"The theory of evolution says that life originated, and evolution proceeds, by random chance."
This is invariably phrased as "It's just like expecting a tornado to go through a junkyard and assemble a 747."
There is probably no other statement which is a better indication that the arguer doesn't understand evolution. Chance certainly plays a large part in evolution, but this argument completely ignores the fundamental role of natural selection, and selection is the very opposite of chance. Chance, in the form of mutations, provides genetic variation, which is the raw material that natural selection has to work with. From there, natural selection sorts out certain variations. Those variations which give greater reproductive success to their possessors (and chance ensures that such beneficial mutations will be inevitable) are retained, and less successful variations are weeded out. When the environment changes, or when organisms move to a different environment, different variations are selected, leading eventually to different species. Harmful mutations usually die out quickly, so they don't interfere with the process of beneficial mutations accumulating.
Nor is abiogenesis (the origin of the first life) due purely to chance. Atoms and molecules arrange themselves not purely randomly, but according to their chemical properties. In the case of carbon atoms especially, this means complex molecules are sure to form spontaneously, and these complex molecules can influence each other to create even more complex molecules. Once a molecule forms that is approximately self-replicating, natural selection will guide the formation of ever more efficient replicators. The first self-replicating object didn't need to be as complex as a modern cell or even a strand of DNA. Some self-replicating molecules are not really all that complex (as organic molecules go).
Some people still argue that it is wildly improbable for a given self-replicating molecule to form at a given point (although they usually don't state the "givens," but leave them implicit in their calculations). This is true, but there were oceans of molecules working on the problem, and no one knows how many possible self-replicating molecules could have served as the first one. A calculation of the odds of abiogenesis is worthless unless it recognizes the immense range of starting materials that the first replicator might have formed from, the probably innumerable different forms that the first replicator might have taken, and the fact that much of the construction of the replicating molecule would have been non-random to start with.
(One should also note that the theory of evolution doesn't depend on how the first life began. The truth or falsity of any theory of abiogenesis wouldn't affect evolution in the least.)