Friday Facts #166 - Combat Revisit

[Mehve]

On the topic of maintaining cleared-out areas, I'm a little surprised it hasn't specifically come back into the conversation this week, but

I forgot to mention that. This will be solved by the artillery train

As seen not too long ago in FFF#160.

[tehroach]

Unending research will have negative consequences on combat ballance

While I believe that the concept of unending research is good for resource consumption and gives the player a way to exhaust resource fields.
It will undoubtedly make the task of balancing combat for the duration of a game an impossible one.

What if instead of research just blanket upgrading every turret on the map the player researched upgradeable aspects of the turret that could be fitted to an individual turret like a module, but instead of only being able to place one module per slot these modules were stackable, with diminishing returns on each additional module.
ie

• Barrel - increases range of turret
• Feed mechanisim - increases rate of fire
• etc

Each turret would have a module slot for each module type it accepted and the slot could be stacked as high as the player liked.
This would mean that a player could create areas of denial with far fewer turrets, but at an exponentially greater cost in resources and the turrets used in "turret creep" would either be way more expensive or considerably weaker than their defensive brethren.

Brain fart or stroke of genius? - you decide
IMO this overall concept of modules would be a superior one to the endless research one, both as a resource sink and from a factory tinkerers perspective.
ie this small factory has a similar output to my previous MEGA-factory, however, it cost many times more to build.

The way to improve biters is to improve their AI, not to add more enemies.

This is generally easier said than done.
From the general idea of GIGO, you can also expect limited input, limited output.
When you look at the current situation there is really only 3 different enemy types with 4 stages of difficulty and 1 objective.

Given these limitations, you can't really expect AI to be interesting and fast at the same time.

Example - instead of following the same path each time, they will send small attacks to different areas, probing for weak spots, and when they find one send a large attack. That would definitely keep you on your toes and make you pay more attention to defense.

I disagree. Making biters search for weak spots will force players to build turret walls around their factories, basically what we have now.

I would have to agree with R3vo on this point, something like this idea would simply add complexity to the AI with no real noticeable effects.

The ability to give your turrets a target priority would also be very welcome - too often turrets let themselves be destroyed by spitters while some other biters are chewing their way through walls that will take them a few minutes to get through.

I believe that your problem here is simply a symptom of the blanket idea of just increasing HP and DPS of enemies to increase difficulty and then repeating the process for the other side.

While I really like the idea of a turret target priority, but a blanket customization approach to this would also be a bad idea, as it would effectively render spitters obsolete or require spitters to be buffed.
ie to the point where they have equal HP to their melee counterparts, which then only creates another set of balancing problems.

IMO the best way to achieve a turret target priority would be to have an interface where the player could set "this turret only targets this enemy type" or "this turret ignores this enemy type" and to have more variety in turret types, turrets that operate in different ways.
(ie open the ideas spoiler in this post), this way a player could customize their defenses for specific purposes and desired effects and not just be always required to build the "Great Wall of Turrets" in every game.

[Johntron]

Any chance the sourcecode for the internal mod will be made public?

[arides]

Landmines
Strange, I always considered landmines to be a very effective addition to the defense. Yes, their damage isn't great, but:

• They are dirt cheap and manufacture fast
• Drones are able to replenish used mines with no effort
• Blueprints with minefield make creating new fields nearly effortless
• They are low tech, making them available around medium biters / spitters, which they blow up nicely

I wonder, why are landmines considered useless?

[DaveMcW]

Landmines are useless without roboports.
Roboports are a late game building.
Landmines are useless against late game biters.
Therefore landmines are useless.

[Boogieman14]

Unending research will have negative consequences on combat ballance
While I believe that the concept of unending research is good for resource consumption and gives the player a way to exhaust resource fields.
It will undoubtedly make the task of balancing combat for the duration of a game an impossible one.

Unending research, by definition, only affects extremely long running games. By the time you're able to devote resources to unending research, any form of balance has left the building weeks ago and biter attacks are a mere nuisance anyway, while clearing biter camps is just a tedious annoyance. Unending research shouldn't play any role at all in balancing combat.

