

#21
Posted 17 June 2016 - 06:33 AM
Last (in)famous words.
I hope you succeed, guys.
#22
Posted 17 June 2016 - 06:45 AM
Mickey Knoxx, on 17 June 2016 - 06:15 AM, said:
Community maps and mods kept CoD1 afloat for almost a decade. Hell, the community developed CoDMW for them. They chose to shut it down in favor if $ DLC, not for "competitive" reasons. And I have to say CoDTools was prehistoric compared to sandbox.
And I'll see you at the next CoD1 tournament on the e-sports circuit with community made maps. Oh wait - it appears they only play BO III...

But I take and agree with your point, big developers dropped community tools to sell homogenised DLC that could be used for e-sports. PGI doesn't have a huge DLC factory department to pump out content that is of the quality required for competitive play - however they won't open their game up to the mod community for fear of losing that polish. So lose on both accounts :/.
#23
Posted 17 June 2016 - 06:51 AM
#24
Posted 17 June 2016 - 06:56 AM
Mechwarrior 4 mercs for example has some very good layouts. Concourse being one, a large walled space port in the center of a large city with towering buildings and everything.
Or a desert with rolling sand dunes
A rolling Prarie map with rivers cutting gradually through it.
So many maps in loved in that game, and not every map needs to be super alien either. Some very well could be basic Terra feelings.
#25
Posted 17 June 2016 - 07:04 AM
TKSax, on 17 June 2016 - 06:51 AM, said:
You mean people exceeding their knowledge of the enigne and creating custom Skins, mods and maps?^^
#27
Posted 17 June 2016 - 07:43 AM
Give someone the SDK for generating maps, and you may find some truly talent amongst the fanbase.
I would love to see new maps, but I do think there would need to be a mechanism to try them out and have them added to or removed from the cycle of available maps.
#28
Posted 17 June 2016 - 07:44 AM
Imagine the same process, but with player maps for PGI.
I spent just enough time mapping in Counterstrike to know that there's more to mapping than most folks think. Everything you do has a performance cost. In that engine (a primitive one compared to CE3, of course), you had to carefully segment the open areas and limit line-of-sight, or the game's rendering demands would spike and the FPS would become unplayable. You couldn't do large open areas at all. One object clipping another in the wrong way would increase the polygon count in a certain area and again drive rendering up.
I don't know how things work in CryEngine, but my guess is that you're still tiptoeing around every decision in order to keep performance up. Bad textures, bad alignment, invisible walls, flawed entities, bad lighting, stuck points...it has to be combed over for all these things. The devs won't just take your word for it. Remember they've mentioned that MWO is especially sensitive to FPS issues because of CE3's inability to use the later, better-performing versions of Scaleform.
I'm sure MWO's community has a few folks who know of all this. But just to find them, you've already required PGI to playtest hundreds of user submissions for thorough technical feasibility. You've also had to throw out the maps with hidden penises on them and the ones that are just "Canyon Network Flooded" or "River City The Way I Like It" that don't really contribute anything new to the game. Any gimmicks like low gravity also get thrown out, because that's a whole new round of design stuff. Those hundreds of destructible houses you threw in? Also out, because performance. That's a lot of dev time.
Now you're down to the 5% of submissions that are unique and playable, but you've still got to test them for 1) compatibility with every gamemode and 2) balance. Examining lines-of-sight, approach corridors, overlooks, exploits, large-scale player behavior (the "stable strategy" once players "figure out" each gamemode), making sure that one side isn't too stacked and that the map doesn't favor one kind of weapon over the other...is weeks of dev time. For just a handful of maps. Then there's art demands; a guy might be able to create a playable AND balanced map, but is he a good enough artist to make the environment look as convincingly incredible as Tourmaline or Vitric Forge? PGI's going to want that for maps entering their public reputation.
Basically, for every map that actually reached a professional level of quality, you'd be asking PGI to test and dump 200 that didn't. It is a lot of work and knowledge required just to create a playable map. With dev resources limited (especially in QA), it's probably something they just don't think would bring enough bang for their buck.
