GW2 and SW:TOR at the forefront of MMORPG?

[Tru Reptile]

The AI and aggro code has to be simple, because although regular machines can run this easily, servers has to do this for everyone playing the game and all monsters involved. This is why the AI and aggro is simplified in online games.

I don't buy that argument.

It was released about 8 years ago, btw.

[Alaris]

How was the combat like? Monster density? Pathfinding?

I'm not trying to disagree... what I said is true, it's more computationally intensive to make multiple checks (vision, sound, smell), especially if the forms that need checking aren't standard. However, if each monster type has only one sense it checks, then it's pretty trivial (still more expensive if the shapes are different, but by maybe a small fraction).

Also, everything is relative. For example, they may spend more server CPU cycles on AI (especially senses) but then economize on pathfinding or combat. Of course, they may also spend more on servers... who knows...

----

tl;dr

Can it be done? Yes.
Is it costly? It's certainly an added cost in server strain.
Is it worth it? Depends on the game & customer.

[Tru Reptile]

Unfortunately, the combat was pretty damn slow and boring (although I think it was a flaw of the combat system, and not because of complex mob AI). Mob density seemed about the same as in WoW, as did the pathfinding.

I'm sure it does take more time and money to implement mob AI as complex, but it really should be done if MMO's are to evolve. I'll be a sad panda if the aggro system in GW2 is the norm...

[Alaris]

It's easy to implement. In fact, I could do it. But it takes more resources to run in real-time.

Part of the issue is that you also need to consider the AI. If you can't see behind you, how do you move? How often do you turn around? Monsters with a huge blind spot would adapt (you'd think) to move in a way to reduce their weakness, no?

[BrotherGrimm]

Wouldn't allowing players to downscale graphics to a GW1 medium level be enough to allow even low-end pc's to compute this?

I tend to disagree with Enion's assessment of this issue. Graphics settings and PC graphics power has nearly NOTHING to do with AI detection calculations. Most graphics these days are done ON the graphics card and the only system resources being used is the PCIE memory bandwidth being used.

A general circular detection AI is VERY simplistic and requires very little CPU processing to calculate. Is something hostile to me within X distance of me and/or has an ally detected an enemy?

The computatoinal power for a line of sight / cone of perception type of detection is orders of magnitude more complex and requires EVEY enemy to be running this much more complex code. While the more complex code MUST take local geometry into consideration, that's something that is based on the pre-programmed / rendered world, not the direct graphics being displayed. The complexity of the area would matter, not the complexity of the PC graphics system or those settings.

[Alaris]

AI is done server-side.

Also, it really depends on the complexity of the algorithm. (1) Do you take obstacles in consideration? (2) Do you compute for several senses? (3) Do you include tracking, i.e. if you leave traces and smells the AI can pick up on?

Some games include some of those things, but they tend to be more single-player games with a heavy focus on stealth.

[BrotherGrimm]

I don't buy that argument.

It was released about 8 years ago, btw.

You don't buy the argument, but you also don't know the target user / server scaling these 2 gamers were designed to handle. If economics of running the game require that GW support 100 users per server (as a FREE game, keep in mind) and yet FF11 was targeted at 25 users per server, then maybe they could afford the difference in AI processing power. There is always much more involved in something like this than, "is that technicallly possible".

[Thalanor Thornhale]

According to PC Gamer, candidates for "next-gen" are RIFT, GW2, SW:TOR and TERA.

Each of them has a few aspects of next-gen but none really are.

http://www.pcgamer.com/2011/03/24/ed...mos-look-like/

[Akirai Annuvil]

Apparently the wrong thread, or something.

I won't consider an MMORPG next-gen until it's seemlessly integrated into Facebook and consists of more than combat.

(Hi box, what are you doing with all those other people?)

[Akirai Annuvil]

No you didn't. You did however disagree with the idea that most games are designed with the average user performance in mind, as I said.

Let's recapitulate our replies in this thread:

Technically yes that would work. However, the fundamental problem with developing on PCs, as opposed to developing for consoles, is that not everyone has the same specs. Increasing the specs of a game for PC means that there are less people that can play it without upgrading and therefore less people to buy the game. With consoles you know the minimum spec for all players and you need to base the game on that. With PCs you try to make the game as low spec as possible so that more people can buy it.

That's untrue. Crysis is a prime example, but so is pretty much every other game ever. Otherwise we'd all still be playing at Pong's graphical quality

You never said average performance, you said lowest specifications. That's a substantial difference from average performance. I didn't make that much of a fuss of your mistake aside from pointing it out since I assumed you knew what you meant, but what you meant wasn't what you said. That's why I ended with this:

I'd expect the stealth feature allowed by a proper aggro system to bring in more sales than it'd lose.

So we could further ignore your mistake and continue with an actual substantive discussion.

Now. Having cleared that up.

@Thalanor, if it's done server side I'm not so sure about my previous statement. If the added cost translate to a monthly fee or a need for a korean webshop, the feature would probably directly translate to lost sales. If they don't all the better; it doesn't affect pc's at all, so sales can't decrease.