Category Archives: Exercise Technology

Are Iron-Distance Courses Really 140.6? Read This Before You Get That Tattoo.


I have been racing triathlons for about ten years.  However, I have only just learned, while writing my most recent post (about GPS and race distances), that, unlike the strict regulations for measurement of running races, triathlon distances can have a great deal of inaccuracy.  I have since scoured the internet looking for specific regulations about how race directors of triathlons are required to measure distances in their courses.  First of all, it is very difficult to find any information at all on this topic.  In other words, while USA Track and Field and the Association of International Marathons and Distance races have easily accessible and highly detailed information about how courses are measured, such information is hard to find for such organizations as the World Triathlon Corporation (the owners of the Ironman and Ironman 70.3 races) and USA Triathlon.  Where instructions about course measurement are available, the level of rigor does not appear to approach running-only races.  For example, the International Triathlon Union (ITU) has a manual for run course measurement inside of the Event Organizer Manual which is nominally based on the rules of the International Association of Athletics Federation.  A cursory review of ITU’s manual reveals less rigor in such critical areas as calibration and repeated measurement of the course.  Indeed, the International Association of Athletics Federation suggests adding 1/1000th of the race distance to every kilometer of the race, to prevent short measurements.  This convention has been institutionalized in certified marathons, such that marathons are always certified at a length 42.2 meters longer than 42,195 meters.  Such suggestions and conventions appear to be absent in most triathlon run courses.  In the manual for Triathlon Ontario:  “The course should be measured with a GPS or Jones Counter.  Measurements by car odometer or bike computers are not recommended.”  (see my discussion about the accuracy of GPS)  According to a quote from Xtri (the link to the source no longer works), the German Triathlon Union allows for a deviation in course length of plus or minus 10%.  Can you imagine the uproar if, after months of training, someone finally runs a sub-two hour marathon, only to later find that it was 10% short?  This amount of error is, apparently, permissible in triathlon.  Practically speaking, however, most deviation is probably well under 10%.

This amount of error in run course distances in triathlon begs the question: why do participants permit it?  First of all, most participants do notice some inaccuracies, but it does not appear to be common knowledge that up to 10% deviation may be permissible in the rules.  However, most importantly from the participant’s perspective, race performances and the rankings that result, like the elegant ranking system of USA Triathlon, are internally controlled by by the race results and not by the actual distances.  In other words, participant’s performances are measured with regard to each other and not with regard to the actual race distances.

So, why do correct distances matter if rankings are internally controlled?  There are a lot of reasons.  For example, if a participant attempts to crunch numbers, like I do, to measure race performances and the success of training plans to improve each of the three disciplines of triathlon, these numbers may be meaningless (along with performances in each discipline, I also usually record my ranking, within my age group, in each discipline – but this is dependent upon who shows up to race).  If a participant uses GPS data, this is vulnerable to the inaccuracy of this type of measurement.  Online programs, like MapMyRun, have their own inaccuracies. Another reason why accuracy matters is the ability to compare performances between different races or between the same race run different years.  There are a few Ironman and Ironman 70.3-brand races that are USATF certified: the World Championship course in Kona, Ironman Louisville, and Ironman 70.3 Poconos.  But this is only 3 out of 41 Ironman and Ironman 70.3-brand races in the United States.  Yet another big reason why correct measurements are important is the ubiquitous “70.3” and “140.6.”  Should the bumper magnets read “63.3-77.3” and “126.5-154.7”?  Imagine the Chicago Marathon selling a hat that reads “23.6-28.8”.

This discussion has not included the inaccuracies in measuring swim and bike courses, but these measurements are treated with much less rigor than run courses in triathlon.  As a personal example, last year I was lucky enough to participate last year in the USA Triathlon National Age Group Championships in Milwaukee, in the Olympic distance.  I am not a very fast athlete and I don’t know if I will ever qualify again.  So one of the benefits, I thought, of participating in this race is to know exactly how I was doing in each distance.  I got out of the water in under 27 minutes, which is fantastic for me.  I learned a week later that the swim course was believed to have been measured too long (not too short).  Furthermore, in a search of USA Track and Field’s certified courses, this USAT championship event has not had its run course certified.

Oh no...This form doesn't exist. Head back to the manage forms page and select a different form.

