Jump to content
Dante Unbound: Share Bug Reports and Feedback Here! ×

Pondering Railjack mission structure


Steel_Rook

Recommended Posts

Let me preface this by saying: I like Railjack. I liked it back when it sucked ass and soft-locked every mission, I like it now. This isn't a "Railjack taxi" mission, as I both understand the technical need for running full Ground missions with the entire squad and am comfortable with the approach. Years ago I might have imagined doing small pieces of instances as Points of Interest, but it seems like Warframe just can't make that happen without tanking the host's performance. Fair enough. However, this creates a single overriding issue...

 

The issue:

As much as I like them, Corpus Railjack missions play out like ground missions with added fluff. Alternately, they play like really short, unfocused Railjack missions. While this isn't unique to Corpus Railjack (Defence missions are really cut-down ground missions), it's particularly prominent there because it holds true across ALL mission types. That's because the issue isn't inherent to either Railjack or Ground missions, but rather the connective tissue between them. The missions aren't really integrated into each other so much as staggered in sequence. You do the Railjack bit, then you do the ground bit. You ostensibly do two missions one after the other across two game modes - rather like a back-to-back Sortie.

In other words, the problem as I see it is the 1-1 ration of Railjack to Ground. EVERY ground mission requires one Railjack mission before it, every Railjack mission requires one ground mission after it. Intuitively, Railjack "feels" like it was originally intended to be a new take on Free Roam maps. Rather than always returning to the Orbiter to start a new mission, you simply travel to a new location and keep doing missions - exactly like Free Roam Bounties. The 1-1 ratio, however, acts more like a "cost."

 

The propositions:

If the problem is a 1-1 ratio, then the solution - intuitively speaking - would be to unlock that relationship. But can that be done, and what form would it take? See... I'm honestly not sure. I've thought about this for a while and I still don't know if I have an "answer" so much as I have idea stubs. Here are some of them:

  • Get rid of "nodes" altogether and treat every Proxima region as a single Free Roam space.
  • Drop the concept of a "primary" objective. Every Proxima region would start as Free Flight.
  • Add Bounty-like missions that players can take while IN the Proxima region, consisting of multiple stages.
  • For each stage of the Bounty-like mission, players would "travel" to a different part of the Proxma region (read: reload the map), where they'd tackle one or more objectives.
  • Objectives could be purely Railjack, purely Ground or a mix of the two.

Basically, what I'm imagining here is the ability to do everything from ground missions to space missions to everything in-between, either through procedurally-generated mission sequences or directly chosen by the player. This would mean that players could do multiple ground missions in the same Railjack instance, or done none of them. Crucially, it unlocks this 1-1 ratio where players HAVE to do a Railjack mission first, then they HAVE to do a ground mission second. Free Roam maps already work like this, with Bounties offering missions in pretty much any order and even having small procedurally-generated stories tying them together.

The Railjack tech should be capable of doing this. We know the game is capable of reloading map instances in real time with players in them (that's what Railjack travel is, you never leave your Dojo). We know Railjack is capable of doing space missions with no ground components (besides ship interiors), we know Railjack is capable of doing ground missions with no space components (besides travel). The game should be able to mix-and-match missions comprised of either in isolation or both together in non-equal ratios. Maybe do ground-Railjack-ground. Maybe do Railjack-Railjack-Ground. Maybe do just Ground - why not?

The Free Roam tech should be able to handle the rest. Removing nodes and going Free Roam allows for the procedural generation of arbitrary mission sequences AND has the side benefit of condensing the player base as everyone wanting to do Railjack on a single planet would be matchmaking through the single Free Roam node the same as happens in Cetus, Fortuna or Deimos.

 

In conclusion:

I'm sure there's something I'm forgetting. I'm sure a lot of what I claim to be technically already possible would actually require brand new code. However, I feel that if Railjack is to have a real future, it needs to move away from its current heavily scripted format. It needs to stop trying to mimic ground missions and embrace its nature as the game's true open world. Railjack shouldn't be comprised of large, monolithic scripted nodes, but rather be procedurally generated on the spot from its component pieces. There is SO much more that Railjack missions could be if they weren't locked to "destroy nodes/crew ships then ground mission."

Link to comment
Share on other sites

A post-factum thought on matchmaking:

While condensing all players in a Proxima region into a single Free Flight "mission hub" should improve matchmaking times considerably, it does come with a new issue - mission selection. With a node-based system, players can "queue up" for just the mission types they actually want. When different missions offer different rewards, that matters.

However, Free Roam maps already have this issue and the Bounties system already solves this issue. Players can commit to a mission before entering the Free Roam zone and thus they can be matchmade with only people who picked the same mission. Bounty missions are themselves (somewhat) procedurally generated and rotate on a time. This offers both a variety of mission objectives and level ranges, so that problem is already solved as far as I can tell.

Thus, players can be allowed to zone into the Free Flight area of a Proxima region, pick a Bounty from there, then travel to that Bounty by using the Railjack Navigation console. To avoid an unnecessary zone load, players could additionally be allowed to access a Proxima region's Bounties directly from the Railjack Navigation console such that they can pick a Bounty and queue for it straight from their Orbiter - or rather, the Railjack instance attached to the Orbiter. This should preserve the current implementation of picking a mission and zoning into it directly.

 

Basically, I believe that Railjack missions should be broken down to their component pieces, then procedurally put together as players queue for them. Free Roam maps already do this quite well, so it feels like that system is a better fit for Railjack.

Link to comment
Share on other sites

My opinion? 

All the corpus proxima missions should be scrapped for railjack, as far as i can see its not only faster, but better to just go do a normal exterminate or defense.

Railjack should not use the same missions or structure that the normal star chart missions do, they should be more like skirmishs, where you need people both in the railjack and on boarding enemy crew.

If i wanted to do an exterminate mission, i don't go to railjack, i do a normal exterminate. If i want to get prime parts while im at it, i don't go to railjack for void storms  that just 1shot you out of nowhere at random, i do a normal fissure.

