The armor system needs work

[Marceror]

It's good to see Josh's thoughts here, and a confirmation that the AI/encounters need work in order to incentive ranged characters to wear armor.

 

That said, when I look at the current state of the beta, then I look at my calendar to see how little time there is between now and the end of Winter 2014, I'm expecting that a lot of problems we're seeing here will persist beyond the game's release. It's going to take a lot of patches, a lot of iteration and polish and I simply don't see that getting done in the next 3 months, alongside everything else that's yet to be done.

 

So my expectation is that naked archers will be a thing for a while to come yet.

[Hiro Protagonist II]

I think it's wishful thinking to think the A.I will be so good in the final game that naked ranged characters will now become unviable in a lot of encounters. I'm quite sceptical that the A.I. can be fixed to such a degree when Josh also confirmed the A.I. won't be as tactical as IWD2.

[Tartantyco]

I think it's wishful thinking to think the A.I will be so good in the final game that naked ranged characters will now become unviable in a lot of encounters.

 

And as was pointed out, there really isn't any reason why "naked" ranged characters shouldn't be viable in many encounters.

[Marceror]

 

I think it's wishful thinking to think the A.I will be so good in the final game that naked ranged characters will now become unviable in a lot of encounters.

 

And as was pointed out, there really isn't any reason why "naked" ranged characters shouldn't be viable in many encounters.

 

Yeah, Josh pointed that out, but it doesn't mean he's right. He's not a god, last time I checked (though I guess he has god-like influence over the game). At any rate, I'm not a fan personally of the idea a character needs to be unarmored to optimize their attacks. This is just way too "gamey" for my taste.

[Matt516]

 

 

I think it's wishful thinking to think the A.I will be so good in the final game that naked ranged characters will now become unviable in a lot of encounters.

 

And as was pointed out, there really isn't any reason why "naked" ranged characters shouldn't be viable in many encounters.

 

Yeah, Josh pointed that out, but it doesn't mean he's right. He's not a god, last time I checked (though I guess he has god-like influence over the game). At any rate, I'm not a fan personally of the idea a character needs to be unarmored to optimize their attacks. This is just way too "gamey" for my taste.

 

 

It depends on what you mean by "optimize". If an archer wants to fire arrows as fast as possible in real life, then technically the best way to do that would be not to wear armor. That said, there are 2 reasons why this isn't the best strategy in real life. Namely, the danger of being shot and/or stabbed. Also, the fact that wearing armor (light armor in particular) probably only slows you down a little bit - maybe 5-10% in total firing speed.

 

Note that neither of these reasons for archers wearing armor is perfectly reflected in PoE at the moment. Once the AI is fixed and the speed penalties are tuned, they will be and the problem will go away. That is - going naked into battle will still give archers a slight speed advantage (as it should), but will come with the disadvantage of being more vulnerable. Just tuning issues.

[Marceror]

 

 

 

I think it's wishful thinking to think the A.I will be so good in the final game that naked ranged characters will now become unviable in a lot of encounters.

 

And as was pointed out, there really isn't any reason why "naked" ranged characters shouldn't be viable in many encounters.

 

Yeah, Josh pointed that out, but it doesn't mean he's right. He's not a god, last time I checked (though I guess he has god-like influence over the game). At any rate, I'm not a fan personally of the idea a character needs to be unarmored to optimize their attacks. This is just way too "gamey" for my taste.

 

 

It depends on what you mean by "optimize". If an archer wants to fire arrows as fast as possible in real life, then technically the best way to do that would be not to wear armor. That said, there are 2 reasons why this isn't the best strategy in real life. Namely, the danger of being shot and/or stabbed. Also, the fact that wearing armor (light armor in particular) probably only slows you down a little bit - maybe 5-10% in total firing speed.

 

Note that neither of these reasons for archers wearing armor is perfectly reflected in PoE at the moment. Once the AI is fixed and the speed penalties are tuned, they will be and the problem will go away. That is - going naked into battle will still give archers a slight speed advantage (as it should), but will come with the disadvantage of being more vulnerable. Just tuning issues.