[Grimakar]

Unending research will have negative consequences on combat ballance
While I believe that the concept of unending research is good for resource consumption and gives the player a way to exhaust resource fields.
It will undoubtedly make the task of balancing combat for the duration of a game an impossible one.

Unending research, by definition, only affects extremely long running games. By the time you're able to devote resources to unending research, any form of balance has left the building weeks ago and biter attacks are a mere nuisance anyway, while clearing biter camps is just a tedious annoyance. Unending research shouldn't play any role at all in balancing combat.

Have you ever seen a graph like this?

This way you can research endlessly without fearing an unbalanced game. Among gamers this is called "cap". For example in a RPG the stats you have/loot in the early game bring lots of progress. The stronger you get, the less progress they bring. Finally and that is the idea behind the graph posted, you will never reach the cap.

So in factorio, you could say, if you are "this(x)" strong, you will never have any problems in killing those biters and of course everything is nicely balanced. Let's say "this(x)" is the cap, so you will never reach that and that resolves that you will always have more or less problems with the biters. But the more you research the closer you get to that cap "this(x)" and the less problems you will have with the biters. If, at the same time, the research gets more expensive, you will reach the point, where you make a megabase naturally just to progress faster.

Sry for the theory.

[Wildejackson]

First of all, fantastic to see that combat is being revisited. I would like to provide some ideas for potential enhancement of the combat experience.
Right now, the combat in the game boils down to two theaters of war, attack and defense. Attack and defense are effectively two entirely different parts of the game, when they should probably be much more closely related. Attack is when the player needs to expand territory and actively attacks bases. Defense is when pollution comes in contact with the biters bases, and they come and destroy the source of the pollution, which is your base.
Attack:
For players expanding territory, attacking is almost always manual, using either manually placed turrets, or handheld/vehicular weapons. I feel that the manual aspect is what makes attacking and clearing out bases a chore. When building, you manually construct parts of the base, then you use automation to build for you, allowing you to scale to the ultimate goal of building a rocket with seemingly gargantuan costs. This concept should be reflected in the combat. Automated soldiers, tanks and automated outpost construction systems should be considered. Instead of the player creeping with manual placement of turrets, allow the base to creep on its own, with programmable automatic placement of blueprints (which may be a bit complicated but will greatly benefit those who learn it, think self-replicating blueprints). Automated expansion has been thought about for quite some time and is often looked at as the "next level of automation" that would go beyond the existing mid-endgame. I mentioned automated soldiers, and there is a mod that does this simply and elegantly that has picked up resounding popularity, take a look at the Robot Army Mod for inspiration on "automated attacks." Another mod that has been doing fantastic work is earendel with his very early but very promising mod Programmable Defense. Automated planes on bombing runs, vehicular patrols, etc are some ideas, with many more great ideas in the forums and the subreddit.
Defense:
In the vanilla game, defense is simply finding the biter's attack route and shoving loads of lasers near a wall. It is rather boring and leaves very little room for strategy. I feel like the defense area of combat should be more focused on "front line defense" where there is a huge (uge!) wall with layers and layers of weapons defending against waves and waves of aliens. There should be ammo shipped out to the defenses and massive waves of enemies being crushed under the combined might of thousands of gun-toting machines. The weapons themselves could also use some sprucing up, ideas such as ARTILLERY TRAINS, long range artillery, rail-guns, poison dispensers, thumpers, electric fences, de-evolution mechanics, training biters to fight and spread for you, self-expanding walls, etc. There are so many great ideas that the community has come up with and these should certainly be looked into. On the alien's side, there could be some development of their own. Worm networks like in Starcraft II that allow aliens to reinforce each other over longer distances, more varied attack waves, more diverse enemy types and so on. There could even be more enemy structures, such as an EMP structure that damages lasers or electric devices, a spawner-type building that leaches acid and damages buildings and walls near by (think creep from SCII) and perhaps more diverse spawner types and even "queens" that build bases. Copy the Zerg from SCII in other words.
TL;DR look into adding new, more diverse military will deepen the strategy and overcome the tedium of the combat in the game. Tell me what you think.

