PC Overdone horde AI balancing

R

RyanX

Refugee
So now everyone is posting videos testing zombie pathing, showing their examples of how to outsmart the hordes. I'll never get why people want to find the easy way out. To each their own, but part of the fun is going in not knowing what the zeds are going to do, not watching the "cheat code" solution by JoeblowO2TG on Youtube.

/rantoff

To my point...

The AI pathing was waaaaay overbalanced to the point of being easy to figure out. A couple of easy small changes could have been more effective imho. I like that zeds can slip through pillars, that's cool. At least it was for a minute or two. The upside down spikes being removed? Excellent choice. Zombies focusing on 1 area. REAL BAD!

I liked using an above ground pillar style base with upside down spikes. Imagine if all you did was change the pillar thing and remove the log spikes. My horde base is now useless, and I'm ok with that. I get into A17, start experimenting with bases and I can basically herd them like cattle. HERD THEM LIKE CATTLE!

 
Well, i still build pillar bases.Multiple layers, one gets overrun, i am moving to the second layer, and so on.If all layers on the ground gets over run, i move to floor one and shot them from above.My safe base starts from floor 2, so for the first time, you have the reason to build big.I love the idea of zombies getting in my base, what i hate is them hitting the same block and the lack of npc to repair.I believe repair has to have a cost, even a big one, but not in time. Maebe, have some NPC to repair, put some survivors it to help the player.Make them eat food, so player needs to go out to hunt, to loot, etc. The game has a lot of potential, each release is almost another game!

 
All I'll say is if you think A17's AI bad, go back and play A16 (or 15, 14, 13, etc.. for that matter!).

People will always find out ways of beating the AI, and with each build the AI will get tweaked. In reality the best way to get the best AI is for people to produce videos demonstrating it's weaknesses.

It would be lovely if TFP could get data from everyone's horde nights (anonymously, of course) and machine learn it, allowing them to work out a killer AI that wont fall for the easy tricks, but the resources (both in time and money) let alone the whole issue of asking people for their data makes this idea a complete non starter - sadly imo.

 
So now everyone is posting videos testing zombie pathing, showing their examples of how to outsmart the hordes. I'll never get why people want to find the easy way out. To each their own, but part of the fun is going in not knowing what the zeds are going to do, not watching the "cheat code" solution by JoeblowO2TG on Youtube.
/rantoff

To my point...

The AI pathing was waaaaay overbalanced to the point of being easy to figure out. A couple of easy small changes could have been more effective imho. I like that zeds can slip through pillars, that's cool. At least it was for a minute or two. The upside down spikes being removed? Excellent choice. Zombies focusing on 1 area. REAL BAD!

I liked using an above ground pillar style base with upside down spikes. Imagine if all you did was change the pillar thing and remove the log spikes. My horde base is now useless, and I'm ok with that. I get into A17, start experimenting with bases and I can basically herd them like cattle. HERD THEM LIKE CATTLE!
Click to expand...
First of all, you're playing Experimental.

By no means did TFP at any point claim this was the finished product.

It's full of bugs and unbalancing issues and it's not finished.

[RWG being a complete mess comes to mind.]

Yet here you are judging and complaining about something that's not finished yet.

More than one TFP staff member has posted that there are changes coming to the AI.

MORE THAN ONE!

So my question is, how many times does it take before folks like yourself get that A17's AI isn't done?

That's a fair question right?

Secondly, why are you complaining about an experimental build at all?

Constructive criticism is good.

You know... like pointing out errors and bugs.

Mentioning that certain things may not be working as intended.... etc.

What I'm getting from your post, however, is just ranting.

What's your objective?

 
well, actually the "AI" in a16 felt better, because it was more zombie like. I would have liked the ai from a16, but improved (the running in circles was really, really bad), not completely changed. Its also quite bad, that up to 25 zombies (from what i have seen, might even be more) zombies can stack together (not one over the other, but one inside the other, seems there is no collision between them anymore) in one space, and still deal amount of zeds*strength*multiplyer_because_buffed_group damage. Thats really. bad. Don't get me wrong, i am all for the multiplyer, but only if they are NEAR to each other, not INSIDE each other.

