If there was a fully random change within all the genes, more can wrong than right, since even important systems would be free game for bad choices. Take any enzyme and tell me how many ways to make it worse and how many ways to make it better? Which has more options?
It is "free game" and "open," but in more ways than you realize. Evolution has a guiding hand which often takes care of "wrong" choices (note that there really is no "wrong" or "right" in evolution, I'm using the terminology to simplify the discussion)--we call it Natural Selection. Individuals (and their collective genetic package) which acquire lethal or too many detrimental mutations have natural selection selecting against them, and hence are less likely to be able to reproduce and pass their detrimental mutations onto future generations. So long as natural selection applies pressure to shape the gene pool of the population through the generations, your supposition is basically moot. In fact, because of natural selection and other factors at play in evolution, this creates a standard for surviving organisms to be well adapted, and thus increasingly successful, at what they do, which are all things reproduction, survival, and transmission and multiplication of their "fit" genes. The end result is a population of individuals carrying genes with mutations/variation who have more right than wrong.