Ranged pet setting

Ok, you just missed my point. You claimed this code already existed. I said it didn't. Your retort was "Well just add it."

Fine. That doesn't change the fact that it doesn't work that way now like you said it does.

The code does exist, witness our pets running in and meleeing, it's more of a matter of tweaking what is there (letting us set our pets to NOT use that attack) rather then reinventing the wheel.

The goto/stay commands should override any movement actions other then follow if thats our choice, and from what I keep reading, that is the case.

The AI is broken, and it would be nice if they would fix it, and the code to do that exists as well, it's a matter of taking some of the code from passive mode and applying that to attacks we want to skip.

Except Protector Bots are notorious for going into melee when they have no melee attack. Letting us set pets to not use melee attacks won't help.

I'm not arguing that it's broken. I think the best thing for pet AI would be to let goto/stay override the henchmen's desire to move. I just know that the coding isn't that simple.

Actually, it IS that simple. If/Then/Else is a very powerful override in any programming language, and most of the time and energy for putting said override in place is in finding the start-point of the pet's attack subroutine.

Turning off Melee attacks is part one (I would just want to do this anyway even if it didn't fix the issue) that should stop some minions from checking for range, Prot bots have another issue which is covered bellow.


Actually it is, if you can program the bots to move closer to a target you can turn that code OFF, it's not even making new code, it's just disabling some.

But yes Goto/Stay/Follow should trump everything and if they can do that it would fix most of the issues (hell, if stay worked properly this thread would never have existed).

It's such a pain to have to hit passive/follow so many times in a fight to get my bots to return to my side and then tell them to attack again, this is not WAI.

Unless someone here has actually seen the code that a) handles the pet AI, b) links the pets with the AI (e.g. is it part of the pet itself? Is it linked to an external library? etc.) and c) knows what the underlying data structure that defines pet state and behavior looks like, any amount of speculative pseudocode is completely worthless. The "start-point" you're talking about may or may not even exist, and even if it does, you'll probably have to do a lot more than just add "if (!lastCommandWasGoto)" at the beginning of it.

Lets put it this way. The end results of the code as it is, is rubbish. The AI is extremely stupid for enemies, controller pets, lore and MM pets alike. I've seen AI code that works almost exactly like the pets in this game written in a small handful of lines. If this code is very large at all that would be very mysterious. The mobs in this game have absolutely no ideas about tactics or positioning even. Just "is this in range, then move toward player until it is" and thats it. There is nothing at all to this code, its just a couple of paragraphs -- it simply doesn't DO anything to justify the belief that there is a "large amount of code".

No one who does any programming HAS to speculate about the AI code in this game because its plainly obvious.

"Simple" does not equal "non-trivial".

I'm not saying that there is a "large amount of code". What I'm saying is that there are tons of variables on what might have to be included/added to this. There may (and probably would) need to be changes to underlying data structures, as well as gutting and replacing the individual bits of code, plus the addition of multiple "types" of AI (melee primary, ranged primary, a mix of the two, etc.). All of the changes would need to be vetted for not causing excessive server load, verified against dozens of maps (and, for masterminds alone, six different sets of creatures), then tested some more. Should there be multiple weightings based on potential damage, secondary effects, AoE vs. single target, recharge speed, etc.? If so, how do those interact? How many modifications do we have to make to the power tables for this to function? Will those break anything that we didn't expect? The list goes on and on. On top of that, the testing would need to be pretty intense or it wouldn't actually solve anything (since people would just complain more loudly).

This isn't something that you just fling out there and hope it's not any worse; if you want to do it right (and if you're not going to do it right, why would you spend precious project manager, developer and QA hours on it?), you do it just as carefully as any either major project.

Hence we need a redname to respond to this issue, but as for what you wrote above, I gotta tell you the way the pets feel now seems like it was "flung out there".

I've played plenty of MMO's that have pet classes and they have never been this bad.