TL:DR; Downgrade mechanism doesn’t work for PvP, domination mode is proposed which can be triggered by attacker after he initially takes down the room.

Introduction

Last week, Heggico, Genovese and me (TNG Alliance) attacked SteeleR (Thunderkittens) and managed to take down 9 of his RCL8 rooms. We enjoyed the PVP aspect of this game while being the underdog very much, however it showed us some imbalances (at least in our eyes) that should be discussed.

Please note that the income of Thunderkittens is about 5 times higher than ours. We’re not saying we should have been able to win a war against them or that the changes we propose would have had effect on the outcome of it. Although we’re pretty sure it would have made the battle more interesting for both parties.

Problem

Imagine your offensive scripts in combination with manual interactions (since you’re the attacking party) are stronger than the enemy defensive code and break down the rampart/walls brick by brick. You’re seeing the hits decreasing to a very low number and all of a sudden your creeps stop attacking. OMG do I have a critical error in my code right now…? Nope, it’s the safe mode that was triggered. Don’t get me wrong, I’m totally in favor of this mechanism and think it’s a great feature within the game.

Of course you’ve anticipated on this, planned your attack in advance and started attacking multiple rooms at the same time. The other rooms eventually fall to your attack code and you’ve breached the enemy defensive zone. A great euphoric feeling enters your body since the attack that you’ve planned for so long finally takes shape and you start crushing the enemy structures within the room.

At this point you’ve put a lot of effort into taking down the room. A lot of players will emphasize that attacking is easy since you’re online, while the defensive party tends to be offline during the attack. I strongly disagree with this opinion. If both the attacking and defending party would have decent code, the defender is strongly in favor because his creeps will be in ramparts. He only needs to boost his ATTACK parts while the attacker needs to boost ATTACK/DISMANTLE, TOUGH and MOVE. Furthermore the defender has the option to enter safe mode which already saves 1 room for around 20 hours.

Even if you disagree with defending while AFK being easier than attacking while being online, please proceed to read the problem statement.

The room has been cleared and the controller starts downgrading. To make sure the enemy room will downgrade with even one level you have to prevent your opponent from putting a single unit of energy in the controller for 150.000 ticks, which is around 7 days in real life with current tick speed. What are the measures we have for this? We can block the controller by performing the attackController with a creep that has claim (thus limiting the range to 10 rooms distance). This means we have to structurally use creeps to attack the controller, and as soon as the defensive party comes in with his attacking forces to upgrade room, you have 1.000 ticks to get this room back into your possession or else he can upgrade OR safemode the room. This process seems very imbalanced. The defending party has to get the room back under his control ONCE in 7 days, or prevent your creep with claim parts to get to the room ONCE in 7 days (to start safemode).

In the current state this mechanism is never going to work and therefore makes the PvP aspect so much less interesting. The only way to really capture rooms from your enemy is by braking morale. If you keep on taking down the room for long enough, or completely wipe a player and his alliance & friends in the surroundings, you might get him to quit the game or respawn somewhere else.

Please consider this from an underdog perspective. You can never capture rooms from your bigger opponent since you won’t be able to break his morale, he has much more resources and therefore will be able to continue this boring resources trade much longer.

Proposal

We think that after you’ve successfully attacked a room and taken it down, there should follow an interesting battle which is only slightly in favor of the defender. The defender already has spoiled his “big” benefit by letting the room go down in the first place. We think that the attacker should be able to trigger a “domination” mode within the room, in which both parties battle for it in a somewhat balanced way.

Let’s say the there’s an additional counter (call it “dominated”), and as soon as this counter reaches a specified threshold based on the RCL, the room gets auto unclaimed. The attacker would have to use claim parts to increase this counter, while the defender would use claim parts to lower it (can’t get lower than 0). The defender would still have the benefit of being able to construct structures, but he would not be able to use safe mode after the domination mode has started. Domination mode should end automatically after a specified time, and probably have a timeout after this period.

Final remarks

I’m very sure a lot of players will not agree with our vision, to counteract some of the arguments that we will get I’ve already listed a few:

Q: This will strongly influence new/small players
A: Nope, new and small players tend to respawn after their rooms are taken out (morale issue). It will only impact wars between bigger players/alliances.

