There is not one Borg scale but two, and they do not share a metric: the 6-20 is described by a linear regression of workload, the 0-10 by a power function. No conversion from one to the other is legitimate, and '15 out of 20' is not '7 out of 10'. The tool below therefore asks which one you are using.

Both scales are registered trademarks under licence, whose agreement forbids modifying the verbal anchors. The detail is below.

Figures taken from the studies listed at the end of this page, every value carrying its source where it is written. The scales themselves are not reproduced.

Two scales, and not one

Under the name 'Borg scale' circulate two distinct instruments, from the same hand. The Borg RPE, for rating of perceived exertion, rates perceived exertion across fifteen whole steps, from 6 to 20. The Borg CR10, for category-ratio, rates on a scale of 0 to 10, and it is the one the respiratory literature calls the 'modified Borg scale'. Both are single-item scales: one value is recorded per measurement, with no subscale and no total.

Here is why confusing them is costly. A direct comparison of the two, conducted in the Borg family's own laboratory, shows that their data are not described by the same metric: those of the 6-20 by a linear regression of workload, with individual correlations around 0.98, those of the 0-10 by a power function with an exponent of about 1.2.

Two caveats accompany that result, and must be carried with it. The same abstract states that the 0-10 data could also be described by linear regressions, the power function being the description yielding the reported exponent. And the often quoted exponent of about 1.4 holds for both category-ratio scales taken together, not for the 0-10 alone.

The essential point holds despite these caveats: two metrics that do not superimpose, and therefore no legitimate conversion formula. '15 out of 20' is not '7 out of 10', and the tool at the top of this page refuses to convert for that reason.

A third scale from the same hand exists, the Borg centiMax, rated out of 100. It is not the subject of this page, but it explains part of the figures one meets: when a source speaks of an exponent for 'both category-ratio scales', it means the 0-10 and that one.

What the rating is really worth against physiology

This instrument's reputation is that of a reliable proxy for physiological load. The figures are more modest, and it is better to know them before basing a prescription on them.

The reference meta-analysis of criterion validity gives weighted mean coefficients of 0.62 for heart rate, 0.57 for blood lactate, 0.64 for percentage of maximal oxygen uptake and 0.72 for respiration rate. Its authors write explicitly that validity may not be as high as previously thought, the usually quoted interval being 0.80 to 0.90.

The very large cohort of 2,560 subjects finds 0.74 with heart rate and 0.83 with blood lactate. The two works do not contradict each other term by term: the 0.74 stays below the interval the meta-analysis declares overestimated. The workable reading of that gap is simple, and the only one this page defends: the rating is an excellent intensity marker at group level and a poor individual proxy for heart rate.

One rule of thumb circulates, finally, 'rating times ten equals heart rate'. Both studies from the Borg laboratory observe a tendency to depart from linearity between ratings and heart rates, and conclude to the need to standardise test protocols. The relationship depends on the protocol; it is not a constant.

On the respiratory side, the reference everyone cites for the validity of the 0-10 scale in dyspnoea deserves to be read in full. Its favourable conclusion rests on correlations of -0.31 in asthma, in 42 patients, and -0.42 in chronic obstructive pulmonary disease, in 60, between change in score and change in peak expiratory flow. That is far weaker than its reputation.

The physiological correspondences, and their standard deviations

The only usable numerical thresholds in this file concern the 6-20 scale, and they come from a cohort of 2,560 subjects of median age 28, undergoing incremental exercise tests.

The lactate threshold corresponds there to a rating of 10.8, the individual anaerobic threshold to 13.6, a fixed blood lactate of 3 mmol/L to 12.8 and of 4 mmol/L to 14.1. The same authors derive prescription ranges: 11 to 13 in an untrained subject, 13 to 15 for more intense but still aerobic work.

These values must be read with their standard deviations, which are 1.8 to 2.1 points. On a fifteen-step scale, two points of standard deviation cover a large part of the useful training zone. These correspondences therefore describe a group, not the patient pedalling in front of you.

