236 replies to this topic

[Mister Blastman]

Zeusus, on 04 August 2015 - 11:48 AM, said:

I have no idea who this Dan fellow is, and detail settings don't cut back in the generation... Doesn't help at all... Weaker CPU will generate slower, and less RAM will slow things as well. And then once done detail settings matter.

Thank you for proving my point. Reference your spelling of "potato," hint, you spelled it, "potatoe."

You should google him and that spelling. You can learn a lot from history.

Detail settings absolutely matter because they determine what is rendered and what is not. The time to generate the world is minimal. No more than loading into the game for a drop. This is a non-issue, especially when including pre-rendered assets into the procedural pipeline.

[Zeusus]

Mister Blastman, on 04 August 2015 - 11:52 AM, said:

Thank you for proving my point. Reference your spelling of "potato," hint, you spelled it, "potatoe."

You should google him and that spelling. You can learn a lot from history.

Detail settings absolutely matter because they determine what is rendered and what is not. The time to generate the world is minimal. No more than loading into the game for a drop. This is a non-issue, especially when including pre-rendered assets into the procedural pipeline.

So he's some dude from the US who mispelled potato like I did while typing on a phone, fail to see how that is relevant and why I should care about him. Also I'm not sure you even have a valid point bringing him up, care to elaborate?

And generation speed is affected by CPU and RAM, read up on how it works. We haven't even got to rendering at this point.... Generation and render are not the same thing.... If you don't have enough RAM it is going to be slow, period.

[Rhaythe]

Mister Blastman, on 04 August 2015 - 11:52 AM, said:

Thank you for proving my point. Reference your spelling of "potato," hint, you spelled it, "potatoe."

You should google him and that spelling. You can learn a lot from history.

Detail settings absolutely matter because they determine what is rendered and what is not. The time to generate the world is minimal. No more than loading into the game for a drop. This is a non-issue, especially when including pre-rendered assets into the procedural pipeline.

I encourage you to play "Sir, You Are Being Hunted". That is a true procedurally-generated game. Every world is unique, and the rules for generating it actually result in very good gameplay.

That said, the initial loading time for map generation is UNBEARABLY long. And you're asking it be done every game. Something has to generate those worlds. If it's the server, the clients are still waiting an unbelievably long time for the world seed to be created, checked, balanced, verified, and then communicated.

You're shoving an idea that doesn't NEED to be shoved. We all agree with you. But you have to understand - procedural generation is computationally expensive. Especially in a competitive gaming environment. There's a reason so few games implement it today.

[lol lol lol lol lol lol lol lol]

It'd certainly be cool, and I love the random generation stuff (why I always liked the Diablo series) but while I'm not saying it's not possible "AT ALL," I don't think it's possible "WITH PGI AT THE HELM." Meaning no offense, the people are more than talented at working with what they have, but...

The perception I get from PGI in how they describe their capabilities are ironically similar to the Battetech Universe lore. They make it seem like the Cryengine they use and how they use it is like that of how the Inner Sphere looks at Lostech. In the Succession Wars of updates and patches they're doing the best they can with keeping the current engine and game modes running. Ideas were set aside, scope was narrowed down and time frames are replaced with words like "eventually" or "Soon." Thinking outside of the box creating new things like random maps would probably not work with how they have things currently set up, and require a massive overhaul of everything they've duct taped and glued together.

It is funny though, that the player mentality of this game is to rip out all of the old junk that comes with the mechs, to make them faster, more powerful and efficient... yet this engine and capabilities seem limited or beyond a functional scope of what players want and what dev's can easily accomplish... but everyone's ok with that.

I know the vicious circle of change and improvements takes money, money takes people investing, and investing takes faith in there will be change and improvements. But in the not to distant past, there was a push for a new game with a new scope... so I know that it's not IMPOSSIBLE... its just I think they're comfortable with not having to do anything more with what they have now because still making the same money, so long as people keep buying mechs. And at the rate of release, that could last for a few years at least.

[DAYLEET]

https://www.reddit.c...map_generation/

There's awesome link in there about procedurally created city in action.

[Fate 6]

Procedurally generated maps would actually be amazing - suddenly scouts are actually vital and protecting your slow mechs is a real effort. We don't need maps as fancy as new River City as long as they look and feel like Mechwarrior maps like RC does now

[DaZur]

Mind you... I'm not poo-pooing procedural.

