Gerhard Richter, randomness and knitting

I went to see an exhibition of Gerhard Richter’s paintings 4900 Colours: Version II.

He has taken 4900 squares in 25 different colours using a computer programme to arrange them at random into 49 paintings of 100 squares. As a statistician I was intrigued to investigate if the randomness were in fact true. If it were you would expect to see several squares of the same colour next to one another, and I did indeed find a number of examples with two identical squares next to one another, both horizontally and vertically. There were some, but fewer, occurrences of three together, and I saw none of four. I will not try to calculate the probability of four identical squares next to one another; it will be larger than 0 and it need not be very small for the likelihood of it not occurring in a sample of 4900 not to be significant. It was interesting, nevertheless, that it did not occur.

When I knit my blankets my main interest is the randomness of the colours, as seeing what colour comes next makes the knitting so much more interesting. As Gerhard Richter I select the colours before I start – he took 25 colours, and I take about the same number, usually at least 25 but often 30 – 35 or more. I am not too concerned about arranging my colours randomly. An even distribution will produce a more attractive product. For me the main motivatgion is that I should not have to make a decision of what colour to use next. I arrange my colours in order, any order, and take the one that comes next. Gerhard Richter uses a computer programme to arrange his. Mine is the simpler method. The objective is the same.