Link to comment
Share on other sites

2 hours ago, Joezone619 said:

Railjack should not use the same missions or structure that the normal star chart missions do, they should be more like skirmishs, where you need people both in the railjack and on boarding enemy crew.

I should probably address this as I skipped over it in the OP, assuming it was common knowledge by this point. Warframe Game Director Steve Sinclair made it abundantly clear that this is not an option. Warframe's technical performance allows either decent space combat or decent ground combat, but not both at once. On Railjack release, we were told that Kosma/Evo Grineer were a significant compromise. Railjack ground combat could only support a very small amount of ground enemies at a time, which is why ground enemies were given vastly inflated stats (to the tune of 20 times the health relative to other Grineer). Corpus Railjack is the same compromise from the other direction. In order to ensure proper ground combat, the space layer has to be temporarily suspended. In order or that to happen, all players must be within the ground layer, where they take part in standard ground missions.

Regardless of what we want or prefer, the above is what we can actually have. I too used to imagine doing ground missions as individual small Points of Interest, but Warframe can't do that without tanking the host's framerate and spiking their CPU usage. We work with what we have, and what we have is a segregation between Space and Ground components. Everything I've proposed above is an attempt to reconcile this limitation and come up with a way to integrate both ground and space activities within a single Railjack mission without having either feel superfluous relative to the other.

I believe that the primary reason people dislike Corpus Railjack is the poor integration of its ground mission instances. They're behave like standard ground missions with a mandatory Railjack "gate" upfront. Every Corpus Railjack mission plays out exactly the same - clear out the Railjack gate, finish the ground mission. I believe that by mixing-and-matching ground and space components by generating missions procedurally, this can be greatly mitigated. When the opportunity exists to do JUST space objectives or JUST ground objectives or some unknown mix thereof, I feel integration will flow a lot smoother. By removing the standardisation, the effect is greatly lessened.

To give you an example - I have zone control Bounty objectives in pretty much every Free Roam zone. I hate them in Cetus, I hate them in Fortuna, I hate racing the Grineer for kills on Deimos. However, I'm still willing to do them if they show up as one part of a much larger Bounty. When it's not a thing I don't like every time all the time, it's less likely to stand out in my mind. Similarly, having a Bounty made up of five Railjack missions (so five jumps) where some of them are pure space, some of them are pure ground, some of them are a mix would feel a lot less samey than the same framework every single time.

In my opinion, Warframe lives and dies on variety, at least as far as gameplay goes. Despite the theoretical variety in Corpus Railjack missions, given that they each have a different ground aspect, they all play like the same mission because the Railjack aspect is always the same and always there. Kill Crew Ships, blow up Nodes, then ground mission. It's this rigid structure which makes the flaws more apparent than they necessarily need to be. Break up the structure, introduce procedural generation and the experience will improve. At least, intuitively speaking based on what I perceive the issue to be.

Link to comment
Share on other sites

39 minutes ago, Steel_Rook said:

I should probably address this as I skipped over it in the OP, assuming it was common knowledge by this point. Warframe Game Director Steve Sinclair made it abundantly clear that this is not an option. Warframe's technical performance allows either decent space combat or decent ground combat, but not both at once. On Railjack release, we were told that Kosma/Evo Grineer were a significant compromise. Railjack ground combat could only support a very small amount of ground enemies at a time, which is why ground enemies were given vastly inflated stats (to the tune of 20 times the health relative to other Grineer). Corpus Railjack is the same compromise from the other direction. In order to ensure proper ground combat, the space layer has to be temporarily suspended. In order or that to happen, all players must be within the ground layer, where they take part in standard ground missions.

Regardless of what we want or prefer, the above is what we can actually have. I too used to imagine doing ground missions as individual small Points of Interest, but Warframe can't do that without tanking the host's framerate and spiking their CPU usage. We work with what we have, and what we have is a segregation between Space and Ground components. Everything I've proposed above is an attempt to reconcile this limitation and come up with a way to integrate both ground and space activities within a single Railjack mission without having either feel superfluous relative to the other.

I believe that the primary reason people dislike Corpus Railjack is the poor integration of its ground mission instances. They're behave like standard ground missions with a mandatory Railjack "gate" upfront. Every Corpus Railjack mission plays out exactly the same - clear out the Railjack gate, finish the ground mission. I believe that by mixing-and-matching ground and space components by generating missions procedurally, this can be greatly mitigated. When the opportunity exists to do JUST space objectives or JUST ground objectives or some unknown mix thereof, I feel integration will flow a lot smoother. By removing the standardisation, the effect is greatly lessened.

To give you an example - I have zone control Bounty objectives in pretty much every Free Roam zone. I hate them in Cetus, I hate them in Fortuna, I hate racing the Grineer for kills on Deimos. However, I'm still willing to do them if they show up as one part of a much larger Bounty. When it's not a thing I don't like every time all the time, it's less likely to stand out in my mind. Similarly, having a Bounty made up of five Railjack missions (so five jumps) where some of them are pure space, some of them are pure ground, some of them are a mix would feel a lot less samey than the same framework every single time.

In my opinion, Warframe lives and dies on variety, at least as far as gameplay goes. Despite the theoretical variety in Corpus Railjack missions, given that they each have a different ground aspect, they all play like the same mission because the Railjack aspect is always the same and always there. Kill Crew Ships, blow up Nodes, then ground mission. It's this rigid structure which makes the flaws more apparent than they necessarily need to be. Break up the structure, introduce procedural generation and the experience will improve. At least, intuitively speaking based on what I perceive the issue to be.

Well, putting aside the fact we already have ground and space in galleons, and lich assassinations, the main reason i dislike railjack is because even 3 reworks later, its still nothing close to what we were shown or promised at tennocon 2018.

Simply put i think DE aimed way too high for what they could deliver, and promised a gamemode they simply can't give us.

Link to comment
Share on other sites

6 hours ago, Steel_Rook said:

Thus, players can be allowed to zone into the Free Flight area of a Proxima region, pick a Bounty from there, then travel to that Bounty by using the Railjack Navigation console. To avoid an unnecessary zone load, players could additionally be allowed to access a Proxima region's Bounties directly from the Railjack Navigation console such that they can pick a Bounty and queue for it straight from their Orbiter - or rather, the Railjack instance attached to the Orbiter. This should preserve the current implementation of picking a mission and zoning into it directly.

