than the old 1.06 exponent did — every pace on this page, not just the finish-time prediction, shifts with it. Two mitigations already in the plan: treat the projection as a range, not a number, and re-test with a 10K at week 12. A 10K-to-half is a 2.1× extrapolation instead of 6.5×, and is inside the range every source here actually validated.
At a 2-mile trial, this pace calculator now predicts a half about 12-14 minutes slower than the old 1.06 exponent did — every pace on this page, not just the finish-time prediction, shifts with it. Two mitigations already in the plan: treat the projection as a range, not a number, and re-test with a 10K at week 12. A 10K-to-half is a 2.1× extrapolation instead of 6.5×, and is inside the range every source here actually validated.
Blythe & Kiraly, PLoS ONE 2016 (PMID 27336162) · Riegel, American Scientist 1981
BTwo hard sessions per week with the rest easy
Directionally holds up, with a caveat worth stating plainly: in the 30-runner trial most often cited for this, both a polarised and a threshold-weighted group improved 10K time by similar amounts over ten weeks — the primary between-group comparison was not statistically significant. The polarised advantage shows up only in a post-hoc comparison of the most-adherent runners in each group. The direction (mostly-easy, not mostly-moderate) is still the safer bet and matches broader elite-training-log observations, but treat "polarised clearly wins" as more confident than the trial evidence actually is.
Muñoz et al. 2014 · Stöggl & Sperlich, Front Physiol 2014
AHills for a hilly course, and strength work to survive it
Central Park's 470-590 feet of climbing across Harlem Hill twice and Cat Hill three times makes the hill block course-specific rather than merely knee-sparing. And the strongest injury-prevention finding in the whole document applies directly here: strength training's own relative risk for sports injury is 0.315 — roughly a two-thirds reduction — outperforming stretching and proprioceptive work, with a clear dose-response relationship. Four lift days through a race build is the evidence-backed choice, not a compromise.
Lauersen et al., Br J Sports Med 2014 & 2018
CDeload every fourth week
No controlled evidence that a fixed four-week cycle beats autoregulating by feel. It's in the plan as scheduled insurance against accumulated joint load, and because it gives the testing protocol a natural home on fresh legs. Convention, honestly labelled.
Convention, not trial data
CTwo weeks of easy running after the race
The "one easy day per mile raced" heuristic has no trial behind it. What is established is that prolonged running measurably degrades the determinants of endurance performance — a study in trained marathon runners found VO2peak dropped 3.1% after 90 minutes and 7.1% after 120 minutes of running, with running economy deteriorating progressively. Recovery is clearly needed; the specific fourteen-day number is convention.
Prolonged running determinants study, 2025 · heuristic otherwise
Corrections things this plan had wrong
FIXEDThe 10% weekly mileage rule was in this plan and shouldn't have been
It has no trial support. A randomised controlled trial of 532 novice runners found no injury difference between a 13-week program using 10% weekly increases and an 8-week program progressing far faster. The rule appears to originate in 1980s coaching convention, not research.
What replaced it: a cohort of 5,205 runners followed 18 months across 588,000 sessions and 1,820 injuries. Week-to-week mileage change predicted injury poorly. Single-session spikes did. A single run exceeding 110% of the longest run in the prior 30 days raised overuse injury risk by ~64% (hazard ratio 1.64); the calendar now builds the long run against the recent longest run, not against last week's total — and checks the ratio dynamically against actual pace rather than trusting fixed targets.
What replaced it: a cohort of 5,205 runners followed 18 months across 588,000 sessions and 1,820 injuries. Week-to-week mileage change predicted injury poorly. Single-session spikes did. A single run exceeding 110% of the longest run in the prior 30 days raised overuse injury risk by ~64% (hazard ratio 1.64); the calendar now builds the long run against the recent longest run, not against last week's total — and checks the ratio dynamically against actual pace rather than trusting fixed targets.
Buist et al., Am J Sports Med 2008 (GRONORUN RCT, n=532) · Frandsen, Nielsen et al., Br J Sports Med 2025 (n=5,205)
Lifting Sun · Mon · Wed · Thu
A10+ sets per muscle per week is optimal — why you now lift four times, not twice
An umbrella review of the hypertrophy literature concluded at least ten weekly sets per muscle is optimal. A separate systematic review put the band at 12-20 sets for trained young men, while finding no significant difference between moderate and high volume for quads and biceps specifically. Your current numbers — 8 to 12 across muscles — sit inside that band. The two-session version sat below it.
