A while back I released a map called fy_span. It's not a particularly special map. It's a simple fy map for CS 1.6 that a cooked up in about 2-3 days. It is, however, an example of what I'm talking about. Have you ever "saved" a decent concept that was turned into a terrible map?
A while ago I was giving someone advice on their first ever map (it was for CS:GO). Not on Source specific things, I've only made one map for Source myself, but on CS mapping theory type things and more basic brushwork detail kind of stuff. His map was an fy map featuring a bridge in the center of the map. However, that's all it was, a bland bridge with only a tiny bit of cover on the sides with the way up consisting of a ladder on either team's end of the bridge. The only other cover was a bunch of big crates; all of them the same size, spread equally apart from each other. The ground was a flat grass texture and it was surrounded on all sides by really tall, flat, brick wall textures.
I was honest with him, I told him the map needed some massive improvements, though it was actually better than other 1st maps I've seen (including my own first map) since at least this wasn't fullbright and players didn't spawn with their head in a ceiling and die. I made some suggestions about detailing the walls a bit and making the cover more varied because he had basically created "suicide lanes" for the first player bored enough to peek. I suggested not making all of the ground one big flat plane and that it didn't have to have extreme height differences but a few dips and bumps could go a long way. I also told him to nix the ladders going up onto the bridge ends for some ramps or stairs or something along with a bit more to make climbing on the bridge and crossing it less suicidal.
To my dismay, he released Version 2 a day later, so I knew he did almost nothing. Sure enough, he did the bare minimum of work based on what I told him, and the "varied" cover was just changing some crates and things to brick walls or other random objects. The "suicide lanes" were still a thing. It was better than version 1 but better in the sense that getting kicked in the butt is better than getting kicked in the nuts.
But I really liked the bridge idea. I play CS 1.6 on a custom map server with over 400 custom maps, but early in the night before the server fills up there aren't a ton of good map options for a small number of players. There are a bunch of fy maps on the server but most of them are the same or similar and are either a glorified aim map or have sort of a "4-square" design like Snow and Iceworld. The bridge was something I haven't seen, so, I "saved" the concept.
I created my own fy map with the bridge, and fy_span is what I produced. It's not perfect, it's not groundbreaking, but it's something different. I did also use it as an example map to try to give that new CSGO mapper a better idea of what I was talking about. I told him about the thought I put in to certain bits of cover, why the stairs going up to the bridge were perpendicular to the bridge so firefights across the bridge would be better, etc. I hope maybe he learned a couple of things.
I'm bringing this up because I thought of a recent terrible map I made years ago. It's not even available online anywhere anymore. I called it de_snowfort. It was terrible on so many levels. Bad brushwork, bad layout (Ts could get to one site before the CTs), bad spawn points, horribly timed chokepoints, and just not enough room for players to move around easily. I didn't even understand how decals worked and I made signs using brushes with the site A and site B decal textures on them. But I liked the concept I had in mind.
Basically, the theme of the map was a big snowdrift up against a house with a snow fort in it that actually extended underground into this network of snow tunnels and stuff. You could enter through an opening in the snow pile or go through the house into the basement and out the bulkhead under the snow. It's a ridiculous but fun theme and I've decided to "save" the concept. I plan on making a new map with this idea (though it's currently behind a few other ideas I have), but with a proper layout, proper brushwork, and without all the horrible noobery that my original de_snowfort was plagued with.
Has anyone else ever done anything like this? Re-visited a concept that they had a long time ago that was ruined by your noob days of mapping? Drew inspiration from the concept of an otherwise terrible map that a bad mapper couldn't be bothered to put more effort into?