I found a scientific (Department of Systems and Biology at Harvard Medical School) blog about evolution the other day. It was an interesting read. The post was a rebuttal of an argument (irreducible complexity) against evolution that is part of the foundation of Intelligent Design.

Let me be clear that I am not a scientist or a mathematician. I am an engineer, and we deal in simplifications. As I read the post this statement struck me:

“…try to get your head around over 3 billion years of a meandering multibranched process of accident, failure and occasional success.  In particular, it’s hard to keep in mind that the advantage that leads to increased evolutionary success can be very very small; often so small that it’s extremely hard to measure.”

The thing is, trying to get my head around something complex to turn it into an algorithm is what I do. So I started to think about it and I realized 3 billion years just isn’t that big a number, ~95 quadrillion (10^15) seconds. For the record the estimated probability for monkeys typing Hamlet is 10^360,783 to 1 (that’s a big number!). The reason that number is so big is because when you stack probabilities they are products not sums and they get ridiculously large pretty fast.

3 billion is ~1/4000th of the US GDP.

3 billion is about 1/20th of Bill Gates personal fortune.

3 billion years is only 750 thousand times recorded human history.

3 billion is only 666 times the period that humans have been on the planet.

To get a feel for the scale of the value 3 billion consider that 3 billion is roughly the number of calculations a typical PC is capable of making in about 1.5 seconds (assume 2 Ghz clock and single clock cycle instruction budget for simplicity). This is relevant because if you are reading this you have a feel for how long your computer takes to start up and within an order of magnitude you now have a feel for how many operations it takes to just get your computer ready to use.

This is relevant because the complexity of this time limited, “meandering multibranched process of accident, failure and occasional success” cannot be significantly less than the complexity of starting a computer (one would assume it is orders of magnitude more complex).

But wait we have 95 quadrillion seconds right? Except that we need to consider the time scale of an evolutionary operation.

My understanding of the process is that some small evolutionary change occurs randomly in a population. This change produces a slight survivability advantage in the organism which eventually causes the distribution of that feature in the population to replace the original population (the speed of this replacement being somewhat proportional to the strength of the advantage). This advantage, it is noted in the blog, may be very very small which (it seems to me) would require a very very long time for population replacement (ever run a life sciences simulation?).

But wait there’s less…

As it turns out the real action didn’t start until the last 20% of that 3 billion years! What that means is it wasn’t until the Cambrian age a mere 600 million years ago that these permutations began in earnest.

So I think I can get my head around 3 billion and it doesn’t pass the sniff test. There just isn’t enough time.

Around the time I became a Christian Lucy was discovered. She was a big deal back then because the charts on hominid development had to be changed and what I had been taught as scientific fact in high school turned out to be just scientific theory. Not only that but I’ve also learned that when new discoveries disrupt existing theories (on which the current academics have built their careers) they are not so warmly embraced for expanding our knowledge as one might think. Apparently we have not evolved very much past teaching that the sun revolves around the earth if that’s what the prevailing wisdom happens to be.

Postscript: I kept thinking about this and did a little more digging (yuk, yuk). As it turns out the time frame we are talking about is around 200 million years (the first mammals). The process of developing humans from the first mammals had to occur in this period. Recently some archeologists claim to have found human remains dating back 400,000 years. If we assume the evolutionary cycle (which has apparently been stable for that long) is on that order of magnitude then we only have enough time for 500 evolutionary cycles in that 200 million year period. Since 500 changes is a fantastically low number the only other explanation to fit the evidence is that the rate of change must have started very fast (short cycle periods) with a fairly rapid reduction of changes and a large expansion of the period for each cycle. A logarithmic function could model this kind of behavior where there is a steep rise that rapidly changes to a long slow increase. With the appropriate parameters this approximates a step function.

The thing is I’m not sure this model (based on the known parameters and very forgiving of the concept of a random walk) is really so much different from creation. If you have a better way to make the data fit I’m all ears.