Guild Wars 2

I played GW2 for a while and enjoyed it immensely. I liked it so much that I abandoned one of my cardinal rules of gaming: NEVER PRE-ORDER.

That was a mistake.

After that came the 5 stages of grief. Or what would have been the 5 stages of grief if the decisions and direction in regards to the game didn’t keep me at stage 2.

It’s their game, so they can do what they want with it. And I’m sure there are people who like the way things have gone. But I just found myself logging in less and less, asking “why are they doing that?” a lot more, and finally just stopped playing altogether.

Lesson re-learned. No matter how great you think a game or a game company is, they can still totally wreck a game for you. Maybe just you, but that doesn’t really matter.

I still check the forums and such from time to time to see how things are going. I still see that they’re on the same path. If it keeps raking in the money I see scant reason for them to stop, but I certainly haven’t come across any incentive to get me playing again.

Creating anything the relies on being able to precisely time something like jumping in a game with variable latency is, in my opinion, not very well thought out.

If it’s something short, and only causes a minor inconvenience if you screw up, that’s not really much of an issue. It could even be fun. But if you spend thirty minutes trying to get somewhere only to fall to your death because the network decided to randomly have a latency spike when you tried to jump, that’s more than a little irritating to say the least.

I don’t hate the jumping puzzles themselves. I hate the fact that whether or not I succeed depends just as much on my network latency as my reaction time. For some of the content, it is a huge PITA to have to start over.

Norn Toon: “Better than moon loot!”
Me: “What is?”
Norn Toon: Holds up helm with no color border
Me: “You’ve been drinking again, haven’t you?”

Bags in bags? Just use the same strategy for your inventory! Have your last bag slot be your “swap space” and your next to last bag slot hold multiple bags. When a bag fills, swap it with an empty one. Sure it’s inconvenient but now you can have 400 slots for BAGZ!

You can even nest this process so you can have infinite slotz for infinite BAGZ! w00t!

All joking aside, it is kind of annoying when you spend 20 minutes doing something and then have to spend 30 minutes cleaning out your inventory. A configurable auto-open/auto-salvage/auto-sell set of options would be pretty useful.

Hmm. Max unsigned integer. Definitely a bug. :P

Good AI is hard. Developing a learning AI is very hard.

However, if you limit the number of potential variables it could be manageable. For example, a “slow” genetic algorithm could be implemented using parameters such as skill timing and chaining. Basically, over the course of time the AI tries random permutations of actions. Variations that are considered “successful” are bred with other successful variations. The resulting offspring are then used to generate new permutations of behavior that should, hypothetically, allow the the AI to perform better.

I call this slow since typically when you run genetic algorithms you do millions of iterations in minutes. Since this would be gated by encounters, the process would be slower. And since players are free to use any number of tactics it could take some time before optimums vs. different tactics are determined. But hypothetically, this would add a fairly large amount of unpredictability to encounters and the AI would grow more and more difficult over time.

I’m glossing over a lot of complexity here, and determining the parameters, mutation rates, evaluation criteria, etc. can be difficult just by itself. But it would certainly make things interesting. Careful attention would also need to paid to prevent the AI from becoming “too good”.

fireflyry.7023:

Capitalism is the most embraced financial model as it keeps players playing, and that’s what most game makers want.

It also creates a level playing field.

Hardly. Capitalism in no way shape or form guarantees a level playing field. The first rule of capitalism is they who have the gold make the rules, and history is replete with examples of exactly how “unlevel” the playing field can be made.

Now apply that to game economies where you have everything in an infinite supply. Obviously you can’t have that be the case as you’d rapidly wind up with Zimbabwe like inflation and the whole game economy collapses. So instead you create artificial scarcity. Certain things are harder to get or harder to make, rewards are scaled so that you don’t get millions of gold in matter of minutes, time gated rewards, etc. But since the economy is essentially unregulated you can still manipulate the market. And clearly, people do.

Eliminating gold would just shift focus from garnering reward for doing pretty much anything to reward for farming the mat that everyone has replaced gold with.

This I agree with. If you take out gold, the problem would simply arise in another medium of exchange. Worse, a barter system would be open to all kinds of potential exploitation.

The core game client looks to be C++, with a DSL (domain specific language) built on top of it for handling things like AI and such. There are several options for DSLs, but most games these days seem to go with C# as a basis. I haven’t tried examining their VFS like DAT file so I can’t say for sure what thy use for scripting.

The back end could be a number of things. Someone earlier said that it has to be C++ on Linux, but it certainly doesn’t have to be. Linux is almost certain, but there really isn’t any reason why the server itself needs to be written in C++. The bottleneck for performance in a game servers is I/O. Any language with a half decent JIT and threading can handle the throughput (I’ve written Java applications that required processing millions of messages and GBs of data for enterprise systems and the language was never the performance problem).

For the amount of data this game needs to track and the number of connections, your talking distributed databases. Most modern DB’s support clustering of one form or another, including Postgres and Oracle. I don’t think there using something like Mongo on the back-end as cool as that would be.

That being said, anything we say is just speculation about the back-end. Unless a dev comes on here and says “Here’s what we used…” there’s no way to tell.

What’s the difference between a Norn man and a Norn woman?

A couple of pints.

I like the idea of optional finishers on certain enemies in PvE, but it would be way too much if it were not optional or if you had to use it on every enemy.

Some players may not play PVP or don’t really like PVP play.