Making too big of interior maps. I did it too, as I made a mansion which was 90 x 90 tiles. It was ridiculous in the end. I think the kitchen if you tried to translate it to real world would have been 36 feet by 42 feet. No one has a kitchen that big, for any reason. As to how to prevent it, most adopt the convention that since the bed is 2 x 1 tiles and most beds are near 6 - 7 feet long, then a tile is about 3 - 3.5 feet for interior maps. Since many rooms in the US are 10 - 12 feet, that means if you were mapping US stye rooms it would be 3 - 4 tiles in each direction, total.
Mis-setting the To Hit/EVA ratios. I did it too so I can testify to this. I tried 75% to hit in my demo initially, and people complained about how often they missed. Now I do know really old games had hit rates this low, and I went back and played one, and you know what? It wasn't fun. Battles were miss, miss, miss, miss, miss, just hit that stupid Goblin already! And even the games that used low hit rates well it worked because you could manipulate things to improve it. For example, Jagged Alliance is a game where you miss a lot in gun fights. But you can improve your hit rate by getting a better gun, finding a sniper scope, or just moving to a better position. Fire fights across 10 - 12 tiles of open terrain with everyone in cover would have a lot of misses in relaity, so if you want to hit, someone has to get closer and flank the other. Or you can just rush them and pray, which can work too. Or you can grenade them and blow up their cover while at it. Or you can blow up the stairs and the walls and make them fall to their death. Or you could blow up a hole in the wall and go around them and surprise them from the back with another group (I did this many times). There were options. We can't do that in RPGMaker without massively re-writing the battle system though, so for us, we should avoid really low hit rates unless we want to completely redo the battle system to make it fun still.
Making too big a game at first. This has been said already, but I say it too as I did this, and here is how well it worked:
June 6th, 2014: Started game for the IGMC 2014.
June 10th, 2014: Started over as first idea was awful.
July 2014 sometime: Started over
August or September 2014: Started over. Released demo of this version for initial feedback.
November 2014: Started over
February 2015: Started over. This version became the version I'm still working on to this date.
August 2016: Game was playable from start to end, though very buggy.
July 2018: Beta version released to testers.
Today: Still Beta testing. It's slow though as I got to test 20 - 25 hours of playtime (for me) over and over and over. Hope is to release in February or March of next year though. I could probably release it now, but I refuse to release during the holiday deluge of winter sales as it will just get lost in the mass of black friday and steam winter sales. So I'm using the remaining time to polish it up more in little ways, aka fix those little things that work as they are, but could be improved (hint: Make a backup if you do this, sometimes you end up undoing your change as you make it worse).