I wonder why they are not doing it now. I mean, I've been doing some low-level earth missions. Most stuffs were the same. There were maybe different number of ships to destroy. When I pick another, similar mission, why it cannot... just spawn more enemies etc?

6 hours ago, Steel_Rook said:

While condensing all players in a Proxima region into a single Free Flight "mission hub" should improve matchmaking times considerably, it does come with a new issue - mission selection. With a node-based system, players can "queue up" for just the mission types they actually want. When different missions offer different rewards, that matters.

However, Free Roam maps already have this issue and the Bounties system already solves this issue. Players can commit to a mission before entering the Free Roam zone and thus they can be matchmade with only people who picked the same mission. Bounty missions are themselves (somewhat) procedurally generated and rotate on a time. This offers both a variety of mission objectives and level ranges, so that problem is already solved as far as I can tell.

Bounties don't solve similar problem. The problem taken from different perspective. It's "grouping different mission types".  You can have 40+ frames but you will use only 1 that couldn't fail. You have 5 stages. You can finish 4 stages with frames you like but then there is 5th. It's bugged or something and you need special frame. Frame that can finish that stage but it's boring. So what you have to do? Do 4 stages with a frame you don't like. That's kills variety which I loved in old WF (pre open world).

And if all missions were the same then I would rather pick one than game dictate it.

10 hours ago, Steel_Rook said:
  •  
  • Drop the concept of a "primary" objective. Every Proxima region would start as Free Flight.

If it was just some endless mission that you can "pick" (or rather being counted towards some goals). Something like:

- Kill X ships

- Destroy X Crewships

- Destroy... some "complex" stuffs (doing to ships/stuffs, hacking some stuffs, destroying other stuffs, hacking... bla bla bla)

 

For example you need to kill 50 ships. Every 50th ship you get reward. Every next ~50 ships will have higher hp/armor/something.  After you cannot kill it you may want to change to Crewships. That way you won't be doing 1 thing ad infinitum.

ps. changing to Crewships may slightly decrease normal ships' level/armor/something.

 

I guess it won't happen because.... stuffs

Link to comment
Share on other sites

1 hour ago, Joezone619 said:

Well, putting aside the fact we already have ground and space in galleons, and lich assassinations...

There's a caveat to that. Galleons in Grineer Railjack use "space combat" rules. That is to say, they use Kosma/Evo Grineer with boosted stats and spawn very few of them. I'm uncertain if Lich Assassination missions do this as well, but the two Liches I've killed since the update dropped have shown me incredibly sparsely-populated Galleons which feel like running through an empty map most of the time. The same applies to the Murex, incidentally, and explains my initial AWFUL experience of it. I'd run around for hours through giant empty chambers, occasionally fighting one or two Sentients before returning to my glorified museum tour. Corpus Railjack is the first time that ground engagements actually "feel" like Warframe combat to me.

Incidentally, Corpus Railjack ground instances all seem to spawn substantially more enemies even during the space segments relative to Grineer Railjack. I absolutely can see this reflected in my framerate. I can play Grineer Railjack at 120 FPS stable no problem. Corpus Railjack drops me to 100, 90, sometimes down to 60 when inside ship interiors. Quality standards aside, that's a substantially more significant load in comparison. And that's with me hosting and solo, so no issues of connection lag. It is a technical issue.

Railjack is not what DE promised in the 2019 TennoCon trailer. I don't think it ever will be and to an extent I don't really want that. What they showed was an unfocused mess of ideas that I strongly suspected wouldn't pan out. However, it COULD still be better, and I don't think it requires massive redesigns or rebalances. It WILL most likely require new code, but I believe that we already have all the systems in place to make Railjack feel organic and open, rather than the "bigger Archwing mission" that it currently is.

 

1 hour ago, quxier said:

I wonder why they are not doing it now. I mean, I've been doing some low-level earth missions. Most stuffs were the same. There were maybe different number of ships to destroy. When I pick another, similar mission, why it cannot... just spawn more enemies etc?

There seem to be technical limitations around what can be loaded into an instance dynamically. All of the ship and PoI interiors seem to be loaded at mission creation during the "Void Tunnel" transition. You have to remember that Railjack is, in fact, a glorified Clan Dojo with teleporters and window view projection. Ship interiors are just rooms elsewhere on the map outside of the miniature "exterior space" room. Remember how your Dojo has to reload every time you place a new room, or every time a room is destroyed? That seems to be the issue with Railjack. When you change the objectives, the map has to reload. The Void Tunnel simply hides spotload so assets don't pop into and out of existence.

This is why I proposed a compromise - instead of trying to cram more missions into a completed instance, let us "travel" via the Nav Console, put us in the Void tunnel and spot-load the new mission around us. Narratively, you can sell it as us travelling to an arbitrary different location in the Proxima region. That bit of Railjack already works.

 

1 hour ago, quxier said:

Bounties don't solve similar problem. The problem taken from different perspective. It's "grouping different mission types".  You can have 40+ frames but you will use only 1 that couldn't fail. You have 5 stages. You can finish 4 stages with frames you like but then there is 5th. It's bugged or something and you need special frame. Frame that can finish that stage but it's boring. So what you have to do? Do 4 stages with a frame you don't like. That's kills variety which I loved in old WF (pre open world).

Fair point, though I should explain. When I say "Bounties," I'm referring to the Bounties SYSTEM here - the ability to pick a mission before entering the mission instance and have mission objectives procedurally generated after the fact. The reason Bounties come in stages is two-fold. For one thing, it allows procedural generation of larger missions from smaller pieces. For another, the individual pieces are just too small to justify hosting them as individual missions. Because the Bounty system was developed before the in-zone mission-selection consoles, a small mission (like find 3 hidden caches) would require the team to zone back into Cetus upon completion. Zoning would take up the majority of uptime, so missions are padded with staggered objectives.

