If you are a novice and have completed all the training necessary to run a half-marathon, you should be ready to start training for a marathon, should you so wish. It is necessary to emphasise that if you started as a complete novice with no recent running experience, you should have undergone at least 25 weeks of training. If you start on the next phase of the programme, without an adequate base, you will be at greater risk of injury once you start running more intensively with less rest between long runs.

The table below details the programme that I suggest to ensure that the runner (who is training for 160 minutes per week and who has successfully completed at least a 10km race) will be able to complete a standard marathon in a further 26 weeks. The programme is a slight modification of the one we used successfully to train 26 novices to complete a marathon within 36 weeks of their first 20-minute walk.

The key to the programme is the gentle extension in daily training volumes, with emphasis on the long runs, which increase by 10 minutes every second week.
This programme is clearly for those runners wishing to complete a marathon comfortably with a low risk of injury, and with the highest possible probability of success. It does not include speed or hill training which, if done properly, will undoubtedly improve your race time substantially.

Many programmes advise on the exact mileage that runners should cover when training for a marathon. This begs the question of what science tells us about the optimum training distances for marathon runners. In fact, there are few studies of the actual distances people run in training for a marathon. Thus, we do not really know what the optimum training distance is for the majority of novice marathon runners.

The distances advocated in this programme have been arrived at empirically, but are compatible with the findings of recent sports science research. When evaluating the training patterns of 88 marathon runners it was found that that average distance run in training was 60km per week for the 12 weeks prior to the race, and this ranged from 24 to 103km. This study also debunked two important myths. Firstly, there is no relationship between weekly training distance and marathon time. Secondly, despite their apparent inadequate training, the runners did not slow down dramatically after hitting their predicted ‘collapse point’ at about 27km.

Thus, they could find no evidence to support the collapse-point theory proposed by Ken Young. This theory holds that runners who do not train more than 101km per week ‘collapse’, and are reduced to a ‘shuffle’ when they race more than three times their average daily training distance for the last eight weeks before the marathon. Finally, these novice marathoners were unable to predict their marathon times accurately. However, the accuracy of their predictions did improve the closer they were made to race day.

During my marathon running career, I achieved personal best times of 2:50:20 (42km/marathon), 3:59:49 (56km/35 miles) and 6:49:00 (90km/56 miles). I achieved these times on the training programmes described here. I present them as an option for those with a similar physiology and training capacity. A measure of my physiological capacity were my best times for certified courses of 60:59 for 16km/10 miles and 81:39 for 21km/half-marathon.

My personal training approach was similar to the legendary Arthur Newton’s. It included plenty of long, slow distance to the exclusion of speedwork. This was because I originally switched to running (from rowing) with the express intention of completing the 56mile Comrades Marathon, regardless of finishing time. For the first six to eight years of my running career, I trained exclusively by running long, slow distances. However, I now firmly believe that this training approach, which emphasises distance training to the virtual exclusion of speedwork, although very safe, is not the best way to train for any distance, including ultra-marathons. I endorse Roger Bannister’s view that high mileage distance training increases the athlete’s speed of recovery from effort, but does not increase racing speed. The athlete must achieve a balance by doing just the right amount of speed training.

Thus, the evidence is that the fastest middle-distance and cross-country runners are the best at all distances, even the very long ultra-marathons. However, there is one important proviso – they need to have superior fatigue resistance. But this alone will not make a world-class marathon or ultra-marathon runner. For that, both speed and fatigue resistance are required.

With this background, I include details of the training practices I followed when running marathon races on a regular basis between the ages of 22 and 36. After that, I found that I could no longer train as hard as the programme required.

The initial goal of my hard training programme (see table above) was to condition myself to be able to run 110km per week, a distance that I have also found to be optimal for the majority of recreational runners who have major time constraints. This break-in phase lasted for 10 to 12 weeks, during which time my long weekend runs would not be less than 24km and not longer than 32km. The major indication that this phase had had its desired effect was that I started to finish the long runs so fresh that I wanted to run farther on the following long run. At the same time, my average training speed increased and the hills that I ran became much easier. When this happened, I was ready to move on to the second phase of my programme, the so-called ‘peaking phase’.

