The Long Run for Marathon and Ultramarathon races

The Long Run is a widely accepted cornerstone of endurance training and virtually every marathon and ultramarathon training program emphasizes the Long Run. However, while there are some areas of consensus, much of the advice around the Long Run is contradictory and The Science of the Long Run is limited. This page tries to balance the need to give usable, specific direction with the need to impart the rationale and limitations of that advice.

1 How long is a long run?

There is no clear definition of how long a run has to be before it is considered a 'Long Run'. Some coaches use the term to describe the longest run of the week as the Long Run, and they consider even a 5 mile run as a Long Run. However, the term 'Long Run' is most frequently used in the context of marathon and ultramarathon training. In this context, a Long Run is often longer than a given distance, such as 16 miles, or a given time, such as 2 hours. At these distances and times Glycogen depletion becomes a significant factor, as does muscular damage.

2 Benefits of the Long Run

While there are many Endurance Adaptations, it is unclear which are specifically enhanced by Long Run training. There are a number of factors that are specific to the Long Run.

• The stress of a Long Run may be linked to adaptations that occur as a result of training with depleted Glycogen store. It is unclear how the level of Glycogen at the beginning of a run changes the impact of a Long Run. Beginning a run with low Glycogen stores, either through a low carbohydrate diet or prior training, certainly makes a Long Run much harder.
• The muscular stress of a Long Run can lead to Delayed Onset Muscle Soreness (DOMS). One of the effects of DOMS is protection against future training that would otherwise have produced DOMS. It seems likely that the benefits of the Long Run are linked to DOMS.
• The patterns of Glycogen depletion suggest that during a Long Run different Muscle fibers become activated. Therefore a Long Run may train muscle fibers that would otherwise not be stressed by shorter runs.
• The psychological benefits of the Long Run should not be underestimated. The Central Governor Theory suggest that Long Run training provides critical mental adaptation.

3 Estimating the Stress of a Long Run

The table below uses an estimate of the relative difficulty for different length and pace Long Runs. This particular table is for a 4:00 marathon runner, but you can create a customized table using the VDOT Calculator. This approach uses a number of assumptions.

• The level of Glycogen depletion is used as a proxy for the training stress.
• It is assumed that a marathon race (marathon distance at marathon pace) uses all available Glycogen.
• It is possible to calculate the relative amount of Glycogen used at different exercise intensities. (For those interested, I used the work of Romijn to create the formula y = 0.0021x² + 0.7896x - 21.031, where X is the percentage of V̇O₂max and Y is the relative percent of Glycogen used.)
• Pace at 100% of V̇O₂max (vV̇O₂max) is assumed to be the fastest pace that can be maintained for 6 minutes.
• The energy cost of running a given distance is assumed to be constant, allowing the percentage of V̇O₂max to be determined by running speed.
• The calculator assumes that the rate of glycogen consumption remains constant for a given pace.

