Originally by Jim Propp; see the main SRAT page. Adapted as an HTML form by Eddy: each question gets a mutually-exclusive family of radio buttons, one per answer; each answer gets a check-box alongside its radio button. Without JavaSpit, the check-boxes can serve as some space to note which answers can't be right, which is all the help you should need in working out the right answer. (If your browser supports the CSS3 :checked pseudo-class, selecting a check-box should hide its radio button.)
If you have JavaSpit enabled: selecting a checkbox will disable (and
uncheck) the accompanying radio button; and there will be some check-boxes at
page's end to turn on some extra assistance. One tells you how many of each
answer you have given or could give (to save you assorted counting); the other
tells you whether your existing answers contradict or agree with your
selection of a radio button to indicate an answer to a question – pink
says your answer contradicts some other, white says your answer agrees with
the rest of the form and greys say there's some doubt. Note that the test
result is only valid at the time it's done; no attempt is made to keep it up
to date as you change answers to other questions (aside from #20). Since
the checking script merely encodes the questions, it
has no knowledge on #20, so selecting any answer to #20 provokes a check
across all questions and uses its result to decide #20's. You can, thus, get
an all white
answer without actually being right (see Jim's comments on
#20): in fact, there are four consistent sets of answers. They only differ in
#14 and the last three questions, so it is fairly easy to work the others out,
given any one of them.
Jim said:
The solution to the following puzzle is unique; in some cases the knowledge that the solution is unique may actually give you a short-cut to finding the answer to a particular question, but it's possible to find the unique solution even without making use of the fact that the solution is unique. (Thanks to Andy Latto for bringing this subtlety to my attention.)
I should mention that if you don't agree with me about the answer to #20, you will get a different solution to the puzzle than the one I had in mind. But I should also mention that if you don't agree with me about the answer to #20, you are just plain wrong. :-)
You may now begin work.