Q: The defender should have time to adjust his code and that wouldn’t be fair with the “domination mode” since the attacker is probably online.
A: We agree! There should be a delay after which domination starts to make sure the defender is aware of the situation and is able to respond. I personally suggest a 3 day period (to survive offline weekends), and then have domination mode ongoing for 2 days.

EDIT

In stead of the "domination" mode there's also a much easier implementation: Have the attackController method increase the upgradeBlocked period and put in a "defendController" method to decrease the upgradeBlocked period. The max upgradeBlocked period should be high (around 20k ticks). This would enforce a battle between defender and attacker that is slightly in favor of the defender. This idea got quite some support within the Slack channel.

Sorry, couldn’t find a creep tomato

That attacker and defender need to rely solely on their code.

A delay would solve no online/offline problem.

Also, I am in the role of an underdog and not gotten so far yet. But can't you already cancel a safe mode by nuking the room? (From the docs it reads as a nuke cancels the safe mode?)

So, the only real way to take over a room is to destroy the player completely, which is very hard to do. Even more so if they are in an alliance. I too "lost" rooms, but in fact, I didn't lose any. I still have those rooms, they just don't have anything build in them. I can just keep dropping a little energy in them to prevent them from downgrading. Then I can just rebuild a terminal en start sending creeps to build the defences in the room and everything else. 

This is what happend with one of the rooms I destroyed from steeler. I did send a few guards in that room, but he could just send stronger ones. We didn't send a claimcreep to block the controller, since its pointless. If the defender enters the room with a large force, you have 1000 ticks to spawn a counter, walk there, destroy the enemy creeps en start attacking the controller again. It is not sustainable to keep a large enough force in the room to block a player from retaking it. 


The proposed solution of a domination means that both attacking and defending gets more gameplay than just throwing resources at each other. Lets say the threshold is 10000 or even 100000, it still takes a long time for a player to lose a room, but the defender doesn't immediately get 150000 ticks of additional time to reclaim it. The defender also has to act in order to not lose the room and not wait till there is only a few 1000 ticks remaining before sending an army and dropping 1 energy into the controller to gain another week.

In total, you have to defend for 400,5K ticks for a single creep to successfully "claim" a room from someone else. This extreme long duration, which is counter-able by a single creep, seems to be the main issue. The amount of resources an attacker has to spend to maintain control is disproportionate to the amount a defender has to spent:

• Defender: 1 creep in 400K creeps with 1 WORK and 1 energy
• Attacker: multiple creeps defending said controller from any creep reaching it

A quick short-term "fix" is preventing upgrading while an enemy non-npc creep is inside the room, but I think taking over rooms as a whole has to be reconsidered. I do not have any idea's yet on how to tackle the issue. When I get one I'll drop it here.

Attacker: multiple creeps defending said controller from any creep reaching it

Not quite true. You need only one creep to attackController once per 1000 ticks, and only if it’s killed, then you need to spawn a military squad.

Also, you can nuke to end a safe mode. After you do this, there is a 200 tick period in which the defender cannot safe mode again. You might get a demonstration of how one might use these mechanics together to their benefit. Planning for the strong likelihood of using safe mode would have meant building up some nukes and using them strategically.

My biggest complaint about safe mode is that it doesn't seem to fit well with any of the other game mechanics. It is just a deus ex machina that doesn't really add to the enjoyment of the game. It also doesn't fulfill its original purpose, which was to stop a player from getting steamrolled a short time period due to defense bugs. If the attacking player has the resources and the code to support it, they can completely negate any benefit from safe mode. 

I like the root controller proposal but I think those are separate issues. Resetting the downgrade timer is just too easy right now and that is only partially fixed by removing the RCL saving, although it would already be a step in the good direction.

"You need only one creep to attackController once per 1000 ticks, and only if it’s killed, then you need to spawn a military squad." 

That's completely true, however the the "defending" party just has to make sure that your new creep doesn't reach the controller within 1k ticks. This is insanely imbalanced as you will probably understand yourself as well.