[malecord]

If you don't, you have to put a single roboport there and a crate with supplies. And then have to check on it occasionally or massively oversupply. This is kinda unfun and really not in the spirit of the game regarding automation.

There are solutions to problems like these if you look deeper: https://wiki.factorio.com/Circuit-netwo ... fied_items

Thanks for the suggestion, but tbh that seems ridiculously overkill for a base that consists of like 40 drills, a small train station and a couple of walls with turrets. Would have to run an extra train line for it or even put some extra waggon onto the train that gets the ore. Aside from that, i would also welcome a solution which fixes the "rectangular bases" problem.

Personally I find the logistical challenge more interesting that the production one. Supply outposts is funny and find a scalable way is even more fun.

What I did in the past was having a supply train and then use an expanded system like the one above. Then all I have to do is deploy the supply station plus roboport with personal roboports and setup the supply train. The rest of the outpost, from drills to walls is automatically built with the train supplies. Same thing with dismantling. I can dedicate a wagon to dump and let robots dismantle everything and send back to main base all the stuff for reuse (not to mention the convenience of unloading artifacts, stone and wood in whatever outpost I walk by).

It worked until I researched stack inserters update. In 0.13 they broke the inserters by making them upgradeable and that broke the supply train with it. I'll try to see if I can do something better with the Logistic train mod now. That should fix everything and make it perfectly efficient by keeping trans smaller and sending out only what is needed where is needed.

[tehroach]

Unending research will have negative consequences on combat ballance
While I believe that the concept of unending research is good for resource consumption and gives the player a way to exhaust resource fields.
It will undoubtedly make the task of balancing combat for the duration of a game an impossible one.

Unending research, by definition, only affects extremely long running games.

Are long running games less important than short ones?

By the time you're able to devote resources to unending research, any form of balance has left the building weeks ago and biter attacks are a mere nuisance anyway, while clearing biter camps is just a tedious annoyance.

But should this be the case, should long running games be left only to the realm of play peaceful or just expect to be annoyed?

Unending research shouldn't play any role at all in balancing combat.

This highlights my point
How much should research effect combat and is it simply a good idea or just a bad one, to allow a mechanic that makes combat really really easy, if a player sets up one particular configuration of machinery and heavily penalizes the ones that don't.

Have you ever seen a graph like this?

Well yes
It highlights an idea of:
"if I was to cross a road and my first step was half the width of the road, then all consecutive steps were half the width of the previous step, how long would it take me to cross the road?"

This way you can research endlessly without fearing an unbalanced game. Among gamers this is called "cap". For example in a RPG the stats you have/loot in the early game bring lots of progress. The stronger you get, the less progress they bring. Finally and that is the idea behind the graph posted, you will never reach the cap.

Sorry, I am not questioning the mechanics of a region bound equation.

So in factorio, you could say, if you are "this(x)" strong, you will never have any problems in killing those biters and of course everything is nicely balanced.

But is it nicely balanced?
The point that you make "you will never have any problems in killing those biters" means that point x fall deep, in the realm of boredom as described by Jenne Nakamura and Mihaly Csikszentihalyi (hope I spelled that correct ) on the concept model of optimal experience flow.

Therefore x as you describe it, lies directly on the x-axis, meaning that x is the definition of unbalanced.

Let's say "this(x)" is the cap, so you will never reach that and that resolves that you will always have more or less problems with the biters. But the more you research the closer you get to that cap "this(x)" and the less problems you will have with the biters. If, at the same time, the research gets more expensive, you will reach the point, where you make a megabase naturally just to progress faster.

This describes a solution that is only intended for expert players, what about the "other" player situations?
Should the "megabus" be the only way to play Factorio?