This, I believe, reveals the core issue in the lack of accurate measurement of triathlon courses: confidence.  I really don’t care if the run or swim are a little long or short, just inform me.  I understand that adding another two miles to a bike course may land the course in another jurisdiction with additional fees and regulations, or adding a little loop to the run may detract from the aesthetic of the run, or the transitions have to be in certain types of locations that, then, limit the arrangement of the swim, bike, and run.  But if I think I am running a 10K Olympic run course, it must be 10K.  If I am one of the thousands (tens of thousands?) of people who have 140.6 apparel or tattoos, the course had better be 140.6 miles.  Indeed, the distances are part of the branding of the events.  Finally, as my wife, who is an 11-time marathoner and 3-time ultra-marathoner says: “it is just wrong for a marathon to not be a marathon, ask for a discount!”


Race Distances and GPS


When I participate in a race, I try to set a pace I think I can maintain and then, desperately, try to stick to it.  Often, I eventually reach the point at which I watch my GPS count down the remaining fractions of the course.  “Only ¼ of the run remains!, only ⅕ of the run remains!, only ½ mile left!!”   But, so often, my GPS says I am done while the finish line is off in the distance.  What is the problem?  Is my GPS off or was the course measured inaccurately?

Oh no...This form doesn't exist. Head back to the manage forms page and select a different form.

GPS (global positioning system), is a technology that has been developed and maintained by the United States government since 1973.  It became fully operational in 1995.  GPS satellites fly in medium Earth orbit at an altitude of approximately 12,550 miles (20,200 km).  Each of the 24 operational satellites circle the Earth twice a day.  This picture is from


The satellites in the GPS constellation are arranged into six equally-spaced orbital planes, each of which contains four “slots”, occupied by satellites.  This arrangement ensures that users can view (use) at least four satellites from virtually any point on Earth.  Since at least three satellites are necessary for triangulation of the position of a GPS receiver, having at least four available at all times assures that a measurement is available.  The US Air Force maintains the satellites and normally flies several extra satellites to maintain coverage whenever the baseline satellites are serviced or decommissioned.  In June, 2011, there was an expansion and reconfiguration of the satellites, so that three of the extra satellites became part of the constellation baseline.  This led to improvement in GPS coverage since, now, there is effectively a 27 slot constellation.  There is a wealth of additional information about the GPS program, the different types of satellites, and anything else someone could need for a cool high school research paper on Along with the US-operated GPS program, Russia, the European Union, India, and China have been developing, or using, GPS-like systems, as well.

GPS receivers (such as in the sport devices many of us wear) all use similar technology.  GPS has been improved over the years by differential GPS (DGPS), which involves using a fixed land-based point, along with GPS satellites, to improve accuracy.   Wide area augmentation system (WAAS) is a further improvement of the use of ground-based stations to improve accuracy. Here is a detailed discussion of this technology.

The following diagram is from


100 m: Accuracy of the original GPS system, which was subject to accuracy degradation under the government-imposed Selective Availability (SA) program.

15 m: Typical GPS position accuracy without SA.

3-5 m: Typical differential GPS (DGPS) position accuracy.

< 3 m: Typical WAAS position accuracy.

What this means is that, with WAAS, your Garmin or similarly equipped device can, ideally, provide accuracy with an error of under 3 meters 95% (NOT 100%) of the time.  The other 5% of the time, the error may be up to 10 meters.  However, these devices are not uniform.  Some are just better.  Here is an interesting recent review of a number of popular devices.  Aside from differences between devices, accuracy can be affected by a number of factors, including mountains, tall buildings, clouds, and interference to the radio signal.  Accuracy can also be affected by the way that the data points are recorded. GPS devices do not collect continuous data.  Instead, they collect a series of data points at regular intervals, up to every second.  If a course involves a lot of hairpin turns, or turns in areas with a lot of trees, for example, it is predictable that data points would have some “misses.”  The GPS unit, itself, or the software program to which the data is uploaded, will then connect these data points and use additional algorithms (data smoothing) to create the best estimate possible of the actual course that was covered.  Since the algorithms are not the same, different, well-respected, software programs can produce varying results.

So, GPS is not a accurate as many of us would like to believe.  But how does this relate to racing?