Frontiers in Sports & Active Living 2022 umbrella review · Baz-Valle et al., 2022
AFour hard sets per muscle per week still grows muscle — your bad-week fallback
A narrative review established a minimum effective dose of about four working sets per muscle per week, performed close to failure, at loads corresponding to 6-15 repetitions. For strength specifically, 2-3 sets per exercise per week at 70-85% of one-rep max is sufficient. This is why dropping to two sessions during a brutal quarter isn't abandoning the plan.
Iversen et al., Sports Medicine 2021 · Androulakis-Korakakis et al.
AStopping 1-3 reps short of failure is enough
A meta-regression across 55 hypertrophy and 67 strength studies found growth improved as sets got closer to failure, but the slope flattened past about 2 reps in reserve — and proximity to failure did not help strength. An 8-week randomised trial in trained adults found similar quadriceps hypertrophy whether sets went to true failure or stopped at 1-2 RIR. Practical translation: 1-3 RIR on hypertrophy work, 3-5 on heavy strength sets, true failure rarely if ever.
Robinson et al., Sports Medicine 2024 · Refalo et al., J Sports Sci 2024
BTwice-weekly frequency per muscle — a defensible default, not a proven optimum
Frequency effects independent of total volume remain genuinely unsettled. A 2024 dose-response meta-regression across 67 studies and 2,058 participants flagged the frequency evidence as mixed. Upper and lower are both hit twice here because it distributes volume sensibly, not because 2× is established as superior.
Pelland et al., resistance training dose-response meta-regression 2024
CDeload every 8-12 weeks
Widely practised, thinly evidenced. There's no strong controlled data showing scheduled deloads outperform simply autoregulating by feel in recreational lifters. It's in the plan as cheap insurance against accumulated joint stress across an 18-week build, not because the research demands it.
Convention, not trial data
Legs specifically why lower stayed 2×
AStrength training is the single best-evidenced way to not get injured while running
A meta-analysis of randomised controlled trials found strength training's own relative risk for sports injury was 0.315 (95% CI 0.207-0.480) — roughly a two-thirds reduction, and the best single number in this program's favour, better than "outperforms stretching" understates it. A follow-up analysis found a similar relative risk (~0.34) with a clear dose-response relationship: a 10% increase in strength-training volume cut injury risk by more than four percentage points. This is the strongest single argument in the document for keeping two lower days.
Lauersen et al., Br J Sports Med 2014 · Lauersen et al., Br J Sports Med 2018
AHeavy lifting makes you faster — and running economy is a lower-body adaptation
A systematic review and meta-analysis found heavy resistance training improved running economy more than plyometric training, with a larger effect on time-trial performance as well. A separate review across 652 runners found high-load work at or above 80% of one-rep max improves economy, particularly at faster speeds. The protocols producing these effects ran 2-3 sessions per week for 8-12 weeks — which is the dose your Monday and Thursday lower sessions hit and a single weekly lower day would not.
Eihara et al., Sports Med Open 2022 · Llanos-Lagos et al., Sports Medicine 2024
BRunning interferes with leg growth more than cycling does
A large meta-analysis of 43 studies found concurrent training did not compromise hypertrophy or maximal strength overall, with interference showing up mainly in power and explosive strength — and mainly when both modalities were trained in the same session. A fibre-level follow-up found a negative effect on type I fibre growth specifically when the aerobic training was running rather than cycling. The picture isn't unanimous — at least one meta-analysis found the opposite modality pattern for HIIT specifically — so call this a lean rather than a settled result. Two consequences kept in your plan regardless: legs get their own dedicated volume rather than relying on running, and the hard session defaults to hills rather than flat intervals.
Schumann et al., Sports Med 2022 (43 studies) · Lundberg et al., Sports Med 2022
Running AM slots · Sunday long run
ASingle-run spikes predict injury; weekly totals mostly don't
See Corrections above. The operative number is 110% of your longest run in the past 30 days. This is the one hard ceiling in the running plan.
Frandsen, Nielsen et al., Br J Sports Med 2025 (n=5,205, 588,000 sessions)
BMostly-easy intensity distribution beats mostly-moderate
A randomised trial of 30 recreational runners over 10 weeks compared a polarised distribution (roughly 77/3/20) against a threshold-weighted one — both improved 10K time by a similar amount (the between-group difference was not significant; see above). A separate multi-sport trial (runners, cyclists, triathletes and skiers) found a polarised group produced the largest gains in VO2peak and time-to-exhaustion — notably, at the highest total training volume of the four groups it compared, not a lower one. No improvement in running economy specifically was found in any group in that trial.
