On March 31st of 2006, JK Rowling released a new addition to her website where by solving a certain puzzle, fans could open a secret door and discover a Harry Potter test known as “W.O.M.B.A.T. Grade One”. For the next five days, fans could take this test and submit their answers.
However this was no ordinary triva test. The questions were all asking things that weren't in the books, and although many of them could be inferred from published material or author interviews, the vast majority of them relied heavily on intuition and guesswork to solve.
On April 4th the door closed again and fans received their results a few days later in the forms of the six in-universe letter grades used for the O.W.L.s – from “O” for Outstanding to “T” for Troll.
Rowling would later follow up the first WOMBAT with a 'Grade Two' in September 2006, and the final 'Grade Three' in June 2007 (one month before the release of the final Harry Potter book). She would reveal later in an interview that the reason she made these tests was that her lawyers had cautioned her against releaseing any actual previews from the book and this was her alternative way to build up hype. However, a few of the questions and answers in Grade Three actually contained some hints towards the seventh book.
Sometime later, on October 31st 2007, all three tests reappeared on Rowling’s website with instant grading. There they remained until February of 2012, when the entire website was taken down and replaced with a more generic author website. During this period where the test had instant grading, a fan was able to take each test multiple times and properly reverse engeneer how the scoring system worked.
In 2018 there were some fan efforts to recreate the test experience, both by tweaking archived copies of the original website's flash files and by recoding that section of the website in CSS. These recreations of the tests have been taken by many since then and the information on this page is built from a dataset of 6,811 of these.
Notes: Grade Three has negative values because it penliazed wrong answers. The cuttoffs for each mark were different between grades so they aren't shown in the overall distribution.