Talent Vs. Training

Talent in literature review

Simple repetition is a necessary condition for progress but a great deal more is necessary as well. Mastery in a field requires innate ability, an eagerness to work and doing a great deal of very laborious work. It is also believed that heredity played an important role, with successful people often having successful ancestors.

There are ways of measuring performance in a wide variety of fields, ELO ratings for chess players, Watkins-Farnum for music sight-reading, GPA for students. None of these are direct measures of talent.

To measure Talent accurately is almost impossible task. Talent is a qualitative attribute of human beings and there are debates to quantify it for the purpose of its measurement.

For the purpose of this analysis, our proxy for talent will be Learning Rate. While one cannot say exactly how talented a person with an 75% learning rate is, it is easy to argue that they are more talented than another person with a 85% learning rate. Faster learning suggests more talent.

There are two views that may or may not be in conflict :

• One states that, with enough hard work, around 10000 hours of focused practice, that anyone can be good at anything.
• The other suggests that everyone has some special talent, and that it is just a matter of finding it.

The idea that everyone has talent was put forward by Buckingham and Clifton after examining over 2 million surveys gathered by The Gallup Organization over a 30 year period they posit that everyone is capable of doing something better than the next 10,000 people.

Maxwell combines these ideas with the speculation that if one’s ability is measured on a 1 to 10 scale, that whatever a person’s initial score on a given ability, effort can raise that score by 2 points. If you initially score a 7 in a particular area, effort will allow you to rise to a 9.

Maxwell claims that effort in the right area will take you from being better than the next 10,000 people to being better than the next 100,000. The question then becomes how much effort?

Coyle also argues that deep practice is not the same as simple repetition. Deep practice occurs at the edge of one’s ability, not in the routine performances of a task. Pilots learn how to fly in bad weather by flying in real of simulated bad weather, not by flying on warm sunny days. They learn by trying, failing, and trying again.

This argues that it is not simply the accumulated hours of practice but also the nature of practice in those hours.

Learning Curves and Adaptations of the Model

The basic learning curve model as it exists in modern text books looks like the following:

y = a * ( x ^ b )

Or

The vertical axis in a learning curve represents time necessary to produce the nth unit, with faster production times being preferred to slower times.

For this model, we are recasting the speed of production to an arbitrary general measure of performance ability. The Performance Score is structured with P1 = 100 and lower numbers being preferred.

The learning rate will serve as a proxy for an individual’s level of talent.

One other change has been made in the presentation of this model. The usual horizontal access is in units produced. Instead of casting this in hours practiced, we will cast this in the passing of calendar time. This is an important difference in our representation of the data. This allows us to compare the progress that subjects have made over comparable time periods when the effort expended in those time periods is not the same.

Simulations and Comparisons

In order to examine the relationship between talent and effort, we will simulate learning by students with differing levels of talent and differing practice habits.We will assume that the average learner learns at an 80% rate. Based on the findings of the researchers, we will use 72% as the learning rate of a very talented individual and 86% for a relatively untalented individual.

Simulation : Equal Talent with Unequal Time

The model : y = ax^b

Now, equal Talent is 80% or .8

b= log(.8)/log(2)
b
## [1] -0.3219281        

Therefore, b is -0.3219281.

Case 1

a = 100 and x= 20 ( i.e.?20 hours of practice / week)

y= 100 * 20^-0.3219281
y
## [1] 38.12079        

Therefore, Performance Score attained by the nth hours of practice is 38.12079.

Case 2

a = 100 and x= 45( i.e.?45 hours of practice / week)

y= 100 * 45^-0.3219281
y
## [1] 29.36192        

Therefore, Performance Score attained by the nth hours of practice is 29.36192

Case 3

a = 100 and x= 90( i.e.?90 hours of practice / week)

y= 100 * 90^-0.3219281
y
## [1] 23.48954        

Therefore, Performance Score attained by the nth hours of practice is 23.48954

So, the conclusion is, if the employees are in the same Talent level, one who has more training, takes lesser time to perform the nth unit or gets lesser Performance Score attained by the nth hours of practice.

y1= 100 * (52*20)^-0.3219281
y1
## [1] 10.68396
y2= 100 * (52*45)^-0.3219281
y2
## [1] 8.229145
y3= 100 * (52*90)^-0.3219281
y3
## [1] 6.583316        

At the end of a year, the student who practices 20 minutes a week has accumulated 1020 minutes of practice and has a performance score of 10.68. The student who has practiced 90 minutes each week has accumulated 4680 minutes of practice time with a performance score of 6.58. Note that all students on this graph have the same amount of talent. All three groups asymptotically approach the same level of performance. The time it takes to arrive depends only on the accumulated experience.

Simulation : Equal Time with Unequal Talent

Given the well documented existence of different learning rates, it seems obvious that equal amounts or work will not lead to equal performance scores.

The model : y = ax^b

Case 1

Let’s look at an Employee who is very talented. This employee would have a talent score or learning rate of 72%, but would only be working 20 minutes per week. This would yield a Performance Score of 3.716497 after a year of practice.

b= log(.72)/log(2)
b
## [1] -0.4739312
y= 100 * (52*20)^-0.4739312
y
## [1] 3.716497