A Practical Training Budget for Runners

A training budget is a simple way to express how much running you do in a week, adjusted for how hard that running is. Every minute you run is converted into easy-running minutes: a minute of genuinely easy running counts as one, and harder minutes count for more. The total gives you a single number you can compare with other runners, with your own past training, and with the training of elite marathon runners. A typical elite marathon training week works out at about 1,000 easy-running minutes.

The budget is deliberately rough. As the statistician George Box put it, "Essentially, all models are wrong, but some are useful". It is an accounting unit for comparing and planning training. It does not measure fitness, fatigue, recovery needs or injury risk, and a minute of threshold running does not literally "equal" any number of easy minutes.

The main points are:

• Why not just count distance or time? Distance flatters fast runners: an elite marathoner covers far more ground per hour than most runners, so weekly kilometres exaggerate the difference in how much running each person actually does. Time alone treats an hour of easy jogging the same as an hour of hard running.
• The rule. Easy running at about 70% of your maximum heart rate counts as 1 per minute. The weight of a minute doubles for every 15 percentage points of maximum heart rate above 70%, up to a maximum of 4. So a minute at 85% counts 2, a minute at threshold (about 90%) counts about 2.5, and a very easy minute at 55% counts 0.5 (see the weighting table).
• Easy running is not all the same. Running just below your first lactate threshold counts about 1.6 per minute, while a very easy jog counts about 0.6. Many runners do their "easy" running much closer to the first threshold than elite runners do, and three-zone summaries hide that difference.
• The elite reference. Typical elite marathon weeks come out at about 850–1,200 easy-running minutes, and peak weeks at about 1,250–1,400. Use 1,000 as the reference, treating it as accurate to roughly ±20% (see the evidence).
• Comparing yourself. Divide your weekly budget by 1,000. A runner doing 5 hours a week of genuinely easy running scores about 32%; the same 5 hours run at a "steady" effort scores about 44%. A percentage describes the scale of your training, not your fitness, and it is not a target.
• Planning. At a fixed budget, every minute of threshold running costs about 2.5 easy minutes and every minute of hard intervals about 3.5. The budget makes the trade-off between more running and harder running explicit.
• What it leaves out. Impact and mechanical load, session structure, long-run length, training history, heat, altitude and other sports all matter, and none of them is captured by a single weekly number (see What the budget cannot tell you).

Disclosure: this article was created using iterative research loops using AI, with AI writing up the final conclusions. ChatGPT and Claude were used.

Practical guide

Quick start

1. Find your maximum heart rate. Use the highest reliable value you have seen in a race or a hard interval session, not an age formula (see Estimating maximum heart rate).
2. Set your watch or app heart rate zones to 65, 75, 82, 87 and 92% of that maximum. Those edges give the six bands in the weighting table.
3. At the end of a week, note the minutes you ran in each band and multiply by 0.7, 1, 1.5, 2, 2.5 and 3.5.
4. Add them up. That is your training budget in easy-running minutes.
5. Divide by 1,000 to see it as a percentage of a typical elite marathon week, and round it: "about 40–45%" is more honest than "42.7%".

If you have no heart rate data, you can use pace relative to your threshold pace, or a simple three-zone summary (see Three ways to calculate).

Why not just count kilometres or hours?

Weekly distance is the usual way runners compare training, but it mixes two things: how much you run and how fast you run. An elite marathoner averaging about 4:00–4:30 per km in training covers 13–15 km in an hour; a runner averaging 6:30 per km covers about 9 km. When both run for 10 hours, the elite runner logs roughly 45–65% more distance for the same amount of running.

Time solves that problem but creates another: an hour of very easy running and an hour of hard threshold running are not the same training. The training budget keeps time as the basic unit and adds a weighting for intensity.

The reference: easy running

The reference intensity, counting as 1 per minute, is genuinely easy, conversational running. Any of these describes it:

• about 70% of maximum heart rate (roughly 65–75%);
• about 75% of your threshold speed, which means a pace about 1.33 times your threshold pace (if your threshold pace is 5:00 per km, easy is about 6:40 per km);
• an effort of 2–3 out of 10, at which you can talk in full sentences.