Pace	13 Mile	14 Mile	15 Mile	16 Mile	17 Mile	18 Mile	19 Mile	20 Mile	21 Mile	22 Mile	23 Mile	24 Mile	25 Mile	26 Mile	26.2 Mile	27 Mile	28 Mile	29 Mile	30 Mile
MP-30 sec 8:39/Mile	54 %	58 %	63 %	67 %	71 %	75 %	79 %	83 %	88 %	92 %	96 %	100 %	104 %	108 %	109 %	113 %	117 %	121 %	125 %
MP-20 sec 8:49/Mile	53 %	57 %	61 %	65 %	69 %	73 %	77 %	81 %	85 %	89 %	93 %	97 %	101 %	105 %	106 %	109 %	113 %	117 %	121 %
MP-15 sec 8:54/Mile	52 %	56 %	60 %	64 %	68 %	72 %	76 %	80 %	84 %	88 %	92 %	96 %	100 %	104 %	104 %	108 %	112 %	116 %	120 %
MP-10 sec 8:59/Mile	51 %	55 %	59 %	63 %	67 %	71 %	75 %	79 %	83 %	86 %	90 %	94 %	98 %	102 %	103 %	106 %	110 %	114 %	118 %
MP 9:09/Mile	50 %	53 %	57 %	61 %	65 %	69 %	73 %	76 %	80 %	84 %	88 %	92 %	95 %	99 %	100 %	103 %	107 %	111 %	115 %
MP+10 sec 9:19/Mile	48 %	52 %	56 %	59 %	63 %	67 %	70 %	74 %	78 %	82 %	85 %	89 %	93 %	96 %	97 %	100 %	104 %	108 %	111 %
MP+15 sec 9:24/Mile	48 %	51 %	55 %	59 %	62 %	66 %	69 %	73 %	77 %	80 %	84 %	88 %	91 %	95 %	96 %	99 %	102 %	106 %	110 %
MP+20 sec 9:29/Mile	47 %	50 %	54 %	58 %	61 %	65 %	69 %	72 %	76 %	79 %	83 %	87 %	90 %	94 %	94 %	97 %	101 %	105 %	108 %
MP+30 sec 9:39/Mile	46 %	49 %	53 %	56 %	60 %	63 %	67 %	70 %	74 %	77 %	81 %	84 %	88 %	91 %	92 %	95 %	98 %	102 %	105 %
MP+45 sec 9:54/Mile	44 %	47 %	51 %	54 %	57 %	61 %	64 %	67 %	71 %	74 %	77 %	81 %	84 %	88 %	88 %	91 %	94 %	98 %	101 %
MP+60 sec 10:09/Mile	42 %	45 %	49 %	52 %	55 %	58 %	61 %	65 %	68 %	71 %	74 %	78 %	81 %	84 %	85 %	87 %	91 %	94 %	97 %
MP+90 sec 10:39/Mile	39 %	42 %	45 %	48 %	51 %	54 %	57 %	60 %	63 %	66 %	69 %	72 %	75 %	78 %	78 %	81 %	84 %	87 %	90 %
MP+120 sec 11:09/Mile	36 %	39 %	42 %	44 %	47 %	50 %	53 %	55 %	58 %	61 %	64 %	66 %	69 %	72 %	73 %	75 %	78 %	80 %	83 %

4 Guidelines

These guidelines based on my interpretation of the scientific and anecdotal evidence. These guidelines attempt to provide specific, usable direction that is balanced by the limitations of the available evidence. Because of the limitations of our understanding, these guidelines use specific words in upper case to indicate the degree of confidence.

• "MUST" or "WILL" indicate a certainty in the guideline and the belief that the advice is generally accepted.
• "SHOULD" indicates a high level of confidence in the guideline.
• "SHOULD PROBABLY" is used where there is moderate confidence in the guideline and I believe the guideline is most likely right.
• "MIGHT" provides only a broad suggestion with limited evidence and confidence.
• "UNCLEAR" is used when there is no evidence and usually reflects a simple observation of a common practice.

I've included the underlying rationale as well as the caveats (limitations) for the guidelines to allow you to evaluate them for yourself. For more details on the available science, see The Science of the Long Run.

4.1 Increasing Distance

The distance of the Long Run MUST be increased gradually.

• Rationale:
  • It is generally accepted that Endurance Adaptations require an incremental increase in training stress.
  • There may be some injury risk to longer long runs, but it seems more likely that the injury risk comes from ramping up the length of the long run too quickly.
• Caveats:
  • Many factors impact the difficulty of a Long Run beyond the overall distance, including pace, the amount of Downhill Running, [[Nutrient Timing| post run nutrition], Massage and Training Monotony.
  • Following a fixed plan may be inappropriate, as individual responses vary.

4.2 Muscle Soreness

Long runs SHOULD produce slight Delayed Onset Muscle Soreness, with recovery in a few days.

• Rationale: Muscle damage appears to be a stimulus for remodeling and improvement, so some muscle damage is beneficial. However, a long run that produces too much muscle damage would take too long to recover from and result in detraining.
• Caveats: The time taken to repair muscle damage from running appears to vary with the level of damage, but it can be tricky to work out what the right level of training is to produce the required soreness.

4.3 Long Run Length for Marathon Training

For marathon training, the longest Long Run SHOULD be over 20 miles, with the details varying with the expected marathon time.