With regard to USA Track and Field (USATF) certified running race courses, there are very strict rules about the measurement of these courses.  Take a moment to follow the hyperlink and skim the pdf about these rules.  Even temperature is a consideration in accurate course measurement.  In the case of marathons, which are defined as 42,195 meters, the international convention, as established by the Association of International Marathons and Distance Races (AIMS), is to add 1/1000th of the distance (42.2 meters) to the race, to ensure that marathon courses are never too short. This is clearly very important when paychecks, sponsors, and running careers can be attached to race times.  Therefore, with regard to certified non-marathon-distance running races in the US (and in other countries with similar strict regulations) and to certified marathons worldwide, participants can be very confident about the distances they have run.  But GPS-based measurements are usually long, compared to certified running courses.  Why is this?

The biggest reason why GPS measurements of race distances are usually long, compared with the certified distances, is that the courses are measured using the shortest distances around corners and using straight lines from point to point.  Every time we, the tired racers, do not round a corner at the apex (and, thus, take the shortest distance around the corner), we add distance.  DC Rainmaker has an excellent post about this topic, in which he illustrates that, in a two lane road, runners add about 12 meters per turn in which they take the longest path rather than the shortest path around the turn.  This adds up quickly.  With regard to running in straight lines, even while training many of us do not run perfectly straight lines.  In a race, every time we dodge around slower runners, every time we move over to high-five spectators, and every time we deviate from a straight line to get water at an aid station, we add distance.  The end result of adding distance is that a GPS watch will often overestimate the speed you are running, since it estimates pace based on time elapsed and distance run.  If you have a set pace in mind for a race, this inaccuracy needs to be considered. Therefore, trust the mile markers and use GPS estimate as just that: an estimate.

The story with triathlons, however, is surprisingly different.   Please stay tuned for my next post.

Published February 1, 2015


Is Your Body Fat Measurement True?


A little over a week ago, Samantha, one of my nurses, approached me and said that her trainer told her that her body fat percentage was 30%.  This surprised and upset her and she, consequently, suggested that I do a post about the accuracy of these measurements.  Since then, I have read a number of sources about this topic.  Similar to measurements of caloric expenditure, it is difficult to get a really clear idea about the accuracy of measurements of body fat percentage.

To frame this discussion, however, it is important to first review body composition and “normal” ranges of body fat percentages.  The human body is composed of fat and lean mass.  Fat mass is divided into essential fat (the fat that is required for normal physiologic function), which is approximately 3% for men and approximately 12% for women, and non-essential fat, which is anything above those baseline figures.  For athletes, typical total body fat percentages are 6-13% and 14-20%, for men and women, respectively.  For fit men and women, typical percentages are 14-17% and 21-24%, respectively.  For average men and women, typical percentages are 18-24% and 25-31%, respectively.  Finally, obese is considered to be above 25% in men and above 32% in women.  Typical percentages rise slightly with aging.  Separate from specific measurements of body fat, measurement of body mass index (BMI) can be used.  This is calculated by dividing weight in kilograms by height in meters squared.  A BMI of 25 or greater is considered overweight, while a BMI of 35 or greater is considered obese.

There are a number of methods available to estimate body fat percentage.  These include skinfold caliper testing, bioelectrical impedance analysis, dual-emission X-ray absorptiometry (DXA or DEXA scan), hydrostatic (underwater) weighing, whole-body air displacement plethysmography (the Bod Pod), near-infrared interactance, ultrasound, CT, and MRI.  Further discussion of the most commonly-used methods is as follows:

Skinfold caliper testing.  This inexpensive and widely available test involves measuring the loose skin that can be pinched in specific areas of the body.  Typical protocols involve measuring 7 areas of the body.  The measurements are entered into a formula and body fat percentage is then estimated.  This method cannot measure deep belly fat.  It is also limited by the expertise of the person doing the testing.  In one reference, compared to DXA (which is considered the research standard), the skinfold caliper method was off by approximately 7%.

