INNACURATE MARBLES

The world is full of imperfections, even if that world is a factory producing parts in series. It is impossible to produce 100% identical parts, so the question becomes how much imperfection can be tolerated in the manufacturing process.

It is up to engineers, using statistics, to understand the margin of error of the parts produced, what the acceptable level of error is and what can be done to adapt one thing to another.

We do not know the process of making marbles and we know that hardly anyone will care about the slightest differences between the diameter of one and the other, but it is interesting to observe how these variations are distributed.

In the video below, we show a way to understand this distribution! If the sticks are well positioned, when we release the marbles from the highest part of the sticks, we will see each marble falling at the point where the distance between the sticks is greater than their diameter.

It is interesting because, like most of the quantitative variables found around the world, the diameter of the balls is distributed under a Gauss curve, meaning that there are very few balls with very small or very large diameters, while most of them have a diameter median.

This phenomenon can also be seen in the heights, weights and blood pressures of a certain part of the population, in the time spent by a certain group of students to take a test, in the duration of a certain light bulb, among other examples.

Image: Valéria Guimarães

Since the margin of error is very much seen in the difference in the size of the marbles, this Wikipedia page is super detailed and contains approximations, formulas used in this type of calculation, related concepts, among other things.

This page of the Institute of Mathematics, Statistics and Scientific Computing at Unicamp also talks about margin of error, presenting the content through podcasts! In addition, ideas are offered on how to show this theme in a class.

The main idea of this piece is to show the normal distribution, so this Wikipedia page talks about exactly that, in an extremely detailed and complete way, showing applications and properties, as well as making connections with other distributions.

This page of the Municipal University of São Caetano do Sul also talks about the normal distribution, in a slightly more visual way, with some graphs and examples to clarify some of the concepts present in a study like this.