Yes, some games get good maps out of it. But those games have hundreds of times the population of MWO, and thus hundreds of times the number of aspiring mappers. Even just setting up the vetting process would require internal design work for PGI that take away from other stuff. And any game studio would do this, not just PGI.
I'm hoping that the new mappers PGI has hired will give us a bump in content rate.
Edited by Rebas Kradd, 17 June 2016 - 07:59 AM.
#29
Posted 17 June 2016 - 09:26 AM
Rebas Kradd, on 17 June 2016 - 07:44 AM, said:
Imagine the same process, but with player maps for PGI.
I spent just enough time mapping in Counterstrike to know that there's more to mapping than most folks think. Everything you do has a performance cost. In that engine (a primitive one compared to CE3, of course), you had to carefully segment the open areas and limit line-of-sight, or the game's rendering demands would spike and the FPS would become unplayable. You couldn't do large open areas at all. One object clipping another in the wrong way would increase the polygon count in a certain area and again drive rendering up.
I don't know how things work in CryEngine, but my guess is that you're still tiptoeing around every decision in order to keep performance up. Bad textures, bad alignment, invisible walls, flawed entities, bad lighting, stuck points...it has to be combed over for all these things. The devs won't just take your word for it. Remember they've mentioned that MWO is especially sensitive to FPS issues because of CE3's inability to use the later, better-performing versions of Scaleform.
I'm sure MWO's community has a few folks who know of all this. But just to find them, you've already required PGI to playtest hundreds of user submissions for thorough technical feasibility. You've also had to throw out the maps with hidden penises on them and the ones that are just "Canyon Network Flooded" or "River City The Way I Like It" that don't really contribute anything new to the game. Any gimmicks like low gravity also get thrown out, because that's a whole new round of design stuff. Those hundreds of destructible houses you threw in? Also out, because performance. That's a lot of dev time.
Now you're down to the 5% of submissions that are unique and playable, but you've still got to test them for 1) compatibility with every gamemode and 2) balance. Examining lines-of-sight, approach corridors, overlooks, exploits, large-scale player behavior (the "stable strategy" once players "figure out" each gamemode), making sure that one side isn't too stacked and that the map doesn't favor one kind of weapon over the other...is weeks of dev time. For just a handful of maps. Then there's art demands; a guy might be able to create a playable AND balanced map, but is he a good enough artist to make the environment look as convincingly incredible as Tourmaline or Vitric Forge? PGI's going to want that for maps entering their public reputation.
Basically, for every map that actually reached a professional level of quality, you'd be asking PGI to test and dump 200 that didn't. It is a lot of work and knowledge required just to create a playable map. With dev resources limited (especially in QA), it's probably something they just don't think would bring enough bang for their buck.
Yes, some games get good maps out of it. But those games have hundreds of times the population of MWO, and thus hundreds of times the number of aspiring mappers. Even just setting up the vetting process would require internal design work for PGI that take away from other stuff. And any game studio would do this, not just PGI.
I'm hoping that the new mappers PGI has hired will give us a bump in content rate.
My 980ti + 4790k on new River City + Forest Colony laughs at your optimisation concepts and says they are non existent to frame rate drops to 60fps due to time of day + water effects on maps.
iE no matter how much hardware is thrown at this game it is poorly optimized so why not let the community try?
Every single map plays out around the same choke point/middle ground due to PGI map design.
This is understandable in Faction Play/Community Warfare but unacceptable in quick play/solo group drops.
They have had more than enough time/attempts to 'improve'/get it right and have failed.
Edited by Better Call Saul, 17 June 2016 - 09:27 AM.
