Idea: Factories Can Repair/Stage Units

Units ordered to do so would automatically stage at factories, repairing and/or rearming. This would apply to all factories, including land and sea, or just air.

Dedicated staging facilities can still exist, but factories would have a dual purpose analogous to engineer repair. Forward factories can repair units near the front. During wreck reclaim the spike in mass can be used to produce units from these otherwise idle factories.

What do you think?

I think that unless factories are an order of magnitude more efficient at repair power than engineers, nobody will use this option. Repair was of limited utility as it was (in SC if a unit was valuable enough for you to consider repairing it, you'd often be better off self-destructing it and building a new one using the reclaim, and in TA few units were big and expensive enough for you to actually bother with the micro of manually repairing them), and when it was used it was from an engineering station or a construction plane on a patrol route because that was micro-free.

Air repair pads were sort-of viable because planes didn't need to pathfind to them, so they could land on the top, get some reps and then take off. Actually having to have units path to a factory and then walk onto it and then walk off once they're done would take so much time that the factory would almost certainly be put to use better simply making another unit.

The efficacy of repair and staging is a matter of balance, using the SupCom example air staging was entirely free - however the short life of combat units rendered staging moot during actual battles.

I agree with the bottleneck concern around single factories and it's one a share with any possible land or naval staging suggestion, luckily from what I understand flow fields will render this far less of an issue. Units in line at a staging facility can simply wait in order for their turn.

Factories should be the cheapest unit build power possible or you will end up with 100-engineer(station)-factory set ups. This was a big flaw in Supreme Commander and one which I hope will be addressed in PA. This is tangential but engineers and engineering stations already have many unique uses besides providing additional unit build power.

Engineers being mobile are the best for cost base build power. Engineering towers being immobile are the best for cost versatile build power. By the same token factories should be the best for cost unit build power. Extending that further, missile silos should be the best for cost missile builders (they are in Supreme Commander). The value of assisting is that it allows you to do everything albeit worse than specialized infrastructure. Zerg in StarCraft has this advantage and it is the reason Zerg is considered the most powerful in long games. It can switch tech faster than any other race.

They already exist. They're called factories.

Hold on. Supcom had a very bogus system:

Code:

Repair: 100% cost for 100% health
Rebuild: 50% cost and starts half way
Reclaim: 80% material returned

When the numbers are put side by side, it doesn't make any sense. It is cheaper to rebuild a damaged unit than it is to repair it, and you get more resources from cleaning up wreckage than you get in rebuilding the exact same thing.

I would certainly like repair to be a viable, used tactic. In SupCom it was so rarely utilised.


I don't really understand bobucles...

It is exactly the same cost to repair as it is to build units!
If you are talking about the ability to rebuild buildings on their wreckage, then it isn't actually cheaper since you can reclaim 80% and then end up only paying 20% of the cost to rebuild it. Alternatively you can pay 50% of the cost but it is built twice as fast (starts at 50%).

That's the trade-off for being built twice as quick. I'm guessing you must be concerned with the games scientific explanation of why rebuilding isn't as efficient as reclaiming? Personally I am not so concerned with this if it gives more strategic choices to improve gameplay.

Even if it's free in terms of resources, no balancing can make it free in terms of factory time, micro and unit time. Any time you're using this option you've got a factory sitting idle, which means wasted mass, and units being pulled off the front, which means wasted mass. Why is anyone going to use a micro-heavy option that only provides extremely minor benefits in situations where you've already messed up your eco somehow?

It actually makes perfect sense. In real life it's often cheaper to completely replace damaged items than to repair them simply because building new things is done on production lines where they can benefit from economies of scale, but repairing broken things has to be done on a per-object basis since things don't all break in the same way every time. Especially with expendable things which aren't even designed to be repaired. A single penetrating hit on an armoured vehicle, even if it doesn't destroy the vehicle, compromises the entire armour plate that was penetrated (which could be a piece about a fifth of the size of the tank), you can't fill in the hole and expect it to be as good as new, because microfractures will have weakened the structure of the rest of the piece and the join where you've patched it up will never be as strong as the original which was probably cast as a single unit with complex layering of different materials which can't just be reapplied.