These three descriptions agree only approximately, so use the one you can measure most reliably (practical approximation). Your threshold pace is the pace you could race for about an hour: roughly half-marathon pace for faster runners, and closer to 10 km pace for slower runners.

This reference was chosen because it is where elite easy running actually sits. In the elite example below, the athlete's easy runs averaged about 69% of her estimated maximum heart rate. It also lies in the lower part of the 65–82% of maximum heart rate that Norwegian elite runners used for their continuous easy running. Choosing a different reference would multiply every budget by the same factor and leave every comparison unchanged.

The weighting rule

For any minute of running:

weight = 2^{(%HRmax − 70) ÷ 15}, capped at 4

In words, the weight doubles for every 15 percentage points of maximum heart rate above 70%, and halves for every 15 points below it. This rule is a simplified, rounded version of the heart-rate weighting used in Banister's TRIMP, which is based on how blood lactate rises with heart rate (see The doubling rule as a simplified TRIMP).

Without heart rate data, use speed relative to threshold. Write S = threshold pace ÷ run pace, with both in minutes per km, so S = 0.75 means you are running at 75% of your threshold speed. Then:

weight = 2^{(S − 0.75) ÷ 0.19}, capped at 4

The pace version is calibrated to match the heart-rate version at easy and threshold intensities, and it rests on the assumption that easy running at 70% of maximum heart rate is about 75% of threshold speed. Treat it as a practical approximation.

The weighting table

Most runners will find the table easier than the formula. The weights are the formula's values at the middle of each band, rounded.

Band	% of max HR	% of threshold speed	Effort (0–10) and talk test	Weight per minute
Recovery	under 65	under 70	1–2, effortless	0.7
Easy (reference)	65–75	70–80	2–3, full sentences	1.0
Steady, up to LT1	75–82	80–88	4, sentences with pauses for breath	1.5
Marathon pace, moderate	82–87	88–95	5, short phrases	2.0
Threshold, up to LT2	87–92	95–102	6–7, a few words	2.5
High intensity, above LT2	over 92	over 102	8–10, no talking	3.5

LT1 and LT2 are the first and second lactate thresholds. The band edges are approximations: the lactate thresholds of individual runners fall at different percentages of maximum heart rate, so if you know your own thresholds, set the steady band to end at LT1 and the threshold band to end at LT2.

Setting up your watch zones

Most watches and training apps report time in heart rate zones each week. Set five zone boundaries at 65, 75, 82, 87 and 92% of your maximum heart rate, and the app's weekly zone totals become the minutes you need for the table. Zones based on heart rate reserve or on a lactate threshold heart rate will not line up with these bands, so either change them or convert between them.

Three ways to calculate

The same weights work at three levels of detail, and the results usually agree within about 10%.

Detailed

Apply the formula to each heart rate sample, or to each lap's average heart rate, and add up the results. Spreadsheets and analysis software can do this automatically. Heart rate lags behind effort in short efforts, so for intervals shorter than about 3 minutes, weight the work periods by pace instead.

Intermediate

Take the minutes spent in each band of the table and multiply each by its weight. This is the recommended default: it needs only a weekly time-in-zone summary, and it lands within a few percent of the detailed method.

Basic

If all you have is a three-zone summary (below LT1, between LT1 and LT2, above LT2), use these weights:

• Zone 1: 1.0 if your easy running is genuinely conversational, or 1.3 if most of it is "steady" running close to LT1. Be honest here: this choice can change the result by 30%.
• Zone 2: 2.2
• Zone 3: 3.5

In the elite example below, Zone 1 running averaged a weight of 1.08 because most of it was easy and some was very easy.

Estimating maximum heart rate

Maximum heart rate is the single largest source of error. A 5 bpm error in maximum heart rate changes the budget by about 9%, because every minute's weight depends on it. Age-based formulas can be wrong by more than that for individuals (see Maximum Heart Rate).

