GB2527113B

GB2527113B - Electronic measuring device

Info

Publication number: GB2527113B
Application number: GB1410524.1A
Authority: GB
Inventors: Howe Robin
Original assignee: Mindful Machine Co Ltd
Current assignee: Mindful Machine Co Ltd
Priority date: 2014-06-12
Filing date: 2014-06-12
Publication date: 2019-06-12
Anticipated expiration: 2034-06-12
Also published as: GB201410524D0; WO2015189574A1; GB2527113A

Description

ELECTRONIC MEASURING DEVICE

BACKGROUND

Many aspects of mammalian performance are governed by exponential variables and physical size, for example, growth from maturation and power output. It is therefore advantageous for any measures and analyses of such aspects of performance to use linear and non-linear scaling for more accurate interpretation.

Prior to adulthood differences in chronological and biological ages in the population obscure many relativistic measures useful in adult environments. As such, analysis of such information can lead to erroneous estimates of performance potential for adulthood.

Consider the example of a class of 14-year old boys. Typically we would find one who is shaving and one who is pre-pubescent, with the physical maturity of all the others lying between these extremes. It is highly likely that the larger power-generating capability of the early-maturing, shaving boy will provide them with a significant advantage in any athletic endeavour or sporting competition.

By the time all the boys have reached prime adulthood, however, any power-generating advantage due to early maturation will have likely been nullified. By this time, the motivation and skill development of normal and late-maturing boys is typically so compromised that they no longer participate in sport. The early-maturing may also not participate in further sport if their athletic potential fails to materialise.

Situations then arise where the motivations of the majority of participants are eclipsed by the erroneous heralding of potential 'champions', only for such aspirations to be proven unfounded in later years. More appropriate coaching should provide appropriately scaled measures to compensate for the variability of development in non-adults.

The above situation is compounded by sporting competition normally being divided into year or two year age bands. In such cases, the spread of biological ages is increased further, effectively being convolved with the spread of chronological ages within the age band.

Instead of a producing a desire to progress with an understanding of the maturation process, young minds produce a much simpler choice - the motivation to avoid participation rather than be cast as a failure or a loser becomes dominant. Proper attention to all participants regardless of performance measures is therefore a vital factor in the maintenance of participation levels in sport and/or other like activities.

Coaching efforts should be directed at producing participants who are capable of producing their best performances in prime adulthood and who have the motivation to still be participating at that age. Since meaningful performance measures prior to adulthood are often obscured by differences in chronological and biological ages, a means to compensate performance measures and render them independent of biological age is therefore a worthwhile objective.

STATEMENT OF INVENTION

According to a first aspect of the invention there is provided a device according to claim 1.

According to a second aspect of the invention there is provided a method according to claim 11.

According to a third aspect of the invention there is provided a device according to claim 20.

According to a fourth aspect of the invention there is provided a computer program or a suite of computer programs according to claim 30.

The invention may employ an electronic device, programmable or otherwise, to provide linear and/or non-linear scaling of measures such as that recorded by a stop-watch. The measures can be scaled according to known or predicted bases, and can be applied retrospectively or performed in real-time. A stopwatch counter, for example, can be scaled to compensate for the variation of power output with biological age prior to adulthood.

Mathematical modelling may be performed of one or more measures of growth and measures of performance in athletic-based sporting activities by simple functions. The difference between an individual's predicted performance versus age measure and the mean recorded measures across a gender may be the same as the difference between the timing of an individual's peak pubertal growth and that of the peak recorded mean in the respective population.

Where measured growth does not yet identify the pubertal spurt, the mathematical functions may permit extrapolation by, for example, normal least-squares error minimisation or such like to predict the growth spurt. The biological age-compensated performance model may serve to both predict performance in adulthood and provide a basis for useful relative comparisons between non-adults at the same event using, for example, multiple weighted counters, and/or historically recorded results.

Embodiments of the present invention may provide .means to compensate for the differences in the biological ages of non-adults in athletic-based sporting competition or performance-based measures. Thus, the relevance of relativistic comparisons between participants may be increased both longitudinally and cross-sectionally. Such comparison may be performed at some chosen nominal chronological age or at an age band and/or may be performed prior to adulthood or in a manner that predicts performance in adulthood.

