Essential Randomness
Keywords: randomness, uncertainty
A lot of what goes on in pseudoscience is the result of people finding patterns where none exist. I think the concept of randomness and how we can find illusory patterns in it warrants a little bit of an exploration here for anybody who hasn't thought about it.
Randomness is one of the hardest things for people to understand. Even those of us who pay lip service to the role of randomness in the universe have trouble with the concept. We're just too good at seeing patterns. We see them everywhere, even when no template for the seen pattern exists. See my recent post about the nebula that looks like a hand.
We typically expect anything that's random to show absolutely no pattern. We think random things should resemble an undecipherable mess. But that's not the case. Patterns in randomness are everywhere, and they can fool you more easily than you think.
I performed a classic experiment by creating a computer program that draws random dots on the screen. I ran it a few times and picked my favourite examples to show you here:
A lot of what goes on in pseudoscience is the result of people finding patterns where none exist. I think the concept of randomness and how we can find illusory patterns in it warrants a little bit of an exploration here for anybody who hasn't thought about it.
Randomness is one of the hardest things for people to understand. Even those of us who pay lip service to the role of randomness in the universe have trouble with the concept. We're just too good at seeing patterns. We see them everywhere, even when no template for the seen pattern exists. See my recent post about the nebula that looks like a hand.
We typically expect anything that's random to show absolutely no pattern. We think random things should resemble an undecipherable mess. But that's not the case. Patterns in randomness are everywhere, and they can fool you more easily than you think.
I performed a classic experiment by creating a computer program that draws random dots on the screen. I ran it a few times and picked my favourite examples to show you here:
The dots don't seem to be random, do they? They seem to be clustered together in places, sometimes forming shapes... and there are a few areas where there are no dots at all. It's not the even distribution that people generally expect out of randomness.
There are a few different ways you could organize this image into patterns. If you're like most people, though, you've probably already noticed a number of shapes, lines, clusters, and mysterious voids (blank spots).
Here's my interpretation:
There are a few different ways you could organize this image into patterns. If you're like most people, though, you've probably already noticed a number of shapes, lines, clusters, and mysterious voids (blank spots).
Here's my interpretation:
- Areas 1, 4, 8, and 12 seem to form lines that are all pointing in the same general direction. Mysterious, no?
- Areas 2, 6, and 13 are mysterious voids. If this is completely random, then why do the dots appear to be avoiding these wide open spaces?
- Areas 5, 7, and 10 all form clusters of dots. Is there some mysterious attracting force in these areas?
- Area 3 is a Llama, and I'm standing firm on that one. Any of you who think it's a misshapen chair are clearly wrong.
- Area 5, in addition to being a close cluster, also resembles a garden spade.
- Area 9 is a wine glass
- Area 11 is a balloon on a string.
(These shapes could be characterized as large clusters as well)
Well, clearly this is all a bunch of silly fun. But there is a point to this, and it's about more than just our ability to find a Llama in random dots or, similarly, to find a dipper in the stars.
Think about situations in which people place dots on a map in order to track certain events.
During World War 2, the Germans aggressively launched their V-1 buzz bombs and V-2 rockets towards London, England. These weapons were not extremely accurate. They were able to consistently hit the general London area, but they were not able to target any specific part of London.
Nevertheless, the resulting map showed certain areas of London to be extremely hard hit (clusters), while other areas seemed little affected or not affected at all (voids).
Many Londoners felt that there were certain parts of the city that were safer than the others. But was this really the case? Would Londoners have been any safer hiding out in these void areas under continued bombardment?
The answer is no. No one place on the map of central London was any less likely to be hit than any other.
But it's a powerful illusion, isn't it? Imagine standing in a pristine neighbourhood and knowing that absolute devastation was evident just a few blocks away in most directions. Would it not seem like the place where you were standing was somehow protected?
If you were a suspicious person you might suspect that there was a Nazi spy in the neighbourhood that was being protected by the Germans. I don't know if any such suspicions came to light.... but it's interesting to think how a random situation could be interpreted in sinister ways.
