All this is a breakdown of a report I received after passing IBM 483. I know this is an OO group but this also has a bearing on 486 as the same problem has been reported by someone else on this group. %exam %right q/per/o q/right J2EE architectu16%0%9.280.00 Web Component17%77%9.867.59 EJB 28%90%16.2414.62 Client 10%56%5.803.25 Connectivity19%83%11.029.15 Assembly10%72%5.804.18 58.0038.78 I sat and passed IBM 483 recently and although I found the test hard I was still suprised to achieve such a low score of 75%. When you have finished the exam you get a breakdown as above. The first column represents the % of question in the exam against each objective. The second column shows the % of questions you got correct against each objective. What stood out was the 0% attained against the J2EE architecture objective. Of all the objectives this is probably the easiest as the question are generally high level and common sense based. Being curious I calculated the rough number of questions per objective based on the report and then the rough number I must have got correct. As you can see from the third column, the total of the questions adds up nicely to 58, the number of questions in the exam. The fourth column is the rough number of questions I got right and as you can see this adds up to about 39. 39 questions correct from 58 gives about 67% correct, which is way short of the reported 75% score. If you take away the 9 J2EE architecture questions I would have achieved a score of 89% across the other objectives so I doubt very much that I have managed to score 0% on one objective while scoring such a high average on all other objectives, especially when they are typically more detailed and harder questions. I suspect the 75% score may actually be correct, but in fact all that has happened is that the report is not showing the correct number of questions passed for the J2EE architecture. If I adjust the J2EE architecture objective to about 50% I get my attained 75% score. So don't be suprised if you get 0%. Perhaps they the old chestnut where adding 1 to 99 gives 0! I have sent this to IBM.