Ghostlight
New member
Did the AI change in either of the last two builds?
PC My honest thoughts on A17 so far
- Thread starter Zorngodofall
- Start date
ricp
New member
I disagree that A16 was "right", and found A16 AI to be positively off putting and made the game far, far too easy, but that's a different argument.
My point still remains the same, though, allow them to work out the AI. Get people finding the faults in it, that way we will get improvements. Stopping half way and "throwing the baby out with the bathwater" gets us nowhere and simply wastes time.
As neither of us have visibility over the actual code there may be significant game breaking reasons why A16 AI couldn't be re-introduced, aside from the fact it would be a huge step backwards (imo, of course).
Chronosphere
New member
yes this is lazy assumption from the writer at hand, we all agree and dissagree hes in the wrong insome sense right?
Ghostlight
New member
Says you. Several people are clearly disagreeing on this! From my point of view the AI is a huge step backwards FOR ZOMBIES. They are not zombies any more. Zombies are mindless, whereas this new AI is smart enough to go round your spikes etc. A zombie, on knowing where the player is, should head straight for the player, nothing else. If it cannot get to him, including by digging or climbing, then it should smash whatever is in the way and what is DIRECTLY in between them both. The one thing it should never do is walk around your line of spikes or run all the way round your base because you have one tiny hole in your wall round there.
Well again, your opinion. In A16 I saw only 2 problems with the AI....running in circles aimlessly and spinning round on spikes. I believe A16 AI with those 2 things fixed, would be far superior to A17. And by superior I mean:
a) More fun to face
b) Less predictable / easy to exploit.
ricp
New member
I at no point suggested it is. What I have suggested is that people stop wanting to throw out all the work done up until now and replace it with what is generally accepted to be a very flawed and bugged AI.
You either continue with A17 AI making improvements, and at present there seems to be only one main exploit that people are getting worked up about - although I fully admit there is more than that to address, or you return to A16 and put in the same if not more improvements and amends to bring it up to the level it needs to be.
If A17 Stable shipped with this AI, I'd be disappointed, but the extremely important point is that it's experimental not stable.
I have also said that people complaining about the AI have every right to. It's not the problem that I have issue with, it's the solution.
ricp
New member
* facepalm *
I think this argument has been roundly debunked as there is not set definition of what a zombie in a zombie game should be like; Romero style shuffler? 28 Days Later manic denizen? A zombie is essentially anything it wants to be.
As I stated several times, people complaining is not the issue, asking for an incomplete system to be removed and replaced by the previous one during experimental builds is simply "daft". You say A16 just needs tweaked. So when does this happen? Are you entirely sure it's just a case of swapping one set of routines out for another. I personally don't know, but from a programming background I would suggest that it's very unlikely to be that simple.
The choice is either improve an AI built for the new world that has been built, or go back and tailor a system that wasn't built for the world we have. A decision made far easier as we are still in experimental.
Ghostlight
New member
I hear this "different" argument a lot. What difficulty did you play A16 on and what gamestage did you take it up to? I played on max difficulty and took gamestage up to 2000 once (3000 was the max in A16 for the record). I can report that horde night (which was roughly day 150) in A16 at those setting and GS 2000, was amazing. Amazing! I lost half a Steel base and just survived it and no more. The horde lasted till after 6am (yup it got light and I was still fighting!!) and was non-stop irradiated Cops (like 3 or 4 at a time). A17, as far as I have been (day 42, GS = 300-odd) is NOTHING compared to that A16 horde. No comparison whatsoever.
Last edited by a moderator:
Ghostlight
New member
I am glad you brought up those 2 examples, as they are surely polar opposites as far as depicting zombies goes. Yet what do both those very different zombie depictions have in common...?
The zombies move in a straight line towards their "breakfast", and they do not problem solve. Even in Land of the Dead where they sort of began to use tools in rudimentary fashion, they still just used them to bash down whatever was directly in between the humans and themselves, with no deviation.
ricp
New member
On max hard, in fact one concern I have about A17 is despite the claims that it would run much better than A16, I find it almost impossible to get it to run at a decent lick which means I can't really have everything set up to max as the FPS needed is just not there - at least it's not on my PC.
As for how long I lasted? I'd say probably about day 80 or so. I am normally utterly bored by about day 25 ish, it's not the game stage or the zombies that stop me, it's the boredom. A16 is forge by day 3, workbench by the end of the week, minibike and steel production by the middle of the 2 week. At no point in time does the AI offer a single shred of resistance.
Postmortem
New member
I couldn't agree more. Although it seems that we are pretty much in the minority with that opinion..
Zorngodofall
New member
Alpha 16 was very exploitable. Playing just on warrior difficulty I managed a gamestage of 1300 by day 7. There was a simple trick to it. Put almost every single point into scavenging. Ended up with a 600 AK and military gear by day 2. It was a really awesome playthrough i want to go back to and work on sometime. I didn't realize what would happen so much. Irradiated spam on day 7 lol, just didn't have the materials.
