PC A solution to the "Run away from hordes" problem.

[RestInPieces]

As a relatively old player of the game (have been playing it since the single digit alphas - don't even remember which one anymore) - dude do you even read before you respond? I know it's not doable in highpoly. It has nothing to do with Unity. It's entirely doable with sprites batching within Unity, just like how it has been doable within 2D era engines with 3D terrain and sprites.https://docs.unity3d.com/Manual/DrawCallBatching.html

And here's an example of pathfinding of 100000 objects in Unity without even writing your own pathfinding algorithm:

-Yes there are always ways to optimize but engine matters

-7D is obviously not low poly

-They definitely won't use sprites for zombies and they shouldn't tbh

-Can't batch draw calls on high poly with models with multiple materials (like you said)

-The experiment in the video is isolated - nothing else burdens the cpu

-PC specs that can handle the zombies can't be on the high-end or TFP severely reduce their target audience

-Minimizing draw calls to lessen cpu burden doesn't solve everything - gpu still has to process mesh vertices

-Cpu still has to process physics like colliders, raycasts etc for each zombie, Unity's weakest point is probably the physics engine. I think the cpu should be the game's main bottleneck atm.

So these links don't mean much for this game at least, since they seem to not be able to add many more zombies at their target specs I guess. It is possible in the future though, if they optimize the game more or maybe rewrite some of their code for ECS.

 

[KucheKlizma]

-Yes there are always ways to optimize but engine matters-7D is obviously not low poly

-Can't batch draw calls on high poly with models with multiple materials (like you said)

-The experiment in the video is isolated - nothing else burdens the cpu

-PC specs that can handle the zombies can't be on the high-end or TFP severely reduce their target audience

-Minimizing draw calls to lessen cpu burden doesn't solve everything - gpu still has to process mesh vertices

-Cpu still has to process physics like colliders, raycasts etc for each zombie, Unity's weakest point is probably the physics engine. I think the cpu should be the game's main bottleneck atm.

So these links don't mean much for this game at least, since they seem to not be able to add many more zombies at their target specs I guess. It is possible in the future though, if they optimize the game more or maybe rewrite some of their code for ECS.

Actually most games are lowpoly, the highpoly was just a joke. If 7dtd was using highpoly meshes during realtime rendering.... it wouldn't be realtime rendering xD

Afaik usually the artists make a high poly model which is then baked onto a lowpoly model using textures.

Furthermore something I didn't realize until now is that batching can apparently be used for meshes as well:

All in all, It would be interesting to know what the showstopper is, as this should be entirely possible (within Unity) and apparently even without resorting to sprites.

 

[RestInPieces]

Actually most games are lowpoly, the highpoly was just a joke. If 7dtd was using highpoly meshes during realtime rendering.... it wouldn't be realtime rendering xDAfaik usually the artists make a high poly model which is then baked onto a lowpoly model using textures.

True, still, what they call low poly for real time rendering can be way higher (depending on the game) than a typical low poly game or that horse example. As far as I know as well, since I am no expert on the subject.

All in all, It would be interesting to know what the showstopper is, as this should be entirely possible (within Unity) and apparently even without resorting to sprites.

They had said that they will focus on optimizing at some point and since larger hordes has always been a very popular request, I am hoping that they will find a way and manage to do it.

 

[AtomicUs5000]

I don't see improving the BM as a waste of developer resources. I've worked on large projects with millions of lines of code for several years. What would be a waste is when you work on some major feature for years and never complete it. Whether it was time constraints, the budget, or beyond the capabilities of yourself or your team, even settling for something lesser than what you were aiming for over years of time with thousands or person-hours invested would be a waste. If one person working on it for another 500 hours brought that vision to life, it's a good 500 hours.

Sure the devs have more pressing issues to deal with, but this does not stop them from improving on the features that they have already invested in. They are playing around with shape selection. This is great! Love it! It wasn't broken though... it worked just fine and we used it that way for a long time... and yet this improvement is still worth it because it solidifies what they've created. It secures that investment.

I wouldn't worry too much about bugs. The only time you really focus on bugs is after you have your next version, unless if for some reason some bad code just broke everything. You can't chase bugs all the time when working on something of this scale or it will never get done. There hasn't been such a thing as bug-free software since the advent of operating systems. Getting this game feature-complete is far more of a priority.