• Use the highest reliable value from a race, a hill effort or the end of a hard interval session.
• Reject readings that are obviously artefacts. Optical sensors in particular can lock on to running cadence and report a "heart rate" of 170–200 that is really your step rate. A sudden jump of 50–70 bpm in a few seconds, or a reading that matches your cadence, is a warning sign.
• Your maximum may fall slowly with age, and it is slightly lower at altitude, so recheck it every few years.

Spreadsheet formulas

Assuming heart rate in column A, maximum heart rate in B and minutes in C:

• Per run or lap, from heart rate: =C2*MIN(4,2^((A2/B2*100-70)/15))
• Per run or lap, from pace (threshold pace and run pace in decimal minutes per km): =C2*MIN(4,2^((Threshold/Pace-0.75)/0.19))
• From weekly band minutes (recovery to high intensity in D2:I2): =SUMPRODUCT(D2:I2,{0.7,1,1.5,2,2.5,3.5})
• As a fraction of the elite reference: =Budget/1000

Using a run's average heart rate instead of second-by-second data slightly underestimates runs with large swings in effort, but for steady runs the difference is about 1%.

Worked examples

An elite marathoner's peak week

This week comes from the publicly visible training records of an elite female marathoner (sub-2:30), recorded during her peak week before an autumn marathon. Most of it was run at altitude. The calculation used second-by-second heart rate for most runs and lap averages for the rest, with an estimated maximum heart rate of 185 bpm. Lactate measurements from one of her threshold sessions put her second threshold at about 168–172 bpm, roughly 91–93% of that maximum, which is consistent with the band edges above.

Band	Minutes	Weight	Easy-running minutes
Recovery	55	0.7	39
Easy	434	1.0	434
Steady	143	1.5	215
Marathon pace	80	2.0	160
Threshold	114	2.5	285
High	6	3.5	21
Total	832 (about 14 hours)	1.39 average	1,154

The continuous formula gives 1,123 and the basic three-zone method 1,070–1,130, so all three methods agree within 8%. If her true maximum heart rate were 181 or 190 bpm, the budget would be about 1,220 or 1,020. The week covered about 185 km. Including her longest run of the build in an adjacent seven-day window gives about 1,250–1,300.

A threshold session

A 60-minute session made up of 15 minutes easy, 4 × 8 minutes at threshold with 2-minute jog recoveries, and 7 minutes easy:

• 15 minutes easy: 15
• 32 minutes at threshold: 32 × 2.5 = 80
• 6 minutes of recovery jogging: 6 × 0.7 = 4
• 7 minutes easy: 7
• Total: 106 easy-running minutes from 60 minutes of running.

A recreational runner whose easy running is steady

A runner who runs 5 hours a week, almost all of it "easy":

• if the easy running is genuinely at about 70% of maximum heart rate, with 5% of the time at 85%, the budget is about 315 (32% of the reference);
• if the same 5 hours are run at a steady 78% of maximum heart rate, it is about 440 (44%).

A three-zone method scores both weeks the same, because both sit in Zone 1. The difference is the main reason the training budget uses a continuous weighting.

Same time, different distances

Two runners each run 8 hours a week, 90% easy and 10% at threshold. One covers 120 km and the other 80 km. Both have a budget of about 550 (55%). Their kilometres differ by 50%; their budgets do not. Whether the faster runner's extra distance carries extra mechanical cost is a separate question (see What the budget cannot tell you).

The author's ultrarunning peak weeks

These are two representative peak-volume weeks from the site author's training for competitive ultrarunning, calculated from Strava records with a maximum heart rate of 190 bpm. The author established that value from interval training at the time. Only running activities are included. The calculation used each run's average heart rate, which is close to exact for these steady runs. Moving times are as recorded and include brief stops on some long runs. Distances are the author's corrected values where he recorded a correction; these were 1–11% longer than the GPS figures.