If there is one contentious issue in training for distance running, it is the exact value of running many miles at low intensities. That the majority of runners spend most of their time training at quite low intensities has been shown by a number of studies. For example, a study of 13 élite New Zealand distance runners found that their average training intensity was characterised by the following: their average heart rate was 145 beats-per-minute; their average percentage VO2max was 64%; their average running speed 15.6km/9.75 miles per hour, which corresponds to 77% of the speed at which the lactate turnpoint occurred. Remarkably, only 4% of their training involved running at speeds greater than that at which the lactate turnpoint occurred.

Another study found that the average training pace of a group of top German female marathoners corresponded to only 60% VO2max, or less than 77% of the running speed at which their blood lactate concentrations reached 4 mmol.l-1.

However, I am not yet ready to conclude that all low-intensity training is unnecessary. Certainly, provided the total training volume is less than 100km/62miles per week, this low-intensity training would not seem detrimental. But its value for running performances, certainly over the shorter distances, has not yet been proven. I have also collated evidence showing how well many elite runners have performed on relatively little training. The major benefits of heavy training volumes in excess of 120km/75miles per week are to increase the strength of the connective tissue in the muscles and the resistance to the eccentric muscle damage that produces fatigue after running 30 or more kilometres (19 miles), which then increases your ability to keep running beyond the marathon ‘wall’.

The aim of peaking is to increase the training load further, by adding speed training sessions, either in the form of intervals, speed-play (Fartlek), time-trials or short-distance races (5 to 16km/3 to 10 miles) for a period of four to six weeks before competitions. This form of training produces dramatic changes in racing speed but if maintained for too long, it can induce early symptoms of overtraining. Thus, it is a high risk/high reward period of your training.

The next phase of my hard training cycle differed, depending on the length of the race for which I was preparing. For shorter distances, I emphasised mostly speed training and maintained the weekly training distance at about 120km/75 miles per week. For ultra-marathons, I emphasised distance training and long weekend runs, only adding speed training when I had completed the heavy distance training.

During the peaking phase of my standard marathon training, I would emphasise speed training sessions, either on a Tuesday or a Thursday, and would run two or three races of 10 to 16km (6 to 10 miles) – but no further. I found that these are the optimum racing distances for preparing for both the 10km race and the marathon. Longer races tend to cause more severe muscle damage from which recovery is slow. Also, from a psychological viewpoint, the marathon breaks up neatly into two 16km/10 miles races and one 10km/6 miles race. Thus, during the marathon race, I would concentrate on running as close to my best times for each of these distances as was possible. When properly prepared, it is remarkable how close you can come to this goal.

During the second-last week before the marathon, I would reduce my training to between 50 and 80km/30 to 50 miles of easy running and would rest and carbohydrate-load for the last three days before the race. During the intervening four days, I would incorporate three days of mild carbohydrate-restriction and runs of 12 to 18km/7.5 to 11 miles, depending on how I felt.

The ideal taper for marathon and ultra-marathon runners has not yet been established in a scientific trial. My bias is to believe that there should be more rest and less running during the tapering phase and certainly more days in which you do no training at all.

I have written elsewhere about the ‘Zatopek phenomenon’ in which élite athletes achieved remarkable performances after a period of reduced training – in the case of Zatopek, even after being hospitalised for two days before his record-breaking performances. Some 30 years since this phenomenon was first recognised, I realise that I ran one of my best 56km/35 miles– ultra-marathon races after a period of enforced rest. I ran the race a mere three weeks after undergoing surgery to my foot, which prevented me from running for two weeks. In the last week before the race, I had only been able to fit in a few jogs.
Without trying, I ran a time that was less than 40 seconds slower than my best, over the distance achieved three years later after a much more intensive training programme, but for which I did not taper properly. The last word on the ideal taper has yet to be written.