I think that it would be interesting to have more of a concept of investing in an unowned or reserved room. Anyone have ideas for types of structures that would make sense? In particular structures that only exist in non-claimed rooms.

• Outpost limited spawn with the ability to recycle or renew (reduce max TTL each time) or even heal inefficiently.
• Palisade Weak walls that block movement or have huge fatigue for all units and decay like ramparts.
• Sentry Range 5 tower-like structure that can attack
• Watchtower Provides vision of adjacent rooms at the cost of energy per tick.

The idea of combat engineers that create things on the battlefield or setting up skirmishing outposts now that we have power creep harassment seems interesting. With the addition of the mysterious factories, and the pull ability it would be cool to have special buildings that you can drag into place.

Also, siege specific non-creeps that you bring with you would be very interesting. Things like

• Siege Towers Lets you bypass a wall, needs to be defended.
• Battering Rams Let units not adjacent to wall use their work parts to damage wall.
• Forcefield Reduce incoming ranged damage in a small region with a constant energy upkeep.
• Undermine Able to slowly construct a tunnel through a wall, let you bypass incredibly high strength walls if you can successfully complete it.

One of the most complicated parts of maintaining sieges in real life is the supply chains. Making it so that there are ways to attack that involve something more complicated logistically than sending streams of units to slowly damage walls could be interesting.

@Fepalito Looks like you're going to get factories of some sort

Personally I liked the idea of renew creep. I worry that renew creep would lose most of its usecase (and interesting long range gameplay) if you don't have a good reason over spawning a creep.

The kind of spawning idea I would be interested in would be a certain creep type you define, eg. you have have ability to spawn large numbers of a particular creep body. Basically like an aircraft carrier that can cheaply produce fighters, rather than a mobile manufacturing plant than can create anything.

However, that just because I think the challenge of going a far distance with renewing along the way sounds interesting. No strong reason against spawning as a power. The idea of building a container to start producing creeps from just outside an enemy base is an interesting concept. Cool thought and good limitation with max energy cost.

@artch Won't it cause the attacker/defender discrepancy you stated could not be allowed?

Everything else in screeps can't be undone. Despawning is a far more punishing road imo. It's a nice thought to prevent a malicious hacker from causing account damage, but no other similar mechanism exists. Someone with access could make you buy everything at max price and sell everything at .001 credit to delete your credits. They could completely respawn you, even copy/delete all of your code. I think adding the delay as "protection" from stupidity/malicious behavior is unnecessary. I think programmers know the cost and care required for accessing as root.

If you want to keep it as a noob protection, I think there should be a param to override it just like clicking the "I'm sure I want to do this button", although I highly doubt anyone would choose to intentionally delay themselves.

@artch Ok, in the end of the day if you are set on punishing PC change in a clear permanent way it does do the trick. However the tokens are going to essentially render it non-existent unless players want to swap multiple times a week. If you really are set ensuring permanent penalty for changing is the 24 hour wait necessary? I feel like it goes against your previous comment.

No delay-based mechanics would solve this unfair feeling. We either have to allow everybody to switch instantly, or disallow everybody to switch at all (or restrict it).

I'm personally in favor of a delay since I feel it makes offense easier. However, I think that as from a game mechanics standpoint, having something based on RL time and not ticks is confusing as you cannot cleanly code for it and varies from shard to shard and day to day. I understand you did this because they are account resources, but the problem still remains that the game is experienced within game time. If you keep it, personally it seems more logical that the PC would be absorbed (die) at the beginning of the time period and at the end of it the levels are unlocked, rather than it suddenly poofing from somewhere on the map.

@artch Significant, permanent loss based on a judgement call on what game state in 24 hours will be? Definitely falls under the difficult to code region imo. If PC are intended as a separate way of playing the game from a GCL focused player, switching a major PC from eco to defense represents a huge loss of your empire for the duration, and then another level loss again when you swap back. 2 level loss and 48 hours of missing a chunk of the engine that is driving your PC based empire is a harsh tradeoff to wrap in code.

That's part of why I suggested having power creep templates and make swapping back to your old heroes be easier. Otherwise you pay a double price.

@artch Thanks for the clarification. Do you feel that this is taken care of with the current implementation of a 24 hour delay and losing a level?

This will always exist if it isn't easy to programmatically respond to threats, because people can't be oncall for screeps 24/7. Any significant barrier to readying defense that requires more player interaction (no api, huge costs, long time, etc) makes this worse, with increased frustration if a player knows they are being attacked and can't respond.

I think lack of PC API is a great example of something that would have made for a terrible user experience in a game about automation. I think having a huge cost or an out-of-game resource experience (some sort of change credit) is a large barrier to being able to code a defense. If you're looking to have people able to react defensively in a programatic way I'm concerned this isn't a good approach.