#30
Posted 17 June 2016 - 09:51 AM
#31
Posted 17 June 2016 - 10:05 AM
Have jobs, experience, and passion (at least the ones coming to the forums instead of just leaving)
There seems to be a lot of relevant talent, especially on the forums, who could contribute
Hell I remember one player, offering advice on the specific font and color choice they make ingame
Advice he normally charges for for other company's
Voted for
Not holding my breath
But would love it
Frankly I don't even know why PGI is sooo difficult about this
Just let the community vote on the best maps
Then only the best get reviewed by PGI
And as "salary" you can choose between MC and a mech pack of you're choosing
Not like digital items for one person actually costs something for PGI
would love it
Don't think it'll come
Though I think the UT engine has a better level editor
Drag&Drop easy in some things
#32
Posted 17 June 2016 - 10:11 AM
Rebas Kradd, on 17 June 2016 - 09:51 AM, said:
You should work for PGi with replies like that. I present a bunch of facts and examples and it's dismissed based on what?
Here are people offering free labour to improve the game using their professional skills and it's dismissed?
What we need is more destructible trees cause that's what people want blocked sight lines and performance impacts for trees they knock over. </Sarcasm>
I really hope "new" frozen city is not as "good" as the other revamped maps.
#33
Posted 17 June 2016 - 10:59 AM
Tom Sawyer, on 17 June 2016 - 05:06 AM, said:
So many threads so many just play the way we say and buy our mech packs.
For a mere $100,000 I'd be willing to LEARN how to use CryEngine, understand map making theory, and produce a map in my spare time. What a deal.
Now imagine a bunch of people designing maps for FREE and VOLUNTARILY (giving up all rights to any "intellectual" properties therein) and creating maps for the game.
They submit them.
A PGI employee sets aside a few hours a week to review them and toss out the garbage, finding the few needles in that haystack.
Now, they don't have to take the user designed map in total, maybe they just use the best parts as inspiration for their own design, maybe they take a particularly good one and tweak it a bit. Sure there will be "errors/glitches" on these user submitted maps that will have to be sussed out and corrected, but THAT HAPPENS ALREADY with PGI designed maps, so there is really no time lost.
Worst case scenario, they have a large volume of creative ideas to mine and make use of.
So, you see....it's obvious why it won't work. PGI insists on doing things the hard way.
#34
Posted 17 June 2016 - 11:20 AM
CK16, on 17 June 2016 - 06:56 AM, said:
Or a desert with rolling sand dunes
A rolling Prarie map with rivers cutting gradually through it.
You mean maps like these?
(pulled from SP since the MP map pictures look worse, also Concourse is pretty much New Avalon: Spaceport, don't have a picture of that)
These pictures are all pretty much attack/defend maps. (Therefore more suited for CW/FW/FP)
Edited by Red Shrike, 17 June 2016 - 11:20 AM.
#35
Posted 17 June 2016 - 11:31 AM
#36
Posted 17 June 2016 - 11:59 AM
Pretty please

#37
Posted 17 June 2016 - 12:59 PM
TLBFestus, on 17 June 2016 - 10:59 AM, said:
You're vastly underestimating how much time it would take. PGI's employees already don't have a few hours a week to set aside.
HOWEVER...
What I'd be interested in seeing, is PGI use a group of "squirrels" (i.e. players with developer experience and playtesting savvy) to do the narrowing-down. Hand them the maps, hand them their set of standards, and tell them to come up with three or four worthy maps that are playable, balanced, optimized for every gamemode, and visually interesting. Allow them to clean up the obvious crap. Then submit those final candidates to PGI for review.
Even then, it would still be a lot of work for PGI. Map creation is usually the intersection of several different types (designer, artist, modeller). PGI would also have to carefully vet the squirrels doing the work. And then there are legal and server issues involved.
But community-generated maps certainly would provide a jolt of player interest.
Edited by Rebas Kradd, 17 June 2016 - 01:00 PM.
#38
Posted 17 June 2016 - 01:02 PM
Kotzi, on 17 June 2016 - 04:11 AM, said:
Kuh-Leerly PGI has never seen other games like Skyrim, Fallout, Men of War and other games that allow player created content.
#39
Posted 17 June 2016 - 01:33 PM

#40
Posted 17 June 2016 - 01:39 PM
13 user(s) are reading this topic
0 members, 13 guests, 0 anonymous users