 

[OzHawkeye]

I don't see improving the BM as a waste of developer resources.

Nor would I, I'd like to see an even more interesting BM horde as much as the next person, but also to see it prioritised within the larger framework of jobs outstanding between now and Gold. To me, it's current avoidability would be a relatively minor issue in that great big jobs list, but if it rates higher for the Pimps, then I'm all good with that to.

I still believe though, that the "best" solution to any sort of "Run away from hordes" problem, isn't so much about preventing the players escape, as about making the Horde so interesting that most people wouldn't want to skip it in the first place.

 

[meganoth]

dude do you even read before you respond?

I noticed your suggestion of using sprites. There have also been other ideas proposed in a few threads in the forum. All ideas have one thing in common: To implement them in the quality TFP probably wants for their release version needs months of work and it still might not work out eventually. Which is exactly what TFP doesn't want to do at the moment. Except for bandits, because that is the last kickstarter-feature they need to implement before release of 1.0.

Even if it actually were easy to implement, there is no indication that anyone from TFP believes there is an easy solution available. To sum it up: "not doable (for them and in a reasonable time and with Unity and for minimum spec PCs and in this voxel game...)". (quote changed for clarity).

But there is nothing hindering you from continuing to propose that idea. There have been surprising turn-arounds from time to time.

 

[Roland]

Actually most games are lowpoly, the highpoly was just a joke. If 7dtd was using highpoly meshes during realtime rendering.... it wouldn't be realtime rendering xDAfaik usually the artists make a high poly model which is then baked onto a lowpoly model using textures.

Furthermore something I didn't realize until now is that batching can apparently be used for meshes as well:

It's not a secret. Madmole mentions it often. Voxels.

If the world was a static background then all of what you say would be possible.

 

[KucheKlizma]

I noticed your suggestion of using sprites. There have also been other ideas proposed in a few threads in the forum. All ideas have one thing in common: To implement them in the quality TFP probably wants for their release version needs months of work and it still might not work out eventually. Which is exactly what TFP doesn't want to do at the moment. Except for bandits, because that is the last kickstarter-feature they need to implement before release of 1.0.
Even if it actually were easy to implement, there is no indication that anyone from TFP believes there is an easy solution available. To sum it up: "not doable (for them and in a reasonable time and with Unity and for minimum spec PCs and in this voxel game...)". (quote changed for clarity).

But there is nothing hindering you from continuing to propose that idea. There have been surprising turn-arounds from time to time.

Are you saying that they're doing the absolute minimum effort to get the game released, without any consideration if it's even playable or good at the end. If that's how it is, that's a rather grim outlook on things you have there. I don't know what else to say to you.

The developers are aware of what you are suggesting and did seriously look at it. They even thought about doing something like what you suggested but having 1000's of zombies as part of a huge horde entity that would stay in the background from which about 8 at a time would emerge to actually come and attack so that the illusion of 1000's of zombies surrounding you would be there.
But they abandoned it after looking into it because it wouldn't work.

It's not a secret. Madmole mentions it often. Voxels.

If the world was a static background then all of what you say would be possible.

What about voxels though? Am I supposed to take that obtuse parroting to heart religiously without any context or supporting information?

I sincerely doubt voxels are the root cause of this issue. If voxels are so evil, why can lots of trees exist? Could it be that trees are simpler than zombes? Where is the voxels in that?

So what is it specifically about zombies? Zombie pathfinding algorithm? Zombie AI? Zombie collision detection? Zombie rendering? Zombie netcode? What is the biggest bottleneck and how are voxels to blame?

 

[pApA^LeGBa]

@Roland it is importnant to talk about these things here. It relates to this. If the game has better and more logic challenges in between the BM hordes you might think very differently about them. Also they are a bigger part of the game hence more importnant for having a challenge. So mentioning this is accurate.

I wanna see how the guys here think about BM when there is no forge for a month, ammo and guns are rare, food is always on the edge and you gotta do pathfinding because you can´t just easily climb up a vertical hill like a goat plus you can´t abuse quest houses for loot anymore. And so on.

I don´t know when i want to skip a horde. It happens randomly. Be it because i forgot to repair my hordebase and realize it on day 28 at 21:50 or be it because i don´t look at the time and end up beeing too far away from my horde base.