One application of the invention provides for the weighting of competition results to compensate for natural differences in the biological ages of the participants. Embodiments may also aid in the assessment of performance potential which may otherwise only have become evident in prime adulthood or passed unnoticed altogether, such performance being desired in 'talent identification' schemes. Both real-time and retrospective analyses are possible.

Further, embodiments aid the analysis of performance measures and the subsequent design of training programmes, and in particular, where those programmes are to be optimally geared to individual requirements. The invention may also be employed as a motivational aid and/or an educational tool when disseminating measures of progress for participants, their carers, coaches or other interested parties.

Whilst embodiments of the invention apply to activities dominated by athletic power output, there are a number of other potential applications of the invention. There are a number of sports and activities, for example, in which athletic power output is not a dominant factor and where some form of proportional weighting would be applicable.

BRIEF DESCRIPTION OF DRAWINGS

Figure 1 is an illustration of a programmable electronic device;

Figure 2 is a schematic representation of the components within a programmable electronic device;

Figure 3 is a flow diagram for a programmable electronic counter;

Figure 4 is a flow diagram for a programmable electronic counter;

Figure 5 is an illustrative example of a real-time scaled measuring device showing scaled and unsealed measures represented on the programmable electronic device of Figure 1;

Figure 6 is an illustrative example of a retrospectively scaled measuring device showing scaled and unsealed recorded measures represented in the programmable electronic device of Figure 1;

Figure 7 is a graph plotting an example (x(t)) and mean (y(t)) male height versus age characteristics constructed using Gompertz functions;

Figure 8 is a graphical illustration of the derivatives (x'(t), y'(t)) of the example and mean (x(t), y(t)) male height velocity versus age characteristics shown in Figure 6 illustrating the pubertal growth spurt and the derivation of biological-chronological age offset, To;

Figure 9 is a graphical illustration of the male and female mean height velocity versus age characteristics illustrating gender age offset, TG;

Figure 10 is an illustrative example of a maximum performance class measure, Pmax(t), (measurement class includes jumping height, jumping length, lifted weight, throwing distance);

Figure 11 is an illustrative example of a minimum performance class example, Pmin(t), and the complement that produces an equivalent maximum performance-type measure (measurement class includes race times);

Figure 12 is a plot of mean male performance versus age characteristic, u(t), and a predicted individual performance versus age characteristic, v(t).

DETAILED DESCRIPTION

An electronic device such as that shown in Figure 1 is able to serially process instructions and data from which an output can be conveyed such as that of a stopwatch or counter. As shown in Figure 2, the device may comprise a processor 1, a user input device such as a touch-screen and an output device such as that of the touch-screen 2. The device may further comprise a non-volatile memory 3 and, optionally, additional temporary memory 4. The device may also comprise a wireless internet connection 5 and/or wired connections, such as USB 6 allowing for the exchange of data to and from the device. One or more clocks 7, such as a nominally fixed frequency crystal, may be provided for synchronisation of serial instructions.

Instructions stored in the non-volatile memory may allow for programming and reprogramming of the device to impart some function, such as that of a counter shown schematically in Figure 3. For clarity, embodiments of the present invention will be described with reference to applications on a device such as that shown in Figures 1, 5 and 6. However, it will be appreciated that embodiments of the present invention may be implemented in software or hardware and/or may be implemented on any other suitable device/platform. For example, embodiments may be implemented using one or more FPGAs, microcontrollers, ASICs or a combination thereof.

Figure 3 shows a flow diagram for a linear implementation of an electronic counter according to an embodiment of the present invention. The electronic counter may receive a regular clock input 8, such as the system clock 7 shown in Figure 2, which may be used to increment a counter "A" at step 9 at (nominally) fixed intervals. The output from step 9 may then be compared at step 10 to a scaling factor "Ω" which may be stored in a register of memory 3, 4 such that a flag is raised if counter "A" output is equal to Ω. Generation of the scaling factor will be described in more detail below. If the flag is raised, counter "A" is reset to zero and the clock pulse is propagated to the output 12. As such, the rate of the clock output 12 is equal to the rate of the clock input received at step 8 divided by the scale factor Ω. The process is then repeated such that the clock output is a periodic signal having a period which is dependent on the scale factor Ω. It will be appreciated that the implementation shown in Figure 3 is one example of many which would be within the remit of the skilled person.