• Rationale:
  • The correlation shown in the available marathon studies suggests that longer Long Runs may be beneficial. The link between the longest training run being over 20 miles and the self-reported incidence of 'hitting the wall' provides the only available guidance for the actual mileage of the longest Long Run.
  • Many runners have found they can build up their Long Run distance to the point they can complete marathons on a regular basis. The Marathon Maniacs club has members, some of them relatively slow runners, who complete the marathon distance on a weekly basis.
  • The details of the longest Long Run depend on the target time, as noted in other guidelines.
• Caveats:
  • Running the Long Run while pre-fatigued may reduce the needed length. This pre-fatigue can be achieved by using two shorter runs with insufficient time to recover between them, either two on the same day, or on consecutive days. This pre-fatigue could also be achieved through higher overall mileage but this should be carefully structured to minimize Training Monotony and the risk of Overtraining.
  • Unless the length of the Long Run is increased gradually, running longer can produce excessive muscle damage and be counterproductive. (See muscle soreness recommendation above.)
  • Focusing on the single longest run is probably inappropriate, and the average of the 3-5 longest runs would be a more useful metric, but one that is rarely used.
  • The pace of the long run is an important factor. Faster paced shorter runs may have similar stresses to longer, slower runs, though it seems reasonable that the resulting adaptations would be somewhat different.
  • Hilly Long Runs will produce more training stress than the same distance on the flat.
  • The need for longer Long Runs might be dramatically reduced if the training runs are continuous running and the race uses a walk/run pattern.
  • Any training stress requires adequate recovery time to enable Supercompensation. Therefore training must be structured to minimize Training Monotony.
  • Individuality and prior experience may influence the optimal Long Run distance. I know of runners who believe that Long Runs over 26 miles has benefited their marathon performance and runners who believe that reducing their longest Long Run has been beneficial.

4.4 Long Run Length for Running Sub-4:30 Marathons

The longest Long Run for sub-4:30 marathons running SHOULD be over 20 miles.

• Rationale:
  • For a 4:30 marathoner, a 20 mile run at a 12 min/mile pace is around 4 hours. This is 30 minutes less than the race at 100 seconds/per mile slower, which is around 60% of the stress of the race. This provides a reasonable training stress without requiring an excessive running time.
    • See VDOT Calculator for details of estimating Long Run difficulty.
• Also, see the rationale and caveats in 'Long Run Length for Marathon Training' above.

4.5 Long Run Length for Running 4:30 to 5:30 Marathons

Long runs for 4:30 to 5:30 marathons that will be raced by running continuously rather than using a run/walk method SHOULD use an approach of combining shorter, faster long runs with longer run/walks. The shorter, faster Long Runs SHOULD PROBABLY be up to 16 miles at Marathon Pace. The longer run/walk Long Runs SHOULD PROBABLY be 21-23 miles. (For run/walk marathons, see the guideline below.)

• Rationale:
  • Marathons in the 4:30 to 5:30 range can be run, but the long run training pace becomes too slow to be efficiently run without walking breaks.
  • Using a run/walk approach for the long runs can build endurance, but need to be balanced by some shorter Long Runs that do not have walking breaks.
  • Running at marathon pace for 16 miles should provide around 60% of the race stress.
  • The Walking Breaks on the long run/walk of 21-23 miles should reduce the stress of the long distance.
  • Also, see the rationale in 'Long Run Length for Marathon Training' above.
• Caveats: It may be better to use a Run/Walk approach to racing a 4:30-5:30 marathon
• Also, see the rationale and caveats in 'Long Run Length for Marathon Training' above.

4.6 Run-Walk 5:30 and Slower Marathons

Marathoners finishing the race in 5:30 or longer SHOULD use a run/walk approach for the race and Long Runs.

• Rationale:
  • At slower paces it is more efficient to walk than to run. For most people, this occurs between a 13:00 and 15:00 min/mile pace.
  • The impact of a walking break on overall pace depends on the paces of the running and walking. For instance:
    • Walking 1 minute at 15:00 min/mile pace and running 10 minutes at a 7:00 min/mile pace results in an average pace of 7:21, a big drop of 5%.
    • By comparison, Walking 1 minute at 15:00 min/mile pace and running 10 minutes at a 12:00 min/mile pace results in an average pace of 12:13, a much smaller relative drop of 1.8%.
• Caveats: It is practical to use a run/walk approach for faster marathon runners as well.

4.7 Long Run Length for Run-Walk Marathons

The longest Long runs for run-walk marathons SHOULD PROBABLY be run/walk of over 20 miles at slower than race pace. Long Runs of 16-20 miles that use the same run/walk ratio and pace as the race MIGHT act as a substitute. Continuous running Long Runs of 12-16 miles MIGHT be a partial substitute for sub-5:30.

