Dismiss Notice
This Section is READ ONLY - All Posts Are Archived

"Elevation abuse prevention mechanism" is broken.

Discussion in 'Release 23 Feedback' started by helm, Nov 4, 2015.

Thread Status:
Not open for further replies.
  1. helm

    helm Avatar

    Messages:
    660
    Likes Received:
    1,282
    Trophy Points:
    93
    I'm not sure, I think I did not move much at that point, but everything is possible. I'll need to go get some more arrows (that large timber wolf put a serious dent in my supply) and will be able to do some more testing in the evening (European time).
     
  2. redfish

    redfish Avatar

    Messages:
    11,366
    Likes Received:
    27,673
    Trophy Points:
    165
    Yea, I'm just saying that in certain circumstances the AI makes sense, while not in others. In many cases, monsters leashing is natural behavior, because they're just guarding their territory. They're attacking you because you're threatening them, and when they drive you off, they go back. This would also cover a lot of aggro-dumping tactics -- the monsters would just have no interest in attacking someone who you dragged them to, since they're out of range from their territory and are no threat.

    In other circumstances, it doesn't make sense. Like in this case, where a wolf might want to flee someone who is abusing them from higher ground. Or, if they're in the middle of running back to their leash point, but you attack them mid-way: I'd think they'd stop again and fight you.
     
    Wilfred, Numa and Miracle Dragon like this.
  3. Browncoat Jayson

    Browncoat Jayson Legend of the Hearth

    Messages:
    6,334
    Likes Received:
    14,098
    Trophy Points:
    153
    That I agree with; the current design is lacking realism if they don't continue to agro because the leash has been engaged. Once the leash is triggered, they can start to heal and run back, but if they are engaged they should stop and start fighting again. Sure, there is some abuse risk (ping-pong in and out of leash range so the mob doesnt have time to heal, easier to accomplish with 2 players), but still better behavior that current.
     
    Wilfred likes this.
  4. Scrambler

    Scrambler Developer Emeritus

    Messages:
    228
    Likes Received:
    860
    Trophy Points:
    18
    So I looked at the navmesh on this scene. It's complete garbage. Essentially there is very little contiguous navmesh area - is all disjointed... I'll rebake the mesh and see if it's better.

    Update-

    Turns out the trees "visible mesh" is turned off on this scene. This is actually a bug - they should be on. Our team is making a fix for that right now.

    Bottom line - the trees are creating the disconnected mesh - confusing the heck out of the AIs pathing ability. This is what I mentioned earlier. Until we fix the trees to play nice with the mesh, this is going to be a problem. We are actively working on the solution, but its taken a while to fix all the trees. When that happens we can rebuild the meshes in the scenes and get pathing fixed up.
     
    Last edited: Nov 4, 2015
    Wilfred, Elwyn, Kara Brae and 5 others like this.
  5. helm

    helm Avatar

    Messages:
    660
    Likes Received:
    1,282
    Trophy Points:
    93
    Thank you, good to know as I was just about to go do testing, will do it later then.

    Speaking of which, what in your opinion should we do with bug reports regarding these issues? Are they actually useful at this stage of development? And should they parhaps be consolidated somehow, I mean, if as an archer I file a separate bug report every time I encounter something like this, I suspect the bug forum might become somewhat cluttered.. :) What approach would benefit you the most?
     
    Miracle Dragon likes this.
  6. Scrambler

    Scrambler Developer Emeritus

    Messages:
    228
    Likes Received:
    860
    Trophy Points:
    18
    Bug reports are incredibly useful. They're actually essential at this stage. The bug forum has a lot in it, but our QA team combs through it, and as individuals we find issues related to our systems. In this -particular- instance there is already a bug/task to fix trees, and get the navmesh data fixed up. After that if we are still having issues, then I'd like to see bugs in particular scenes or instances.

    Thanks.
     
  7. Balec Fares deCani

    Balec Fares deCani Avatar

    Messages:
    1,343
    Likes Received:
    4,015
    Trophy Points:
    125
    Gender:
    Male
    This one
     
    Miracle Dragon likes this.
  8. mithra

    mithra Avatar

    Messages:
    101
    Likes Received:
    169
    Trophy Points:
    18
    I do understand the justifications given, we just disagree on what is a desirable design. I do think it's a matter of perspective ( or corporate determination ) what is exploitative and what is game play.