@artch I strongly disagree with

Right, and this is a big problem. In an MMO like this, the attacker is always at an advantage due to the time he spent planning the attack, and if some mechanic gives him even more time-related advantages over the defender, it's a bad mechanic.

I think defense is definitely at the advantage. Walls, ramparts, towers, closer spawning, safemode, TTL travel distance. They all are 100% in favor of the defender and are huge advantages. The time spent planning and honestly just that you can attack when players are AFK is the only thing that makes it even feasible to attack a properly reinforced room.

I don't think that purely reactive defense should be as strong as a prepared defense. Right now the only difference is a few hundred ticks of spawning to get defenders out, which is negligible in terms of wall damage.

The cost with deleting and rebuilding PC is that you are missing millions of power worth of power creep levels for 24 hours. That's a HUGE opportunity cost, especially for a fully upgraded PC. Also, if your rooms are designed to run with them in mind you're operating at a position of weakness for that duration. I'd put it akin to suddenly restricting yourself by 50 cpu for 24 hours.

I do agree with making power creeps feel special. I see how you've arrived at this, especially given your starting point and I like the idea of a power creep being a commitment. The one thing that having freepass tokens for long periods of time does is that it makes it painful to programmatically control power creep creation. Which I think might be one of your goals based on previous discussions (not even wanting to make an API for power creeps etc), but is a very strange goal for a programming game.

Obviously the punishment of losing a level is way too high, and giving out tokens is weird and unintuitive gameplay that will especially punish newer players. I'd like to propose another approach:

• When a power creep is spawned, you are locking your power levels into that creep for a long duration. After X ticks since that PC was first brought into this world, you can choose to retire it by reabsorbing it over 24 hours. At the end of 24 hours you can freely assign those levels again.
• Leveling a creep adds a fraction of the CD to the timer
• Optional: The CD scales in part with the number of power levels an account has, where having a ton of levels locks a player out of refitting individual creeps for longer.
• Optional: Recreating a power creep you had before offers a discount. (Builds identity)
• Optional: Every time you reassign power creeps, the constant that multiplies the timer goes up. This decreases with time. (Disincentives frequent swapping)

This will make power creeps a large commitment that players will have to think carefully and design around. This means that it's not free or easy to swap between options, but it is possible to do so. If you want to add shortcircuiting logic you can spend a level to immediately reassign, but I'd advise against that.

I think it's clear the tokens are a just a weird way of trying to limit player swapping, and level loss is a definite feel-bad mechanic. I think removing both and making it clear that building a PC is like building a room, a clear commitment of time and resources that you need to think clearly about.

@artch 24 isn't negligible, what it's meant for is so that if you have no offensive/defense creeps you can't immediately make them when you're attacked. It's to make offense more powerful and defense less powerful when its purely reactive.

I disagree that the goal should be to prevent players from being able to change their creeps. I think creeps will naturally gain identities because of the code it takes to run them.

If anything, it is more likely is that all creeps will follow particular patterns if you don't allow changing them once they are made. If you make a bad creep its a high cost to change it, so experimentation is not incentivized.

Why add this artificial cost? Just let it grow naturally, and later if it is an issue you can add a cost to rebuilding creeps.

Also, if respeccing is a high permanent cost, I highly doubt players will build much offense. If we're generous and assume that screeps is 90% eco 10% offense (more like 99%), then that would mean that dedicated offensive creeps would need to be scaled to be 10x as powerful as pure eco creeps to an effective use of resources.

@artch Just a comment on the

Rebuilding a room takes a lot of resources. It's not painless at all.

I think one thing that makes it ok is that no matter what, rebuilding a room is still helping increase your GCL. The biggest resource sink by at least 2 orders of magnitude is the controller, and you know that is going towards permanent progress for you. So it's kinda a pain, but definitely beneficial and it feels good because even though you lost something you're also gaining.

The problem with power is that it's just straight loss. Gone, poof, forever. Recovering it isn't increasing your GCL, no longterm benefits, it's just gone.

TBH I think its an artificial mechanic that isn't needed. People's power creeps will naturally become specialized and semi-fixed, simply because they have to write code for them and design their rooms around them. I don't think you need a punishment for it.

The only benefit I could see is preventing players from swapping to wartime heroes when they decide to attack. And the 24 hour delay + respawn time already does that to some extent.

An idea if you want to limit things: Have limited power creep templates that you spawn from. You unlock more templates by processing power, like getting more room claims. You can trade out power creeps for others whose template you have defined. This gives you hero "classes" but adds flexibility to shift around resourcing. To get more templates you'll need to continue processing power, and maybe to reset templates you could spend power (if you still want to add a sink).