One way that such clusters and voids have been interpreted in sinister ways is in looking at maps of cases of certain diseases. It is not unknown for people to notice a clustering of cases of a disease in their community and to then go looking for a reason.
And they usually find a culprit, and usually with very little evidence to support the accusation. But to the people in the community looking at the data, it seems scary and convincing. People have a hard time believing that such things could possibly be the result of random chance.
But it's very clear that such randomness does play a part in things like V-2 Rockets and disease cases. It's not hard, or rare, to find patterns in random events.
I played around a little more with my random dot generator, modifying the number of dots generated. I got many interesting results. Take a look at these:
- Areas 2, 6, and 13 are mysterious voids. If this is completely random, then why do the dots appear to be avoiding these wide open spaces?
- Areas 5, 7, and 10 all form clusters of dots. Is there some mysterious attracting force in these areas?
- Area 3 is a Llama, and I'm standing firm on that one. Any of you who think it's a misshapen chair are clearly wrong.
- Area 5, in addition to being a close cluster, also resembles a garden spade.
- Area 9 is a wine glass
- Area 11 is a balloon on a string.
(These shapes could be characterized as large clusters as well)
Well, clearly this is all a bunch of silly fun. But there is a point to this, and it's about more than just our ability to find a Llama in random dots or, similarly, to find a dipper in the stars.
Think about situations in which people place dots on a map in order to track certain events.
During World War 2, the Germans aggressively launched their V-1 buzz bombs and V-2 rockets towards London, England. These weapons were not extremely accurate. They were able to consistently hit the general London area, but they were not able to target any specific part of London.
Nevertheless, the resulting map showed certain areas of London to be extremely hard hit (clusters), while other areas seemed little affected or not affected at all (voids).
Many Londoners felt that there were certain parts of the city that were safer than the others. But was this really the case? Would Londoners have been any safer hiding out in these void areas under continued bombardment?
The answer is no. No one place on the map of central London was any less likely to be hit than any other.
But it's a powerful illusion, isn't it? Imagine standing in a pristine neighbourhood and knowing that absolute devastation was evident just a few blocks away in most directions. Would it not seem like the place where you were standing was somehow protected?
If you were a suspicious person you might suspect that there was a Nazi spy in the neighbourhood that was being protected by the Germans. I don't know if any such suspicions came to light.... but it's interesting to think how a random situation could be interpreted in sinister ways.
One way that such clusters and voids have been interpreted in sinister ways is in looking at maps of cases of certain diseases. It is not unknown for people to notice a clustering of cases of a disease in their community and to then go looking for a reason.
And they usually find a culprit, and usually with very little evidence to support the accusation. But to the people in the community looking at the data, it seems scary and convincing. People have a hard time believing that such things could possibly be the result of random chance.
But it's very clear that such randomness does play a part in things like V-2 Rockets and disease cases. It's not hard, or rare, to find patterns in random events.
I played around a little more with my random dot generator, modifying the number of dots generated. I got many interesting results. Take a look at these:
Doesn't that look like there's some sort of dot generating element in the left corner somewhere that's spewing the dots out in a nice fan shape like perhaps a snow blower?
Well, there isn't. It's pure random.
How about this one:
Well, there isn't. It's pure random.
How about this one:
I want you to take a look at this picture and tell me it's not a long tailed giraffe walking towards a tree branch with a single leaf on it.
Of course, I kid. There was no bias written into my program towards drawing giraffe's. The shape of a giraffe that I've outlined comes purely from my imagination.
The fact that I was able to discern a pattern from randomness does not mean that there's anything to that pattern. I think that fact should outline a clear lesson here:
The next time you think you can connect the dots in order to come to a firm conclusion, stop and consider it further. Are you really onto something? Or is your inner dumbass just finding patterns where none exist?
The fact that I was able to discern a pattern from randomness does not mean that there's anything to that pattern. I think that fact should outline a clear lesson here:
The next time you think you can connect the dots in order to come to a firm conclusion, stop and consider it further. Are you really onto something? Or is your inner dumbass just finding patterns where none exist?
