What is Random?
Hey, Vsauce. Michael here and Derek. Generate(!) 78? That's so random. Or is it? What does it mean to be random? Can anything really be random? What's the most random thing ever? Today let's stop being random and become 'ransmart'.
If something is unpredictable, and contains no recognisable patterns, we call it random. So let's begin our hunt for the most random thing with a coin toss. The Australian 50 cent coin is one of the largest coins currently in circulation. Coin flips and rolling dice are not intrinsically random; they are only random because of our ignorance. If we could know every initial condition, the exact forces and properties that play for a particular flip or roll, we could theoretically calculate the result before it even happened.
And, sure enough, researchers have built coin flipping robots that can precisely control a flip to get the result they want 100 percent of the time. So here is our question. Is there anything you couldn't predict, even if you knew everything? A process determined by nothing? And how can you be sure there aren't any patterns in what you're looking at? Maybe you just haven't looked for the right pattern yet. Or maybe you have already seen true randomness but didn't know it because you didn't look for long enough.
As if protected by a sort of camouflage, a random process can, and will, occasionally produce patterns. YouTube URLs are pretty much random. A unique one is made for every uploaded video, but sometimes, like the proverbial monkeys typing away on typewriters, YouTube generates a URL that contains, by chance, a word. The official music video for 50 Cent's 'In Da Club' contains a synonym for bottom. This video was assigned 'hello', there are some 'sexies' and a sauce.
If you want to find out if a particular word has been randomly assigned to a video yet, you can search for it by using this string in Google. The point is, randomness is difficult to identify. It is easier to be certain that something is not random than that it is. But despite this elusiveness, something interesting is going on right now. Increasingly, we, especially young people, are calling clearly predictable things random. Like randomly running into your best friend at a popular restaurant. Or hilariously bizarre combinations of things that we call totally random because they are seemingly unrelated, even though, of course, they were chosen not in random but in a very determined way because they are all unrelated.
Those guys that showed up at the party last night you didn't know, they weren't randos in a mathematical sense. They knew about the party, were in the mood to go to a party, and were in the area. Pretty predictable, actually. This non-statistical use of the word random annoys some people, but it's not that far off from the original meaning of the word. In the 1300s, random meant running or at great speed. Later, it would be used to describe things that have no definite purpose.
It wasn't until the 1800s that random took on a particular mathematical definition. Then in the 1970s, MIT's student paper popularised the use of the word random to simply mean strange. Of course, just because something is strange doesn't mean it has no discoverable cause. Why have we started calling so many predictable things random? Well, many theories revolve around the amount of information and new people we are confronted with at an increasing rate; now more than ever before.
Perhaps it's just easier, almost a bit of a relief to call things random, so that we can move on to synthesise other information. Take a look at this die. Now, as you can see, it rolls a five most of the time, but not all the time. Over time, a sequence of its results will contain less randomness. But it is still random. Any face is possible and I have no guarantee beforehand of knowing which face will show up. The outcomes I'm selecting from make the resulting sequence less full of randomness, but the process is still random.
Even though a die and a coin are extremely sensitive to their initial conditions and, over the course of normal use, are quite unpredictable, they do, over time, exhibit certain biases. Biases that make them a bit more predictable and a little less random than you might think. First of all, dice. Even precision dice are only quality controlled within a few micrometers. You can check this out yourself. Make two stacks of ten or more similar dice. Now, if you orient each die every which way, you should pretty much get two equally sized stacks.
But, if you arrange each stack along a shared axis, so that every die faces the same way, any regular imperfections caused by the manufacturing process may become visible. But what about coins? Well, some fantastic research has been done on what happens when they spin and flip. For instance, it has been found that US nickel is just the right diameter and thickness to wind up landing not heads up or tails up when flipped, but on its side, about once every six thousand times it's flipped.
But what about the fairness of flipping a coin? Well, if you flip a coin like this, for statistical and physical reasons, the side facing up before the flip begins doesn't actually have a fifty percent chance of being the result. Instead, as researchers at Stanford have found, it actually has a fifty-one percent chance. In this case, it was the other side though. Still pretty random. If you want a coin flip to be as fair as possible, you should just catch it in your hand. Don't allow it to hit the ground, bounce, tumble and spin.
That's because researchers have found that when a coin spins larger biases come into play. The shape of its edge, its centre of gravity. The heavier side tends to go down quite often. In the case of some coins, as often as eighty percent of the time. It's been found that a one euro coin will spin and land heads up more often than not, and a US penny will land tails up more often than not when spun. But like I said earlier, theoretically, if we knew everything about the initial conditions of a coin flip or a die roll, we could calculate beforehand their outcome.
Why don't we do that more often? Well, it's extremely difficult. Insane amounts of precision would be required because the smallest difference between two initial conditions can be magnified over time, leading to chaotic, extremely difficult to predict, results. Random.org, the service Derek and I used at the beginning of this video to generate a random number, uses atmospheric noise. It's extremely hard to predict, but technically still a deterministic system. All that noise came from somewhere and if we could just find out those initial conditions, we could, theoretically, predict their outcomes.
If we want a system more random than that, we will need to find one that is determined by nothing, and for that, let's look closer. Quantumly close. Quantum mechanics may have our answer. It describes the properties of quantumly sized things as probabilities; just chances. Not because we don't know enough yet to be certain, or predict, but because, well, the idea is there's nothing there to predict. There is no beforehand we could know.
Whether or not a particular individual radioactive atom will decay or not, or whatever the spin of an electron is, is only knowable once we look. They're determined by a deep-seated randomness woven into the universe itself. Einstein couldn't believe this. He refused to accept, as he said, that "God played dice with the universe." But... experiments with entangled particles have shown violations of Bell inequalities.
Entangled particles are particles that exhibit similar properties even when separated by large distances. Now, if they agreed on those shared properties to have, or are somehow determined beforehand to have them, their behaviors should satisfy Bell's famous inequalities. But experiments have found that instead the likelihood of what a machine will see when measuring one particle determines how the other machine will measure the other particle. It is here, when we look, that the chance is determined.
Explanations for this are even weirder, but what the results suggest is that the chance of seeing particular quantum qualities don't pre-exist. They happen when you look. So, if you are ever feeling boring or predictable, just remember that you are made out of octillions of quantum probabilities. Dice that don't tumble in any analysable way we could ever predict. They are the most random thing. God may play dice with the universe, but they are the best dice in the universe.
And as always, thanks for watching. But what does it all mean? Well, true randomness doesn't mean anything. I mean, for us to have meaning, we need structure, predictability, and that is what I'm exploring over on my channel, Veritasium. Wow, okay. So let's go over there and take a look at what is not random. Veritasium, let's go. I'll see you guys over there. It's going to be awesome, come with us. And as always, thanks for watching.