Write smell code [I have caught myself at every jam - overengineering a system, inventory, movement... Basically I was trying to write reusable code and think of best solutions - which is a waste time during the jam].
Complete whole game in the 1 day [all the other code and work should be polishing after the demo is ready, if you complete the game in 2 or 3 days - you will have a working mechanic, but you won't have a game].
Select as few things(mechanics) that make up your game after which it would be somewhat playable and focus on those until they are complete. Then make a game out of it. After that you can either polish[↑] or you can start adding new mechanics to already playable game [many times I have picked the thing I wanted to work on the most and I got it working, but then I didn't have some mechanic that was essential for *fun* in the gameplay, so the game was just boring].
Forget about story if your game isn't story based [make it fun to play without story, it takes a lot of time to come up with a good one, just don't bother with it if you can avoid it].
Don't do it in a team if you are new or some team advises, reasons:
1. You are setting up people that expect something to be completed in time with a good quality [it's not always true, some newbies are dedicated. But I have seen too many people just giving up or doing almost nothing or unusable].
2. Keep your team small, like 4 people is already a crowd, but manageable if it's not your first time. 2 programmers - only if they are very good in version control system or have worked together for some time.
3. You are more likely to do much much less if it's the first time you are working with someone. It's best to have team consisting of people you already know or have worked with in the past.
4. Game Idea: oof, this is a tough one, you will probably not like it, and if you do - it's either you got very lucky to have the same taste or you are the person whose idea was chosen. The best solution I see is to come up with creative way of everybody adding something to the table, so that everyone is excited. Another good solution is to have an idea beforehand and gather people that like this idea as well.
I have probably taken part in 20+ jams :) About some of them I wasn't serious, so I would say it's around 7. I am not a pro jammer, I haven't won any but this is my failures based experience.
I have originally written this in a comment section but thought that it might be a waste to just leave it there.