Bioelectrical impedance analysis (BIA).  This is the method that Samantha’s trainer used at the gym.  In her case, she held both hands onto a handheld device.  Similar devices that use the same principles are in scales that estimate percentage of body fat along with measuring weight.  These devices operate under the principle that an electrical current passes more easily through lean tissue than fat.  Therefore, by measuring the current, percentage of body fat can be calculated.  According to several studies, this method can be off, compared to DXA, by approximately 3-9%.  This wide range is explainable by the sensitivity of  BIA to hydration (dehydration leads to an overestimation of body fat %), recent exercise (underestimates body fat % due to decreased electrical impedance), recent meals (underestimates body fat %), the variety of manufacturers and models of these devices, and the different ways these devices are used (for example, using two hands or two feet vs. all four extremities for electrodes).

DXA scan: In this method, the body is scanned by X-rays of two different levels of energy.  Since X-rays of one level of energy are better absorbed by fat than the other, software associated with DXA scan machines can calculate the difference between fat and lean mass and, therefore, can estimate the body fat percentage.  Inaccuracy has been estimated at approximately 1% and this method is, therefore, used as the current gold-standard.

Hydrostatic weighing: In this method, an individual exhales fully and then gets completely submerged in a tank of water to estimate his or her mass per unit volume.  Hydrostatic weighing is based on Archimedes’ principle using measurements of the weight of the body outside the tank of water, the weight of the body when completely immersed in water, and the density of water  Through a series of calculations, body fat percentage can be calculated.  Inaccuracy has been estimated at 0.1-1.2%, hence this method used to be the gold standard for measurement of body fat percentage.  However, this method can be difficult to perform perfectly and can be uncomfortable for the individual who is being measured.

The Bod Pod:  This method is theoretically similar to hydrostatic weighing, but uses air rather than water.  In this case, the individual who is being tested gets into a “pod” that looks a lot like a science fiction space-capsule.  There are more variables to control in this method than in hydrostatic weighing, such as clothing and facial hair.  Therefore, it is unsurprising that inaccuracy has been estimated at 1.5-5.3%.

In understanding the different methodologies of measurement of body fat, it is helpful to also understand the concept of body components and how these are divided by the different methodologies used.  For example, in most testing, such as BIA, hydrostatic weighing, and the Bod Pod, the body is essentially divided into two components: fat mass and lean mass.  In DXA scans, the body is divided into three components: bone, fat, and soft tissue that is not fat.  Finally, there is also four component testing, in which, for example, radiologically marked water (with deuterium) is used in combination with a DXA scan to further divide the body into fat, protein, water, and mineral (bone).  This final division into four components can be helpful to get the most accurate information possible.  This is because hydration can be a confounding factor to the accuracy of DXA scans.  For example, there is is study that demonstrated that a 5% change in hydration can lead to a 2.7% change in results from DXA scans.

The measurement of body fat percentage can be very valuable to health care providers to help measure and guide health and can be further valuable as a motivator.  There are numerous methods to measure body fat and all have advantages and disadvantages.  In general, quick and easy methods, or methods that require a lot of expertise to perform well (skin caliper measurement) are more prone to error.  So, in answer to Samantha, your body fat percentage may be much lower (or higher) than the measurement you received from you trainer.  The most important thing, however, is that you were motivated enough to embark on an exercise program and to get a starting point.


Ball, SD and Altena, TS. Comparison of the Bod Pod and dual energy x-ray absorptiometry in men. Physiol Meas 2004 Jun; 25(3):671-678.

Collins, MA, Millard-Stafford, ML, Evans, EM, et al. Effect of race and musculoskeletal development on the accuracy of air plethysmography. Med Sci Sports Exerc 2004 Jun; 36(6):1070-1077.

Duz, S, Kocak, M, and Korkusuz, F. Evaluation of body composition using three different methods compared to dual-energy X-ray absorptiometry. Eur J Sport Sci 2009 May; 9(3):181-190.

Fogelholm, M and van Marken Lichtenbelt, W. Comparison of body composition methods: a literature analysis. Eur J Clin Nutr 1997 Aug; 51(8):495-503.

Kushner RF, Gudivaka R, Schoeller DA. Clinical characteristics influencing bioelectrical impedance analysis measurements. Am J Clin Nutr  1996 Sept; 64(3 Suppl):423S–427S.

Lukaski, HC. Soft tissue compostion and bone mineral status: evaluation by dual-energy X-ray absorptiometry. J Nutr 1993 Feb; 123(2 Suppl):438-443.