Railjack missions are made up of considerably larger pieces. Well, the ground segments are very large, at least. As such, I see no issue with Bounties consisting of just a single ground mission. Indeed, I see no reason not to let players pick individual missions. That's a matter of "gamification," really, but the system should absolutely be capable of hosting small Railjack missions. Indeed, Free Roam bounties themselves vary in length from 3 to 5 objectives. Here, I'm mostly trying to brainstorm what the game is capable of doing. How it's actually being used, I'm not really smart enough to dictate.

The reason I brought up Bounties is because that system allows you to choose a mission before you enter the ONE unified node where missions take place, thus solving the issue of matchmaking for preferred objectives. Or at least, offering tools to solve that issue.

 

1 hour ago, quxier said:

For example you need to kill 50 ships. Every 50th ship you get reward. Every next ~50 ships will have higher hp/armor/something.  After you cannot kill it you may want to change to Crewships. That way you won't be doing 1 thing ad infinitum.

I think it's useful to consider objectives more complex than just "kill enemies." Railjack can keep spawning enemies on you 'till the cows come home even now. That was never an issue. The primary reason it doesn't is to keep us from grinding ships for drops forever, because grind.

Consider pretty much any other objective, though - anything which requires specific ground combat. Anything with ground combat in it needs to be loaded at zone creation because it's a Dojo room off in the distance. That then means that when you switch objectives, you also need to reload the zone, meaning you need to travel via the Nav console. That's that console's primary purpose. What you're asking is very much possible. You'd just need to gather your crew in the ship, "travel" and move on to the next objective.

"Railjack survival" is probably possible as well. I'd be surprised if we never see that. However, like ground survival, that feels more like an individual monolithic scripted objective, rather than anything procedurally generated. I'm looking to break Railjack down into smaller components which can be mixed-and-matched such that missions consist of more than "space combat, then ground." Smaller ground instances, smaller space battles and the ability to string them along has tremendous potential to create varied content without requiring the creation of bespoke "combo missions."

If I could, I'd recommend the same for ground missions. Unfortunately, the ground mission tech does not have the capacity for spotloading another map without needing to change instances and end the mission. Railjack does.

Link to comment
Share on other sites

9 hours ago, Steel_Rook said:

If I could, I'd recommend the same for ground missions. Unfortunately, the ground mission tech does not have the capacity for spotloading another map without needing to change instances and end the mission. Railjack does.

What about (E)SO? It loads another maps without "end of mission" screen & such.

9 hours ago, Steel_Rook said:

I think it's useful to consider objectives more complex than just "kill enemies."

I think such objectives should be relatively simple.

Doing "more complex stuffs" increases time needed to finish objectives. However it doesn't player enjoyment that much. I prefer single, well thought mission than group of more time consuming little missions. If you want complexity like: "go to ship/hangar, hack some stuffs, go back to RJ, destroy a thing, go to ship/hangar, hack something.... repeat it few times" then I don't want it. At all.

Link to comment
Share on other sites

2 минуты назад, quxier сказал:

What about (E)SO? It loads another maps without "end of mission" screen & such.

SO's "maps" are just rooms stacked on top of each other.

Easy to see when you get through conduit and someone dies in the previous location. You'll see the bleedout marker and timer for that person in about 500-1500 meters away either up or down.

Link to comment
Share on other sites

Just now, Darth_Predator said:

SO's "maps" are just rooms stacked on top of each other.

Easy to see when you get through conduit and someone dies in the previous location. You'll see the bleedout marker and timer for that person in about 500-1500 meters away either up or down.

Isn't really the case? I get huge loading at the end of mission (before conduit appears) so I assume it loads some stuffs. It may loaded it into "500-1500 meters" away but it still loads it at the end of stage. That's my understanding.

Link to comment
Share on other sites

1 hour ago, quxier said:

Isn't really the case? I get huge loading at the end of mission (before conduit appears) so I assume it loads some stuffs. It may loaded it into "500-1500 meters" away but it still loads it at the end of stage. That's my understanding.

Sanctuary Onslaught seems to spotload enemies more so than terrain. I get the same slowdown you're talking about in the middle of rounds quite often, usually just before a large number of enemies spawn into the map. It could be loading their graphics, it could be prepping their AI. I'm not sure. Little else in the game seems to have enemies spawning in such large numbers so quickly, so that's my guess. I don't think ESO is procedurally loaded in, however, as the layout of the run is not affected by player action. In Railjack, what map loads next depends on what the player chooses from the Nav console. There's no way to just load all assets for a Railjack run during initial zone load because what's in a run depends entirely on what players do. No so for ESO. That basically behaves like a standard mission map, with tiles connected by teleporters.

The reason I don't think ground missions are capable of dynamically spot-loading different map assets and changing on the fly is simple: if that existed, DE would have used it. If they could dynamically reload the map around us, then I expect we'd have seen similar "choices" in mission design before - the ability to stay on the map and do another mission without returning to the Orbiter, the ability to transition from one planet to another, etc. Crucially, Railjack's ability to spotload different maps in place is inherently tied to the Dojo - it's why a Dry Dock is absolutely necessary. None of the game's other missions require a specific shared instance like that.

As such, I'm mostly trying to speculate what we could use this tech for moving forward. As it stands right now, DE make very limited use of this tech since all Railjack missions are monolithic. Being able to break them down into pieces and assemble them procedurally would offer tremendous power to the system, I think.

 

1 hour ago, quxier said:

Doing "more complex stuffs" increases time needed to finish objectives. However it doesn't player enjoyment that much. I prefer single, well thought mission than group of more time consuming little missions. If you want complexity like: "go to ship/hangar, hack some stuffs, go back to RJ, destroy a thing, go to ship/hangar, hack something.... repeat it few times" then I don't want it. At all.

I can't really respond to general statements like that, I'm afraid. You seem to have a mind's eye view of both what I'm saying and what you want that I can't really see.