The implementation in Figure 3 provides a clock output having a constant period provided the period of the clock input 8 is constant, i.e. a linearly scaled clock output. However, in practice, power is supplied from both short-term, high power anaerobic and lower power aerobic energy conversion processes. Short duration (sprint) events exploit the greater power output capabilities supported by the anaerobic processes, whilst longer endurance events are dominated by the aerobic power output capability. A measure of power output capability may therefore vary non-linearly with time and that non-linearity may be modelled by a function ί(Ω,Β). Figure 4 shows a further embodiment of the counter, where like steps have been given like numbering, in which feedback is provided in the flow path, operable to update the value of Ω such that it is dependent on the clock output 12 and therefore can be scaled non-linearly with respect to time. Adjusting the rate of the counter in this manner can therefore take into account non-linearity in power output with respect to time. As with the counter of Figure 3, a clock input 8 is received and at step 9 the counter value Ά' is incremented. The value Ά' is then compared at step 10 to the scale factor Ω and propagated only when A= Ω. In which case, a further variable 'B' is incremented at step 11 and output both to output 12 and to step 22 of a feedback loop. The feedback loop may consist of a processing stage or look-up table 22 operable to update the scale factor Ω based on the value of 'B' such that Ω, and as such the period of the clock output 12, both vary with time. Again, as with the implementation shown in Figure 3, it will be appreciated that the non-linear counter implementation shown in Figure 4 is one or many within the remit of the skilled person.

Event specific look-up tables' may provide as suitable a method as analytic expressions. Different functions and/or tables for ί(Ω,Β) may also be employed for males and females due to their different physical manifestations.

Prior to puberty, both males and females generate power via predominantly aerobic processes. Only close to puberty do the anaerobic processes begin to develop sufficiently to contribute to power output. The non-linear model f(O,B) is preferably also then referenced to the biological age (or estimate thereof) of the subject. Once again a set of 'look up tables' may be used to model the development of the anaerobic processes using biological age as the index. Additionally or alternatively, numerical models can be used to determine the time domain function used to scale the counter rate. Suitable non-linear mathematical models may, for example, employ the linear terms that are related to (the first order of) time but with the addition of terms related to higher powers of time such as time-squared (second order), time-cubed (third order) and so on. Such higher powers aid the modelling of the effects of performance that are dependent on, say, volume of a body part that may not grow linearly with time. More simple non-linear models, such as that adopting a Taylor series may also be used where, for example, processing power is at a premium. Alternatively, models utilising prior values in the model such as Volterra series may prove more accurate, whilst being more processor intensive.

In the example of the invention presented, the scaling of the timing compensation can be performed in real-time, such as in the stopwatch shown in Figure 5, or the information can be represented retrospectively in the analysis of race results as in Figure 6, for example. The scaling applied may vary, separately or combined, linearly with real-time to compensate for the differences in biological and chronological ages, or non-linearly to represent the magnitude profile of the output of the different physiological systems that contribute to motive power.

Methods for calculating counter scaling will now be described in more detail. Generally, the scaling factor may be proportional to a performance metric, i.e. a metric relating to the likely performance of the person being timed by the electronic device. The performance metric may be based, for example, on estimated physical maturity of the user. In other embodiments, the performance metric may be proportional to the general fitness level of the user relative to some mean fitness level. It will be appreciated, therefore, that the scaling factor generated by the performance metric may either increase or decrease the timing rate depending on the performance metric of any particular user. In some embodiments, the performance metric may be equivalent to a handicap of a user relative to other users in one or more groups.

In some embodiments, calculation of the counter scaling can be achieved by predictive modelling. Growth during adolescence is dominated by the so-called pubertal spurt superimposed on an underlying but decreasing trend stemming from early childhood. The timing of the pubertal spurt can vary widely amongst the population but on average occurs approximately two years earlier for girls than boys. The pubertal spurt is caused by the release of sex hormones that result in the development of adult characteristics and therefore gains in athletic power.