 

Yeah, and I guess I'm generally good with that. My main point is that I don't expect the game to be tuned to that level for a while yet, certainly a good while after the scheduled release. So I expect we'll be seeing a lot of this for a while to come yet.

 

Personally, I've opted to play the game on normal difficulty, and keep all of my characters in armor of some form. The game isn't overly difficulty like this, so I'm playing the game the way that I prefer -- even if I'm not using what is currently the "optimal" tactic.

[Matt516]

Oh certainly - I do the same thing, playing about half optimal and half RP. But I think they can certainly get those two things fixed before release. Armor speed penalties can be adjusted in about an hour of work (if that), and AI, while a bit more difficult to fix, is something they've already said they're working on. So I'm fairly hopeful they'll get it at least mostly sorted before release.

[Marceror]

AI, while a bit more difficult to fix, is something they've already said they're working on. So I'm fairly hopeful they'll get it at least mostly sorted before release.

<--------- The hopeless cynic hopes you're right.

[Matt516]

 

AI, while a bit more difficult to fix, is something they've already said they're working on. So I'm fairly hopeful they'll get it at least mostly sorted before release.

<--------- The hopeless cynic hopes you're right.

 

 

Well, I should preface this by saying that I do have little "real" programming experience (I work mostly with MATLAB). But - from what understanding I do have of the logic and mathematics behind this kind of AI, I can't imagine that it should be too complicated to implement. Assuming it will work with a weighted probability system like Josh Sawyer laid out here: http://jesawyer.tumblr.com/post/79942961685/answering-about-ai-enemy-types-and-their-abilities , implementing it is easy - only tuning should be difficult. And even then, that shouldn't be too hard (from how I would imagine doing it if it were my job).

[Namutree]

I've said this before, but I'll say it again.

 

The speed penalty on armor needs to be reduced by 20%. When I say 20% I mean that a 30% penalty becomes a 24% penalty. Right now I feel that being naked is too good; even if the ai gets better.

 

I'm not trying to say that being naked shouldn't be viable; only that it is too good right now. 

[Marceror]

@Matt516

 

I guess where I'm coming from is that while you may be right that it's not incredibly complicated, I worry that the sheer volume of stuff that still needs to get done will render even not super complicated stuff impossible to polish in a 3 month timeframe. And of course, almost nothing goes as smoothly as you expect. You end up running into unanticipated challenges that turn your 40 hours of projected work into 80. That's probably happening all over the place right now at OE, and causing programmers to have to spend more time on the more critical issues, leaving this sort of polish type stuff lower on the priority list.

 

We'll see of course. As I said, I think I'm finding a sweet spot that is working okay for me in the game, so I just do what I have to do until things are a lot more polished. 

[PrimeJunta]

Well, I should preface this by saying that I do have little "real" programming experience (I work mostly with MATLAB). But - from what understanding I do have of the logic and mathematics behind this kind of AI, I can't imagine that it should be too complicated to implement. Assuming it will work with a weighted probability system like Josh Sawyer laid out here: http://jesawyer.tumblr.com/post/79942961685/answering-about-ai-enemy-types-and-their-abilities , implementing it is easy - only tuning should be difficult. And even then, that shouldn't be too hard (from how I would imagine doing it if it were my job).

 

I don't think the AI algorithm itself ought to be particularly hard. I think the tricky bit would come from its interactions with the other systems. Pathfinding for example. Ranged attacks are simple to target, but a melee unit would have to determine if it's able to reach the unit it's targeting, and revise its decisions as the situation changes. If this doesn't work well you'd get some really weird behavior. And of course you don't have unlimited CPU to play with.

[Marceror]

That sounds right. I remember when Total War Rome 2 launched, after promises if the most advance AI yet, and the AI appeared to be the most inept yet.

 

In many cases it wasn't the AI that was truly the problem. It was pathing and other systems that weren't playing well with the AI. It's probably not nearly as complex of an issue in PoE as in R2TW, but the point rings true.