Any Railjack objective which invovles more than just combat is going to have ground mission aspects. As far as I can tell, those require an instance reload to change. Even something as simple as switching from Grineer to Corpus enemies would require a zone reload because the Crew Ship interiors of one faction need to be unloaded and those of the other faction loaded in. You could theoretically keep both loaded, but that's a lot of wasted memory on resources which are typically not going to be needed. Thus, if the new mission objective involves quite literally anything other than "spawn more ships," it's going to require an instance reload, which means it's going to require the player to "travel" through the Nav console. Again - we don't actually go anywhere. The modified Dojo instance just changes around us while we're locked within the "Railjack interior" room with the windows obscured.

I'm also going to have to disagree. Warframe in general doesn't have "well thought out" missions. It has procedurally generated missions comprised predominantly of randomly-assembled assets. Ground missions are 90% identical because only a few isolated tiles or a few isolated enemies are unique to the mission type, if any at all. As it stands right now, Railjack is substantially more rigidly-designed than ground missions because it only really has two missions. All Grineer Railjack missions are the same - they're all Skirmish, which plays out identical regardless of what Points of Interest spawn. All Corpus Railjack missions are the same because they all have identical space objectives, transitioning to a standard ground mission. This is what makes Railjack feel samey and get old fast. Procedurally generating missions by stacking disparate objectives together is a good way to add variety to the experience.

Railjack remains my favourite activity in Warframe, both thematically and gameplay-wise. However, I've ostensibly stopped even looking at my mission objectives by this point because they're always the same no matter what mission I pick.

Link to comment
Share on other sites

1 hour ago, Darth_Predator said:

Can you explain this please?

Cause "end mission" and "conduit appears" are like entirely different event as far as I understand.

I mean that (E)SO is like:

  1. Load some tileset/map (e.g. void) with Griner
  2. Kill some enemies
  3. Timer reach 0.
  4. Warframe start loading some stuffs (games getting laggy - low fps).
  5. After some time Conduit (the thing you transfer to another round/stage of (E)SO) appears
  6. You go through it.
  7. Load another tileset/map (e.g. earth) with Infested
  8. go to point 2.

By "end of mission" I meant single stage, as described in point 5.

Link to comment
Share on other sites

3 minutes ago, Steel_Rook said:

Sanctuary Onslaught seems to spotload enemies more so than terrain. I get the same slowdown you're talking about in the middle of rounds quite often, usually just before a large number of enemies spawn into the map. It could be loading their graphics, it could be prepping their AI. I'm not sure. Little else in the game seems to have enemies spawning in such large numbers so quickly, so that's my guess.

For me it happens when there are no more enemies loaded in the current stage of (E)SO. I have a feeling that's loading maps (different textures etc) but hey, I don't work for them.

8 minutes ago, Steel_Rook said:

I'm also going to have to disagree. Warframe in general doesn't have "well thought out" missions. It has procedurally generated missions comprised predominantly of randomly-assembled assets. Ground missions are 90% identical because only a few isolated tiles or a few isolated enemies are unique to the mission type, if any at all.

I mean mission types not maps.

PCG created maps have different room locations that adds "something different". However after thousands of mission you will find them the same. That's the "grind aspect" problem. PCG maps are fine. Could they be better? Sure? But is it worth the effort? I'm not sure.

 

Leaving topic of PCG maps, I meant mission types are "well thought out". You get Spy where you want to solve some puzzle. You get (Mobile)Defense where you need find a way to defend a point of interest. Some are easier. For example Exterminate requires you to kill N enemies. It still provides you with a well defined goal.

Of course you can cheese many content. If you pick Nuker that can melt enemies then you can use it on any mission that requires you to kill enemies. However that's a problem of player/enemy balance, not in mission type itself.

 

27 minutes ago, Steel_Rook said:

Procedurally generating missions by stacking disparate objectives together is a good way to add variety to the experience.

Sure, but you cannot just mash different stuffs and expect it to be good. That's not how PCG works. It may work for maps because it requires maps not to be 100% the same. However for PCG-missions, is not that easy. You can create:

- adds too little to missions

- over complicate missions

Striking right balance between simplicity and complexity is very hard. To be honest I prefer non open-world/RJ missions because hit good point between complexity & simplicity. Open-world/RJ, while not "very PCG" are not good.

Link to comment
Share on other sites

39 minutes ago, quxier said:

Leaving topic of PCG maps, I meant mission types are "well thought out". You get Spy where you want to solve some puzzle. You get (Mobile)Defense where you need find a way to defend a point of interest. Some are easier. For example Exterminate requires you to kill N enemies. It still provides you with a well defined goal.

Consider the actual structure of those "missions," however, and you'll realise they aren't missions. They're tiles. What is a "Spy" mission? It's three unique tiles in a procedurally-generated map. Literally everything aside from those three tiles is generic. What is a Mobile Defence? It's three modified generic tiles in a procedurally-generated map. You could easily have Spy with just a single Spy vault tile. And we do - that's how Invasion Spy missions work. There's no inherent mechanic which makes a mission with 1 Spy tile, 1 Mobile Defence tile and 1 Sabotage tile from existing. These are all mission design pieces which the game just so happens to bundle together for ease of matchmaking and programming.

Now, you might argue that some mission types aren't just a tile. You bring up Extermination as an example, and it's good one. That's actually even less "designed" because Extermination is simply an objective draped over a fully procedurally-generated map. There's nothing unique to one Extermination mission relative to another besides the map. The same goes for Survival - it's an objective and a map-wide modifier over a procedurally-generated map. But think about the above again - is there any reason we can't have a map with 1 Sabotage tile, 1 Mobile Defence tile and 1 Exterminate objective?

Well, there are a couple of reasons we don't have either of these examples. One is fairly obvious, and that's matchmaking. DE have chosen to handle matchmaking via nodes, so the contents of a node need to be fixed for all players across the game. It's easier, then, to bundle mission objectives together into mission "types" such that players will have access to roughly the same experience everywhere, plus it's easier to matchmake. Bounties solve this issue by offering a list of missions that players can run or queue for before zoning it. The other issue is one I speculate about, but it's "tech." I don't believe DE have the capacity to procedurally generate missions with non-standard objective makeups. Not ground missions, anyway. Their code requires that missions conform to their specs. Bounties, again, solve this issue by allowing designers to procedurally string disparate objectives together without limitation. Railjack further solves this issue by allowing designers to have multiple objectives active simultaneously while still retaining the ability to lock objectives behind each other.