There's a stunning lack of balance for early mid game and a mismatch of the amount of zombies in POI's. TFP are at an odd place with trying to balance MP and SP.
ricp
New member
They were brought up to demonstrate the gamut, that and to show how flawed the "they aren't even zombies" argument is.
Again you just seem to be sticking to a zombie trope that supports your argument.
In regard to zombies being clever, I honestly thought this was blindingly obvious; performance. Want a zombie AI which goes through a checklist of options every second or so? Sure, fine, makes the game run at a snail's pace. Want to remove that? Use the Unity navmesh and give them a direct route path. I obviously would prefer the former, and I do have issues regarding the latter, but it pretty much comes down to a trade off of "realism" or performance.
You are getting away from the initial point, it's not about people criticising the AI, it's that they somehow want development completely stopped and replaced by what we all agree is an AI that is equally unsuited for the latest build. What's more this change should be done half way through experimental?
Zorngodofall
New member
Mate I think the point is lost on you. What everyone is saying is what we had before was OK. What we have now isn't acceptable and is less preferred than that. The point is, it's god awful and it needs to be changed, whether that's a reversion or a heavy reconstruction is up to TFP. Note this. In 28 days, it took up to the last 3 zombies to do more than 1 point of damage to my base aside from random cop spit, the AI is so bad.
As for performance issues. I've been speaking about this for a long time. Let me give you the excuse I was given; deal with it, it's an early access game and blah blah won't be blah blah optimized until blah. You know what the primary problem is? People just accepted the performance issues for far too long and wait until now to complain about it. The problem has actually been worse and much so, before this point.
Ghostlight
New member
I understand that. I have said all along that A16 had a major balance issue in that the gamestage increased far too slowly. This meant when the player hit max firepower (which I define as having level 600 of everything), the zombies were only at half strength, if that. I played A16 a LOT and I typically maxed out around day 100. However GS was typically 1500 to 1800 meaning the zombies were nowhere near max strength (GS 3000). This leads to 3 major problems:
1) Players got bored in A16, mistakenly thinking they had hit endgame and "won", when in fact they were NOWHERE near endgame.
2) People these days talk about A16 as being too easy when in fact they have never seen anything like the end-game for A16. (The GS 2000 horde I faced in A16 would put *ANY* horde A17 has ever thrown at me to shame.)
3) People think A17 is a step up in challenge from A16, when in fact it is the opposite. After the first week or so in A17 (which is frustrating more than challenging let's be honest) A17 soon becomes the least challenging alpha since, well, ever. It is also represents the EARLIEST that the player truly maxes out (level 100, or day 30) when compared to any alpha before it.
Last edited by a moderator:
Ghostlight
New member
By my calculations, this is impossible. Since 2 of the main variables that fed the gamestage calculation in A16 were
- What day it is
- Number of days the player has been alive
Neither of these could be higher than 7 on day 7, making GS 1300 impossible on day 7.
Ghostlight
New member
No. I am sticking to THE zombie trope, and it happens to support my argument.
I welcome you to name any book or movie where zombies are depicted as problem-solving to the extent they are shown recognizing danger as danger and actively circumventing it. The closest I can think of is Bub from Day of the Dead, and I am pretty sure even he would blunder straight onto the spikes that lay between me and him.
Now this made me chuckle. I wonder have you played A17 far enough to face a reasonably large horde when you have a large and complex base?? Every single time something gets destroyed (say a spike or some barbed wire) there is a huge and horrible "pause" where the game literally freezes for about 1/4 of a second. People have worked out this is the new AI algorithm recalculating the path for several dozen zombies.
You can even demonstrate how bad it is yourself, as follows...
I had a simple 4-sided base with spikes and whirly blades on all 4 sides. Every horde night was basically the same, they came from one direction and more or less attacked one complete side of my base. Then for day 42 I built a 3-block-high Iron Bars fence round the entire base. The difference on horde night was eye-opening. They would run at one side of the fence and attack all along its length. However the moment any block in the fence got destroyed.....horrible pause, as the AI recalculated the pathing*.
Surely if they just single-mindedly headed for the player at all times, the calculations would be much quicker and simpler? In addition, the player should be able to go round the other side of his base and have all the zombies follow him round, hungry for blood (as you could in A16 with a little effort). In A17 they will ignore that and stay entirely focused on the path of least resistance, even if that path is somewhat ludicrous, impossible or exploitable.
Why exactly do you consider A16 AI (with needed tweaks) as unsuitable???? It's definitely not performance, I assure you.
* this was not down to my PC being underspecced. I have every setting (except reflections ) maxxed out and I get 80 to 120 Fps solidly when playing, AND I am the host of a MP server. My PC can easily handle the game.
Last edited by a moderator:
HungryZombie
New member
Ok fair enough. I read too much into your comments maybe.
I'll agree it's probably futile to ask them to drop all their hard work, and perhaps they will flesh things out. This is experimental after all.
Ghostlight
New member
A good developer knows when to give up on a certain path, and throw it all away, no matter how far down the path you got. If that path was a bad choice.