    I have always felt that, as a rule of thumb, if it's FUN then you as a designer have succeeded.


    Everything else is neither here nor there. Balance is largely a chimera and balance will BE whatever you implement. What will aggravate your players more than anything is a constant change of design, nerfing this or that. It's more important that players feel they understand the game for what it is than any particular system be perfect.

    One of the games I love to play to this day is Age of Empires II. And isn't because I enjoy standing toe-to-toe and matching my opponents 1:1 with equal chances of victory, but rather because I enjoy outwitting the AI. I build fortifications that lure them through a gauntlet, focusing maximum firepower simultaneously on a single-file line of units. Am I exploiting the game mechanics? All I know is 3 hours later I felt like I had fun. And as far as I know, no one else out there is particularly concerned with my score. The same should be true of MMOs. My opinion.

    So just a thought : I see you guys talking about leashing here, which I didn't have a name for, but just playing Devil's Advocate, isn't that somewhat un-sandboxy? You guys don't want surprises where players can train mobs onto players but isn't that part of playing a dynamic and unpredictable game? I personally find it annoying, having invested some time fighting a mob, to see it retreat like the Flash and re-heal. But on the other hand there are occasions I want to retreat and not be forced to flee the zone. But am I being exploitative here? Once you understand aggro and leashing distances, you can pick, pull and abort when required - to the point that I really don't have to engage a mob that has any chance of defeating me. How is that different from standing on a building? You've mastered the AI in either case.
     
    Last edited: Nov 4, 2015
    Wilfred, Bluefire and helm like this.
  9. Miracle Dragon

    Miracle Dragon Legend of the Hearth

    Messages:
    2,957
    Likes Received:
    6,313
    Trophy Points:
    165
    Gender:
    Male
    Location:
    Currently: Zhongxian, Chongqing, China
    @Browncoat Jayson , @Scrambler

    I think it would be cool if the 'leash' could be edited from NPC to NPC.. so some leash back to a den, or a treasure, maybe others leash to the last player they saw (or that player's last location before going off grid), or to the nearest spawn point (not just their own, but among a series of spawn points), maybe packs could leash to their leader, so the leader's leash determines the location of the rest..

    As for sniping, and single player.. lets remember that people play single player (not in parties) all the time, even while in multiplayer online mode. Maybe AI should just check to see if there are any other players in the NPC's range, and if no, assume the player is alone and allow sniping.

    If sniping in SotA is not, and will not be a thing.. I don't know what I'm doing here. I don't really enjoy melee, and I don't like running wildly around.. that's not fun for me. I would naturally never engage a monster that wants to kill me without finding a strategic vantage point where I know I'll win and it will lose. If this isn't something I can do in SotA, and is even seen as 'exploitation and cheating'?? I'm not sure, but that feels wrong to me.

    There has to be some point that players can think like their characters, and not like game designers. For this reason, I'm all for adding in collision while in combat, so we can't run passed groups of mobs in the underground, etc. But as ranged attackers, we need to be able to snipe. It's unintuitive not to.

    Speaking of AI, I'd love to see mobs at various levels of 'alertness', where their reaction to getting attacked is faster or slower at varying degrees..

    @helm I apologize if I went off topic. I, from a player standpoint, agree with your OP in that it just doesn't feel right at the current state. Maybe the leashes just need to be extended.. I think fighting most things requires more space for range-fighters than we're being 'allowed' right now.
     
    Wilfred, Bluefire, helm and 2 others like this.
  10. Numa

    Numa Avatar

    Messages:
    2,891
    Likes Received:
    5,620
    Trophy Points:
    153
    Gender:
    Male
    Location:
    Breaker's Landing
    I still play this hybrid RTS-RPG game game called Knightshift, there the Archer class is vulnerable in melee range so destructible wooden towers are provided to balance the equation. These towers are the only places you can snipe from ( you can build them btw) but a couple of bears and wolves will destroy them before long. At which hypothetical point given SOTA's falling mechanics, you would take damage then the mobs close in on you.