Many changes during adolescence reflect the pubertal spurt and standing height is one example. Also well-known is the ability to model the pattern of growth with simple mathematical functions, for example, Gompertz and logistic functions. Embodiments of the invention presented adopt, but are not limited to, Gompertz functions.

The superposition of N Gompertz functions is able to form a sufficiently accurate model of growth prior to adulthood. With N=3, the three Gompertz functions model the contributions of the prenatal, early years and pubertal spurts respectively. Embodiments described are primarily concerned with growth just prior to and beyond puberty where N=2 is often sufficient and the effects of the prenatal and early years spurts can often be amalgamated in a single component function.

Figure 7 shows a model of height versus age, x(t), for an example male constructed from Gompertz functions where N=2. Figure 8 illustrates the derivative with respect to time of x(t), x'(t), which represents the rate of change of height with age, or more formally, height velocity. The height velocity shows more clearly the peak height velocity 14 where the pubertal spurt and therefore growth is at its maximum.

In general, such a model may be described by the sum of Gompertz functions as,

where the Nth component (i=N) is reserved to describe the pubertal spurt. (Note the component series actually runs from 0 to N, but bx(0)=0 and cx(0)=0, such that ax(0) can be usefully represented separately). Rewriting the expression yields,

where

TX(N) is then the age at which the peak height velocity of the pubertal spurt occurs 14.

The known wide range of values of TX(N) produces the (longitudinal) mean male growth model denoted by y(t) as illustrated in Figure 7 with the corresponding (longitudinal) mean height velocity, y'(t), shown in Figure 8. Due to the variation between genders, males and females are preferably treated as separate populations with their own respective mean models shown in Figure 9.

The mean growth model y(t) may also be expressed as a sum of Gompertz functions where,

and,

TV(N) is the mean age at which the pubertal spurt has its peak height velocity 15. The difference between TX(N) and TY(N) represents an index for the difference in biological and chronological age of an individual and is denoted by To 16 where Γ0 = Γ»-Γ»

as illustrated in Figure 8.

As stated previously, many measures can be used as indicators of growth. Anthropometric measurements such as, but not limited to, height, seated height, arm span and weight can be used in embodiments of the present invention, or some weighted sum thereof. Such anthropometric measurements are easy to carry out, measure-to-measure accurate (particularly if using the same apparatus) and are non-intrusive. In embodiments described herein, standing height is used as an example.

It is preferable to avoid using bodyweight measures. This is because many sports, such as swimming, show a negative correlation with bodyweight. Such correlations can lead to an emphasis on weight control that subsequently can act to promote eating disorders and the like. Since embodiments of the present invention are intended to aid positive behaviours, body weight measures are preferably avoided.

In other embodiments, other measures, such as blood lactate measures and bone analysis by x-rays, may serve in giving accurate measures of physical maturity and the progression to adulthood. Such measures may, however, be less feasible to implement since they are difficult to perform, expensive and intrusive. The effect of errors in assessment of maturity is somewhat attenuated given that embodiments of the present invention produces a relative performance correction measure.

Calibration of the models x(t) and y(t) in embodiments presented may be via, but not limited to, least squares optimisation with the respective targets of individual, measured height, g(t), and a measured or known dataset, h(t), of the mean height of the appropriate population. The optimisation of x(t) is more complex since, in general, g(t) is not defined for the whole range oft.

The optimisation of y(t) is generally easier since the dataset is known for all relevant values of t. The optimisations process preferably seeks to minimise the total squared error, E, where,

Values for Tmin and Tmax of 6 and 22 years respectively are used in the embodiment described, but an increased or decreased age range can be adopted as required. The range is likely to be sport specific, though one component model of height will suffice for all ranges and sports. Several methods are available for obtaining optimal solution to the calibration process. In general, the

smooth nature of the target function ensures convergence on an optimal solution for y(t) without undue complications.