Are you suggesting that such a system couldn't be automated like it has in other games or other systems within Supreme Commander itself, e.g. air staging? I didn't want to delve into particulars because as it would obfuscate the OP but the existing aircraft return-to-base command could easily be applied here. When ordered to do so units would then return to the nearest factory/staging facility just as aircraft. Zero-K has a much more developed retreat AI that would work as well. In Zero-K you can set a retreat threshold for units thus repairing nearly dead units and saving mass and energy. I was expecting to hear why this was overpowered not why it was underpowered.

As you pointed out factories have a cost, just like existing staging facilities. They would pay for this cost many times over by repairing heavily damaged units, just like existing air staging facilities. The dual utility of factory staging would be spending resources when you reclaim wrecks after a battle.

This is more a result of capitalism than realities of economy. Cars are a great example, repairing a Ferrari is extremely expensive because the company makes more money by selling more cars and then overcharging on repair parts.

The existing aircraft RTB command is poorly optimised because it doesn't need to be efficient. If you have a group of units all return to the nearest factory when ordered, they'll all queue around the nearest and ignore any others if they're even slightly further away, you actually need to evaluate the cost of the current queue length and the pathing cost of reaching each factory because land units are slower and can't always take a straight line so inefficiencies are more problematic. Then you have to deal with designating factories as repair stations since you don't want units just inserting themselves in the middle of your order queues of your main factories. And then you'll need some way of determining a holding area for units to wait (they'll need to form a queue and where they need to queue depends on what's obstructing this particular factory), and an algorithmic way of making units not retreat through known enemy-held areas to stop them from merrily marching to their deaths if there's a shortcut through an enemy base (a lot more of a problem on a sphere, due to the geometry). Getting things to queue by themselves is a remarkably difficult problem. And all of this work is for what? So you don't have to have an engineer on patrol?

How is the factory going to pay for itself when it can't even pay for the opportunity cost of pulling units off the front instead of having them fight to the death? If you're pulling units back you're not pushing, which means you don't have momentum. By pulling half of your units (on average) off the front you create a weak spot which will probably be exploited by your opponent's army. If you can afford to pull all those units off and still defend, you probably had enough units that the correct thing to do was to push against the enemy while you still had numerical superiority. If repair is to be viable at all, the way to do it is to repair units in place with engineers, not mess around fixing things one at a time in factories which have better things to do.

So, how is capitalism to blame for this: http://static.ddmcdn.com/gif/pg-bt-dest ... 22x419.jpg ? That's a T-55, a Soviet tank design. Pretty not capitalist. Turret still cast as a single piece. You're not fixing that hole in a satisfactory way for less than the cost of a whole new turret (assuming the hit didn't write off the entire vehicle).

So make it efficient? These issues have been tackled before, notably by Zero-K.
http://zero-k.info/Wiki/Retreat

It seems like you and I share a similar concern. You should not be spending a great deal of time microing individual units - or even using the RTB command, your focus should be on managing battles, production and strategy. Much like Zero-Ks retreat command this could be automated such that it is a 'set and forget'.

What factories get used for staging? Likely only idle factories without other orders. I agree, units staging and interrupting your queues would be silly. I think a better way to think about this would be you have 2 structures that have the same cost - 1 is a factory and 1 is a staging facility, they just happen to have the same cost.

I assume that's less of an issue when you can manipulate mass and energy

Programming is not free, citizen. You can't just go "okay, optimise this difficult queueing problem" and magically have working code. You're requesting a lot of unit AI for a feature that's of very limited utility.