Consider a basic Corpus Railjack mission. It starts with two simultaneous objectives - kill ship, destroy nodes. Only when both are complete does the mission advance to the next objective - enter ship, then ground mission. The ground mission itself is still monolithic, however, as DE are reusing their ground mission scripts which - unlike Bounties - don't seem to like deviating from the script. Now imagine being able to mix-and-match all of these objectives. Obviously, space and ground missions have to be segregated, but think of the following:

Zone into a Railjack instance, board a ship without combat. On board, do one Spy objective followed by a Mobile Defence objective. Leave, then fight off waves of ships. Once that's done, take out two Points of Interest. Then leave. I suspect you might dislike this for being overly long and complicated, but you have to recognise the merit of how it was put together. I literally pulled a few objectives at random, stuck them together in sequence and gave you the backbone of a mission. If it's too long for your taste, then there's no reason shorter missions can't be created, consisting of just one or two objectives instead of the four I listed. All of this can be done within a single Railjack instance, after which players can "travel" to another Railjack instance and do another procedurally-generated mission without ever leaving the Proxima region. No need for nodes, no need for "space then ground" every time.

The freedom to create missions from component parts allows a vast variety of missions. Sure, not all of them will be good, but they will all certainly be a lot more varied. It's true that playing a game long enough will make us familiar with all of the component pieces of procedural generation, but procedural generation makes it easy to just add more pieces. Take Volatile, for instance. That too is just a single tile. With a mission design system which allows the game to just drop in tiles, adding Volatile would have done more than just add four nodes to the game. It would have affected all nodes in the game by throwing more pieces into the procedural generation pile. Add enough pieces and you pass a "critical mass" of variety where pieces start feeling new simply because we haven't seen them in a while. With enough pieces in the mix, every individual piece shows up less frequently and so feels less repetitive.

I'm convinced that Railjack has the capacity to do all of the above. Indeed, I believe it already has the capacity to do most of it.

 

56 minutes ago, quxier said:

Striking right balance between simplicity and complexity is very hard. To be honest I prefer non open-world/RJ missions because hit good point between complexity & simplicity. Open-world/RJ, while not "very PCG" are not good.

What makes it so hard, do you figure? I've played plenty of Bounties and those generally feel pretty well-balanced in terms of simplicity and complexity. Individual problems do exist, but they're usually restricted to individual pieces which can be improved in isolation. Deimos is a good example. The "Protect the Archaeologist" mission used to be absolutely terrible because it was slow and very easy to fail. DE improved it first by adding regenerating shields to the Shield Osprey (this making the secondary objective less volatile) and then streamlining the sample acquisition and transmission process. As much as I dislike Deimos, I thought that more or less fixed the uneven complexity and difficulty of those specific Bounties.

Just because a design is difficult to pull off doesn't mean we shouldn't try. Ground missions exist as they do largely because they date back to 2013 and Warframe's original launch. They're a very heavily limited technology, albeit generally a pretty stable one. That design works for them, as Ground missions are typically fairly quick and simple. They don't work nearly as well when more elaborate systems are thrown into the mix, however, because their rigid structure itself starts to become an issue.

Take Archwings, for instance. Uranus is an attempt to mix Archwings into ground missions, and it's TERRIBLE. Underwater sections are awful not just due to the slow movement but because the attempt to mix them into the mission seamlessly is not seamless. The idea is solid, but the technology prevents it from being implemented very well. Imagine that instead of floating through tiny claustrophobic coral tunnels, we left the undersea labs entirely and could swim freely in open ocean, able to access other facilities organically. That would be a far better use for the Archwing tech, but that... requires Railjack technology. It requires mixing a proper Archwing layer with a proper Ground layer procedurally, which ground missions just can't do.

Sure, ground missions as they are work fine. However, they can't do anything more complex THAN just ground missions. If Archwings, Necramechs, K-drives and especially Railjack are to have a point in the game, though, we need a system that's than JUST ground missions. We need a system which can seamlessly integrate ungrounded flight, large outdoor locations and small indoor locations all within the same map. Railjack's system can do this for the most part, but it's not doing that yet. Right now, Railjack is limited by trying to use the old node-based, ground-based mission system where all objectives are predetermined and the only thing that changes is the map. That's fun for a while, but it gets really old really fast once players recognise that they're doing the same mission every time.

I don't for a second doubt that what I'm proposing is going to be easy to implement. However, I feel it has the potential to be dramatically better than what we already have, and that the risk - along with the work - is worth it.

Link to comment
Share on other sites

2 часа назад, quxier сказал:

I mean that (E)SO is like:

  1. Load some tileset/map (e.g. void) with Griner
  2. Kill some enemies
  3. Timer reach 0.
  4. Warframe start loading some stuffs (games getting laggy - low fps).
  5. After some time Conduit (the thing you transfer to another round/stage of (E)SO) appears
  6. You go through it.
  7. Load another tileset/map (e.g. earth) with Infested
  8. go to point 2.

By "end of mission" I meant single stage, as described in point 5.

Ok, I'm a bit late for the answer but just my 2 cents.

@Steel_Rook mentioned having same laggy bit in mid mission. I have no idea on both of your specs so can't really compare, but I usually run warframe on somewhat medium settings cause otherwise my eyes are starting to bleed. And I don't see any performance impact during both spawn and conduit opened. Also ssd helps a lot.