 
First of all, you're playing Experimental.
By no means did TFP at any point claim this was the finished product.

It's full of bugs and unbalancing issues and it's not finished.

[RWG being a complete mess comes to mind.]

Yet here you are judging and complaining about something that's not finished yet.

More than one TFP staff member has posted that there are changes coming to the AI.

MORE THAN ONE!

So my question is, how many times does it take before folks like yourself get that A17's AI isn't done?

That's a fair question right?

Secondly, why are you complaining about an experimental build at all?

Constructive criticism is good.

You know... like pointing out errors and bugs.

Mentioning that certain things may not be working as intended.... etc.

What I'm getting from your post, however, is just ranting.

What's your objective?
Click to expand...
Dude you completely misinterpreted my post. Yeah I used caps, but you know how emotion doesn't come across well in text. It was supposed to be more of a funny "herd them like cattle!" not a serious "damn you and your zombies!" My rant was at the "let's figure out the AI and exploit it" mentality. It's all good fun. I have well over 1k hours in this game through several alphas, so I understand what the devs are doing and maybe I just worded it poorly. So yeah, it's all just constructive criticism and I was just suggesting to them that they overshot and to maybe do a little rewind with just a couple of the new changes implemented.

Shiana just above said it well that the a16 AI felt better except the "running in circles" was bad. Well put!

 
well, actually the "AI" in a16 felt better, because it was more zombie like. I would have liked the ai from a16, but improved (the running in circles was really, really bad), not completely changed. Its also quite bad, that up to 25 zombies (from what i have seen, might even be more) zombies can stack together (not one over the other, but one inside the other, seems there is no collision between them anymore) in one space, and still deal amount of zeds*strength*multiplyer_because_buffed_group damage. Thats really. bad. Don't get me wrong, i am all for the multiplyer, but only if they are NEAR to each other, not INSIDE each other.
Click to expand...
I agree with everything you typed.

 
Dude you completely misinterpreted my post. Yeah I used caps, but you know how emotion doesn't come across well in text. It was supposed to be more of a funny "herd them like cattle!" not a serious "damn you and your zombies!" My rant was at the "let's figure out the AI and exploit it" mentality. It's all good fun. I have well over 1k hours in this game through several alphas, so I understand what the devs are doing and maybe I just worded it poorly. So yeah, it's all just constructive criticism and I was just suggesting to them that they overshot and to maybe do a little rewind with just a couple of the new changes implemented.
Click to expand...
Yeah okay I can agree with that.

I think TFP agree to some extent as well since we're hearing rumours about them adding a RND element to Zombies.

Hopefully once Stable is released we've got an AI that's challenging and fun to deal with.

My fear is that there's always going to be a way to exploit an AI in a game like this.

Not enough variables are controllable.

 
a16 AI needed improvement so that it didnt get stuck on infinite loops. It didnt need a total rework. It was a good zombie AI. Trying to make the AI more intelligent, only makes it even easier to exploit.

Before a17, the AI just went for you, directly. That's how zombies behave, and it's even easy to implement. But, it needed to be better because it sometimes got stuck in a loop.

Instead of just improving it for fixing those moments, they tried to make it more intelligent like real human. They didnt realize/didnt care that:

1. they are zombies, going for the direct line to you wasnt bad AI (altho infinite loops was)

2. The smarter they try to programm it, the easier it will be to exploit. The more the AI tries to act like a human, the more patterns it has and more exploitable it is.

Theres so many simple constructions that just make the new AI crazy, making real defenses not needed.