For the 0-10 scale, this file publishes no physiological correspondence. The only often quoted value, a minimal difference of one point, comes from work on dyspnoea, by distribution method across pulmonary rehabilitation trials, whose abstract gives neither the trials nor the sample sizes and does not even say which of the two Borg scales it means. It does not hold as a change threshold for perceived exertion during exercise.

One last property of the 0-10, almost never said: its upper end is open. The abstract carrying its bounds does speak of a scale rated 0 to 10, but the publisher's official demonstration accepts ratings beyond. A rating above 10 is therefore not a recording error.

Two registered trademarks, under licence, with no validated French version

Here is a point almost nobody knows in France, with immediate practical consequences: both scales are registered trademarks, whose use goes through a licence granted by a Swedish company, exclusive rights holder for their author.

The licence agreement was read live. It grants a limited, non-exclusive and non-transferable right of use, with no transfer of ownership. It expressly forbids modifying the scale, preparing derivative works from it, distributing or sublicensing it, and removing its proprietary notices. On publication it imposes a literal copyright line, different depending on the scale used, followed by the phrase 'Scale printed with permission'.

One piece of news, on the other hand, and it is good for teachers and researchers: the academic and educational licence page today carries, in full, 'There is no fee for this license.', and French is among the languages offered. The licence still has to be requested, and its conditions remain those described above, but it is no longer paid for. That is a change from what earlier versions of the site carried.

What remains essential for a French-speaking reader: no validated French version is indexed. Five distinct search formulations were run against PubMed, and the closest result validates another instrument. Only two languages have a published validation, Mandarin Chinese for the 6-20 scale and Brazilian Portuguese for a 0-10 adapted to vocal effort. French therefore exists as a licensed translation, never as a validated version.

Three administration pitfalls

Converting from one scale to the other

The 6-20 behaves linearly with workload, the 0-10 is described by a power function. The two metrics do not superimpose, and no conversion formula is legitimate.

Applying the rule 'rating times ten equals heart rate'

Both studies from the originating laboratory observe a tendency to depart from linearity between rating and heart rate, and conclude to the need to standardise protocols. The relationship depends on the protocol.

Modifying the verbal anchors to 'adapt' the scale

The licence agreement expressly forbids it, as it forbids preparing derivatives or distributing it. A retouched scale is no longer the instrument the figures on this page come from.

Frequently asked questions

Is there one Borg scale or two?

Two: the 6-20, which rates perceived exertion, and the 0-10, called modified, used mostly for dyspnoea. They do not share a metric and do not convert into one another.

What physiological intensity does a rating correspond to?

On the 6-20 and in 2,560 subjects, the lactate threshold corresponds to 10.8 and the individual anaerobic threshold to 13.6, with standard deviations of nearly two points.

What intensity should be prescribed with the Borg scale?

The authors of that cohort recommend 11 to 13 in an untrained subject and 13 to 15 for more intense but still aerobic work.

Does one point of change count?

That figure comes from work on dyspnoea whose abstract does not say which of the two scales it means. It does not hold as a change threshold for perceived exertion during exercise.

Is the Borg scale free of rights?

No. These are registered trademarks under licence, whose agreement forbids modifying the scale, preparing derivatives or distributing it. The academic licence is today free of charge, but it still has to be requested.

