OUTRN

Marathon Training Plan: Real Plans vs What the Research Says

Training | Published 2026-10-03 | By OUTRN Team

The OUTRN team compares real marathon plans from Higdon, Sydney and Melbourne, then checks each convention against the research, and finds where it runs out...

Open any three marathon plans and you get three different answers on weeks, long runs and rest days. Ask the research and the honest reply is that most of those numbers are conventions, not findings, with a few principles that hold up and a long list of questions nobody has tested...

The short answer: published marathon plans run 12 to 18 weeks, build to a longest run of about 32 to 36km and taper for about 1 to 4 weeks. The research supports gradual progression, a taper of about two weeks and strength training. It does not support any specific plan length, any specific peak long run, or the weekly 10 percent rule. This guide is for someone who already runs 10km comfortably and is building toward a first marathon.

What do real marathon plans prescribe?

Here are the plans we could open, with what they actually say. These are conventions: nobody has tested one against another. The sources are Higdon's Novice 1 and Intermediate 1 pages, the TCS Sydney Marathon first-timer and advanced plans, and the Melbourne Marathon training series page.

PlanLengthRuns a weekLongest run
Hal Higdon Novice 118 weeks4About 32km
Hal Higdon Intermediate 118 weeks5About 32km, twice
TCS Sydney Marathon first-timer plan (2024 edition)16 weeks3, time-based80 minutes continuous, or 5 x 30 minutes with walks
TCS Sydney Marathon advanced plan (2024 edition)16 weeks532km
Melbourne Marathon group series (2026)12 weeks2 group sessions34 to 36km

Hal Higdon's Novice 1 is described as for first-time marathoners or those who want a gentle, low-volume approach, and its schedule has no speed sessions. Its Saturday long runs go about 10, 11, 8, 14, 16, 11 and 19km, with a half marathon race in week 8, then about 16, 24, 26, 19, 29, 23 and 32km in week 15, then a three-week taper. The first-timer plan from the TCS Sydney Marathon (coached by Brett Robinson) is aimed at "someone who runs once or twice a week or is new to running" and is entirely time-based, opening with walk-run intervals. Elite coach Benita Willis, writing on the Gold Coast Marathon site, suggests building the long run to 34 to 38km.

Notice the spread: 12, 16 and 18 weeks, and peak long runs from about 32km to 38km. That spread is the point.

How many weeks should you train for a marathon?

Long enough to build gradually, and the research cannot tell you more precisely than that. We found no trial or review comparing 12, 16 and 18 week marathon plans. The only plan-length trial we found (Buist and colleagues, American Journal of Sports Medicine, 2008) put 532 novices through either a graded 13-week programme or a standard 8-week one for a 6.7km event, and injuries were the same: 20.8 and 20.3 percent. That is a short event and a novice group, so it is only a hint.

If you are starting a few weeks late, the Sydney plan's own advice is to complete every other week until you catch up.

How long should your longest run be?

Plans disagree by up to 6km, and the research says only that more is associated with faster. A meta-analysis of 85 studies and 137 cohorts (Doherty and colleagues, Journal of Science and Medicine in Sport, 2020) found that weekly distance, runs per week, the biggest single week, the number of runs of 32km or longer, longest run and weekly running hours were each associated with faster marathon times (R-squared 0.38 to 0.81). That compares groups of runners, not what happens when one runner adds volume: faster, more experienced runners simply train more.

One mechanism study fits: in 26 well-trained men matched on 10km time (Zanini and colleagues, Medicine and Science in Sports and Exercise, 2026), runners who regularly ran 90 minutes or longer lost 3.1 percent of their running economy over a 90-minute run, against 6.0 percent for runners whose runs were all under 70 minutes. That supports long runs as useful, but it is small, cross-sectional and does not prescribe a distance.

How do you build without getting injured?

Progress the long run gradually, and be sceptical of the weekly 10 percent rule. In 5,205 runners (Frandsen and colleagues, British Journal of Sports Medicine, 2025), a single session more than 10 percent longer than your longest run in the previous 30 days was linked to more injuries: a hazard rate ratio of 1.64 for a 10 to 30 percent jump, 1.52 for 30 to 100 percent and 2.28 for more than 100 percent. The week-to-week ratio showed no relationship. The study used self-reported injuries and mostly experienced recreational runners.

The weekly rule itself is weak. A one-year study of 874 novices (Nielsen and colleagues, 2014) found no significant difference in injuries across weekly progression groups, and the Buist trial above found a graded plan built on the 10 percent rule did not reduce injuries. A 2022 systematic review of 36 articles and 23,047 runners (Fredette and colleagues, Journal of Athletic Training) found 26.2 percent were injured and the evidence on which training variable causes it was conflicting, so it advised caution about recommending optimal progressions.

So, as coaching convention and not a tested rule: step the long run up gradually, step it back every few weeks as Higdon does, and treat the 10 percent rule as a rough guard, not a law.

How should you taper?

About two weeks, with the volume cut and the intensity kept. A meta-analysis of 27 studies (Bosquet and colleagues, 2007) found the most effective taper was 2 weeks, cutting training volume by 41 to 60 percent while keeping intensity and frequency. A 2023 meta-analysis of 14 studies (Wang and colleagues) found a taper of up to 21 days at that same cut improved time-trial performance. Bosquet pooled competitive athletes and Wang endurance athletes, not first-time marathoners.

