PorkyThePaladin
Arcane
- Joined
- Dec 17, 2013
- Messages
- 5,198
Looks kinda gay.
Question is... will it be real PvP with real consequences or will it be just another FPS grind fest with retarded children who think they are "hardcore" PvPers?
I honestly have no real hope for a real PvP game these days that isn't a gimmick, fad hype, or designed to facilitate unskilled hacks through RMT.
I read the FAQ, but it really didn't have much that I could see in terms of details to the consequences. Maybe I missed it, but I didn't see anything like what happens when you die, what happens to your gear, do you lose exp/skill, etc...
As for the complete emphasis on PvP, That is the only way to do it imo. Games that attempt to do PvE and PvP are always failures and always run into the problems these guys mentioned in the FAQs. You can not balance PvP and PvE, one side always ends up getting the boot to the head. All we have had over the years are PvE/PvP games and they all are nothing more than gimmicks for PvE players to feel like they are skilled.
The temporary campaign world makes sense as well for a solid PvP game. I mean, at the end of the day, PvP is about... well... PvP, not PvE. So having all the PvE designs and content just gets in the way of a soild PvP experience. This gives them a chance to focus on the singular aspect of PvP entirely with no PvE to get in the way of their efforts.
They haven't described all of those things yet but to me, it sounded like serious PvP based on what they did mention, such as being able to erect structures (with appropriate material requirements), control of strategic resource points and transporting (and guarding) resources, involved player crafting with room for experimentation, which coupled with item decay and monsters only dropping reagents means wars will have to be backed up by industry. Plus one of their main guys comes from Shadowbane.
Well, by PvE, I don't mean stuff like raids, or the typical MMO dumbbell mobs. I just mean, more generally, AI controlled creatures and/or factions that allow the player to do something interesting in the world. I've played a lot of MMOs over the years with the sandbox approach and promise of great PvP (Eve Online, Darkfall Online, Roma Victor, Mortal Online, Salem) and one thing that struck me in all of them was how boring the world was minus the PvP. So you were either in the middle of some PvP conflict, or you were bored out of your mind. That led to most players leaving, many before they even got the chance to try the PvP. So an MMO needs something interesting to keep the players in it and to grow its playerbase in order to even get to the point where you have enough for PvP conflicts. And that's why I believe MMOs should offer some interesting PvE content.
Now, you are right in that in the past, many MMOs have stumbled by trying to keep two separate things together, PvP and PvE, but I think they just took the wrong approach. Instead of creating more advanced AI to mimic the player, they gave the braindead mobs more hitpoints or more damage than the player or different abilities, in order to compensate for their stupidity. This led to different numbers in PvP and PvE and all kinds of balance headaches. But that doesn't have to be the only way, in my opinion. I really don't think it's that difficult to create AI that performs similar combat actions as players, and if it does that, it can have similar numbers as players, and you can balance PvE and PvP together.
The reason AI for MMOs is typically quite poor is not because of the length of fights: You can make AI just fine for fights that last only seconds, as is typical of many FPSes: Good AI, very short time to kill. The reason MMO AI is quite poor is because MMO AI is EXPENSIVE. With a single-player game, the AI runs on the user's computer, so you don't care what the it costs the individual user, who isn't doing anything productive with his computer anyway, or he wouldn't be playing your silly game. But with an MMO, every AI must run on the server. But the thing is, running AIs that can occupy the attention of a single user typically can and does consume the bulk of a computer's resources. This isn't a problem with a single player game where the AI runs on the user's computer. But with MMOs, since hundreds to thousands of players must be crammed onto a server instance not really significantly more powerful than your computer, if that, this is going to be an issue. AI thus tends towards the simplistic: Run at the player and attack until dead.No point in making advanced AI for people who think the average fight should only last seconds. When you have mob encounters that take 5+ or more minutes to kill, then... well... the time spent making intelligent AI becomes meaningful, or rather actually practical because you now have time to execute strategy in play. The average PvP encounter in games today are simply fleeting encounters that are more important of who got the first shot off than anything else. Since most of the player base these days are impatient and immature, I doubt you can implement AI of any real meaning as people wouldn't accept fights taking longer and as I said, AI focus is pointless if the fights don't last long.
The reason AI for MMOs is typically quite poor is not because of the length of fights: You can make AI just fine for fights that last only seconds, as is typical of many FPSes: Good AI, very short time to kill. The reason MMO AI is quite poor is because MMO AI is EXPENSIVE. With a single-player game, the AI runs on the user's computer, so you don't care what the it costs the individual user, who isn't doing anything productive with his computer anyway, or he wouldn't be playing your silly game. But with an MMO, every AI must run on the server. But the thing is, running AIs that can occupy the attention of a single user typically can and does consume the bulk of a computer's resources. This isn't a problem with a single player game where the AI runs on the user's computer. But with MMOs, since hundreds to thousands of players must be crammed onto a server instance not really significantly more powerful than your computer, if that, this is going to be an issue. AI thus tends towards the simplistic: Run at the player and attack until dead.No point in making advanced AI for people who think the average fight should only last seconds. When you have mob encounters that take 5+ or more minutes to kill, then... well... the time spent making intelligent AI becomes meaningful, or rather actually practical because you now have time to execute strategy in play. The average PvP encounter in games today are simply fleeting encounters that are more important of who got the first shot off than anything else. Since most of the player base these days are impatient and immature, I doubt you can implement AI of any real meaning as people wouldn't accept fights taking longer and as I said, AI focus is pointless if the fights don't last long.
