Evolution takes millions of years to get much of anywhere, except with bacteria and such. A bit off-topic, but when people screw up evolutionary theory...it bugs me.
Newer theories on evolution state that it happens in leaps and bounds, typically due to cataclysmic events, or mass migrations. E.G. a small population of an existing species has a mutation where it can digest some fungus, which the rest of the population cannot. Then a disaster occurs, killing most of the species food source, except that fungus. So, the only surviving members of the species are the ones who can digest that fungus. Their offspring are likely to have that same mutation, and of course, those who don't have it will die. In a couple generations, that population is comprised only of the members with that mutation.
The older theories on evolution being slow seemed to rely on some magical genetic knowledge of what is "better". Evolution is a product of random mutation, coupled with some reason why that mutation is seperated, either by death, or migration. At that point you have a new species.
If you want more than a single little mutation, you need to wait a while. Whether you move a foot an hour or an inch every 5 minutes doesn't matter if you need to go half a mile.
Still despite that, due purely to the number of dwarfs in worldgen and the small time scale, it would be very rare to play in a world where evolution was visible.
I wouldn't say it'd be rare, domensticating animals or selective breeding could change things in term of decades or centuries, depending on creatures. Environmental changes can drive it a lot through selective pressure, though it's more likely to lead to extinction.
And that's not considering microorganisms that's got it's generations in hours or less 
While those aren't really good example of long-term evolution, it's still similar pressures and changes.
http://en.wikipedia.org/wiki/Domesticated_silver_fox
This is pretty notable, at least to me, for half century of breeding something wild into tame.
On other hand, you could come up with very different strains of crops within decades with dedicated effort and enough of variety of related plants to work with, as I believe have been done with tomatoes and corn.
Aside from breeding, which is different than natural evolution and not what I was talking about, any kind of genetic change would probably need more than a thousand years to get much past normal variation and therefore worth simulating.