 - Also, you can nuke to end a safe mode

A nuke, when launched, takes 50K ticks to land, a safemode lasts for 20k ticks. It's not a viable counter for safemodes. You only get 200 ticks to "attack" the controller, this is way too short.

This is insanely imbalanced as you will probably understand yourself as well.

It’s not “imbalanced”, it’s balanced in defenders favor. Taking over rooms is not supposed to be easy at all. It has a huge grief potential which can scare away some players who are not able to monitor their Screeps every day.

In the case of our war i was away for about 3 days with very limited net access - so the result - a well structured attack based on the weaknesses of my defence lead to a loss of many rooms without the option of me countering it, still - a week later i don't have time to do so.

Due to the fact that everything in this game is done at a really, really slow pace and you need to automate everything so it plays by itself there is rarely the case where u have prepared code for all possible scenarios or the time do it in a couple of days when being attacked and have to defend the rest of your rooms - my experience there shows that each next attack on you exploits something you failed to oversee, which is totally normal and the best way to find it and patch it

The proposed ideas are good, but not enough in favor of both parties, so i would suggest something in addition:
- if we're going with the domination counter - trigger an automatic safe mode for the duration of that domination to the nearest room of the defender so he can actually engage in that domination fight

But this is not the problem, the problem is more the ease they get for "resetting" the timer. I'm fine with defending a room from being taken over for 14 days. I'm not OK with 13 days of invested time being reset in 1 tick.

People who are looking for a PvP experience in which you can completely decapitate another strong opponent are probably going to be less satisfied with screeps.

Yes, Screeps is not intended to be a truly PvP-based game. It is the Arena feature that we have in mind when we think about all the PvP fun in the first place.

But this is not the problem, the problem is more the ease they get for “resetting” the timer. I’m fine with defending a room from being taken over for 14 days. I’m not OK with 13 days of invested time being reset in 1 tick.

This can be solved by removing downgrade progress savings, right?

- A nuke, when launched, takes 50K ticks to land, a safemode lasts for 20k ticks. It's not a viable counter for safemodes. You only get 200 ticks to "attack" the controller, this is way too short.

I was under the impression that the cancellation happens upon launch instead of upon landing, is it the other way around?

Problem is that you still got about ~6 days of downgrading to do on a RCL 8 room.

Obviously that can be done quicker with CLAIM creeps, but due to their expensive nature you can only create 1 for 500 ticks at a cost of 10500 energy (15 CLAIM 15 MOVE). Attacking a controller is currently a bit too expensive.

I think that another part should be able to block controller upgrading ( WORK ) and should cost energy. if you can "build up" blocked time like reserving a controller while it costs energy you have an extra "cheaper" way of blocking the resetting of the downgrade timer more efficiently.

Personally I think there should be a clear difference between blocking upgrades and downgrading controllers.

But this is not the problem, the problem is more the ease they get for "resetting" the timer. I'm fine with defending a room from being taken over for 14 days. I'm not OK with 13 days of invested time being reset in 1 tick.

I strongly agree with this, it took a lot of effort to make that progress on the downgrade, doesn't seem fair to be able to wipe it out so easily.

Problem is that you still got about ~6 days of downgrading to do on a RCL 8 room.

Why is this a problem? A lot of Screeps players only check their colonies once a week. Such a player may have spent many days to upgrade this room to RCL 8, why do you want him to lose it in a single moment?

@artem, its insanely balanced to the defenders favor, causing it to be unbalanced. I get the idea that the only goal this game has is reaching the number one spot in the ranking. There is currently no point to attack a single room of another player, unless you're prepared and capable of wiping said player. Even if you manage to completely wipe a room and destroy all of the other players structures, it doesn't downgrade to lvl 7 for a week. (well, ok, less if you use attackcontroller creeps). If the defender has a single hour, he can reclaim it easily, causing possibly 7 days of progress to be instantly removed. I don't mind it taking 2 weeks or so for the "domination" to unclaim/transfer ownership of the room, as long as that progress isn't instantly stopped. If I want to prevent myself losing a room thats been wiped I can just send a massive army to keep the room clear for just 1000 ticks. There is no way the attacker can spawn enough creeps in time to defend against that, thus the attacker will lose all progress he made on downgrading the controller within a single creep's lifetime.