You say this, but this is how combat works in real life, long moments of boredom punctuated by moments of terror, chaos, and death. And given the timeframes involves, AIs are actually ideal in these situations. What is "too quick to react" for a player is well within the capabilities of a competent AI, which can crunch all the possibilities of relevance and issue a quick response faster than the player can blink. That's why meatbags fear machines in games, because the machines are smarter and better than they are at them.I don't think I agree. I mean, I see the logic in such, but I don't think this is an issue of technical limitations. My point is that there are only so many actionable moments in limited amount of time that strategy is more of a pre-planned template approach with little time to apply reactive measures.
Have you seen people who have made custom Starcraft unit AIs? These AIs are completely superior in every way to the stock and anything normal, non-Korean humans can do. Matched against a human opponent, they will destroy him with half the units with little to no losses. There is no reason that you CAN'T make good AI.In many ways, action games are like the units that carry out the orders in an RTS game. Their role is simplistic, their application of the orders are direct and purposeful while the overall implementation of the combat given by the RTS controller is where the strategy is applied.
That is not true at all: The less time you have in an action system, the better the AI will perform over the player because machines are superior.So, naturally the less time you have in an action system, the less time you have to apply meaningful strategy over that of simple reactive play.
You say this, but this is how combat works in real life, long moments of boredom punctuated by moments of terror, chaos, and death. And given the timeframes involves, AIs are actually ideal in these situations. What is "too quick to react" for a player is well within the capabilities of a competent AI, which can crunch all the possibilities of relevance and issue a quick response faster than the player can blink. That's why meatbags fear machines in games, because the machines are smarter and better than they are at them.
Have you seen people who have made custom Starcraft unit AIs? These AIs are completely superior in every way to the stock and anything normal, non-Korean humans can do. Matched against a human opponent, they will destroy him with half the units with little to no losses. There is no reason that you CAN'T make good AI.
That is not true at all: The less time you have in an action system, the better the AI will perform over the player because machines are superior.So, naturally the less time you have in an action system, the less time you have to apply meaningful strategy over that of simple reactive play.
However, here's the catch: All these AIs are expensive. Running such an AI dramatically increases the CPU overhead of the game, all of which must be borne by the servers in an MMO, since entrusting anything to the client = security fail. On top of that, players ultimately actually prefer their lame, hitpoint-sink bosses. As much as they claim they want a better enemy AI, they will whine bitterly when said superior AI destroys them utterly like the pathetic meatbags they are on even terms or even favorable ones. So essentially, you're increasing your operating costs by demanding processing servers dedicated just to hosting AI, so that you can scare off most of your non-masochistic players. Few people want an AI that will repeatedly remind them to GIT GUD, SCRUB. For the typical Western weakling, being pwned by the AI is like playing on a PvP server populated entirely by Koreans.
Interesting discussion. I also think that AI can technically do a good job of mimicking player behavior in MMORPGs. I am not sure how the length of combat is relevant to that, because whether the combat is lengthy or short, players will have a certain series of steps that they do in it (a human algorithm if you will). So if you see another player on the horizon, you'll approach cautiously or set an ambush, if you get stunned, you'll use some anti stun ability, if your health is low, you'll heal, and so on. All of these should be easily implementable into a computer algorithm for the AI.
You ever train in martial arts or in boxing? The majority or your time is spent instilling reflexive movements. Strategy exists, but they are pre-planned templates that are executed based on a given action or counter action. In MMORPGs, what you describe is "the drop". The person who sees the other player first then executes a pre-planned combination (maybe a set of skills, spells, etc...) and the person who defends either dies, or throws out a reactive in the same manner. Most of the time, people just die and the game is a frag fest. Rarely do fights go on very long as battle these days in MMOs are treated like FPS games where you zerg over and over. Because of limited time, your responses and actions are limited by time and so most things revolve around raw DPS contests. Why do you think classes whine constantly on those forums and why they keep making classes more and more the same with simpler and simpler abilities? Having a bunch of skills and abilities to apply various forums of strategies are meaningless when your fights last under 10-15 and usually opens up too much "gimmick abuse" that is a nightmare to balance.
Now imagine how much you could do, how many more abilities you could find use for, how many more traditional battle strategies could be applied if it took minutes to kill someone? No longer would having "the drop" on someone be sufficient to win. No longer would a pre-planned reflexive combination of key strokes be sufficient for victory. I mean, if one simply wants to kill in a flash, why bother playing an MMO at all? Seems like a complete waste to put that time into your characters development when little of it is of use, might as well play an FPS don't you think?
So my point in how it concerns AI is that with all of that extra time in the battle, more skills can be applied and in different combinations and ways that would not be feasible in short fight. There are many strategies that take time to setup, but they are useless in the traditional FPS style of combat due to the near instant kill style of play.
