well the basic most important rule is probably: focus on something simple, be clear what to put in and leave out. don't assume you can create a huge project within such a short time, especially if you have never done that before. try to set yourself a deadline for brainstorming your idea, and then go for it. be sure to focus on the core idea, don't have complexity in several areas (best to have no complexity at all, but if you have, then keep it to one area). you can always add stuff later (after the jam - or during the jam, if you finish the core of your idea early). for example, if the game will absolutely need to have several levels, cut complexity on everything else (e.g. just one type of enemy, no AI, etc). if the game will absolutely need to have several types of enemies though, make that the core focus but try to fit your project so it is fun with just one level. if the game will absolutely need to have AI, focus on the AI but just have one enemy and one level. don't try to do "everything".
as for a time schedule, you could set milestones for each evening. on the first evening, the milestone will be to have an idea or a concept, on the second evening, the core game should be playable, on the third evening of course the release will be due. if necessary, add some in-between milestones. at some point during the jam, you might have to decide which features to scrap in order to still submit the game. also, it's not a bad idea to package your game in-between, let's say you have a good version at some point that could theoretically be released as the final submission, then package it all up in order to a) fullly be aware of what stuff you need to provide in the end (readme, screenshots, etc), and b) have a ZIP ready to submit in the case that during the final hours of the jam, all your additions fail to complete or you end up in a weird state of incompletion / half-finished stuff that you don't want to submit.
hope this is some good advice for you :)