Mooney, A, Kelsey, L, Fellingham, GW, et al. Assessing body composition of children and adolescents using dual-energy X-ray absorptiometry, skinfolds, and electrical impedance.  Meas Phys Educ Exerc Sci  2011 Jan; 15(1):2-17.

Prior, BM, Cureton, KJ, Modlesky, CM, et al. In vivo validation of whole body composition estimates from dual-energy X-ray absorptiometry. J Appl Physiol (1985) 1997 Aug; 83(2):623-630.

Oh no...This form doesn't exist. Head back to the manage forms page and select a different form.

Wheeler, LA. Validation of hand-held bioelectrical impedance analysis for the assessment of body fat in young and old adults. Theses and Dissertations 2012 :Paper 208.

Woodrow, G, Oldroyd, B, Turney, JH, et al. Four-component model of body composition in chronic renal failure comprising dual-energy X-ray absorptiometry and measurement of total body water by deuterium oxide dilution. Clin Sci (Lond) 1996 Dec; 91(6):763-769.



The Facts About Weight Loss From Using A Fitbit-Like Device


Most people would like to lose weight and have a healthy lifestyle.  As simple as these goals sound, they require commitment and organization.  This is where wearable devices like Fitbits and Jawbones can potentially be helpful.  Aside from recording activity, these devices estimate caloric expenditure (please see my previous post for more information on this topic), encourage activity by giving feedback, and offer, through apps and websites, a host of additional information and support.  But, what is the evidence?  Can the use of Fitbit-like devices really lead to weight loss and a healthy lifestyle?

There was a nice review published in August, 2014 about the “behavior change techniques” that current research about Social Cognitive Theory have shown to be effective and whether electronic lifestyle activity monitors (like the Fitbit and Jawbone) utilize these techniques.  The “behavior change techniques” that are supported by research are as follows: prompt practice, prompt self-monitoring of behavior, goal-setting/intention formation, barrier identification/problem solving, provide feedback on performance, prompt review of behavioral goals, provide information on consequences of behavior in general, action planning, prompt rewards contingent on effort of progress towards behavior, facilitate social comparison, provide instruction, self-talk, self-rewards, social support, and teach to use prompts/cues.  The authors of this review then evaluated 13 devices, and their supporting apps and websites, with regard to which of these techniques were employed.  All 13 had goal-setting, feedback on behavior, and self monitoring of behavior.  10 used the technique of reviewing behavior goals, while 8 used social support and social comparison.  6 used information about health consequences and 5 used action planning.  There was minimal use of any of the other techniques and, certainly, no Fitbit-like device, with supporting apps and websites, used all of the techniques listed.  This does not mean that Fitbit-like devices are destined to fail but, instead, that the manufacturers of Fitbit-like devices have a considerable opportunity to improve if these devices.

For those individuals who are trying to lose weight and improve their lifestyles, and for their coaches and physicians, the most important question about Fitbit-like devices is “do they work to help reach these goals?”  There is some limited research on this topic.  For example, Pellegrini et al recently published the results of a study in which three groups of subjects were compared over 6 months.  One group attended weekly meetings about weight loss, the second group attended weekly meeting as well, but also used a wearable Fitbit-like armband, and the third group only used the Fitbit-like armband.  As part of the protocol of the trial, all three groups reduced caloric intake and progressively increased physical exercise of “moderate” intensity.  The researchers found that the three groups collectively lost an average of 6.4 kg (more than 14 pounds), decreased hip and waist circumference, decreased percentage of body fat, became more physically fit, increased physical activity, and ate less calories at the end of the 6 months, but that there were no statistically significant differences between the three groups. This means that using a Fitbit-like device can lead to the same improvement in weight, fitness, and lifestyle as can be achieved by attending weekly meetings.  In another study, by Shuger et al, subjects were divided into four groups over 9 months: self directed weight loss with a manual, a group-based weight loss program, a Fitbit-like armband alone, and group-based weight loss plus a Fitbit-like armband.  The researchers found significant weight loss from baseline in all three intervention groups (between about 4 and 14.5 pounds), but not in the self-directed group.  However, only the group that used the combination of the group program and the Fitbit-like device achieved significant weight loss compared to the self-directed group. These results suggest that a variety of interventions can lead to weight loss, but that, at least in this study, only the combination of a group-based weight loss program and a Fitbit-like armband is more effective than reading and following a manual.  Taken together, these two studies showed that the use of Fitbit-like devices can lead to weight loss and can be as effective as in-person interventions.