Muñoz et al. 2014 · Stöggl & Sperlich 2014
CHill repeats as the default hard session
The reasoning is sound but assembled rather than tested: uphill running reduces impact loading relative to flat running, running interferes with strength adaptation more than cycling does, and downhill running produces the highest eccentric knee loads. No trial has compared hill repeats against flat intervals for someone in your specific situation. It's the option that stacks the fewest known risks, not a validated prescription.
Inference from impact-loading and interference literature
Mobility and why there's no long stretching block
AFull-range lifting improves flexibility about as much as stretching
A meta-analysis of 11 randomised controlled trials with 452 participants found no significant difference between strength training and stretching for range of motion, across hip, knee, shoulder, elbow, trunk and cervical spine, with sub-group analyses by active versus passive ROM showing the same. A 2024 controlled trial matched training time and stretch intensity between groups: both improved sit-and-reach equally, and only the lifting group also gained strength. This is why your RDLs and squats are counted as mobility work in this plan.
Afonso et al., Healthcare 2021 (11 RCTs, n=452) · Rosenfeldt et al., BMC Sports Sci Med Rehabil 2024
BStatic stretching after training, dynamic before
Pre-exercise static stretching produces a temporary reduction in force output, but the actual dose-response is more forgiving than the old caution here implied: a multilevel meta-analysis (83 studies, 2,000+ participants) found bouts under 60 seconds sit in the trivial-effect range, and found no evidence for a blanket ban on stretching in warm-ups. The pre-squat calf and hamstring prep here runs 45-second bouts — safely inside that window, not an exception to the rule.
Warneke & Lohmann, J Sport Health Sci 2024
Hip & glute work why it's in every lower session
ADirect hip and glute work is injury prevention, not injury management
Strength training generally — and hip/glute strength specifically, since glute weakness is one of the most consistently implicated factors in overuse running injuries — is the best-evidenced way to keep a runner uninjured (see Lauersen below). The 6-minute block after both lower sessions is dosed for maintenance, not rehab: nothing here is restricted or dosed around a symptom.
Lauersen et al., Br J Sports Med 2014 & 2018
Desk two breaks a day
BInterrupting sitting has measurable acute effects — but the effective dose is walking, not stretching
A randomised crossover trial had adults complete 6-hour conditions of prolonged sitting, prolonged standing, or sitting broken by 2 minutes of brisk, inclined walking every 30 minutes. The walking-break condition raised popliteal artery blood flow by around 80% and shear rate by around 72% at 60 minutes, sustained across the full 6 hours; prolonged standing raised blood flow only transiently. A related trial found the same walking-break protocol did not trigger compensatory eating.
Honest caveat, twice over: the trial's own primary outcome (flow-mediated dilation) was not statistically significant — the vascular case here is stronger on mechanism than on the trial's own headline result. And the active ingredient was ambulation, not a stretch — which is why the desk blocks lead with a lap or the stairs rather than a chair stretch.
Honest caveat, twice over: the trial's own primary outcome (flow-mediated dilation) was not statistically significant — the vascular case here is stronger on mechanism than on the trial's own headline result. And the active ingredient was ambulation, not a stretch — which is why the desk blocks lead with a lap or the stairs rather than a chair stretch.
Peddie et al., PLOS ONE 2021 · Mete et al., Nutrients 2018
Nutrition
BHigher protein preserves lean mass in a deficit
The standard review on athlete protein needs recommends 1.3-1.8 g/kg/day generally, rising to as much as 1.8-2.0 g/kg/day specifically during energy restriction — without adequate protein and resistance training, a meaningful share of weight lost comes from lean mass rather than fat. Not implemented as a tracked target anywhere in this app; treat it as a ballpark, not a number to hit exactly.
Phillips & Van Loon, J Sports Sci 2011
The premise underneath everything
CAdherence beats optimisation
This is why the fallback on a bad week is two sessions rather than four, and why every session has an explicit "if you miss it" instruction rather than silence. It is also the least rigorously supported claim in this document. Nobody runs ten-year randomised trials on training volume. It's an inference from dropout rates and from the flattening shape of the dose-response curve — a reasonable one, but an inference. Worth knowing that the load-bearing assumption of the whole structure is the softest thing on this page.
Inference, not a controlled result