For the optimisation of x(t), a similar process suffices. The initial guesses at x(t) can be set equal to the values of y(t), with the exception of ax(i), such that the least squared error for minimisation, E, is,

Since an important factor for prediction is TX(N), absolute height, or any other measure of growth, is not of prime importance and the need to guess ax(0) can be removed from the initial guesses such that,

The difficulty in optimising x(t) is because in many cases, at least for a few years, it will entail extrapolation into the future in order to predict when T„(N) will occur.

For limited numbers of measures, at least six measures will be required to solve each component of x(t) (plus one more for non-zero ax(0)). Again, the smooth nature of the growth curve aids convergence of optimisation algorithms. Furthermore since early stage growth monotonically decreases with age, the non-pubertal spurt components can be optimised before the pubertal spurt becomes evident. Once the pubertal spurt becomes evident, optimisation and even solution of all the components of x(t) can be obtained in the next three measures.

Accuracy of the optimisation of x(t) is dependent on the number and timing of measures of g(t). In general, measuring g(t) every one or two weeks prior to the pubertal spurt is sufficiently accurate for the optimisation procedure. Once the pubertal peak growth has passed, the requirements for the frequency of g(t) measures is relaxed somewhat. In many cases, however, a sufficiently accurate 'solution' to x(t) is available a few measures after the pubertal spurt has occurred. Some caution should be exercised, however, since for non-adults engaged in physically demanding programmes, growth can be delayed until sufficient rest becomes apparent.

Prior to the onset of the pubertal spurt being detectable, and when differences in maturation become most evident, chronological age can serve as a useful indicator. In a year group of infants, for example, the year between the oldest and youngest has a proportionally greater effect than

beyond puberty. Using chronological age as a starting point for the weighting of measures prior to the onset of the detection of the pubertal spurt effectively merges these chronological age referenced measures into the individualised biological age estimate as puberty arrives and makes the weighting useful prior to the onset of puberty also.

Embodiments of the present invention may require prediction of the performance measures in question based upon previously recorded data that provide the appropriate mean performance measures versus (nominal chronological) age. This mean (and therefore likely smooth and monotonic) dataset can be used directly as the basis for prediction or as the basis for a model of mean performance that is guaranteed to be smooth and monotonic. In embodiments described herein, the basis for a predicted mean performance model is disclosed.

The modelling of performance measures can be performed in a similar manner. There are, however, two classes of performance measure as shown in Figure 10 and Figure 11. Pmax(t) describes a class of measures that increase with athletic performance, including, for example, jumping height and throwing distances. Pmin(t) describes a class of measures that decrease with athletic performance, such as race times for example.

Embodiments may use, but are not limited to, a method where a performance model, u(t), is constructed that is consistent for both classes of measure. For measures of the type Pmin(t), a mathematical complement is performed prior to modelling such that the complement of Pmin(t) 18 shows the same trend with increasing age as Pmax(t). A useful compliment for Pmin(t) is to use the final and therefore maximum value of Pmin(t) such that,

Both Pmax(t) and the modified Pmin(t) then show a general trend of exponentially decreasing improvements as age progresses towards adulthood. Both classes of performance versus age characteristics can then be modelled similarly as a series of one or more mathematical functions, such as, but not limited to Gompertz functions.

For such measures, a model of mean performance, u(t), can be constructed of M components where,

and

The peak improvement in performance denoted by Tu(i) may be real or just notional since it may occur prior to the age at which analysis starts. Theoretical or otherwise, however, the ages represented by Tu(i) are used as a means of reference.

The example in Figure 12 shows a model for u(t) with M=l, using boys age-group swimming results (100m short-course freestyle). Much data concerning improvements with age, such as championship qualifying targets and the like, typically require M>1. However, on closer inspection such data is shown to be dominated by early-maturing participants and skewed by large decreases in participation as age increases. Since the invention sets out to improve participation, simple embodiments of u(t) with M=1 may be sufficient and more useful.

The reference model, u(t), can be calibrated via least-squares optimisation against some known or measured dataset denoted by f(t). There will likely exist a distinct f(t) for each sport, each event in that sport and for each gender. The optimisation procedure is via the minimisation of the error E, where,

Once again, convergence on an optimal solution for u(t) can be achieved using known procedures and is aided by the smooth monotonic nature of the performance versus age characteristics.