The research literature, therefore, has shown that Fitbit-like devices use some, but not all, techniques that are supported by Social Cognitive Theory and can be useful in efforts to lose weight and achieve a healthy lifestyle.  However, there does not appear to be research about the duration of these benefits beyond 9 months.  Ideally, the use of Fitbit-like devices will lead to lifelong improvements in lifestyle and won’t be a passing fad.


Lyons, EJ, Lewis, ZH, Mayrsohn, BG, et al. Behavior change techniques implemented in electronic lifestyle activity monitors: a systematic content analysis. J Med Internet Res 2014 Aug 15; 16(8):e192

Pellegrini, CA, Verba, SD, Otto, AD, et al. The comparison of a technology-based system and an in-person behavioral weight loss intervention. Obesity (Silver Spring) 2012 Feb; 20(2):356-363.

Shuger, SL, Barry, VM, Sui, X, et al. Electronic feedback in a diet- and physical activity-based lifestyle intervention for weight loss: a randomized controlled trial. Int J Behav Nutr Phys Act 2011 May 18; 8:41.



Oh no...This form doesn't exist. Head back to the manage forms page and select a different form.

Fitbit-Like Devices And Calorie Burn


In the previous post in this blog, I discussed the accuracy of measurement of energy expenditure by fitness machines.  In short, it appears that machines that use more data to estimate energy expenditure are likely to be more accurate, but that it is very difficult to know for sure.  With wearable devices, which include, but are not limited to, devices made by Fitbit, Garmin, Polar, Timex, Motorola, Nike, and Suunto, however, there is more information available about the accuracy of this measurement.  This article will be limited to the accuracy of wearable accelerometer devices like the Fitbit.

These devices work by measuring the movement of the device, which can be on a wrist, upper arm, waistband, or elsewhere.  This then leads to estimates of number of steps taken and, by using estimated or measured stride length, estimates of distance traveled.  There is a nice article in Berkeley Science Review about the accuracy of steps taken and confounding factors (for example, Fitbits are more accurate with walking at slow speeds and less accurate with walking at faster speeds and running, especially if the devices are kept in a pocket).  But we are interested in the accuracy of measurement of energy expenditure.  Here is the text from the Fitbit website about how they measure this:

Your tracker and dashboard show an estimated number of calories burned based on your BMR (Basal Metabolic Rate), which we calculate using the height, weight, age, and gender information that you provided when you set up your Fitbit account. If your tracker measures heart rate, the calorie burn estimate also takes heart rate into account. Note that calorie tracking for the following day begins at midnight and incorporates the calories you burn while sleeping.

When you sync your tracker, Fitbit replaces your estimated calorie burn with your tracker’s data. If you manually log activities, the calories burned by during those activities are taken into account as well.

When you haven’t synced your device or logged any activities, Fitbit tries to guess how many calories you have burned if you got out of bed, got dressed, went to your day job, came home, and did nothing much more than walk to your car.

Once you start logging activities, Fitbit stops estimating and uses the data you’ve provided instead. The more you wear your tracker, the more accurate your calorie burn data will be.

In the literature, there have been reports that show that, under controlled conditions, accelerometer devices, such as the Fitbit, can be surprisingly accurate.  For example, in a study published in September, 2014, sixty volunteers (average age 26.4 years, 30 male, 30 female), wore eight different types of activity monitors simultaneously while completing a 69-minute “protocol.” An indirect calorimeter was also worn and the measurements of this device were used as the reference.  The authors concluded that:

Oh no...This form doesn't exist. Head back to the manage forms page and select a different form.

For overall group comparisons, the mean absolute percent error values (computed as the average absolute value of the group-level errors) were 9.3%, 10.1%, 10.4%, 12.2%, 12.6%, 12.8%, 13.0%, and 23.5% for the BodyMedia FIT, Fitbit Zip, Fitbit One, Jawbone Up, ActiGraph, DirectLife, NikeFuel Band, and Basis B1 Band, respectively.