Sry for the theory.

Please never apologize for stating a well-researched idea
I like it when someone challenges my ideas with demonstrable facts, rather than simple opinions.

However, the way I see it moving the resource sinkhole away from research and delegating it to modules is superior, because it will affect gameplay on a local scale rather than a global scale, meaning that the player can create regions on the map where dealing with biters is trivial, while maintaining the overall biter threat.

[Thynix]

I just got Factorio. It's fun! It's been prompting me to stay up way too late. It's late now so this is rambly but hopefully helpful feedback. I was having a blast with the campaign until hitting New Hope mission 3, so I'm glad to see y'all focusing on combat at least a little.

Attacking a base, even the first one in mission 3, is a chore. I tried driving past in the car and throwing grenades (heavily implied approach by the mission text boxes!) but they hardly did any damage, even on direct hits. The car turret was okay but getting that close ran a very high risk of getting swarmed and killed, especially if I got caught on the spawner's hitbox, which seems larger than the spawner looks. The only viable thing I found was turret creep because approaching on foot was suicide, but before looking for tips on how to approach the mission because I kept dying when attempting to Rambo it or running out of ammo, I felt like I was missing something fundamental and doing it wrong. Turret creep is tedious - select a turret, place it, select ammo, ctrl-click it into the turret, hold down the mouse button to remove the previous line of turrets, repeat. I don't find it fun. It's a chore, and the interface doesn't help. For instance, why once I select an item to place must I reselect it to put it away? Why doesn't right-click clear the selected item? I'm vaguely aware of Q doing this but I suspect it's a side effect of switching weapons and not intended functionality. Have you all considered looking at what something like The Sims 4 does in its building interface?

In the presence of any biters I can only fire on them instead of focusing the worms that are the longer-term concern; I can easily kite the biters into turrets. The primary tactical choice available to me is positioning which is impaired by the low movement speed, especially without power armor as in that mission. Others have suggested a combat mode; being able to manually aim could help but seems liable to mess with balance by allowing shots to miss. Setting attack targets could be a middle ground?

Anyway this is a hard problem but I hope you all find improvements to make, because in its current form I dread offense. Defense is okay so far, and combat in multiplayer when not hosting (I have yet to try a dedicated server) is in my experience effectively nonfunctional due to the latency and warping behavior pressing the "shoot nearby enemies" button prompts. 2D combat can be done quite well - Hotline Miami for top-down (though not swarm-based) and Risk of Rain for swarm-based (though sidescrolling) come to mind. It's just not fun here yet in my experience.

This is what mission 3 was like for me; at this point despite sinking all my copper into ammo I was running out, and still not done:

[IronCartographer]

Why doesn't right-click clear the selected item? I'm vaguely aware of Q doing this but I suspect it's a side effect of switching weapons and not intended functionality.

This bears repeating, in light of the subject coming up again:

Tab should really be the default hotkey for switching weapons.

Q is definitely intended to clear the cursor. The weapon switching is almost always an annoying unintended side-effect of trying to clear the item in hand!

Some people may even have avoided using Q for clearing their cursor as a result, judging by Let's Play videos.

[Andrzejef]

I basically select-deselect items, so I didn't even know there was a hotkey for it

[mergele]

That moment when someone makes a lamp setup that probably took more time than your entire ingame time.

[R3vo]

Why doesn't right-click clear the selected item? I'm vaguely aware of Q doing this but I suspect it's a side effect of switching weapons and not intended functionality.

This bears repeating, in light of the subject coming up again:

Tab should really be the default hotkey for switching weapons.

Q is definitely intended to clear the cursor. The weapon switching is almost always an annoying unintended side-effect of trying to clear the item in hand!

Some people may even have avoided using Q for clearing their cursor as a result, judging by Let's Play videos.

I'd love to see a radial menu opened with tab to quickly select weapons. Not sure if that's possible though.

[JCU]

My vision of a Big Enemy Revisit.