An important factor associated with embodiments of the invention is that certain aspects of athletic performance increase linearly or non-linearly with biological (rather than chronological) age. A prediction of individual performance may therefore be obtained by offsetting the mean performance model by the respective value of To obtained from the individual's model of growth as shown in Figure 12. Thus a predicted model of performance, v(t), may be constructed, where,

where for i=0 to M, av(/)=au(i) and for i=l to M,

Where the predicted performance is assumed to be a simple time-shifted version of the mean performance model, for i=l to M,

such that a solution to the model is found from, bJiLZ·1'1·1’··'"-’·1

Note that for M>1, one or more component functions may be related to the timing of an individual's pubertal spurt and so is preferably similarly time-shifted. Any component functions not related to the individual's pubertal spurt undergo no such time-shift. A modified solution to v(t) can be envisaged whereby the values of cv(i) are chosen, with or without modifications to bv(i), to compress v(t) such that u(t) and v(t) converge more rapidly in time. Generally, however, solutions of this type will converge at an age past the entry to 'open' adult competition where no account of biological age is taken and so will be of limited use. In this simple embodiment therefore no such measures are applied.

Notably once u(t) is known, then for measures of the class Pmin(t), the required complement can be formed by subtraction from v(°°), where,

Embodiments of the present invention may use a weighting w(t2,t2) that defines performance corrections for the differences in biological and chronological age. ti and t2 represent two nominal chronological ages for comparison. In many circumstances, such as when performance at a nominal age is compared with a mean, ti is equal to t2. However, as will be explained below, there may be situations where comparisons between two different ages may be required.

For performance measures of the class Pmax(t), w(t 1 ,i2)=v(fi)-u( f2)

For performance measures of the class Pmin(t), w(t1,i2)=u(t2)-v(i1)

For a performance correction/compensation at a time ti 20, such as a nominal age ti in a competition, t2=t2, such that for measures of the class Pmax(t), and for measures of the class Pmin(t) w(ti,ti)=u(ti)—v(ti)

The resultant correction, w(ti,t2), can then be added to the measure Pmax(ti) or Pmin(ti) as appropriate. For example, a compensated race time, Tc, is obtained from the recorded race time, TR, as,

Direct comparisons between individuals A and B can be expressed via w(ti,t2) since,

It is also possible to relate the performance of girls to boys, and vice versa. There will exist two distinct models for u(t) for each of the genders that may or may not be related by a simple time difference. Comparing performances between genders then requires using the appropriate mean performance reference in calculating the appropriate weighting. For example, comparing the performance of a boy to the mean performance of girls at a time ti,

And for comparing the performance of a girl to the mean performance of boys,

It is also possible to define the advanced pubertal spurt in girls relative to boys, TG, 17, as shown in Figure 9,

However, it should be noted as expressed above, that differences in the models of u(t) for girls and boys mean that using TG in isolation to compare girls and boys directly will likely lead to erroneous results - performance expectations of the two genders are different. A more valid comparison will be the weightings with the appropriate population mean.

A further application of the present invention is to predict performance improvements likely in the future at a time t2 from a performance measure at ti. Of particular interest is the predicted performance as an adult, where w(ti,t2) can be written with t2=°° . This is because characteristics ti,t2 converge into adulthood as the effect of differences in pubertal growth fall to zero. For measures of the class Pmax(t),

and for measures of the class Pmin(t),

The resultant prediction, w(ti,°°), can then be added to the measure P max(ti) ΟΓ P min(tl) 3S appropriate. For example, a predicted race time as an adult, T°°, is obtained from the recorded race time, TR, as,

Alternatively, the same results in weighing and/or predicting may be achieved by recasting the process and assigning each sporting discipline and event with a 'peak performance velocity' offset time Tp, where Tp=Tu(M)-Ty(N). The values of Tp may also be used in designing training programmes and the optimal timing of the exposure to certain elements thereof.