So I believe this lag can not be attributed to game "creating" new rooms. In addition some of my observations for (E)SO, make of them what you will:

  • When entering the mission loading into the "intro" zone where Simaris cosplays talking sun takes longer than teleporting by entering a conduit. Battle zones are more sparse and assets heavy thus they probably should be generated longer (e.g. asset load), not faster.
  • Don't know if that's a thing these days, but some time ago when host migration occurred in SO (and maybe ESO as well), your next zone was guarantied to be the first one in that particular run, followed by the second, etc. You were basically looped through the entire run again, still jump times are the same, but logically they should be smaller if the game was not generating them again, or the zone/enemy combo should have not been a carbon copy of what you already did.
  • ESO is fixed through the entire week. This means the same tiles, nothing is randomized => load all of that should be better than run a generator each conduit. What's the point if the result is always the same?
  • Not SO related, but, adds to my points. Before upgrading to ssd I was having this "freezes" of sorts when I was getting a mod. Asset was read from the disk and then shown to me. This caused the game to freeze cause 4yo old hdd was slow and stuff. Same goes for some explosions.

All that plus having people up and down mentioned before  is the basis of my believe SO is not generated on the fly. But ofc none of us works at DE, so we are speculating at best.

 

Link to comment
Share on other sites

1 hour ago, Steel_Rook said:

Consider the actual structure of those "missions," however, and you'll realise they aren't missions. They're tiles. What is a "Spy" mission? It's three unique tiles in a procedurally-generated map. Literally everything aside from those three tiles is generic. What is a Mobile Defence? It's three modified generic tiles in a procedurally-generated map. You could easily have Spy with just a single Spy vault tile. And we do - that's how Invasion Spy missions work. There's no inherent mechanic which makes a mission with 1 Spy tile, 1 Mobile Defence tile and 1 Sabotage tile from existing. These are all mission design pieces which the game just so happens to bundle together for ease of matchmaking and programming.

You are mixing things up.

Tilesets are about maps, generating rooms etc. You can have 3 Mobile defenses in 3 rooms next to each other. Those 3 MD can be in different rooms located far away from each other.

Missions are set of rules that you need to fulfill to finish a mission (and get rewards). For example Exterminate requires you to kill N enemies. Spy requires you to avoid lasers/traps/enemies/etc so you won't trigger timer, and you go to to vault to hack console.

Of course how map is designed affects missions. For example New corpus tileset were (not sure if it's still now) terrible for survival/defense as enemies were slow to enter center of the map. However that's not what I wanted to talk (I can talk about it but that's DIFFERENT topic).

1 hour ago, Steel_Rook said:

The freedom to create missions from component parts allows a vast variety of missions. Sure, not all of them will be good, but they will all certainly be a lot more varied. It's true that playing a game long enough will make us familiar with all of the component pieces of procedural generation, but procedural generation makes it easy to just add more pieces. Take Volatile, for instance. That too is just a single tile. With a mission design system which allows the game to just drop in tiles, adding Volatile would have done more than just add four nodes to the game. It would have affected all nodes in the game by throwing more pieces into the procedural generation pile. Add enough pieces and you pass a "critical mass" of variety where pieces start feeling new simply because we haven't seen them in a while. With enough pieces in the mix, every individual piece shows up less frequently and so feels less repetitive.

I'm convinced that Railjack has the capacity to do all of the above. Indeed, I believe it already has the capacity to do most of it.

You can add new pieces but you need to be careful to make those pieces good enough.

1 hour ago, Steel_Rook said:

What makes it so hard, do you figure? I've played plenty of Bounties and those generally feel pretty well-balanced in terms of simplicity and complexity. Individual problems do exist, but they're usually restricted to individual pieces which can be improved in isolation. Deimos is a good example. The "Protect the Archaeologist" mission used to be absolutely terrible because it was slow and very easy to fail. DE improved it first by adding regenerating shields to the Shield Osprey (this making the secondary objective less volatile) and then streamlining the sample acquisition and transmission process. As much as I dislike Deimos, I thought that more or less fixed the uneven complexity and difficulty of those specific Bounties.

For example Cetus bounties. One stage is just walking around Drone, defending it (killing enemies). You can do it probably in very different way (I haven't tried Banishing it but it might be possible). On other hand you have "protect the camp" were you are required to kill enemies that can be far away from you (I heard that's a bug). It's much harder than other one I mentioned.

2 hours ago, Steel_Rook said:

Just because a design is difficult to pull off doesn't mean we shouldn't try.

We should try some different stuffs but I'm very skeptical (last 2 frames were bad. Sevagoth got some issues, not power-wise; and Yareli... yeah... month after release and we get nothing).

2 hours ago, Steel_Rook said:

Take Archwings, for instance. Uranus is an attempt to mix Archwings into ground missions, and it's TERRIBLE. Underwater sections are awful not just due to the slow movement but because the attempt to mix them into the mission seamlessly is not seamless. The idea is solid, but the technology prevents it from being implemented very well. Imagine that instead of floating through tiny claustrophobic coral tunnels, we left the undersea labs entirely and could swim freely in open ocean, able to access other facilities organically. That would be a far better use for the Archwing tech, but that... requires Railjack technology. It requires mixing a proper Archwing layer with a proper Ground layer procedurally, which ground missions just can't do.

In my opinion Water-archwing is bad because slowness and because it's not well developed as other systems (frames/weapons). It feels like it restricts you. Water parts seems long because of this.

Hmm.... haven't they procedural generated underground caves in Deimos? I'm not sure if it's generated before open world is loaded or before we enter a cave.

2 hours ago, Steel_Rook said:

I don't for a second doubt that what I'm proposing is going to be easy to implement. However, I feel it has the potential to be dramatically better than what we already have, and that the risk - along with the work - is worth it.

I simply disagree - I don't think it's worth... however I don't know their Tech, people and stuffs so I can be proven wrong some day.

1 hour ago, Darth_Predator said:

I have no idea on both of your specs so can't really compare

I run on the almost lowest settings (I have some 2 maybe 3 effects raised/on) in window mode (Open world were laggier so I changed to window and haven't maximized it).

Link to comment
Share on other sites

16 minutes ago, quxier said:

Tilesets are about maps, generating rooms etc. You can have 3 Mobile defenses in 3 rooms next to each other. Those 3 MD can be in different rooms located far away from each other. Missions are set of rules that you need to fulfill to finish a mission (and get rewards). For example Exterminate requires you to kill N enemies. Spy requires you to avoid lasers/traps/enemies/etc so you won't trigger timer, and you go to to vault to hack console.