Again, we don't mind it takes a long time to get a room. Rooms shouldn't be easy to lose, even if you're on a holiday or something. What we do mind is the enormous amount of energy and boosts you need to commit to get a room down to 0, while the defender can undo all of that within a single creep's lifetime. Taking over a single room from a player is impossible, you have to wipe the entire player to get it, since he can then no longer rebuild that room (or he respawns, unclaiming said room instantly). The attacker has a constant and pretty big drain in his energy to block a room, whilst the defender only has small short burst of energy needed to keep the room. No way the attacker can prevent a room for resetting if you send 8 fully boosted creeps. The attacker probably can't even spawn a counter fast enough in those 1000 ticks, let alone spawn, move, attack and reblock the controller.

The changes propose on this thread sound pretty good. Building up a "block" counter like a reserver would give the attacker time to attack the room again, before the defender can reset the entire progress. Downgrade is good for inactive players, not for attacking. With the blocking you both have to work to either reclaim the room or get it down.

It isn't I encourage it. The downgrade process being reset in 1 tick is the thing that seems off. Here is a small example of why I think it should be revisited:

Player D attacks player X
Player D is online 24/7 because he has no life (me)
Player X is online 2/1 because he has a life (the guy I attack)

Day 1: Player D starts an attack on Player X

•  Player X doesn't have time and thinks "I'll check this weekend"
•  Player D is trying to get through the walls but they're quite high level walls ( 200 million )

Day 2: After 24 hours of sustained attack Player D breaches the wall and demolishes all buildings. The timer starts to downgrade.

• Player X still knows something is up but has no time
• Player D start by placing a defensive squad around the controller to prevent reupgrade

Day 3: 24 hours of downgrading have been ongoing, only 100k ticks left!

• Player X still not online
• Player D is happy, no attempt has been made yet to claim back controller

Day 4: Same as the day before, 50K ticks downgrading left

Day 5: Player D has "work" day today and can't be online / fix bugs / alter code

• Player X has time, sees attack squad, performs a single attack and does 1 tick of upgradeController after keeping the room for 1k ticks.

In this scenario in the end player D spent about 4 full days of attacking, which was negated in maybe 1 hour by Player X.

This puts a HEFTY burden on the attacking player. A way to to downgrade controllers is not enough. If an atatcker could attack a controller for 4 days, it should take the defender at least 2~4 days to get back to the old levels using the same effort. The immidiate reset to 150k cooldown for RCL8's seems off.

Possible solutions:

• attackController mechanism which allows for "damage" to controllers. This can be repaird by doing upgradeController (hits / hitsMax like walls, but for ControlPoints)
• allow buildup of blockedController time to at least 50% of controller tick downgrade levels (150K ticks can be blocked to 75K ticks). The defender should also perform attackController to remove said blocked property, maybe even twice as fast. 1 tick of attack on defender = 2 ticks off of the blockedController property.
• allow the attacker to build up a structure over the existing controller which locks down a controller, and has to be attacked by the defending player, can be build up like walls

There might be more solutions to this problem, but time-wise compensation seems more elegant.

To look at it another way: let's say you have a GCL 15 player you want to attack with no allies to help them. All of their rooms are pretty close together, reachable by you, and have comparable defenses. Will it cost you more resource / time / effort to attack one of their rooms and push it down to RCL 0, or to attack all 15 of his rooms and remove his creeps / spawns so he can't rebuild? Assuming the player decides they don't want to lose that one room, this is not an simple question to answer, and that seems really silly, that it might be less expensive to wipe a big player than to take one room from them.

The defender/attacker balance could be fixed by signficantly increasing the upgradeBlocked period from 1K ticks to X ticks, what do you think is the appropriate value?

allow the attacker to build up a structure over the existing controller which locks down a controller, and has to be attacked by the defending player, can be build up like walls

Looks like a nice skill for power creeps.