Week beginning	Runs	Running time	Distance	Typical heart rate	Budget	% of reference
Monday 14 October 2013	4 runs of 41–43 km	14 h 0 min	about 171 km	75–82% of max	about 1,215 (1,130–1,300)	113–130%
Monday 16 June 2014	13 runs of 5–34 km	16 h 15 min	about 174 km	mostly 61–68% of max	about 790	79%

• October 2013 consisted of three outdoor runs of 40–43 km at average heart rates of 143–155 bpm (75–82% of maximum), with weights of 1.29–1.71. The fourth run was 42 km on a treadmill at a −12% gradient, a deliberately damaging downhill session, recorded without heart rate. It was given an assumed weight of 1.5, and the range in the table assumes 1.0–2.0 instead.
• June 2014 had more running time, but almost all of it was at 61–68% of maximum heart rate. The week's 975 minutes were 582 recovery, 337 easy, 54 steady and about 2 high-intensity minutes. The high-intensity minutes came from short fast efforts in one run, which were estimated from speed because the heart rate signal stuck at a constant value partway through.

The two weeks look almost identical by distance: 171 and 174 km, about 90% of a typical elite marathon week of around 190 km. By time, both are above the elite range, at 14 and 16 hours. By training budget they are very different: the October week scores about 1.5 times the June week, despite having 2¼ fewer hours of running.

The comparison also shows the budget's limits. The October week contained four runs of marathon distance or more in five days, one of them a 42 km downhill treadmill run intended to cause muscle damage, and the budget does not capture that eccentric and mechanical load. In the June week, the author's notes describe recovering from three 100-mile races. Heart rate can be lower than usual for a given effort when someone is fatigued, so a heart-rate-based budget may understate how demanding a week like this was.

Comparing yourself with the elite reference

Divide your weekly budget by 1,000 and treat the result as a rough description of the scale of your training:

• A runner at 40% of the reference is training at roughly 40% of the scale of a typical elite marathoner. That says nothing about their fitness, capacity or ability to recover.
• Elite athletes are not necessarily training at their maximum sustainable load, and their training is not necessarily optimal. The reference is simply a well-documented high-volume programme.
• Use several weeks, or a rolling seven-day total, rather than a single week. Individual weeks vary a lot; in one case study, weekly training load varied more than twofold over a build.
• A percentage is not a target. A budget that your training history has not prepared you for is a large jump, however it is counted.

Using the budget to plan

Budgets at 25%, 50% and 75%

A week shaped like an elite marathon programme costs about 1.22 easy-running minutes for every minute run. The table shows how such a week scales at different budgets. These are illustrations of the arithmetic, not recommended loads.

Budget	% of reference	Running time	Recovery	Easy	Steady	Marathon pace	Threshold	High intensity
250	25%	3 h 24 min	20	133	20	16	10	4
500	50%	6 h 48 min	41	265	41	33	20	8
750	75%	10 h 12 min	61	398	61	49	31	12
1,000	100%	13 h 36 min	82	531	82	65	41	16

The intensity columns are minutes per week. The shape is 10% recovery, 65% easy, 10% steady, 8% marathon pace, 5% threshold and 2% high intensity, which is close to the elite intensity distributions summarised in the research section. Peak weeks often contain more threshold work: the elite example above had about 14% of its time at threshold.

Trade-offs at a fixed budget

At a budget of 500:

• All easy: 500 minutes (8 h 20 min) of running.
• Threshold-focused: two sessions of 25 minutes at threshold (125 easy-running minutes) plus 375 easy minutes, giving 7 h 5 min of running.
• Interval-focused: 30 minutes above LT2 (105) plus 20 minutes at threshold (50) plus 345 easy minutes, giving 6 h 35 min of running.

As a rule of thumb, 20 minutes of threshold running "uses" 30 more minutes of the budget than 20 minutes of easy running would.

Scaling a programme up or down

Multiplying every session's duration by the same factor keeps the shape of a programme, and the budget scales with it. Three things break simple pro rata scaling:

• Minimum useful sessions. A threshold session needs perhaps 15–20 minutes of work to be worth doing. At 25% of the reference, a pro rata elite week leaves only about 10 minutes of threshold, so it is better to do one real session or lengthen it and cut some easy running.
• Long runs. A 2.5-hour elite long run scales to about 40 minutes at 25%. A runner preparing for a marathon at that level still needs a longer run, so the long run will take a larger share of a smaller budget.
• Short easy runs. At low budgets, short recovery runs may be better replaced by rest days (see Recovery Runs).

