Copious studies support the theory of evolution, which rests on the insight that natural selection leads to complex innovations.New computation models demonstrate the theory's limitations, however.

In their book, Introduction to Evolutionary Informatics, authors Robert Marks, Winston Ewert, and William Dembski seek to highlight the limits of Darwinian evolutiuonary theory.

No evolutionary process, they say, can produce entirely new outcomes like a new body plan.Albatrosses can develop different beak sizes, but the authors say it is for them to develop whole new forms that swim or have opposable thumbs.

Computer programs like Tierra and Avida can simulate broad-scale, biological evolution.This kind of software was highly anticipated when it was first introduced to researchers, but its results were disappointing to evolutionary biologists because their variability would always hit what's called Basener's Ceiling.

Basener's Ceiling is a point of convergence in a narrow range of results.The phenomenon is found in my fields, not just evolution.Ask the program to produce a smartphone and it will simulate a series of development stages that eventually end with a best smartphone design that optimizes all the features of all smartphones.

The end result is a smartphone, but the program simulates no further because there's no way for it to generate a completely different result.In other words, it cannot predict something that consumers would want instead of a smartphone.It can only produce the designated outcome.

In simulating the evolution of albatrosses, therefore, the outcome can only be a known albatross or an amalgam of known albatross features.

Marks, Ewert, and Dembski use this metaphor to argue that biological evolution depends on blind natural selection and, therefore, has no means to generate wholly new outcomes.In other words, they argue that a human being is a mathematically impossible outcome for the evolution of monkeys.