I don´t wanna avoid them because i don´t like to fight them and it might be week 10 when i avoid my first BM or it might be week 2. Or 30. Or never.

Also it´s a bit about immersion. Knowing you have the freedom to simply drive away from zombies (wich is "realistic") just makes the atmosphere in the game much better.

I can see it already, if TFP is going to do what is suggested they gonna dumb down other things for it because people would cry that it is unmfair now that you have to prepare for the BM that you are forced to face. And we end up with even less of every genre other than FPS. (unless the jump and run part wich has no place in a game like this....)

Wich reminds me, the jump and run part is obsolete. People will just destroy everything in a certain area where they can circle. And everything is back to where it was.

@KucheKlimza Well trees don´t move and don´t have pathfinding. Way less of a workload for the PC.

 

[meganoth]

Are you saying that they're doing the absolute minimum effort to get the game released, without any consideration if it's even playable or good at the end. If that's how it is, that's a rather grim outlook on things you have there. I don't know what else to say to you.

Yes, if you twist my words in the most negative way. Maybe add that they are rushing to finish the game after ONLY 7 years in development :cocksure:

If you think the game will not be playable because you get less than a hundred zombies around you then I wonder how you still can endure playing the game? I still have fun after >1000 hours in the game and A18 sounds like it will be great.

What about voxels though? Am I supposed to take that obtuse parroting to heart religiously without any context or supporting information?

You may have no choice. Beyond such general information the developers seldom go into detail. It is their game and we are not the supervisory board. Ask and you may or may not get a more detailed answer. Ask again and you probably will get the answer that they have no time to answer the same questions again and again to anybody who thinks they are idiots

 

[theFlu]

What about voxels though?So what is it specifically about zombies? Zombie pathfinding algorithm? Zombie AI? Zombie collision detection? Zombie rendering? Zombie netcode? What is the biggest bottleneck and how are voxels to blame?

G) All of the above. (Why are you asking when you clearly seem to have gotten the idea already?)

3D pathfinding is by nature an order of magnitude more expensive than the 10.000 ball example shown, and that alone would limit target amount by tenfold. Also, the perfomance in that example was absolutely floored. There's a reason it ended with a 100 animated horses, not a thousand. 100 animated zombies? That seems to be a goal (supported 8 players, default setting 8 zombies per player in a horde, that's 64 already)

Add in the dynamic nature of the environment and you can't use too lazy methods for your pathing. Your target moves, blocks are broken etc, not really practical to calculate anything in advance.

Add in figuring out if there's a potential route to the target via destroying blocks, your 3D A* is getting expensive.

And then you should customize it to make it act like a zombie.

None of that is cheap or easy.

 

[meganoth]

I don´t know when i want to skip a horde. It happens randomly. Be it because i forgot to repair my hordebase and realize it on day 28 at 21:50 or be it because i don´t look at the time and end up beeing too far away from my horde base.

I don´t wanna avoid them because i don´t like to fight them and it might be week 10 when i avoid my first BM or it might be week 2. Or 30. Or never.

Madmole has said a few times now that they plan to add the option that you can stay at the trader for a fee (he recently brought up 5000 dukes as example, but I think they will make the fee gradually increasing from a lower value so you can use it even on day 7). Just keep a stash of dukes around for when you want to skip the horde, and you should be fine.

 

[Roland]

Are you saying that they're doing the absolute minimum effort to get the game released, without any consideration if it's even playable or good at the end. If that's how it is, that's a rather grim outlook on things you have there. I don't know what else to say to you.


What about voxels though? Am I supposed to take that obtuse parroting to heart religiously without any context or supporting information?

I sincerely doubt voxels are the root cause of this issue. If voxels are so evil, why can lots of trees exist? Could it be that trees are simpler than zombes? Where is the voxels in that?

So what is it specifically about zombies? Zombie pathfinding algorithm? Zombie AI? Zombie collision detection? Zombie rendering? Zombie netcode? What is the biggest bottleneck and how are voxels to blame?

Trees are not entities except, I believe, for the brief time they are falling after being chopped completely.