In more than 158,000 recreational marathoners (Smyth and Lawlor, 2021), strict three-week tapers with volume falling progressively were linked with a median saving of 5 minutes 32 seconds (2.6 percent) over a minimal taper. That is observational: people who taper strictly may be more organised in general. The plans we opened taper for about 2 to 3 weeks (Higdon three, Sydney's first-timer plan two), and Melbourne's group series peaks about four weeks out.

Is 80:20 the right way to train?

It is a common practice, not a proven one. Elite endurance athletes self-organise to about 80 percent of sessions at low intensity and 20 percent hard (Seiler, 2010), but that describes elites training 10 to 13 times a week. In a 10-week trial of 30 recreational runners (Munoz and colleagues, 2014), a polarised group improved their 10km from 39:18 to 37:19 and a between-thresholds group from 39:24 to 38:00, and the 41 second difference was not statistically significant. A 2025 network meta-analysis of 13 studies (Rosenblat and colleagues) found no difference between polarised and pyramidal training on time trials or VO2max, though recreational athletes may gain more VO2max from pyramidal.

What we can say: mostly easy running with a little hard running is the norm and has not been shown to be worse. Higdon's Novice 1 schedule has no speed sessions, and Intermediate 1, which Higdon says has "no speedwork", adds only marathon-pace runs.

Is strength training worth it?

Yes, as a reasonable bet. A meta-analysis of 5 trials in 93 highly trained runners (Balsalobre-Fernandez and colleagues, 2016) found strength work 2 to 3 times a week for 8 to 12 weeks improved running economy with a large effect (SMD -1.42). A 24-study review (Blagrove and colleagues, Sports Medicine, 2018) found economy gains of 2 to 8 percent in most studies, and 1.5 to 10km time trials that tended to improve, in runners with at least six months' experience. Those are trained runners and not marathon times, so we hold it loosely, but the Australian physical activity guidelines also list muscle strengthening on 2 or more days a week.

Should you run-walk?

It is a reasonable strategy. In the only marathon trial we found (42 non-elite runners, Hottenrott and colleagues, 2016), run-walk finishers took 4:14:25 against 4:07:40 for continuous runners, a difference that was not statistically significant, and the run-walk group had less muscle pain and fatigue afterwards. It is a single small trial. Both Higdon and the Sydney plan endorse walking at aid stations.

Where does the evidence run out?

Is a marathon plan for you?

A comfortable 10km is where we think preparation starts, not where it ends. Sydney's first-timer plan is aimed at runners who run once or twice a week, but we would build a base first: a marathon is a long way to learn what your legs can do. That is our view, not a finding. If you are chasing a place in the Sydney Marathon, see our Sydney Marathon guide.

Inspired? Rehearse a long day with a crew

If you want a long trail day with support before your marathon block, our Blue Mountains trip has one built in: the Sunday Six Foot Track run, about 28 to 30km with more than 1,000m of climbing on the return, or a shorter 15 to 16km turnaround. It is not a race and there is no race clock, just a turnaround time so everyone is off the hill in daylight. Dates are to be confirmed and the price is an estimate of about $1,650 per person. Booking is not open yet, so register your interest and we will contact you first.

Register interest in Run the Blue Mountains

Further reading

References

Register interest in Run the Blue Mountains

Frequently asked questions

How many weeks should a marathon training plan be?

Published plans range from 12 weeks (the Melbourne Marathon's group series) to 16 weeks (the TCS Sydney Marathon plans) and 18 weeks (Hal Higdon's marathon plans). We found no trial comparing plan lengths for the marathon, so plan length is a convention. The only plan-length trial we found, in novices training for a 6.7km event, found no difference in injuries between an 8-week and a 13-week plan.

How long should your longest run be before a marathon?

Plans disagree: Hal Higdon's Novice 1 plan peaks at about 32km, the TCS Sydney advanced plan at 32km, the Melbourne group series at 34 to 36km, and elite coach Benita Willis suggests 34 to 38km. A meta-analysis of 85 studies found longer longest runs went with faster marathon times, but that is an association between groups, not proof that a longer long run makes you faster.

How long should you taper for a marathon?

A meta-analysis of 27 studies in competitive athletes found a 2-week taper that cut training volume by 41 to 60 percent while keeping intensity and frequency was the most effective, and a 2023 meta-analysis supported up to 21 days. In more than 158,000 recreational marathoners, strict three-week tapers were linked with a median 2.6 percent faster finish, though that is observational.

Is 80:20 training proven for marathon runners?

No. Elite endurance athletes tend to do about 80 percent of sessions at low intensity, but a 2025 network meta-analysis of 13 studies found no difference between polarised and pyramidal training on time trials, and recreational athletes may gain more VO2max from pyramidal. Mostly easy running with a little hard running is common practice and not shown to be worse.

Should I run-walk a marathon?

It is a reasonable option. In the only marathon trial we found (42 runners), run-walk finishers took 4:14:25 against 4:07:40 for continuous runners, which was not a statistically significant difference, and the run-walk group reported less muscle pain afterwards. Published plans from Higdon and the TCS Sydney Marathon both endorse walking at aid stations.