That said, while I know it can populate environments beyond just trees to include structures etc... That same mechanic can generate some awfully convoluted / unbalanced maps.

Baring into consideration our present play-mechanics of Assault, Conquest and Skirmish and the need to drop-points and capture points to be reasonably balanced... I fear procedural would be more trouble than it's worth.

... Not even taking into consideration the quantitative development cost to accommodate the diametrically differences between our existing topology, it's collision maps and environmentals.

[0bsidion]

They really need something like this for CW. There's a whole galaxy of planets to cover, but right now it's like, hey, Misery looks exactly like Terra, which looks exactly like Otho, etc, etc. It kind of ruins the immersion. Especially if you know the planet you're dropping on is primarily tropical and you get a desert map, or a mostly desert planet and you get an arctic map.

So yeah, I'd love to see it. Even if they just used a procedurally generated map as a foundation, and then went through and balanced it for gameplay. Then we could see maybe 2 or more new maps a month, instead of 1ish. I'm just not sure how practical it is for them to implement.

[Mister Blastman]

Rhaythe, on 04 August 2015 - 12:01 PM, said:

I encourage you to play "Sir, You Are Being Hunted". That is a true procedurally-generated game. Every world is unique, and the rules for generating it actually result in very good gameplay.

That said, the initial loading time for map generation is UNBEARABLY long. And you're asking it be done every game. Something has to generate those worlds. If it's the server, the clients are still waiting an unbelievably long time for the world seed to be created, checked, balanced, verified, and then communicated.

You're shoving an idea that doesn't NEED to be shoved. We all agree with you. But you have to understand - procedural generation is computationally expensive. Especially in a competitive gaming environment. There's a reason so few games implement it today.

How long does it take to generate?

They could be generated say twice a month by PGI and downloaded via the patch updates. This way we have a constant rotation of new maps, all the time.

[ApolloKaras]

Im sure the hit dection meshes would be 100% dead on too!

:|


Plus you are inevitably have maps which are completely slighted one way or the other, anything having to do with comp - would get zero support. We'd still be playing the same maps we have currently.

Edited by Saxie, 04 August 2015 - 12:24 PM.

[Zeusus]

Mister Blastman, on 04 August 2015 - 12:21 PM, said:

How long does it take to generate?

They could be generated say twice a month by PGI and downloaded via the patch updates. This way we have a constant rotation of new maps, all the time.

See this could work, as they could take one high end machine, generate a bunch and keep the best.

[Rhaythe]

Mister Blastman, on 04 August 2015 - 12:21 PM, said:

How long does it take to generate?

They could be generated say twice a month by PGI and downloaded via the patch updates. This way we have a constant rotation of new maps, all the time.

I don't remember if I played that on my i7 or my Phenom II, but upwards of five minutes a pop. Just to generate the seed and verify it.

That might be a happy medium. A collection of pseudo-randomly generated maps per archtype that can be vetted prior to being put in rotation. It still requires an initial investment of coding time to 1) create the procedure that generates maps, and 2) the procedure that vets them. A heavy investment, actually. But it could work in theory.

Zeusus, on 04 August 2015 - 12:24 PM, said:

See this could work, as they could take one high end machine, generate a bunch and keep the best.

You'd have to farm this. This kind of number-crunching requires some dedicated horsepower. Keep in mind all the hit-box problems and falling through the map this game gives us using CREATED maps. Automating the process just asks for kinks.

But again, could work.

[Mister Blastman]

Saxie, on 04 August 2015 - 12:22 PM, said:

Im sure the hit dection meshes would be 100% dead on too!

:|


Plus you are inevitably have maps which are completely slighted one way or the other, anything having to do with comp - would get zero support. We'd still be playing the same maps we have currently.

I beg to differ. I'm sure there are many competitive players that would prefer to play on random maps. Thinking on your feet is the ultimate test of skill and prowess.

[Yellonet]

Zeusus, on 04 August 2015 - 11:26 AM, said:

Is your argument to have them pump out maps using procedural or to have new each match? First I can agree with.

Second won't work. I'm not waiting 20 minutes to DL/generate a map each match...

