by Lael Ewy, MFA
In a recent essay in the statistics blog FiveThirtyEight, Walt Hickey attempts to use statistical analysis to learn something about the work of Bob Ross, whose show The Joy of Painting can still be seen on PBS nearly 20 years after his death. While the Hickey's piece does help us understand statistics, it does little to illuminate painting and is an object lesson in how statistics can muddle common sense when misapplied.
Hickey's use of the word “given” indicates that some certain image in a Bob Ross painting can be statistically correlated with another, showing conditional probability. Here's an example:
“The biggest pitfall people often face is assuming the two probabilities are the same. The probability that Ross painted a cloud given that he painted the beach — essentially, how many beach paintings have clouds — is (0.07)/(0.09), which is 78 percent. The vast majority of beach scenes contain clouds. However, the probability that Ross painted a beach given that he painted a cloud — or, how many cloud paintings contain a beach — is (0.07)/(0.44), or 16 percent. So the vast majority of cloud paintings don’t have beaches.”
In common usage, of course, “given” represents not a statistical correlation but a real relationship—one in which we reason deductively, not inductively. We might say “Given that car's charging system is working properly, the problem probably isn't the alternator.” So far so good: Hickey is just helping us understand how statisticians use the word.
But here's where things start to get sort of silly. Bob Ross is creating paintings, not taking snapshots. Since this is art, the first thing we might want to consider is Ross's aesthetic, which is representational. He presents lots of clouds because he's painting landscapes, and, interestingly enough, you kind of have to depict the sky in a landscape. And one way that people know it's a sky and not, say, blue water, is by putting clouds in it.
And while his paintings are distinctly representational, he's also presenting a Romantic idea along with his paintings, not just putting random objects together on a canvas: clouds are a much more pleasing aspect of, say, a mountain scene than they are of a beach scene. People associate good times at the beach with cloudless days; mountains are impressive because they're up among the clouds. In order for Ross to meet the expectations of the aesthetic situation he is presenting, he's going to be doing things this way as a matter of course. There's no need for a complex statistical analysis in this case; all it requires is a pretty simple aesthetic one.
Along these lines, if Ross is going to present things representationally, he is bound also to present, for example, more than one tree (since they tend to hang out together in forests) and more than one mountain (since they tend to hang out together in ranges). As Hickey puts it,
“What is the probability, given that Ross painted a happy tree, that he then painted a friend for that tree?There’s a 93 percent chance that Ross paints a second tree given that he has painted a first.
What percentage of Bob Ross paintings contain an almighty mountain?
About 39 percent prominently feature a mountain.
What percentage of those paintings contain several almighty mountains?
Ross was also amenable to painting friends for mountains. Sixty percent of paintings with one mountain in them have at least two mountains.”
Again, all this really reveals is that Ross was a representationalist, something patently obvious to anyone who has ever seen his show. It also may suggest that Hickey has never seen a forest or a mountain range. Perhaps someone should give him permission to back away from his spreadsheets for long enough to go visit some of these places—or at least look them up on Google Images. He might be surprised to find that they look a whole lot like the paintings of Bob Ross.
Hickey goes on to break down how often Ross painted certain types of clouds (“Given that there is a painted cloud, there’s a 47 percent chance it is a distinctly cumulus one.”), water (“About 34 percent of Ross’s paintings contain a lake, 33 percent contain a river or stream, and 9 percent contain the ocean.”), and, notably, cabins:
“About 18 percent of his paintings feature a cabin. Given that Ross painted a cabin, there’s a 35 percent chance that it’s on a lake, and a 40 percent chance there’s snow on the ground. While 72 percent of cabins are in the same painting as conifers, only 63 percent are near deciduous trees.”
This again fails to take into account the Romantic aspect of Ross's work. Cabins on lakes and cabins in the snow are simply more pleasant from a Romantic point of view than cabins, say, surrounded by the dirt and grime of the city or, for instance, cabins engulfed in flame because they've just been bombed by the Luftwaffe.
Hickey goes on to speculate about how these things showed what Ross did or didn't like in a painting. One may presume that a person's overall aesthetic approach is based on personal preferences, but there are also a set of expectations within that aesthetic approach that help guide the artist. “Like” isn't a terribly deep description of an aesthetic approach, but once that approach is understood, it's generally predictive. A statistical analysis works the opposite way: by trying to figure what is generally predictive from looking at masses or classes of individual instances. This can be quite useful, but it sort of misses the point when it comes to painting and art.
Hickey was so interested in this idea of figuring out what Ross liked that he went and asked Ross curator and business partner Annette Kowalski about his work. One the one hand, this is good: Hickey realized the limits of what his data could actually reveal. Kowalski noted that what's striking about the work from Ross's show was what they largely left out, which was people.
Here's where we look at the other hand. First, we can go back to Ross's clearly Romantic aesthetic: being surrounded by people rather than happy trees and majestic mountains and such is a total downer from the Romantic point of view. Wordsworth didn't get all poetic about bankers and bricklayers; because he was a Romantic, he trended toward daffodils buffeted by breezes.
But even though he mentions it, perhaps the most obvious thing Hickey and Kowalski seem not to be taking into account is that Ross's show was a teaching program, in which he created a fully-realized oil painting in a 30 minute format, and almost in real time. In other words, Ross is going to choose the forms that a novice painter could easily replicate and that he could easily demonstrate within the confines of the program. He may have avoided people simply because figure painting is notoriously tricky, expensive, and time-consuming, which is why they have separate classes on it in art school. The show was, after all, the Joy of Painting, not What a Pain in the Butt It Is to Paint People. Further, figure painting often requires live models, frequently nude ones, which adds a layer of logistics and cost, and which would have been difficult to get by PBS censors.
What the statistics don't show, then, are the more-or-less obvious aesthetic, pedagogical, temporal, and financial realities Bob Ross faced.
As helpful as statistical analysis can be, it's misused if it's a substitute for common sense. It's also ill-used if applied by people who simply don't understand their subjects and the principles by which those pursuing those subjects act. These principles are generally obvious to those who understand the subject matter, or they are clearly stated by the people engaged in them. Statistics might reveal when a person is being deceitful—for example when a politician's voting record, statistically, doesn't match her rhetoric. But there's no indication that Bob Ross was doing anything other than painting joyfully, and with his audience squarely in mind.