    I also imagine that skeleton fire mages would be the worst mobs to fight in towers. Just throwing out ideas, I understand the balance concerns.
     
    helm and Miracle Dragon like this.
  11. John Markus

    John Markus Avatar

    Messages:
    305
    Likes Received:
    324
    Trophy Points:
    43
    Gender:
    Male
    Location:
    Tokyo, Japan
    IMHO, there should not be any of these "Elevation abuse prevention mechanism" in a well established game.
    Using ranged weapons from the high grounds is called "Tactics".

    Instead, you should be spawning creatures with better mobility (capable of flying, capable of throwing rocks, capable of navigating walls and ceilings) that will target campers and make such way of combat unfeasible.
    It's like having a officer present on a theme park. When a visitor abuses the cast member, you dispatch the officer to handle the abuser. If one officer is not enough, send 10 more (depending on the agro) to squish the abuser.
     
    Last edited: Nov 4, 2015
    Wilfred, Bluefire, helm and 3 others like this.
  12. Kara Brae

    Kara Brae Avatar

    Messages:
    3,872
    Likes Received:
    12,189
    Trophy Points:
    165
    Gender:
    Female
    Location:
    Bavaria
    I don't like the fact that I cannot damage anything within range and within plain sight when I have climbed up out of range of retaliation, because it is unrealistic.

    I am glad the wonky Solace Forest scene is undergoing repair. I have noticed the inexplicable strange behavior of wolves there too.
     
    Bluefire likes this.
  13. helm

    helm Avatar

    Messages:
    660
    Likes Received:
    1,282
    Trophy Points:
    93
    @Miracle Dragon, I think what you wrote is most definitely in topic :)
    There are two ways to interpret the subject line; either the EAPM is "broken" in the game mechanics level (i.e.it's bugged) or in the concept level (i.e. it's questionable design to begin with). I think it's currently a little bit of both, acknowledging, as @mithra observed, that especially the latter is a matter of (personal) perspective. So personal perspectives are certainly most welcome here.

    In project management there are two rather commonly used ways to prioritize the complex mess of tasks; namely urgency and importance. I can fully appreciate and agree with the urgency of fixing the bugs in the current design, like navmesh issues (and I'm eager to help in it as much as possible, now that I understand the problem better) -- because Episode 1 needs to deliver at some point down the line.

    At the same time, what I regard as important are the design decisions, and their ability to support the story and immersion in the story. Even if it comes out properly implemented only in the later episodes. And here I think you totally nailed it when saying that "[t]here has to be some point that players can think like their characters, and not like game designers."

    I digress here a bit, but I'm thinking of some studies about how music listeners and (professional) musicians use their brains - with more than a little (almost new-age-hippie level, my apologies to hippies) oversimplification, enjoying music appears to be primarily right-brain activity, whereas with pro musicians most music-related things are predominantly left-brain activities. And the curious point is, that the pros mostly can't seem to help it, being analytical about music, sometimes even to the point of diminishing the enjoyment.

    And why would this little digression be relevant in this context? Because I'm trying to understand, to be empathic, about the question "Gee, who in his right mind could come up the idea that 'if the enemy can't fight back, let's make it invincible', and how exactly?" I can't help wondering if there might be some similar phenomenon at work in here: from the perspective of a professional 3D game developer, well versed in the limitations of game mechanics and quirks of the game engines, it might well feel like a sensible, reasonable, efficient solution that is easy and cheap to implement. I can understand it well. But I can equally well understand that for someone who is only interested in playing, in the story, appreciating the "3D aspect" of modern games mostly for their "natural feeling" and beauty, this kind of thinking might even feel a bit crazy.

    This made me very curious about two questions; namely
    1. What kind of "worst case scenario" did the developers actually have in mind that made "abuse" such a problem, and under what assumptions?
    2. What alternative solutions were considered?
    It is, of course, impossible to consider these questions separately; they form a so-called hermeneutic circle in the sense that the answers to one question form a set of premises for the other question. So one must necessarily take a hermeneutic, adaptive approach to these questions. For instance one might ask "what is the worst case scenario regarding abuse, provided that we've implemented enemy backing away to a safe distance?" This would largely rule out, for example, the scenario mentioned by @Scrambler where players climbed on top of towers and rained fireballs on helpless NPC:s (assuming that the fireballs behave like "normal" projectiles and not like heat-seeking missiles with infinite range).