Rereading you want the second. Yeah no.....

There would not be any additional waiting involved.

[ApolloKaras]

Mister Blastman, on 04 August 2015 - 12:29 PM, said:

I beg to differ. I'm sure there are many competitive players that would prefer to play on random maps. Thinking on your feet is the ultimate test of skill and prowess.

Having a map one sided no matter what I do if teams are of equal skill, if one map has an advantage - the other side will be hard pressed.

We also have this issue with bandwidth. Assuming you can get the map down to 20-30 mb, thats 24 clients you have to send it to, thats 720mb per match.... The load times would be atrocious, because you would have to wait for that guy who is running on a potato.

[Rhaythe]

Saxie, on 04 August 2015 - 12:32 PM, said:

Having a map one sided no matter what I do if teams are of equal skill, if one map has an advantage - the other side will be hard pressed.

We also have this issue with bandwidth. Assuming you can get the map down to 20-30 mb, thats 24 clients you have to send it to, thats 720mb per match.... The load times would be atrocious, because you would have to wait for that guy who is running on a potato.

It could be alleviated by downloading just the seed. Then you're talking an integer-based string, no real issue. The real problem is what generates that seed and what vets it, but I've preached long enough on that.

[ApolloKaras]

Rhaythe, on 04 August 2015 - 12:33 PM, said:

It could be alleviated by downloading just the seed. Then you're talking an integer-based string, no real issue. The real problem is what generates that seed and what vets it, but I've preached long enough on that.

Thats a good point. Then you throw the rendering work on the users PC. I would be okay I have a CPU that could crank it out, however if you have someone running a dual core and just barely running the min specs....

[Zeusus]

Rhaythe, on 04 August 2015 - 12:27 PM, said:

You'd have to farm this. This kind of number-crunching requires some dedicated horsepower. Keep in mind all the hit-box problems and falling through the map this game gives us using CREATED maps. Automating the process just asks for kinks.

But again, could work.

Not sure you'd need a farm, depends on how many you want to make. Can get up to 6TB of RAM in a server, but you raise a good point and I'm being picky

Yellonet, on 04 August 2015 - 12:29 PM, said:

There would not be any additional waiting involved.

The delay has been explained several times.

[KraftySOT]

http://www.no-mans-sky.com/

Your argument is invalid.

[Mawai]

Mister Blastman, on 04 August 2015 - 11:38 AM, said:

Unfortunately I believe firsthand experience more than a video. I have been playing around in procedurally generated worlds for several years now. And what I've come to see is awe-inspiring.

The naysayers will knock them, but us, the explorers, the ones who actually visit them, we know otherwise. We have no fear. We know how well they can be made and every machine will render them the same.

See, in Space Engine, if our detail settings are identical, I can tell another person to go to x planet around y star and at b co-ordinates and ... it will look the same. We don't share the map. We don't share anything. We both use the same code/engine and the random seed takes care of the rest. There's no uploading or downloading anything. There is no pre-rendering. It is all done real-time.

It all SOUNDS awesome.

I don't know enough about it to know for sure.

The video someone posted made it clear that the challenge is developing the "correct" rules or algorithms that would appropriately describe the worlds that you want built.

Different rules would be required for different biomes and settings including a wide range of texture resources for the different flora, buildings, terrain and other artifacts. Buildings would need to be logically linked with roads and connected to wilderness areas. In each case, a playable world would need to be generated so that there are no sections of the map that could trap a mech, not too many dead ends, or maze like constructs which would just make people wander around hoping to see an opponent.

There is a huge amount of work involved in building an engine where the only information you supply is a random number seed and out pops a playable world.

A world that is playable with the existing game engine.

I also don't know what polygon resolution is used in the current maps or how the lines of sight are calculated by the server during hit registration. Do the worlds generate identically on 32 and 64 bit machines? Would a numeric representation in 64 bit block a line of sight that was clear in 32 bit? (or vice versa).

As I said at the beginning it would be amazing if it could be made to work reliably with MWO mechs but I do suspect that it would be far more work to implement than PGI is ever likely to have time for. The only way I could see this working would be if someone wrote a procedural world generator for MWO and supplied it free of charge or very cheap ... or such a product was available commercially at low cost ... such that engine integration would be straight forward.