Friday Facts #335 - Scenario changes, Damage effect filtering

[FactorioBot]

Here it is! (beep boop)

https://factorio.com/blog/post/fff-335

[steinio]

Why not just add the colored concrete to the base game, please? This step shouldn't take that much coding time but let us make the base a bit lesser grayish.

[Henry Loenwind]

While you're optimising things, may I remind you that researching an arty range upgrade is still a great way to have your game grind to a halt? Even placing a new arty 2 roboport ranges forward from your existing arties is enough to take a lengthy coffee break.

[Rseding91]

While you're optimising things, may I remind you that researching an arty range upgrade is still a great way to have your game grind to a halt? Even placing a new arty 2 roboport ranges forward from your existing arties is enough to take a lengthy coffee break.

That has already been severely optimized. The fact it works at all now instead of just locking up the game shows that. Unfortunately that's just a numbers thing; biters moving is slow and will always be slow. There's nothing that can be done about that. Shooting at a massive amount of them causing them all to run at 1 point just takes a lot of CPU time.

[Pinga]

A problem with scaling wave defense scenarios is that the only way to scale attacks is to increase the ammount of biters. As you increase the number of players, and they get gear, waves become trivial, even when you have thousands of enemies. But at that point, half the server stopped playing anyways because everything is so laggy and few can even handle the simulation.

A better solution would be the ability to increase the health of biters (during the scenario, outside of its entity definition), so you have fewer entities. This way you could support a lot more players and even have stuff like boss waves (I'm sure the community will get creative about it).

[bman212121]

Why not just add the colored concrete to the base game, please? This step shouldn't take that much coding time but let us make the base a bit lesser grayish.

I actually wouldn't want that. It would be yet another item that doesn't add value to the game outside of ascetics. I think colored concrete is much better suited to being put into a mod, so the mod developer can decide how they want to handle it.

For instance, you'd probably need to put that behind a tech upgrade, so there's another unlock in the tree. Then you need to decide if you want to change the recipes for each color. It would be a bit lame if you can magically make all color types of concrete without having to change the recipe. The base game doesn't really have anything obvious for this, where other mods might have colored items in them already. (Geodes come to mind)

Then the 3rd problem is how does it clash with game items. Does blue concrete work with a blue factory on it? Is it too hard to see? Does blue belts cause the same issue? So there are a lot more things that could cascade of out simply adding the item into the game.

[Henry Loenwind]

That has already been severely optimized. The fact it works at all now instead of just locking up the game shows that. Unfortunately that's just a numbers thing; biters moving is slow and will always be slow. There's nothing that can be done about that. Shooting at a massive amount of them causing them all to run at 1 point just takes a lot of CPU time.

Maybe it would make sense to apply the increase over time, so the newly angered biters have a chance to die. But I'm not sure if this can be balanced in a way to actually work well enough. A dirty cheat could be to prevent arties from firing into sleeping biters while there are more than x biters moving on the map already. Or to simply disable auto-firing while there's an UPS slowdown and biter movement makes up a more than a certain percentage of the UPS.

Sorry, just throwing around random ideas...

[Rseding91]

That has already been severely optimized. The fact it works at all now instead of just locking up the game shows that. Unfortunately that's just a numbers thing; biters moving is slow and will always be slow. There's nothing that can be done about that. Shooting at a massive amount of them causing them all to run at 1 point just takes a lot of CPU time.

Maybe it would make sense to apply the increase over time, so the newly angered biters have a chance to die. But I'm not sure if this can be balanced in a way to actually work well enough. A dirty cheat could be to prevent arties from firing into sleeping biters while there are more than x biters moving on the map already. Or to simply disable auto-firing while there's an UPS slowdown and biter movement makes up a more than a certain percentage of the UPS.

Sorry, just throwing around random ideas...

We already do logic like that for construction robots and get complaints about them all the time when people build crazy amounts of things at once. There's no solution so we just leave it as is.

[Ranakastrasz]

Why not just add the colored concrete to the base game, please? This step shouldn't take that much coding time but let us make the base a bit lesser grayish.