What the budget cannot tell you

A single weekly number cannot capture everything that makes training demanding or appropriate. Treat these as things to weigh alongside the budget, not as reasons to abandon it:

• Training history and current fitness. A budget that is normal for one runner may be a large increase for another.
• Injury history and tissue tolerance. The budget is cardiovascular. Impact, eccentric loading and tendon load depend on distance, speed, terrain and body mass, and are not weighted.
• Age and recovery capacity.
• Session structure and long-run length. One 2-hour run and four 30-minute runs score the same but are different training.
• How the week is spread. The same budget in four days loads you differently from the same budget in seven.
• Heat, altitude, hills and illness. These raise heart rate at a given pace. A heart-rate-based budget counts that as extra load, which is often appropriate; a pace-based budget does not.
• Other sports. Keep the budget for running only. Cycling, swimming or strength training affect your recovery, but there is no sound basis for converting them into running minutes; note them separately.
• Fatigue. Heart rate can be suppressed when you are very tired, which makes hard weeks look easier than they are. Combine the budget with the checks in Training Readiness.

The science and supporting research

How to read this section

The training budget is a practical tool, not a validated physiological measurement, and most of its parts rest on evidence of modest quality. Claims are labelled where it helps:

• Strong evidence: supported by substantial, consistent research.
• Limited evidence: some supporting research, but important uncertainty remains.
• Consistent elite practice: common among successful elite athletes, without necessarily being experimentally validated.
• Practical approximation: a deliberate simplification chosen for usefulness.
• Personal hypothesis: a plausible interpretation not established by the evidence.

How much should a minute count?

Several established methods weight training time by intensity. None was designed to answer exactly this question, but together they give a reasonable basis.

Heart rate and lactate: Banister's TRIMP

Banister's training impulse (TRIMP) multiplies duration by the fraction of heart rate reserve and by an exponential weighting factor. The commonly used constants are "A = 0.64 for men and 0.86 for women, B = 1.92 for men and 1.67 for women", so that each minute is weighted by HRR × A × e^{B × HRR}, where HRR is the fraction of heart rate reserve. The exponential term reflects the expected relationship between heart rate and blood lactate, and the method is usually traced to Banister's modelling work with Morton and Fitz-Clarke. Once the weights are normalised to easy running, the constant A cancels out and only the exponent matters (limited evidence: the weighting comes from group lactate data, not from each runner's own curve).

Zone-based methods: Edwards and Lucia

Edwards' method multiplies the time spent in five heart rate zones, "based on 50–60%, 60–70%, 70–80%, 80–90%, and 90–100% of HRpeak", by multipliers of 1 to 5. The multipliers are simple and linear, which has two consequences: weight rises much more slowly at high intensity than in Banister's method and stops rising at 90% of maximum, and each 10-point band treats very easy and fairly steady running as identical.

Lucia's TRIMP divides heart rate into three phases using the ventilatory threshold and respiratory compensation point, and multiplies the time in each phase by a phase multiplier. The multipliers are usually given as 1, 2 and 3; the full paper was not accessible for this article, so those values were not confirmed against it. Like any three-zone method, it cannot distinguish easy from steady running within Zone 1.

Session RPE

In the session RPE method, training load is "computed by multiplying the duration of the exercise bout by the session RPE for that bout". It needs no device, and a review found that it generally correlates well with heart-rate-based methods, with reported correlations of roughly 0.5 to 0.97 against Edwards' and Banister's TRIMPs across sports, though some intermittent sports showed much weaker relationships (strong evidence for its practicality). Because it rates whole sessions, warm-ups and cool-downs dilute the hard parts. It is still a useful cross-check: an easy run rated 2–3 out of 10 and a threshold session rated 6–7 imply a ratio of about 2.5, in line with the weighting table.

Pace-based stress scores

Running Training Stress Score (rTSS), developed for the commercial TrainingPeaks platform, combines "the elements of the normalized graded pace ... relative to the functional threshold pace ... in conjunction with the duration of the workout". It is industry-developed and not described in peer-reviewed literature in a form that could be checked for this article. The training budget's pace formula is therefore calibrated to the heart rate rule rather than to rTSS.