• Rationale:
  • The rationale noted above for a longest Long Run over 20 miles at slower than race pace applies.
  • These three options provide similar training stress, but they probably produce somewhat different adaptations:
    • A 20 mile long run at 15% slower than marathon pace.
    • A 16 mile long run at marathon pace.
    • A 14 mile long run at 7% faster than marathon pace. However, because this is continuous running the training stress is higher.
  • Runners slower than 5:30 should use run/walk even in training.

4.8 Long Runs and Walking Breaks

Taking walking breaks WILL reduces the stress of a long run or endurance race.

• Rationale: There is strong anecdotal evidence that walking breaks on a Long Run allows for recovery, reducing the training stress. This may be because the muscle fibers have time to absorb more nutrients (glucose, fat) from the blood, or because it allows more blood to flow to other part of the body, mobilizing nutrients.
• Caveats: These breaks can be intentional and structured, as in the Galloway approach to training, or due to stops for logistical reasons such as getting a drink.

4.9 Post Long Run Nutrition

The benefit and stress of a Long Run WILL be dependent on the post-run nutrition.

• Rationale: Post-exercise nutrition is an important aspect of training. The correct Nutrient Timing will optimize the benefits of exercise, and Protein intake immediately after DOMS inducing exercise makes a significant difference to the subsequent recovery and adaptation.

4.10 Training for Novice and Experienced Marathoners

Plans for novice marathon runners SHOULD be different to those for those with recent experience. Novice plans SHOULD focus on building up the length of the Long Run gradually, while experienced marathoners SHOULD PROBABLY be back to doing 16+ mile long runs soon after previous marathon.

• Rationale: A key part of training for a first marathon is building up the length of the Long Run. A runner who has just built up their endurance for a marathon race can lose endurance if they use a novice plan that starts the buildup process from scratch.

4.11 Tapering the Long Run

The longest Long Run SHOULD be 3-4 weeks before the race and there SHOULD be no Long Run within 2 weeks of the race.

• Rationale: There is some evidence that the benefits and damage of the Long Run last for many weeks. Therefore it is prudent to avoid long runs too close to the race.
• Caveats: This assumes that the objective is an optimal performance. It is quite possible to do many marathon or longer races in close succession.

4.12 The Long Run Distance and Weekly Mileage

The length of the Long Run SHOULD not be limited to a percentage of the weekly mileage.

• Rationale:
  • Some coaches limit some or all of the Long Runs to a percentage of weekly mileage, often 25-50%. However, there does not appear to be any obvious rationale or support for this limit.
  • Sometime there is a concern that exceeding a given percentage of the weekly mileage is more likely to result in overtraining or injury, but I've found no supporting evidence or possible mechanism. On the contrary, there is evidence that more evenly distributed mileage is more likely to result in Overtraining, injury and reduce the benefit of training. This is because the more even the spread of training, the higher the Training Monotony (average weekly training stress divided by standard deviation).
  • Plans such as FIRST only have 3 runs per week, so the long run is a large portion of the weekly mileage.
• Caveats: Too little training between Long Runs can result in detraining.

4.13 Run Distance Rather Than Time

For marathon running, Long Runs SHOULD PROBABLY be based on distance not time.

• Rationale:
  • While a Long Run of a specific distance takes longer for a slower runner, the same is true of the marathon itself and in many ways racing a slower marathon is harder than faster one. A 20 mile run at 10 min/mile pace is different to the same distance at a 7 min/mile pace, but both runners are preparing for a fixed distance, not a fixed time.
  • When running by time, changes in pace have a compound impact on training stress. For instance, a 2 hour run at 7:30 min/mile pace would cover 16 miles, but running at 7:00 min/mile would cover 17 miles, compounding the faster pace with a greater distance. By one measure, the 7:00 pace could be around 20% harder than the 7:30 pace (assuming a 3 hour marathon runner and using Glycogen usage as a measure of difficulty).
• Caveats:
  • There is no evidence to suggest that there is an optimal time for endurance adaptations to occur, but further research could change this conclusion.
  • For timed races (6 hours, 12 hours, 24 hours, etc.) a timed long run may be appropriate as it mirrors the objective of the race.
  • Using Long Runs to prepare for specific events may be based on the timing of those events. For instance, preparing for a race that will be overnight may use Long Runs that start before dusk and terminate in the night, or even the next morning.