This means that, for most of the devices used, in a controlled test, the inaccuracy of measurement of caloric expenditure was between 9 and 13%.  However, there are limitations.  For example, in another recent study, 23 subjects were fitted with two Fitbit and Fitbit Ultra accelerometers, two industry-standard accelerometers, and an indirect calorimetry device (which served as the reference).  These subjects participated in 6 minute bouts of treadmill walking (3.5 mph), jogging (5.5 mph), and stair stepping.  The results of this study indicate that the Fitbit and Fitbit Ultra are reliable and valid for monitoring steps and for determining energy expenditure while walking and jogging without an incline (the Fitbit devices produced estimates of caloric expenditure that were 88-113% of those values that were measured by the indirect calorimetry device).  However, with inclined activities, the Fitbit and standard accelerometers under-estimated energy expenditure by 40% or more.  Finally, in another recent study,20 participants (10 male, 10 female) wore a Fitbit on the hip and also wore a separate indirect calorimeter as a reference.  These subjects performed walking and running trials on a treadmill and a simulated free-living activity routine.  The authors found that:

The Fitbit significantly underestimated EE [energy expenditure] for cycling, laundry, raking, TM [treadmill] 3 mph at 5% grade, ascent/descent stairs, and TM 4 mph at 5% grade, and significantly overestimated EE for carrying groceries. Energy expenditure estimated by the Fitbit was not significantly different than EE calculated from the Oxycon Mobile [the indirect calorimeter] for 9 activities.

The authors then conclude that:

The Fitbit worn on the hip significantly underestimates EE of activities. The variability in underestimation of EE for the different activities may be problematic for weight loss management applications since accurate EE estimates is important for tracking/monitoring energy deficit.

My conclusion is that accelerometers, like the Fitbit, can serve a useful role in an individual’s fitness program.  These devices do not cost as much as more sophisticated “athlete-grade” devices, but can provide, with limitations, a decent snapshot into “calories burned.”  For limited fitness activities, which many people use for their fitness, such as walking the dog around a park or walking with co-workers over the lunch hour, the Fitbit-like devices appear particularly well-suited.  If you are planning to use such a device, please discuss it with your health-care provider to best incorporate it into an overall fitness program.


Adam, NJ, Spierer, DK, Gu, J, et al.  Comparison of steps and energy expenditure assessment in adults of Fitbit Tracker and Ultra to the Actical and indirect calorimetry. J Med Eng Technol 2013 Oct; 37 (7): 456-462.

Lee, JM, Kim, Y, and Welk, GJ. Validity of consumer-based physical activity monitors.  Med Sci Sports Exerc 2014 Sep; 46(9): 1840-1848.

Sasaki, JE, Hickey, A, Mavilia, M., et al. Validation of the Fitbit Wireless Activity Tracker for prediction of energy expenditure. J Phys Act Health 2014 Apr; epub ahead of print.



Exercise Machines: How Many Calories Are You Really Burning?


IMG_1569Many of us look for motivation and information from fitness machines as we try to reach and maintain a goal weight.  I participated in an indoor triathlon this morning and, without me having entered any information about myself into the computers associated with the stationary bike and the treadmill, I got readouts reporting that I had “burned” about 350 calories in 30 minutes on the bike and about 250 calories in 20 minutes on the treadmill.  Are these numbers accurate and, if not, how far off are they?  Should I smile and have a cheeseburger?

When I began to look for information on the question of accuracy of measurement of calories, I expected to find a few scientific papers and reviews on the subject with easy-to-understand numbers.  For example, I was hoping to find statements such as: “treadmills are off by 20 percent plus or minus, elliptical trainers are off by another percentage plus or minus,” and so on.  As it turns out, it is very difficult to get good information on this topic.  This is in part due to the large degree of variation between individual people.  For example, would a 70 kilogram professional marathoner expend the same amount of energy to run 8 miles in an hour as a 70 kilogram novice runner?  There are so many variables to consider, including percentage of body fat, resting metabolic rate, efficiency of movement, VO2 max (this is a measure of the efficiency with which an individual uses oxygen to perform strenuous activities and is defined as: the maximal oxygen uptake or volume of oxygen that can be utilized in one minute during maximal exercise), altitude, hydration, stride length, underlying medical conditions, and medications (prescription or not).  Furthermore, different models of exercise equipment may not all follow and record the same types of data, such as distance covered, heart rate, resistance (such as an incline on a treadmill), weight of athlete, and age of athlete.  Accuracy can be affected, of course, if the equipment is not well-maintained and calibrated.  Entering inaccurate personal information and holding on side rails (or, in the case of elliptical machines, deciding whether or not to use the hand levers) may affect results.  The manufacturers of exercise equipment may use a variety of algorithms (which are usually not publicly available), to use the data that has been recorded to estimate calories expended.  Good algorithms would need to be tested against known controls with large numbers of individuals.