Recovery after easy and hard running

In trained athletes, "When HT ran 60 or 120 min below VT1, HRV returned to pretraining values within 5-10 min". Training at threshold or above VT2 delayed autonomic recovery much more. Doubling the duration of easy running had little effect, while running at or above threshold had a clear one. This supports a weighting that rises steeply from LT1 upwards, and counting long easy running at close to one per minute (limited evidence: a small study using one recovery marker).

The doubling rule as a simplified TRIMP

Normalised so that 70% of maximum heart rate counts as 1, the methods compare as follows. The Banister values assume a resting heart rate of 50 and a maximum of 185.

% of max HR	Doubling rule	Banister, men's exponent	Banister, women's exponent	Edwards (÷ 2.5)
55	0.50	0.44	0.46	0.4
60	0.63	0.59	0.61	0.8
65	0.79	0.77	0.79	0.8
70	1.00	1.00	1.00	1.2
75	1.26	1.27	1.25	1.2
80	1.59	1.60	1.55	1.6
85	2.00	2.00	1.90	1.6
90	2.52	2.48	2.32	2.0
95	3.17	3.05	2.80	2.0
100	4.00	3.74	3.37	2.0

The doubling rule tracks the men's Banister curve within 0.12 from 55% to 95% of maximum heart rate. It sits about 8% above the women's curve at 90%, a difference that is small compared with typical errors in maximum heart rate. So the rule is best understood as Banister's weighting, normalised to easy running and rounded into a form that can be applied without a calculator (practical approximation).

Why weight differences within Zone 1?

Three-zone models put everything below LT1 into one zone, but that zone spans a wide range of effort, from walking pace jogs to steady running just below the first threshold. On a lactate-anchored five-zone scale, zone 1 is 60–72% of maximum heart rate at under 1.0 mmol/L of lactate, and zone 2 is 73–82% at 1.0–2.0 mmol/L. The doubling rule weights that whole span from about 0.5 to 1.6, a more than threefold range.

This matters for comparisons. In descriptive studies of well-trained endurance athletes, "about 80% of training sessions are performed at low intensity", and successful marathoners do most of their running genuinely easy (consistent elite practice), while many recreational runners do their "easy" running close to LT1. In the elite example above, 489 of 632 minutes in Zone 1 were recovery or easy running, averaging a weight of 1.08. A recreational runner whose Zone 1 running averages 78% of maximum heart rate has a Zone 1 weight of about 1.4. Treating both as 1.0 would understate the recreational runner's budget by about 30%.

Finer resolution than the six bands adds little. In the elite example, moving from six bands to the continuous formula changed the total by 3%.

How much training do elite marathoners do?

Few studies report elite marathon training in hours, and fewer report time spent at each intensity. The estimates below therefore combine reported weekly distances with assumed average training paces and intensity splits. Only the first row was measured directly as time and heart rate.

Source	Athletes	km per week	Running hours	Intensity split used (Z1/Z2/Z3)	Average weight	Easy-running minutes
Elite example (this article)	1 elite woman, sub-2:30, peak week, mostly at altitude	about 185	14.2, measured	heart rate	1.35	1,120 (1,020–1,220)
Haugen et al. 2022	World-class marathoners, mid-preparation	160–220	10.7–14.7, estimated	80/15/5, assumed	1.32–1.35	860–1,160
Casado et al. 2022	4 elite marathoners (2 men, 2 women), two pre-competition weeks	195	13.3, estimated	76/16/8 by time, reported	1.44	1,145
Stellingwerff 2012	3 elite men, 2:11–2:16	182 (peak 231)	11.9 (15.1), estimated	74/11/15 by sessions, reported; 80/11/9 by time, assumed	1.40	1,000 (peak 1,265)
Billat et al. 2001	5 top-class women	166	12.0, estimated	80/12/8, assumed	1.38	995
Enoksen et al. 2011	3 Norwegian elite marathoners, preparation period	187	12.8, estimated	80/20 by distance, reported; 83/17 by time	1.25	950
Karp 2007	US Olympic Marathon Trials qualifiers	men 145, women 116	10.3 / 9.3, estimated	80/15/5, assumed	1.35	830 / 755
Muniz-Pumares et al. 2024	620 Strava users finishing in 2:00–2:30	about 107	8.0, estimated	62/23/15, assumed from reported Zone 1 of 57–67%	1.68	810