4.14 Plan Length for Novice Marathoners

First time marathoners SHOULD PROBABLY use 6 month or longer plans that increase their mileage more gradually.

• Rationale:
  • A novice marathon runner will take longer to adapt.
  • To build up the Long Run distance slowly, and have the flexibility for unexpected setbacks is likely to require more time.
• Caveats: The length of the plan depends on the prior experience of the runner. Someone who has completed several fast half marathons has a much better endurance base than someone new to running.

4.15 Plan Length for Slower Marathoners

Slower marathoners SHOULD PROBABLY use longer marathon plans.

• Rationale: While both fast and slow runners have to cover the same 26.2 mile distance, the stress on the slower runner is greater (assuming the same relative intensity). This will probably require a longer adaptation period.

4.16 Flexible Plans

Training plans SHOULD be flexible and individual.

• Rationale: Increasing the length of the Long Run should be based on how an individual responds and recovers. A runner that experiences too much muscle soreness after a long run should not continue to increase the distance further, but a runner who experiences no soreness may be able to increase the distance more rapidly.
• Caveats: It is common for runners to choose a race, then plan the training accordingly, making this flexibility difficult.

4.17 Frequency

Long Run frequency SHOULD be every 7-14 days.

• Rationale: The right level of muscle soreness is probably where recovery is complete in a few days.

4.18 Ultramarathon Run-Walk

Most ultramarathoners SHOULD practice a run-walk pattern.

• Rationale: The vast majority of ultramarathoners do not run the entire distance and walk some portion of the course. It is important to practice both walking and the transitions between running and walking, and walking and running. These transitions can produce unexpected stresses if not practiced.

4.19 Long Run Pace

The ideal pace for long runs is UNCLEAR, but almost all marathon runners train at slower than race pace for their long runs, while ultrarunners tend to train at a pace that is faster than race pace.

• Rationale:
  • Jack Daniels "Easy" pace is 59-74% of V̇O₂max or 65-78% of Maximum Heart Rate.
  • FIRST uses a number of seconds slower than marathon pace, such as MP+30.
  • Galloway uses a training pace of 2:00 min/mile slower than marathon pace.
  • It's unclear what paces Lydiard recommends. There are indications of quarter, half, and three-quarter effort, as well as running 160 Km/week (100 miles/week) at maximum steady state (maybe Lactate Threshold) plus the same again at a "jogging" pace.
• Caveats: Running at or faster than marathon race pace for some of a Long Run may be useful. The Jack Daniels programs use this approach.

4.20 Ultramarathon Long Runs

The ideal approach to Ultramarathon Long Runs is UNCLEAR, but it is common for Ultrarunners to do multiple shorter Long Runs than a single longer Long Run, and to run a much smaller fraction of the race distance in on their Long Runs than Marathoners, but at a much faster pace than race pace.

• Rationale:
  • The approach of using multiple Long Runs is due to practical limitations of finding the time for a longer Long Run.
  • The length of the Long Run for an ultramarathon may not need to be a large percentage of the target race because of the faster pace in training.
• Caveats:
  • Ultrarunners tend to race more frequently then marathon runners, and ultramarathon races can act as Long Runs for ultrarunners.

4.21 Fueling

Long runs SHOULD be fueled with a regular carbohydrate intake using Gels or drink, aiming to get about 60 grams of carbohydrate per hour.

• Rationale:
  • Taking carbohydrate will reduce muscle damage and spare glycogen.
  • It is valuable to practice race fueling.

4.22 Hills

Long Runs on hills are different to those on the flat as Downhill Running increases the muscle damage that occurs from running, producing more training stress for a given distance. Therefore the terrain of the Long Run SHOULD mimic that of the race.

• Rationale: There is good evidence that both Downhill Running and the marathon races produce Delayed Onset Muscle Soreness.
• Caveats:
  • Some of the Long Runs should be performed on a route that mimics the profile of the target race. This not only means that preparation for a hilly race should include Long Runs on a course with similar hills, but also that preparation for a flat race should include flat Long Runs.

4.23 Walking After a Long Run

A short walk to Cooldown immediately after a Long Run SHOULD PROBABLY be used to improve recovery.

• Rationale: There is anecdotal evidence that using a short (5-15 minute) walk after a Long Run reduces soreness and improves recovery.
• Caveats:
  • Post-race recovery nutrition should be during the walk rather than afterward.