In a majority of missions, those rules are attached to either the map or individual map tiles. A Spy mission doesn't require you to avoid anything until you get to the actual Vault tile. You can raise as many alarms as you want and engage in all of the firefights. That doesn't affect the Spy tile. It's only after you open the Spy vault and initiate that specific mission script that avoiding alarms and cameras and lasers and so on actually matters. The Spy tile is fully independent of the the rest of the mission. You could stick a Spy time inside an Extermination mission and nothing of substance would change. The same goes for Sabotage, Mobile Defence, Hive, Assault, Rescue and many others. The tile is what all of the mission objectives are attached to.

Now granted, it's entirely possible that the overall mission needs bespoke scripting to actually operate the tile, meaning that sticking a Spy tile into an Extermination mission may require manually editing the mission script. If that's the case, then that's a limitation of the ground mission tech, not a design consideration. Again - Bounties show us that the tech exists to string objectives together procedurally and even bracket them with a framing device narrative.

The only objectives I suspect might not play well with others are the ones with bespoke map requirements or with their own unique escape sequences. Defence, Interception, Infested Salvage, Defection, etc - those may be problematic. Certain other mission objectives may be difficult to mix together, such as for instance Survival and Spy. Spy requires a fairly lengthy time of not fighting enemies while Survival requires fighting enemies quickly and constantly. Then again, DE have done this before, mixing Vampire/Timed Nightmare stipulations with Spy, or mixing Fissures with Spy ensuring that pre-aggroed enemies will continually spawn within the Spy Vault.

My point here is to show that mission objectives are not inherent to the "mission" most of the time, but are inherent to just a small piece of it. It's entirely possible to mix-and-match objectives and objective tiles together to procedurally generate mission objectives. Considering how long the typical Railjack mission is, it would benefit from such procedural generation. Without it, content is revealed for what it is - two missions across the entire game.

 

28 minutes ago, quxier said:

I simply disagree - I don't think it's worth... however I don't know their Tech, people and stuffs so I can be proven wrong some day.

Do you figure we have a choice? Unless you're of the opinion that Railjack should be abandoned as a failed project, it's going to need to rely on new tech to survive. The entire Railjack concept can't keep going like it has. It doesn't have nearly enough content to be sustainable and it's pretty clear by this point that DE don't have the resources to make enough content for it. They've been spinning their wheels for two years and only just about managed to make it viable. If the mode is to have any sort of traction and any sort of effect on the broader game experience, it can't keep trying to fit itself into standard ground mission design.

Despite the initial loud feedback, Railjack also can't survive as a pure space fighter. It needs proper integration with ground combat or it's not going to get played. The integration we have now is better than nothing from the perspective of "it works mechanically." However, the format is far too limited to actually be fun to play long-term. DE have done a decent job integrating a plurality of ground missions into the Railjack layer, but the integration is poor. The Railjack segment is always the same and it's not particularly compelling.

Warframe doesn't work with monolithic content. DE can't produce enough content fast enough. Their only option is to design procedural systems with which to generate content from component pieces. Trying to create various combinations of Railjack and ground objectives is a fool's errand. Too many combinations, too many nodes, too much work creating redundant bespoke content. Their current approach to creating Railjack content simply does not appear viable given their ability to create it. I don't see another option besides procedural generation.

Link to comment
Share on other sites

39 минут назад, quxier сказал:

I run on the almost lowest settings (I have some 2 maybe 3 effects raised/on) in window mode (Open world were laggier so I changed to window and haven't maximized it).

I thought running windowed had worse performance, but if that works for ya.

Anyways unless DE dev comes down here and confirms the way it works, the argument we are having here is futile.

Link to comment
Share on other sites

16 minutes ago, Darth_Predator said:

I thought running windowed had worse performance, but if that works for ya.

Windowed and Borderless Window have worse performance than Dedicated Fullscreen, yes. I don't think the performance is significant, but it is there. Luckily, Warframe can switch between Windowed and Fullscreen via Alt+Enter in real time without destabilising.

Link to comment
Share on other sites

3 hours ago, Steel_Rook said:

Then again, DE have done this before, mixing Vampire/Timed Nightmare stipulations with Spy

Kill few guys and you are set. In Vampire mode you can even heal yourself. That's not good mixing. It works but... it's not interesting.

3 hours ago, Steel_Rook said:

My point here is to show that mission objectives are not inherent to the "mission" most of the time, but are inherent to just a small piece of it. It's entirely possible to mix-and-match objectives and objective tiles together to procedurally generate mission objectives. Considering how long the typical Railjack mission is, it would benefit from such procedural generation. Without it, content is revealed for what it is - two missions across the entire game.

I don't disagree with you in this. You can proceduraly generate a map, set some timer when you trigger an alarm and locate some console that, after it gets hacked, shows "N Vault hacked" (giving you some rewards). That's easy.

Maybe it would be beneficial for RJ maybe not. Who knows, it's not easy.

4 hours ago, Steel_Rook said:

Do you figure we have a choice? Unless you're of the opinion that Railjack should be abandoned as a failed project,

Now that you mention... I don't like RJ. If I could delete one feature it would be RJ. But to be serious:

4 hours ago, Steel_Rook said:

DE can't produce enough content fast enough. Their only option is to design procedural systems with which to generate content from component pieces. Trying to create various combinations of Railjack and ground objectives is a fool's errand. Too many combinations, too many nodes, too much work creating redundant bespoke content. Their current approach to creating Railjack content simply does not appear viable given their ability to create it. I don't see another option besides procedural generation.

PCG takes time as well. If they can afford a lot of time to it to make it good then sure. Otherwise I don't think they should even try.

4 hours ago, Steel_Rook said:

Railjack also can't survive as a pure space fighter. It needs proper integration with ground combat or it's not going to get played

I wonder about this. I don't like RJ (space fights with ships). Some people like that stuffs (ships & stuffs). Some like to have "less boring" missions so they would take both (space & ground combat combination).

4 hours ago, Darth_Predator said:

I thought running windowed had worse performance, but if that works for ya.

I had to check it, but thanks.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...