How the estimates were built

Running hours are weekly distance divided by an assumed average training speed. The assumed paces are 3:55–4:05 per km for elite men, 4:20 per km for elite women, 4:50 per km for the Trials women and 4:30 per km for the sub-2:30 group (practical approximation). The elite example's measured average was 4:34 per km, but altitude slows pace at a given effort. The basic-method weights (1.05, 2.2 and 3.5) were then applied to each intensity split.

The main sources of uncertainty are:

• Pace assumptions. Each 15 s per km of error in the assumed average pace changes an estimated row by about 6%.
• Different intensity methods. Studies report intensity by heart rate time, by session goal or by distance, and these give different splits. Using an 80/15/5 split instead of the 76/16/8 reported by Casado and colleagues lowers that row by about 6%.
• Small samples. Most elite studies describe 3–10 athletes, mostly from the 2000s and 2010s.

The figures cluster: typical elite weeks come out at about 950–1,150, and the wider range of 850–1,200 covers the world-class datasets. The reference of 1,000 is a round working figure inside that cluster. Volumes in the cited sources support it: Haugen and colleagues report that "Typical weekly running volume in the mid-preparation period is ~ 160–220 km for marathon runners", and Karp's Trials qualifiers ran "145.3 ± 25.6 vs 116.0 ± 26.5 km/wk" (men vs women). The Trials qualifiers were a broader, mostly sub-elite group, and they fall below the elite cluster, as expected.

Typical and peak weeks

Peak weeks are roughly 10–25% higher than typical weeks. The athletes in Stellingwerff's case study averaged 182 km per week and peaked at 231 km, and the elite example's highest seven-day window scores about 14% more than the week shown.

Men and women

Running hours, and therefore budgets, appear broadly similar. Elite women in these datasets run somewhat fewer kilometres (116–187 km per week) at slower average speeds. In a large dataset of amateur marathoners, "for a given marathon finishing time, female runners accumulated lower training volume compared to males" (limited evidence for elite athletes).

Does the model give sensible results?

Test programmes

The hypothetical programmes below were defined by percentage of maximum heart rate, using a maximum of 185 and a resting heart rate of 50. Every method is normalised so that 70% of maximum counts as 1.

Programme	Hours	Doubling rule	Bands	Basic 3-zone	Banister	Edwards	Lucia (1/2/3)	% of reference
A. Elite, high volume, 80% of time under 75% of max HR	13	942	956	974	938	958	913	94%
B. Elite, threshold-heavy (22% at threshold)	12	984	1,008	991	975	916	922	98%
C. Low volume, interval-heavy (25% above LT2)	4	432	456	440	424	350	396	43%
D1. Recreational, "easy" running at a steady 78% of max HR	5	442	458	404	447	366	315	44%
D2. Recreational, easy running at a genuine 70% of max HR	5	315	315	332	315	366	315	32%
E1. 8 h, 90% easy / 10% threshold, covering 120 km	8	547	552	559	546	595	528	55%
E2. The same time and intensity, covering 80 km	8	547	552	559	546	595	528	55%

• Volume dominates. The two elite programmes score about twice the 4–5 hour programmes, even though programme C has nearly six times programme A's share of hard running.
• Threshold work is not free. Programme B has an hour less running than A but scores slightly higher.
• Differences within Zone 1 matter. D1 scores 40% more than D2 for the same 5 hours. Lucia's three-zone TRIMP cannot see the difference, and neither can Edwards' 10-point bands, because 70% and 78% fall in the same band.
• Edwards undervalues hard intervals. Its weight stops rising at 90% of maximum, so programme C scores about 19% below the other methods.
• Distance is neutral by design (E1 = E2).
• The basic method lands within about 10% of the detailed rule when the Zone 1 weight is chosen honestly (1.3 for D1, 1.05 for the others).