References

9 sources, PMIDs included

1. Borg G. Perceived exertion as an indicator of somatic stress. Scand J Rehabil Med 1970;2(2):92-8. PMID 5523831. The original publication of the 6 to 20 scale. A reading caveat: PubMed indexes no abstract for this article, and this page therefore attributes no figure to it. It is however the year the publisher still carries today in the copyright line imposed on its licensees.
2. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc 1982;14(5):377-81. PMID 7154893. The original publication of category-ratio scaling, presented in its abstract as a new method combining the category method with ratio properties. A caveat: the abstract never uses the acronym CR10; identifying that method with the CR10 is the literature's and the publisher's doing. It does state the founding distinction: ratio scaling describes perceptual variation well, category methods serve better to compare individuals with one another.
3. Borg E, Kaijser L. A comparison between three rating scales for perceived exertion and two different work tests. Scand J Med Sci Sports 2006;16(1):57-69. PMID 16430682. The direct comparison of the three scales, conducted in the Borg family's own laboratory. It establishes that the two are not described by the same metric: the 6-20 by a linear regression of workload, individual correlations around 0.98, the CR10 by a power function with an exponent of about 1.2. Two caveats not to skip: the abstract states that CR10 data could also be described by linear regressions, and the exponent of about 1.4 in the second study holds for both category-ratio scales taken together. Both studies also observe a tendency to depart from linearity between ratings and heart rates.
4. Chen MJ, Fan X, Moe ST. Criterion-related validity of the Borg ratings of perceived exertion scale in healthy individuals: a meta-analysis. J Sports Sci 2002;20(11):873-99. PMID 12430990. The meta-analysis of criterion validity, and its figures sit lower than the instrument's reputation. The weighted mean coefficients are 0.62 for heart rate, 0.57 for blood lactate, 0.64 for percentage of maximal oxygen uptake, 0.63 for oxygen uptake, 0.61 for ventilation and 0.72 for respiration rate. Its authors conclude that validity may not be as high as previously thought, the usually quoted interval being 0.80 to 0.90. The abstract does not give the number of studies aggregated.
5. Scherr J, Wolfarth B, Christle JW, Pressler A, Wagenpfeil S, Halle M. Associations between Borg's rating of perceived exertion and physiological measures of exercise intensity. Eur J Appl Physiol 2013;113(1):147-55. PMID 22615009. The very large cohort, 2,560 subjects of median age 28, and the source of the numerical thresholds used by the tool on this page. The lactate threshold corresponds to a rating of 10.8 ± 1.8, the individual anaerobic threshold to 13.6 ± 1.8, a blood lactate of 3 mmol/L to 12.8 ± 2.1 and of 4 mmol/L to 14.1 ± 2.0. Its authors recommend prescription ranges of 11 to 13 in untrained subjects and 13 to 15 for more intense but still aerobic work. The standard deviations of nearly two points are worth remembering as much as the means.
6. Kendrick KR, Baxi SC, Smith RM. Usefulness of the modified 0-10 Borg scale in assessing the degree of dyspnea in patients with COPD and asthma. J Emerg Nurs 2000;26(3):216-22. PMID 10839848. The reference everyone cites for the validity of the 0-10 scale in dyspnoea, and also the abstract carrying its bounds, this 0 to 10 rated scale. The figure grounding its conclusion is far weaker than its reputation: correlations between change in score and change in peak expiratory flow are -0.31 in asthma, in 42 patients, and -0.42 in chronic obstructive pulmonary disease, in 60.
7. Ries AL. Minimally clinically important difference for the UCSD Shortness of Breath Questionnaire, Borg Scale, and Visual Analog Scale. COPD 2005;2(1):105-10. PMID 17136970. The source of the famous 1 point minimal clinically important difference, and the context regularly omitted alongside it. It concerns dyspnoea, it is established by distribution method across pulmonary rehabilitation trials, its abstract gives neither the number of trials nor the sample sizes, and it does not even say which of the two Borg scales is meant. It therefore does not hold as a change threshold for perceived exertion during exercise.
8. Ding W, You T, Gona PN, Milliken LA. Validity and reliability of a Chinese rating of perceived exertion scale in young Mandarin speaking adults. Sports Med Health Sci 2020;2(3):153-8. PMID 35782285. One of only two indexed language validations in this family: two Chinese versions of the 6-20 scale, with validity tested for both and reliability for one. It serves here to measure an absence: French has no equivalent.
9. Camargo MRMC, Zambon F, Moreti F, Behlau M. Translation and cross-cultural adaptation of the Brazilian version of the Adapted Borg CR10 for Vocal Effort Ratings. Codas 2019;31(5):e20180112. PMID 31691744. The second, and the only one for the 0-10 scale: a Brazilian cross-cultural adaptation of a version adapted to vocal effort. The field therefore has, in all, only two indexed language validations for two instruments used worldwide.