About Neroxis map generator...

3

yeah,i totally agree with Blodir,those maps are just really good,i have been testing for the last 10-12 games and it feels nice

0

Don’t see much point in doing a fusion of random map gen maps and authored maps. Either it’s good enough to stand on its own or it isn’t.

On that note, I’d be willing to use it in pools if it passes the minimum quality level for a pool of authored maps, which hovers around 3-3.5/5 across the whole pool these days.

0

@FtXCommando said in About Neroxis map generator...:

Don’t see much point in doing a fusion of random map gen maps and authored maps. Either it’s good enough to stand on its own or it isn’t.

On that note, I’d be willing to use it in pools if it passes the minimum quality level for a pool of authored maps, which hovers around 3-3.5/5 across the whole pool these days.

What @Wainan said, you don't have a set strategy for the random map like you do with every other map. For noobs that means you have less of a knowledge gap compared to players who either watched a replay on a map, or just played it 10 times already and developed a standard build for it.

@Wainan said in About Neroxis map generator...:

@cocAurico I think it might be more noob friendly in a way because there's no meta to learn (apart from MAKE MOAR TANKS)

0

I don't understand why you quoted me for that as that has no impact on whether random map gen maps are 100% of the pool or 5%.

1

It would take some code changes in the lobby server and we’d have to think about how to add random maps to map pools as they aren’t saved in the database.

0

I played some 1v1s on some mapgen maps that honestly had great layouts and led to very interesting games. @Sheikah would it be possible to rehost maps on mapgen maps you've played on previously (using some unique identifying code) even after you have refafed, or share the code with someone else and give them the ability to host the map too. If that could happen then @Askaholic it seems like it should be able to go into the ladder map pool as well.

@FtXCommando how would mapgen maps be rated? The average/median score of some sample of maps generated with the same initial settings?

0

I would make a sample 20 map pool and rate said pool same as I would rate a ladder pool.

3

The map name for each generated map is a unique seed so if you put the same map name into the generator you will get out the exact same map. That is required for it to work since everyone generates the map locally when joining a game.

0

The awkward thing is that maps which go into the map pools need to exist in the vault database (there needs to be an entry in the map_version table). I’m not entirely sure what the consequences would be of doing this, they might become searchable in the vault, but then they won’t have any preview images or any map files on the server. And this also doesn’t allow for having actually randomly generated maps. After a while you would learn the generated maps that are currently in the pool.

Btw there is this thread for cool seeds that you find: https://forum.faforever.com/topic/635/neroxis-map-generator-hall-of-fame-seed-vault/3?_=1613805662428

0

@FtX what's wrong with having 1 map in the pool to be randomly generated at time of play? Don't follow your objections. Certainly worth a try for a month I would think. Correct me if I've misinterpreted your comments.

0

To do 1 map in 20 as random map gen I would need to do a sample check to see if the map generation is any good. What would this require? Me doing X amount of maps to see if a broad range of maps reaches the level of quality I expect of a ladder map.

In other words, regardless of it it being 1 map or all the maps, I’m essentially doing the same amount of work and expecting the same thing from the map gen. Therefore, either it’s good enough to replace everything or it isn’t good enough to be in ladder.

0

Maybe a mapgen queue is the solution? With the ability to multiqueue everyone can basically adjust if he wants all, none, or some of the maps to be mapgen maps.

0

I just think you can afford to take a risk with 1 random map for a month and see how it goes

0

@BlackYps said in About Neroxis map generator...:

Maybe a mapgen queue is the solution? With the ability to multiqueue everyone can basically adjust if he wants all, none, or some of the maps to be mapgen maps.

This can be tested but I still want it to meet the quality of a ladder pool before doing it.

4

Luckily a huge amount of mapgen samples will be provided today. We gonna see first hand what the mapgen is capable of and what kind of games it provides - If it can reliably create reasonable maps.
That said, I dont think the quality assasment of current hand made ladder maps can be directly transfered to mapgen maps. The will likely never reach the aesthetic level of hand made maps - especially if you look at the astounding progress in the last maps made by Jip, Svenni, biass and other map creators.
The huge advantage of mapgen maps is, well, they are generated and no one has seen the map before. In short, its their novelty. And novelty is a quality of its own. Mapgen trains a player to adapt, to asses what a map has to offer, to think what might be good and what not so much. Hand made maps with established meta simply cannot provide that feature.
My point is, even if mapgen maps lack in comparision to hand made maps in some regards (which they do), they bring some new quality into the game which current ladder maps simply cannot provide.
Therefore I would love to see mapgen maps in ladder to some extend.

1

Just fyi. I quit ladder few years ago after the map pool became huge.
My feeling was that bo whoring was more important than general game knowledge.
And It was just too many maps and bo to learn to be good.
So a map generator, with no perfect bo possible, would probably take me back to ladder.

2

I don't think rng maps need to be vetted manually, they are generally so good. Just throw em in ftx.

0
This post is deleted!
2

Overall from looking at the tourney maps, I think it's nearly decent enough for me to be interested in doing a separate 1v1 queue. I mainly want to work some more with Sheikah to make the 10x10s more varied in structure first as I think over half of the 10x10 games in the tournament followed a near identical template for the map size.

I didn't get too many 5x5 maps (too many 2-0s) so I'll also need to check those out some more on my own.

1

I can agree that 10km need some more variation. From the tournament today I am pleasantly surprised that the 20km were well received as I was worried they would be terrible.

With regards to the 10km I think that the hills/ramps can be reduced in terms of their slope so that there is less terrain issues. Additionally with trying to find more structures for the maps as FtX mentioned. Of course that all happens the fastest with more feedback and ideas. So feel free to suggest.

5kms are probably in a similar place as 10km a little bit although there I did try to add some special cases not make it all open all the time but needs to be looked at.