Sensitivity

Changing how quickly the weights rise changes budgets, but comparisons between programmes hardly move.

Doubling interval	Weight at 90% of max HR	Elite example	Programme C as % of reference	Programme D2 as % of reference
12.5 points (about 20% steeper at threshold)	3.0	1,215	46%	30%
15 points (recommended)	2.5	1,123	43%	32%
18 points (about 15% flatter at threshold)	2.2	1,056	41%	33%

The reference was rescaled along with the weights (to 1,081 and 941). A moderate change in weighting shifts a runner's percentage of the elite reference by 3 points or less; an error of 5 bpm in maximum heart rate matters more.

Limitations

• The weights are not physiological equivalences. Banister's curve describes how lactate rises with heart rate. It does not show that a minute at threshold is equivalent to 2.5 easy minutes of fatigue, adaptation or injury risk.
• Heart rate misbehaves. It lags in short repetitions, drifts upwards on long runs and in heat (see Heart Rate Drift), rises at altitude (see Altitude Training), can be suppressed by heavy fatigue, and optical sensors can report cadence instead of heart rate.
• The pace and heart rate versions are only roughly aligned. For some runners they may differ by 10–20%.
• The elite reference is partly estimated. Only the elite example and one study provide time-in-zone data; the rest needed assumed paces and splits.
• Mechanical load is ignored. Two runners with equal budgets can have very different impact and eccentric loads, as the author's October 2013 week illustrates.
• Weights above LT2 are the least certain, so comparisons between interval-heavy track programmes and marathon programmes are the least reliable.
• The budget has not been validated against performance, fatigue or injury outcomes. Its usefulness rests on plausibility and consistency with established load methods, not on outcome studies.

Evidence table

Claim	Basis	Evidence label
Elite marathoners run about 160–220 km per week in mid-preparation	Haugen et al. 2022; Casado et al. 2022; Enoksen et al. 2011; Billat et al. 2001; Stellingwerff 2012	Consistent elite practice
Elite marathon training is mostly low intensity, with a pyramidal distribution	Casado et al. 2022; Enoksen et al. 2011; Haugen et al. 2022; Seiler 2010	Consistent elite practice
Heart rate weighting should rise exponentially with intensity	Banister's TRIMP and its lactate basis	Limited evidence
Easy running below VT1 causes little autonomic disturbance, even when long	Seiler et al. 2007	Limited evidence
Session RPE correlates with heart-rate-based TRIMPs	Haddad et al. 2017	Strong evidence
The doubling rule and its 1,000-minute elite reference	Derived in this article from the sources above	Practical approximation
Pace-based weights match heart-rate weights	Calibrated assumption	Practical approximation

Confidence summary

• Elite reference of 1,000 easy-running minutes (±20%): moderate confidence. The distances are well documented; the hours and intensity splits are mostly estimated.
• Doubling rule as a weighting: moderate-to-high confidence that it ranks programmes sensibly, and high confidence that it reproduces a normalised Banister TRIMP. Low confidence that the weights represent true physiological cost.
• Differences within Zone 1: moderate confidence that weighting them is worthwhile, based on lactate-anchored zones and on how differently elite and recreational runners use Zone 1.

Unresolved questions

• Does a training budget predict anything useful, such as performance changes, illness or injury, better than distance or time?
• What are the right weights above LT2, where heart rate is least reliable?
• How closely do the heart rate and pace versions agree for runners of different abilities?
• Should walking within runs, run–walk, and hiking in trail and ultra races count, and at what weight?
• How should altitude and heat be handled when comparing weeks?
• Could a separate mechanical-load budget, based on distance, speed and terrain, sit alongside this cardiovascular one?

See also

• TRIMP
• Training Readiness
• Recovery Runs
• Tempo Runs
• Toxic Miles
• Pace Range Ratio
• Lactate Threshold
• Maximum Heart Rate
• Heart Rate Drift
• Heart Rate Variability
• Modeling Human Performance
• How Often To Run