That hasn't really been my experience in MMO PvP. When I played WoW, PvP would go on for a while unless one party was just much higher level than the other. Lots of abilities, counter-abilities, cc, etc were used. Darkfall Online/Mortal Online PvP was shorter, but it still went on for a bit, it wasn't instantaneous and it wasn't about who lands the first shot if the parties were somewhat equal in skills/equipment.
But ultimately, I think you are arguing about a different thing. You seem to be arguing for longer combat, and that's fine, we are not arguing against it. We are just saying you can code AI to behave like the player, whether the combat is long or short.
Like I said: Long moments of boredom punctuated by moments of terror, chaos, and death.Actually, combat in real life is extremely quick and strategy is as I said, a higher level element placed by those directing the combat. Those in combat are following orders and "strategy" during action is just reflexive training. There is really no drawn out element to apply relevant tactics at the micro level, everything is done at the macro level while the micro level (ie an individual in the heat of battle) carrying out those plans is the action. Now if you are saying combat as per a large over all strategy, for instance the AI actually pre-planning for an encounter and have various contingency lines for reaction, then sure... but that is as I said, all pre-planning and that involves time.
The irony of your statement is that the aforementioned demonstration was done with Protoss, the exact opposite of Zerg.Most of it comes down to getting the drop on someone and the battles become more of a zerg to achieve that.
That is...very much not true. Like I said: I've seen it. The gloriousness of organic meatbags dying horribly against the superiority of the machine. It is, however, true that with human on human play, it often does come down to the ambush, but this is precisely because it exploits a meatbag weakness: People OPTIMIZE for this kind of attack because it WORKS. If it did not work because the AI was capable of effectively countering this, people would stop doing that.That is not what I was saying though. I said that the less time you have, the less "meaningful strategy" over that of "simple reactive play". You can make an AI act faster than a human. Yes. You can create an AI with 100's of thousands of decisions trees to apply reactive play during combat in perfect execution, but when your fights last mere seconds, it all comes down to that frag fest, who got the first drop type of play.
All true, but yet games are designed to cater towards generation ADHD. Gone are the days when a minor space skirmish was an hour-long duel between two skilled captains. People no longer have the attention spans for that kind of thing, and your game won't sell, so the logic goes.I prefer long and drawn out measured battle where strategy over trained reflex is superior. Honestly though, most of the trash talkers that I have known in the past have been gimmick players who rely on "the drop" style of play as at the end of the day, any fool can get good with hand and eye coordination (what do you think Nintendo began as?), but few can really apply successful strategy in play.
Yes, I do, actually. I've done a fair amount of work in this area, probably more than many developers have. I've done good AI, and the result took up three computers to run back then, on a game which only needed one server to serve the game. Fielding an AI population of sufficient size to engage the entire playerbase required 3 times as many computers as it did to just run the game without an AI. The AI was far dominant over the entire rest of the game in terms of resource consumption. So, to run a good AI for a game will involve dedicating the vast majority of the game's computing resources towards AI. This shouldn't be surprising: It's the part of the game that actually thinks!Do you understand hardware architecture, networking code and systems to be able to make such a claim or are you going off of what someone else says?
The graphics are voxel based, that's why it looks kinda cartoonish. What this means is that you'll be able to destroy anything you want in game, it's like everything in the game is made of these little bricks. It's same thing as EQNext...
No it isn't, you're thinking in terms of how things were done in ~2004. One way to circumvent the cpu problem for AI is to remove all AI routines from the server and make 20 headless clients - say 1 headless client per 1 big area, and let them control all AI in that area. In WoW there are 500 mobs per area? 500 AIs on a single modern PC, even if every single one of them gets engaged at the same time - peace of cake.The reason MMO AI is quite poor is because MMO AI is EXPENSIVE.
If the server isn't running the AI routines, then the server has no idea what the mobs are doing, where they are, and how to authoritatively regulate them. If you're furthermore entrusting this operation to the user's client, you've just created a massive security hole. I've thought about the idea of offloading AI onto some random user's computer, so he lacks an incentive to attempt to cheat for himself, but this seems like it would be rather tricky to pull off given the possibility that said user could disconnect or lag out, causing your AI creation to grind to a halt until the failure is detected and the gamestate can be transmitted to another user. Offloading the AI to multiple users simultaneously and working off a consensus could minimize both the risk of tampering and the probability of user loss, but I can't think of anyone who has actually attempted this yet.No it isn't, you're thinking in terms of how things were done in ~2004. One way to circumvent the cpu problem for AI is to remove all AI routines from the server and make 20 headless clients - say 1 headless client per 1 big area, and let them control all AI in that area. In WoW there are 500 mobs per area? 500 AIs on a single modern PC, even if every single one of them gets engaged at the same time - peace of cake.
>no idea what the mobs are doingIf the server isn't running the AI routines, then the server has no idea what the mobs are doing, where they are, and how to authoritatively regulate them.
That's not what headless client means.If you're furthermore entrusting this operation to the user's client