The weightings such as w(ti,ti) and w(ti,°°) as calculated above are sufficient for the basis of the programme element in the electronic device as shown in Figure 6. For the real-time elements such as shown in Figure 5, a weighting factor O(ti,t2), or the reciprocal thereof is preferably calculated. For the example of a the stopwatch shown in Figure 5, the counter value Ά' in Figure 4 is preferably multiplied by O(ti,t2), such that the output counter 12 is advanced or retarded as required.

For real-time measurements of an individual where compensation for the difference in biological age of that individual compared to some nominal chronological age, ti, is required,

Where real-time measurement requires compensating for the biological age at some nominal chronological age, ti, to some other nominal chronological age, t2,

So for a real-time prediction at some nominal chronological age, tx, for performance potential in adulthood can be observed using the counter multiplier ratio with t2=°°, such that,

In a further application, a nominal counter weighting may be applied where no account of the differences in chronological to biological age is required. In this case, the counter weighting compensates for performance based solely on nominal biological age for use in performance prediction. Such a normalised counter weighting, Q„(ti,t2), is defined as,

where for t2=°°, the real-time compensated measures will be representative, if not predictive, of a comparison those as an adult where,

The level of learned skill also presents a significant advantage in any sporting competition. Such skill advantages can be modelled as an orthogonal component to the power-based component described in the simple embodiment of the invention presented herein. Furthermore, as an orthogonal component, it can be compensated with an additional series component to w(tx,t2) or a multiplicative series component to Q(ti,t2). A model of skill acquisition can be adapted from adult performance results that are devoid of the distortions due to growth spurts.

An implementation of the invention is likely to adopt hard storage of the reference models y(t) and u(t) for each activity, together with the specific variable Ω and any intermediate function or variable used to calculate Ω. For example, referring to Figure 2, models, functions and variables may be stored in non-volatile memory 3 and/or additional temporary memory 4. Additionally or alternatively such data may be transferred via the wireless internet connection 5 and/or wired connections to a remote location. Remotely stored data and/or data stored in the memory 3, 4 of the programmable electronic device shown in Figure 2 may then be analysed and/or used for further calculation in real time and/or after the event has taken place. Data already stored may be

updated at the end of an event performed by the user. For example, a user may have a profile stored in the local memory 3, 4 or at a remote location. This profile may include history of previous events and data concerning their physical condition, e.g. one or more dimensions. The profile may be updated as and when the user completes an activity and records the results of that activity using the electronic device. The user's performance may then be used to update or adjust one or more variables or models linked to the user and may be stored alongside other data linked to that user.

Analysis of actual performances versus predictions made as time progresses allows adaption of the reference models and the subsequent calculation of v(t).

Claims

1. A device, comprising: an input operable to receive a time variant scaling value which represents a performance metric of the user; and a timing unit having a variable timing rate set to be proportional to the variable scaling value received at the input, wherein the timing unit is operable to measure a scaled time taken for the user to complete the whole or part of a physical activity, wherein the performance metric is an estimate of the physical maturity of the user based upon one or more physical measures of the user thereby mitigating against the effects on performance in the specific activity due to the difference between the specific biological and chronological age of the user.

2. A device according to claim 1, wherein the rate of the timing unit is further dependent on the type of physical activity that the user is undertaking.

3. A device according to claim 1 or 2, wherein the rate of the timing unit is further dependent on the event in the sport or type of physical activity that the user is undertaking.

4. A device according to claims 1 or 2 or 3, wherein the timing rate of the timing unit is time variant over the course of the event.

5. A device according to any preceding claim, wherein the performance metric of the user is estimated based on one or more physical dimensions of the user.

6. A device according to claim 5, wherein the one or more physical dimensions includes one or more of standing height, seated height, arm span and weight.

7. A device according to any preceding claim, wherein performance metric of the user is estimated based on measures of body tissue or blood lactate.

8. A device according to any preceding claim, wherein the performance metric of the user is estimated using the superposition of one or more Gompertz or logistic functions.

9. A device according to claim 8, wherein the performance metric of the user is estimated based on a model comprising the product of two or more superpositions of one or more Gompertz or logistic functions.

