A random deterministic process

People often look at me with disbelief in their eyes when I start going on about determinism in random events. I'm not exactly certain why people at large have such a difficult time understanding that a series of random events can amount to a measurable and predictable outcome.

Let's take evolution as an example. There are many reasons as to why it's deterministic because there's obviously many similiraties between living organisms. For example all vertebraes have a heart and such. Why? Well because the heart is the ultimate blood pumping device for this kind of animal. Why does everyone have ears based on the same principles (thin vibrating membrane and so on)? Because it just works.

But then, if evolution is random as so many people tend to believe, how is it that time and again it comes to the same solutions to similar problems? Wouldn't a completely random process create MUCH more variation? Well yes, so evolution cannot be random.

But isn't it that mutations are random you say? Well yes, they are. So how is it that evolution can be deterministic then?

It's really quite simple. Evolution isn't a process, it's actually two intetwined processes. There's mutation, which is random and then there's natural selection, which is selective. Together they form a process that through random events reaches a perfect solution.

Because people still had a hard time understanding me I created a short algorithm that displays a dumbed down process of evolution on real numbers. It's based on units with ten children each. The way it works is actually fairly simple. It starts with a single unit that is a number 1. Then produces its ten children by applying adding a random number between -0.300 and 0.300. There is a 50% chance of this delta being negative. Once the ten children are produced it applies the same "breed" function to those that are nearer the ideal number (1000.000 in my case). This is to display natural selection which favors those that are better to those that are worse. Within 6600 generations this algorithm became such that all still living units were within 0.3 from the ideal and if I kept it running for long enough the population would probably die out as it couldn't improve anymore. Real evolution solves this by always changing the ideal.

So you see, a random process, by help of a selective process, can easily become deterministical and move towards a "goal".