I actually wouldn't want that. It would be yet another item that doesn't add value to the game outside of ascetics. I think colored concrete is much better suited to being put into a mod, so the mod developer can decide how they want to handle it.

For instance, you'd probably need to put that behind a tech upgrade, so there's another unlock in the tree. Then you need to decide if you want to change the recipes for each color. It would be a bit lame if you can magically make all color types of concrete without having to change the recipe. The base game doesn't really have anything obvious for this, where other mods might have colored items in them already. (Geodes come to mind)

Then the 3rd problem is how does it clash with game items. Does blue concrete work with a blue factory on it? Is it too hard to see? Does blue belts cause the same issue? So there are a lot more things that could cascade of out simply adding the item into the game.

I would just make it like hazard concrete rotation. Except probably a different option because blueprint rotation would alter the colors. No need for more items, but allows for asthetic changes.
Could allow that in general for a lot of things, though it may cause problems, since changing color is the basic way to copy a model and use it for something else.

I would also go and add a mirror option, for liquid assemblers, splitters, burner drills and so on, but that is out of scope.

[deef0000dragon1]

I think the colored concrete is cool. Would like the ability to use it in the base game, but I dont know if it would make sense as a normal item.
1: are the colors shown the only colors or does it work like the trains where you can have any arbitrary color
2: if not an item, would getting a blueprint that allows you to set the color work?

[0111narwhalz]

This would probably not help much overall, but maybe you could spread out the arty-aggro'd biter pathing across multiple updates in a wave? The cost of checking whether each biter is within the appropriate range may still be significant, but it's probably less than the cost of actually doing the pathfind all at once. Wouldn't scale great with lots of turrets, I guess.
It could be interesting if biter reaction times were limited to the speed of sound or something.

[ptx0]

This would probably not help much overall, but maybe you could spread out the arty-aggro'd biter pathing across multiple updates in a wave? The cost of checking whether each biter is within the appropriate range may still be significant, but it's probably less than the cost of actually doing the pathfind all at once. Wouldn't scale great with lots of turrets, I guess.
It could be interesting if biter reaction times were limited to the speed of sound or something.

yep, it's called batching, and they refuse to do it.

[Rseding91]

yep, it's called batching, and they refuse to do it.

We already do it: https://factorio.com/blog/post/fff-317 and wayyyy moree.

[ptx0]

yep, it's called batching, and they refuse to do it.

We already do it: https://factorio.com/blog/post/fff-317 and wayyyy moree.

yeah, it's much better in .18.x but still, firing 7 nukes into a vanilla mega nest is a good way to drop FPS to 0. any fewer nukes and we die instead.

[Ubertwink]

To be honest, I am a big fan of DoT being applied every game tick, which makes it feel much more natural. I do understant that it comes at a perfornance cost, but it's a big "feels good" moment.

[Rseding91]

yeah, it's much better in .18.x but still, firing 7 nukes into a vanilla mega nest is a good way to drop FPS to 0. any fewer nukes and we die instead.

That has nothing to do with biter path finding, artillery, or biter movement. You seem to just want to complain.

Maybe you can make your own game and show us how to do it properly?

[milo christiansen]

To be honest, I am a big fan of DoT being applied every game tick, which makes it feel much more natural. I do understant that it comes at a perfornance cost, but it's a big "feels good" moment.

At 60TPS you won't notice a difference.

[ptx0]

yeah, it's much better in .18.x but still, firing 7 nukes into a vanilla mega nest is a good way to drop FPS to 0. any fewer nukes and we die instead.

That has nothing to do with biter path finding, artillery, or biter movement. You seem to just want to complain.

Maybe you can make your own game and show us how to do it properly?

you seem to have woken up on the wrong side of the bed today.

i'm sure firing nukes into a massive nest that lags players who aren't even standing nearby or with enemies visible on screen has absolutely nothing to do with biter path finding, artillery, or biter movements. it has nothing to do with them clumping in multiple thousand large swarms of enemies that consume CPU time. which we reported last week...

[milo christiansen]

you seem to have woken up on the wrong side of the bed today.

I don't think Rseding was the person who woke up pissed.

[<NO_NAME>]

You could also limit the number of particles per amount of damage done.
Particles would double as damage indicator, then.