Along with weapon rebalance I think biters bases also must be changed. We needs weapons to protect from biters and we need biters as a challenging target for weapon industry. There are connected topics. But fighting against biter bases is slightly boring now as their behavior is not changing from battle to battle. Also we don’t need a lot of different weapons if enemy abilities and tactics remains the same for all battles. Also weapons with low rate of fire are nearly useless with high enemy spawn rate.

Spawn rate control

Current spawn rate is constantly high. Only "level" of biters is changed along with evolution factor.
I suggest every spawner to have limited and slowly regenerating “internal power buffer”. Each spawn cost energy, so in a start of battle player will fight against a big horde or two. Power buffer capacity and recharge rate depends on evolution factor. In other words, power cost of each spawn become "cheaper" depending on evolution. When hordes are defeated, the enemy base becomes less protected by moving beasts for some minutes. But player may be caught if enemy base regenerates power and respawn a new large horde near a player. Horde spawn may be coordinated by a group of spawners or by nearby Queen's nest.

Queen's nest

It is an enemy building with high HP level. It is not spawning mobs and have not weapons. But it can heal nearby structures, "buff" them and even create new spawners and worms. It may be more then one queen's nest per a large enemy base. Small bases (below 7..10 buildings) will have no queen's nest and use no advanced tactical logic.
Queen's nest may have different levels (level of building) depending on evolution factor and some random magic. Larger nests have more energy and HP and also a larger influence radius. If a queen's nest remains alive after player's attack, enemy base will slowny restore itself. Also a queen's nest may coordinate offensive and defensive actions of nearby buildings. Give them command to stop spawning and recharge within battle and after some time spawn mobs horde simulteneously to got player trapped.

Possible enemy base types

Base type may be bound to its queen's nest.

1. Fortress. Dozens of worms on sides with spawners near to a base center.
2. Zerg Rush. Mainly biter nests producing low-level biters in high rates. Some spitters and worms.
3. Behemoth’s nest. Producing small hordes of slow and powerful behemoths. Small in count but powerful. Player may use armor-perching rockets against them.
4. Artillery snipers. Spitters with a low fire rate and a high range.
5. Generic mix. “Tactical” spitters with high firerate and biters mix with mainly medium-level biters.

[tehroach]

"Along with weapon rebalance I think biters bases also must be changed."
"There are connected topics."
"fighting against biter bases is slightly boring now as their behavior is not changing from battle to battle."

Agree with this.
IMO the biters themselves are the key to the combat solution.

• 1. Fortress. Dozens of worms on sides with spawners near to a base center.
• 2. Zerg Rush. Mainly biter nests producing low-level biters in high rates. Some spitters and worms.
• 3. Behemoth’s nest. Producing small hordes of slow and powerful behemoths. Small in count but powerful. Player may use armor-perching rockets against them.
• 4. Artillery snipers. Spitters with a low fire rate and a high range.
• 5. Generic mix. “Tactical” spitters with high firerate and biters mix with mainly medium-level biters.

I really like the ideas of different types of biter bases, that would require the player to change up and use different types of tactics to overcome different obstacles

We need weapons to protect from biters and we need biters as a challenging target for weapon industry.

IMO not only would the need for this kind of infrastructure make the game more interesting, but it would also create a resource sinkhole itself and alleviate some of the need for endless research, just to keep the machines in your factory running.

IMO combat in Factorio can't be made interesting without an interesting enemy to fight.
"Boring opponent = boring combat"
To me, the general stereotype that the biters seem to fit into is longing for a queen type of entity.
And while I like the idea of a queens nest building,
I think that having a queen unit that could defend itself would have several advantages.

• As it would represent a "boss" fight, ie something to avoid in certain circumstances and a target of opportunity in others.
• A queen unit being mobile could add a little surprise every now and again to the unwary traveler.

But maybe both would simply be better

• While in the nest the queen would be a formidable foe; even for the most tech-savvy characters
• While outside the nest the nest itself would be very vulnerable, but the queen could help coordinate attacks and "buff" her minions

These queen coordinated attacks could be somewhat rare, but much more dangerous that the every day biter attacks.

