a list of badly structured websites
You must be joking, right? What kind of need is it? If you are looking for some issues in a badly formed website, design one!
which is better; String based wrappers, or tree based wrappers?
Depends on scenario to scenario. Both have their own advantages and disadvantages.
Have a look at this PDF for comparison:
Expressive Power of Tree and String Based Wrappers[
^]