10. A device according to any preceding claim, further comprising a display operable to display the scaled time at the completion of or over the course of a physical activity.

11. A method of generating a scaled measure of the time taken for a user to complete the whole or part of a physical activity, the method comprising: receiving a time variant scaling value which represents the performance metric of the user; and measuring the time taken for the user to complete whole or part of the physical activity using a timing unit having a variable timing rate which is proportional to the received scaling value, wherein the performance metric is an estimate of the physical maturity of the user based upon one or more physical measures of the user thereby mitigating against the effects on performance in the specific activity due to the difference between the specific biological and chronological age of the user.

12. A method according to claim 11, wherein estimating the performance metric of the user comprises comparing measurements of one or more physical dimensions of the user taken over a predetermined period with a predefined growth model.

13. A method according to claim 11 or 12, wherein the performance metric of the user is estimated using the superposition of one or more Gompertz or logistic functions.

14. A method according to claim 11 or 12 or 13, wherein the performance metric of the user is estimated based on a model comprising the product of two or more superpositions of one or more Gompertz or logistic functions.

15. A method according to any of claims 11 to 14, wherein the physical measurements include one or more of the standing height, arm span and weight of the user.

16. A method according to any of claims 11 to 15, wherein the physical measurements include one or more measures of body tissue or blood lactate.

17. A method according to any of claims 11 to 16, wherein the scaling value is time variant over the course of the performed task.

18. A method according to any of claims 11 to 17, wherein the scaling value depends at least partially on the type of activity being performed by the user.

19. A method according to any of claims 11 to 18, wherein the scaling value depends at least partially on the event being performed by the user.

20. A device, comprising: a timing unit operable to measure the actual time taken for a user to complete whole or part of a physical activity; an input operable to receive a time variant scaling value which represents the performance metric of the user; and a scaling unit operable to generate a scaled time based on the actual time taken to complete the activity and the scaling value, wherein the performance metric is an estimate of the physical maturity of the user based upon one or more physical measures of the user thereby mitigating against the effects on performance in the specific activity due to the difference between the specific biological and chronological age of the user.

21. A device according to claim 20, wherein the rate of the timing unit is further dependent on the type of physical activity that the user is undertaking.

22. A device according to claim 20 or 21, wherein the rate of the timing unit is further dependent on the event in the sport or type of physical activity that the user is undertaking.

23. A device according to any of claims 20 to 22, wherein the rate of the timing unit is time variant.

24. A device according to any of claims 20 to 23, wherein the performance metric of the user is estimated based on one or more physical measures of the user.

25. A device according to claim 24, wherein the one or more physical measures includes one or more of standing height, seated height, arm span and weight.

26. A device according to claim 24 or 25, wherein the one or more physical measures includes one or more of body tissue or blood lactate.

27. A device according to any preceding claim, wherein the performance metric is an estimate of biological age and physical maturity of the user.

28. A device according to any preceding claim, wherein the performance metric of the user is estimated using the superposition of one or more Gompertz or logistic functions.

29. A device according to claim 28, wherein the performance metric of the user is estimated based on a model comprising the product of two or more superpositions of one or more Gompertz or logistic functions.

30. A device according to any of claims 20 to 29, further comprising a display operable to display the scaled time and/or the actual time at the completion of the event or over the course of its duration..

31. A computer program or a suite of computer programs operable to perform the method of any of claims 11 to 19.

32. A computer program or suite of computer programs operable to perform a method comprising: receiving a time variant scaling value which represents the performance metric of the user; and measuring the time taken for the user to complete whole or part of the physical activity using a timing unit having a variable timing rate which is proportional to the received scaling value, wherein the performance metric is an estimate of the physical maturity of the user based upon one or more physical measures of the user thereby mitigating against the effects on performance in the specific activity due to the difference between the specific biological and chronological age of the user.

33. A counter arranged to measure the relative physical performance of a user, wherein the rate of change of the counter is dependent on a calculated or estimated performance metric of the user, wherein the performance metric is an estimate of the physical maturity of the user based upon one or more physical measures of the user thereby mitigating against the effects on performance in the specific activity due to the difference between the specific biological and chronological age of the user.