But their official response is that we just cant handle the new AI. Last week a dev told me that I was confortable in a16 thinking that my defenses are good and now Im just struggling with the new AI bc I can not adapt. So it seems they just dont read what we write, they automaticly transform every negative but constructive post into a [wining bc of difficulty post] in their minds.

So instead of spending a week building a cool fort, we spend half a day building a ramp where they go up and drop, because improved AI we cant handle.

 
Last edited by a moderator:
a16 AI needed improvement so that it didnt get stuck on infinite loops. It didnt need a total rework. It was a good zombie AI. Trying to make the AI more intelligent, only makes it even easier to exploit.
Before a17, the AI just went for you, directly. That's how zombies behave, and it's even easy to implement. But, it needed to be better because it sometimes got stuck in a loop.

Instead of just improving it for fixing those moments, they tried to make it more intelligent like real human. They didnt realize/didnt care that:

1. they are zombies, going for the direct line to you wasnt bad AI (altho infinite loops was)

2. The smarter they try to programm it, the easier it will be to exploit. The more the AI tries to act like a human, the more patterns it has and more exploitable it is.

Theres so many simple constructions that just make the new AI crazy, making real defenses not needed.

But their official response is that we just cant handle the new AI. Last week a dev told me that I was confortable in a16 thinking that my defenses are good and now Im just struggling with the new AI bc I can not adapt. So it seems they just dont read what we write, they automaticly transform every negative but constructive post into a [wining bc of difficulty post] in their minds.

So instead of spending a week building a cool fort, we spend half a day building a ramp where they go up and drop, because improved AI we cant handle.
Click to expand...
While I understand what you're trying to say, I'm not sure a rework wasn't necessary.

It may be that the A16 AI was at the limits of what it could do.

At least from TFP's perspective.

However, A17's AI has just begun and we have yet to see what it can achieve with some tweaks.

I'm not ready to write it off just yet.

YES you are right that's it's not working atm, but who knows?

Maybe within a few more patches we'll see some really great things.

I can see a Hybrid coming.

Something that has all the strengths from A16 and 17.

[That might take time.]

 
I'm not ready to write it off just yet.

YES you are right that's it's not working atm, but who knows?

Maybe within a few more patches we'll see some really great things.

I can see a Hybrid coming.

Something that has all the strengths from A16 and 17.

[That might take time.]
Click to expand...
But I agree with you. That would be awesome. The problem is, every dev response I read about it is just saying the new AI and the new defense building that this new AI encourages is fine, that the problem is players not adapting.

Actually I read 1 response from the very start a17 where one of them said they are going to add some random zombies that behave like before, so that the mix would mean less conga and more mixed attacks. But I never again saw them talk about this and several times just disregard our comments about it.

 
But I agree with you. That would be awesome. The problem is, every dev response I read about it is just saying the new AI and the new defense building that this new AI encourages is fine, that the problem is players not adapting.
Actually I read 1 response from the very start a17 where one of them said they are going to add some random zombies that behave like before, so that the mix would mean less conga and more mixed attacks. But I never again saw them talk about this and several times just disregard our comments about it.
Click to expand...
Where's Roland?

We need some old posts from when A16 was first released on the same issue to make a comparison.

He's likes doing that kinda thing anyway. =P'

 
Where's Roland?
We need some old posts from when A16 was first released on the same issue to make a comparison.

He's likes doing that kinda thing anyway. =P'
Click to expand...
Don't bother Roland. He's busy wading through the trolls and fanbois to get to the real discussions and constructive criticism. It's a full-time job really.

 
Don't bother Roland. He's busy wading through the trolls and fanbois to get to the real discussions and constructive criticism. It's a full-time job really.
Click to expand...
I don't know how he's got the patience to do it.

Even more so, Clare in the Console section.

That woman must have a set of wings hiding or something.

Her patience is amazing!

Those console folks taking ranting and whining to a WHOOOOOLE new level.

Make the slightest change and it's....

cNVuiOD.gif


 
You must log in or register to reply here.
Back
Top