I don’t have the technical knowledge to be able to give you what you want. I know the developers really do want more enemies on screen at once so they have no reason to lie about why they can’t do it. In looking at other voxel games they all seem to have the same limitations on number of entities alive at the same time compared to non-voxel games. So I trust the explanation since it seems consistent and makes sense with what I do know.

I think they could probably handle more zombies than they currently do but I think they want to be able to add bandits and survivor settlements without having to reduce the number of zombies in the world. If they upped the zombie counts now to the limit of what they can do they would then have to cut back for bandits which would be disappointing.

Better to add bandits and survivor npc settlements and then if there is room increase the zombies as well.

There are mods that increase the zombies 2x 4x 8x or 16x what is currently in the world. My computer runs well with the 4x option selected as long as I lower some of the settings which I’m willing to do for more zombies.

But TFP have higher standards than modders do for what they can do because of all the variety of configs of all their customers. So mod it to 16x and see how your rig handles it.

 

[Roland]

Madmole has said a few times now that they plan to add the option that you can stay at the trader for a fee (he recently brought up 5000 dukes as example, but I think they will make the fee gradually increasing from a lower value so you can use it even on day 7). Just keep a stash of dukes around for when you want to skip the horde, and you should be fine.

Which also will make it even more important to fix easy button avoidance of the horde. Why would anyone pay the fee when they can tread water all night for free? Running away has to carry enough risk to it that the player needs to weigh the cost of sitting in a room all night waiting it out with 100% safety for a fee with the worst case scenarios of defending a base, running away, swimming all night long, etc.

 

[RipClaw]

Which also will make it even more important to fix easy button avoidance of the horde. Why would anyone pay the fee when they can tread water all night for free? Running away has to carry enough risk to it that the player needs to weigh the cost of sitting in a room all night waiting it out with 100% safety for a fee with the worst case scenarios of defending a base, running away, swimming all night long, etc.

Not everyone sits with the calculator in front of the game. To be honest, no player I know has ever calculated how much the horde costs him. They just enjoy fighting the horde and don't think about what they could save if they avoid the horde.

And it wasn't that expensive what madmole said. I think the amount was 5000 Dukes. With quests this amount is quickly earned.

 

[meganoth]

Not everyone sits with the calculator in front of the game. To be honest, no player I know has ever calculated how much the horde costs him. They just enjoy fighting the horde and don't think about what they could save if they avoid the horde.
And it wasn't that expensive what madmole said. I think the amount was 5000 Dukes. With quests this amount is quickly earned.

Do you need a calculator to decide whether avoiding the horde on a specific night is worth 5000 Dukes to you? I don't.

Nobody said the amount has to be prohibitive. It just needs to be enough that you might have to decide between hotel and buying a level 5 military vest a week earlier.

 

[RestInPieces]

Wtf what calculator, it's just a decision. Anyway, MM said this:

Its an ongoing effort to remove loopholes. Gas has been nerfed significantly but I need to get a vehicle on a long term save and see how it feels.

 

[RipClaw]

Do you need a calculator to decide whether avoiding the horde on a specific night is worth 5000 Dukes to you? I don't.

No, because I wouldn't wonder if it's worth 5000 coins to spend the horde night at the trader's. I would just see if I had the money or could earn it by then. Spending the horde night in the trader is convenient.

By the way, I probably spend more money on buying ammunition than 5,000 coins a week to fight the horde. I would actually save money if I didn't fight the horde and accept the trader's offer instead.

 

[RipClaw]

Wtf what calculator, it's just a decision. Anyway, MM said this:

Its an ongoing effort to remove loopholes. Gas has been nerfed significantly but I need to get a vehicle on a long term save and see how it feels.

And that's a good decision from madmole. If the fuel is generally scarce, a player can adjust to it. It's nothing that only exists to stop a specific behavior. And it is not something that only affects the gameplay only during horde night.

The player can save fuel by riding a bicycle and then using the fuel in the horde for the generator or for the motorcycle. Both ways of spending the horde are equally affected. And the player can mine oil shale in the desert to overcome the shortage. He can improve his situation through work.

 

[RestInPieces]

He can improve his situation through work.

Yes, it can be either cost or risk (it's cost here) - the point is that the player will have a choice that is "well-weighted" between fighting the horde or working/risking to avoid it. An obvious choice, like being able to walk away or drive forever with overabundant gas, is neither a fun choice to make nor a real choice in the first place.