Also how the player deals with the queen could make for interesting gameplay within itself

• The player kills the queen - The biters under her control become less coordinated and attacks are significantly weaker for a period of time until a new queen grows.
• Trap the queen - The biters under her control focus all their efforts to free her and waste no effort on other targets; allowing the player to create a focal point for dealing with the biters.

The idea of being able to trap the queen could lend itself to additional gameplay

• Just leave the queen and surround her with turrets and she will eventually die, meaning that the trap was just a temporary solution and the game eventually continues on as if the player had just killed her.
• Feed the queen - requires a resource chain to be built and maybe some research, but prolongs and allows the player to maintain control of the focal point of combat.
• Drug the queen - requires the infrastructure to feed her, plus additional resource chain, but biters under her control behave peacefully to the player.
• Frankinstein experiments with the queen - on this point I think the sky is the limit.

With a Frankinstein queen the player could setup various resource chains that would control the queens biters in various ways
ie. With Alien artifacts collected from biters from a different queen the player could coordinate biter to attack, ignore or defend against them.
ETC.

[Atomicking74]

Would it be weird if I said the best way to make combat worthwhile is to make nests much, MUCH harder to kill, but make them not respawn, while keeping the "attack waves" more or less on the current difficulty level? I think the main problem with combat is that it has become more and more trivial, with the artifacts going the way of the dodo, now the only reason to clear a nest is because it's incredibly close to some resource, or because it's on the way to/from a resource, that's it.

The combat is not skill based, and the plans to confront the enemy are dull (usually you just run circles around them or turret creep, that's the entirety of your plan). Maybe if it was possible to send small flying scouts to spy the nests, and if the nests had more of a structure to them (using something similar to your walls, for example), then there would be a need to strategize, and perhaps use different weapons (explosives might do extra damage to structures). The truth is that zerg enemies aren't fun: clearing them temporarily will always feel like a chore, and having the biters respawn so quickly makes them annoying, not challenging, and this I believe applies to the combat on the micro and macro level. Since the map is infinitely large, I feel there is no need to make it impossible to "tame" the wilderness a bit, you'll have to venture outside eventually, and meet new nests.

But then again, I don't care about the aliens in Factorio at all, they really are just a nuisance to the core gameplay. Maybe if every now and then there was a "mega-wave" and its size depended on the number of nests you killed, it would be more of a logistic challenge (get everything ready for a large scale attack). Maybe if the aliens actually were more aggressive, and destroyed anything they found (but this should only be possible if permanent nest destruction is possible), so that hunting down nests becomes more of a necessity when you start getting attacked, as they might damage electricity lines and rails.

Honestly, anything would make the combat better for me, I think.

Completely agree, the "enemy" has become more of a nussense (spelling) than a threat, and with purple science disappearing there is no need to hunt them.
I like the idea of changing the bitter bases, make them tend to huddle on resource patches or in larger groups. I would personnaly prefer a shift to more enemies rather than harder eniemes.
Id also like to see what the Artillery Train can do and hope that it has a much longer range that anything currently. Also increase the range of tank and add a wider variety of player weapons and ammo types that have strengths and weaknesses against different biters, specifically to fight the fact that player weapons are useless mid game.

[TheJogMan]

One thing that I think has been ignored about turret creeping is that it comes in multiple forms, the particular form that I use is to just stuff my inventory completely with laser turrets and then give absolutely no concern to picking them back up or preventing them from being destroyed, I find that this form of turret creeping works well with my play style and doesn't need any nerfing to be balanced as it gets to be very expensive as it is. I do understand that this is only one form of turret creeping and that most forms do need to be nerfed in some way and that there would be no practical way of detecting which method the player is using in game and also that I should probably just look into a different strategy. I just want to point out that trying to nerf one play style that is overpowered could very easily end up nerfing another play style that was already well balanced.