    Regarding the second question, some things that come to mind, that would also rule out a lot of the abuse scenarios, while being relatively easy to implement, these include but are not limited to
    • enemy backing away to a safe distance, possibly at accelerated speed, and possibly temporarily lengthening the leash if necessary (already mentioned above)
    • enemy backing away to cover (cover should be at least partially effective also against "heat seeking missile" type attacks)
    • ranged attack distance being a factor when determining accuracy (as it should anyway, so that accuracy is diminished at greater distances, less so with higher level skills)
    Note I'm not discussing the more elaborate AI changes as they would certainly take too much development resources at this stage, only the simpler ones that would take the edge out of most "abuse" scenarios, rendering them roughly as harmless as" gustball training". I might well be wrong about it, but in that case I would love to know where exactly I am wrong.
     
    Miracle Dragon likes this.
  14. Scrambler

    Scrambler Developer Emeritus

    Messages:
    228
    Likes Received:
    860
    Trophy Points:
    18
    Good ideas. We want to expand this - I've talked about this internally - but needed to get sign-off/buy-in from leads.

    We will get there - tech. support for this exists to some extent.
     
    Last edited: Nov 5, 2015
  15. redfish

    redfish Avatar

    Messages:
    11,366
    Likes Received:
    27,673
    Trophy Points:
    165
    Yea, I think it would be nice to see a dynamic leashing system, which was adaptive, and changed depending on the goals of the NPC.

    It would also be nice to see leashing interruptible, too.
     
    Scrambler, smack and Miracle Dragon like this.
  16. GreyMouser Skye

    GreyMouser Skye Avatar

    Messages:
    1,163
    Likes Received:
    1,971
    Trophy Points:
    125
    Location:
    Wherever the Moongates take me.
    I really like the idea of 'leash to the leader' for the subordinates. This allows easier decisions for the AI.
    1. Sub/animal - I see you, but the leader did not. I will go a certain distance to attack but turn around quick if you string me out.
    2. Sub/intelligent - I see you and I'm telling the leader. I follow his lead.
    3. Leader/patrol - I can pursue and then acquire a new leash point in some sort of pattern. Maybe I will kill the other mob there or decide to move on to the next point.
    4. Leader/guard - this is my turf. I will defend it. Get back in line here you subs!

    Etc.
     
    Scrambler and Miracle Dragon like this.
  17. redfish

    redfish Avatar

    Messages:
    11,366
    Likes Received:
    27,673
    Trophy Points:
    165
    @GreyMouser2,

    Though, if the leader dies, the subordinate needs to change its leash. It needs to be a kind of adaptive leashing.
     
    Miracle Dragon likes this.
  18. redfish

    redfish Avatar

    Messages:
    11,366
    Likes Received:
    27,673
    Trophy Points:
    165
    Also in regard to groups, I think it would also make sense if the groups could break up depending on the challenges they're facing. And also for individuals : if they had some type of intelligence to pair off with attackers.
     
  19. GreyMouser Skye

    GreyMouser Skye Avatar

    Messages:
    1,163
    Likes Received:
    1,971
    Trophy Points:
    125
    Location:
    Wherever the Moongates take me.
    Yes. I have a lot in the Etc. part. This was just a quick blurb during lunch. I never know how much to think something out without at least an 'I'm interested' from a dev. :)
     
    Miracle Dragon likes this.
  20. John Markus

    John Markus Avatar

    Messages:
    305
    Likes Received:
    324
    Trophy Points:
    43
    Gender:
    Male
    Location:
    Tokyo, Japan
    Here's another reason this needs to go.
    1. Start as ranger.
    2. Somehow get a lot of arrows (like go to dev server and buy arrows in Soltown).
    3. Go back to Solace Bridge, get up on the surrounding walls. Look for a Zombie in the courtyard.
    4. Hooray ! Practice Dummy ! (You will not be able to deal damages, but your skills will raise anyways).
     
Thread Status:
Not open for further replies.