Instead of relying on the readouts from exercise machines, it may be more helpful to take another approach.  The Compendium of Physical Activities, last updated in 2011, is a research-based approach to estimating the energy costs of various activities (they even have rodeo sports, juggling, and curling).  The authors emphasize the limitations of these estimates:

Oh no...This form doesn't exist. Head back to the manage forms page and select a different form.

‘When using the Compendium to estimate the energy cost of activities, investigators should remind participants to recall only the time spent in movement.  The Compendium was not developed to determine the precise energy cost of physical activity within individuals, but rather to provide a classification system that standardized the MET [metabolic equivalent] intensities of physical activities used in survey research.  The values in the Compendium do not estimate the energy cost of physical activity in individuals in ways that account for differences in body mass, adiposity, age, sex, efficiency of movement, geographic and environmental conditions in which the activities are performed.  Thus, individual differences in energy expenditure for the same activity can be large and the true energy cost for an individual may or may not be close to the stated mean MET level as presented in the Compendium.”

MET is derived from estimates of resting metabolic rates (RMR).  RMR, in turn, is higher in men than women and increased with height, weight, and lean mass.  It decreases with age.

It is very difficult to find online calorie calculators that use the, above, 2011 updated figures.  The sites I have found that name their references, even the sites of highly-respected academic centers, use pre-2011 formulas.  However, there is another way to get this information.  The equation to determine RMR is the Harris Benedict equation:

Male: 66.4730 + (5.0033 * height in cm) + (13.7516 * weight in kg) – (6.7550 * age in years)

Female: 655.0955 + (1.8496 * height in cm) + (9.5634 * weight in kg) – (4.6756 * age in years)

 Fortunately, you don’t have to do the math.  This website uses this equation (I checked) in calculating RMR (called BMR on the website).

RMR is set at one MET unit.   You can refer to the tables on the Compendium website to get MET units for various activities.  Divide your RMR by 12, then multiply this figure by the MET units  for the selected activity to get an estimate.  Using these tables, I was disappointed by the accuracy of these figures for myself.  For example, 60 minutes of bike racing would cost me about 1040 kcal while 60 minutes of running at 8 minutes per mile would cost me 780 kcal.  Based on my perceived need for calories during and after cycling and running, I would predict that the caloric costs of these activities would be flipped (i.e. higher calories for running than cycling).  This discrepancy may be explained by the fact that I am a much more efficient cyclist than runner.  This is where individual variation starts to significantly affect the value of equations such as these.

So, considering the limitations of readouts on fitness machines and of population-based estimates of caloric expenditures, what can you do?  Well, you can go to an exercise lab and have measurements taken with indirect calorimetry.  This is a technique that provides accurate estimates of energy expenditure from measures of carbon dioxide production and oxygen consumption during rest and exercise.  Since this is not practical for most people, for any number of reasons, please allow me to make some suggestions.  Use information from the Compendium, above, to get ballpark estimates.  Then try to select exercise equipment that have been well maintained and that ask you for as much information about yourself as possible.  Equipment that uses more personal information in its algorithms is more likely to provide better data.  Then, compare your estimates from the Compendium with the readout to get a sense if the readout is realistic  However, please remember, unless you are in a fitness lab using sophisticated equipment to measure your caloric expenditures, all you can really hope for is a rough estimate.  Skip the cheeseburger.


Ainsworth, BE, Haskell, WL, Herrmann, SD, et al. 2011 Compendium of Physical Activities: a second update of codes and MET values.  Medicine and Science in Sports and Exercise, 2011;43(8):1575-1581

Harris, JA and Benedict, FG. A biometric study of human basal metabolism. Proc Natl Acad Sci USA, 1918;4(12):370-373