That wiki page doesn't suggest any of those problems have been solved. Units retreat to the nearest retreat zone, which doesn't work when you replace the words "retreat zone" with "factory" because it completely ignores the queuing aspect of it. Units don't automatically return to the battle (or even remember where 'the battle' was) once they're repaired, they don't determine which retreat zone is best based on pathing cost and current number of units in each queue, they don't form up in lines so they can go through the tiny chokepoint that is the factory without playing bumper-cars. And retreat zones still need to be manually defined and units need to be manually told that they can use them. Trying to use that will just have all of your damaged units milling around one factory, blocking the exit and not getting repaired. And you still haven't addressed the fact that whenever you have units that need repairing, having them leave the battlefield is the last thing you want them to do. The only situation when you actually care about not having damaged units around is when your opponent, either through micro or targeting priorities, is targeting your most damaged units to bring your firepower down as fast as possible. And guess what? If that's happening, you won't have any damaged units to send back for repair because they're being specifically targeted. It's a self-obsoleting feature.

And this is exactly the problem. Any staging facility that costs as much as a factory is massively overpriced.

What does that even mean? I can manipulate mass and energy right now. It's how everything interacts with the world. Just because they have futuretech doesn't mean they magically gain the ability to disassemble armour which is specifically designed to resist weapons made of the same futuretech. Because if they did, said futuretech armour would be worthless. Armour is fundamentally a thing designed not to be reconfigured by external forces. Repairing something involves reconfiguring it with external forces. Surely you can see why these two things lead to the nontriviality of repairing armoured vehicles?

What do you mean, retreat to the nearest non-loop production factory? Units waiting to stage park in formation near the facility - when it's their turn all the units behind them move forward in formation and queue, just like the existing air staging in Supreme Commander. Perhaps you should take another look at Supreme Commander staging. I should note land/sea staging isn't a new concept - it exists in a couple Spring mods.

why not, this is done in many existing Spring games and flow fields will make pathing a non-issue. even engineers in supreme commander could deviate from their patrol or alt-walk to reclaim/repair/assist and then remember their original command.

simple as setting a toggle and %hp threshold, say units at 10% or less HP retreat to the nearest established zone or staging facility. and of course if you don't want units to retreat you can toggle it like RTB.

If that's the most efficient way for your opponent to spend their attention micromanagement might already be over emphasized.

I'm not sure how you can say that without knowing what staging facilities, factories and units will cost.

Are you even reading my posts? The way you're saying this would be implemented would have vast numbers of tanks showing up at one factory at the same time (because battle casualties are temporally clustered) overwhelming it and you couldn't rectify this by building more factories because all the units that get damaged in a single battle will all choose the same factory to go to, even if there are a hundred other factories sitting idle. Or, in the other extreme, you have two idle factories, the first unit that gets to the closest one marks it as 'occupied' and all the other units traipse off to the second idle factory on the other side of the map, even though it would have been faster to just wait at this one. Or in a third case, the nearest unoccupied factory, due to spherical geometry, is actually through the enemy base and all your damaged units suicide. This is not how air staging works in SC because air units don't have to care about the physical location of the staging facility they go to and the case of the nearest facility being on the other side of an enemy base is much rarer because maps are square, not spherical. And don't say "you can toggle it" as if it's a solution to everything. Unit AI should predictably not do insane things, and your suggestion requires special-casing a whole bunch of difficult-to-define situations to prevent player rage from things like this.

Okay, give me a link to a Spring mod which algorithmically determines the location of a battle and has units which are repaired return to that battle (not their previous location, since battles can move and it's as much micro to order a unit to return to a battle from 100m away as from 1km away, even assuming the unit doesn't blindly wander into enemy fire because the enemy front has advanced since it was last there) regardless of where it now is.

I repeat: And you still haven't addressed the fact that whenever you have units that need repairing, having them leave the battlefield is the last thing you want them to do. If you cannot demonstrate that repairing units is a tactically useful thing whether or not it can be implemented is irrelevant.

That is not a counterpoint, and if you'd actually read the entire passage you were quoting, you'd know why what you just said was irrelevant. Read what I wrote again.

A staging facility is like a factory but worse, because it cannot build new units. Thus, if they cost the same, either the staging facility is overpriced or the factory is underpriced, because one is strictly superior to the other. What you just said was equivalent to responding to "I will sell you one widget for ten zorkmids, or one widget and one gadget for ten zorkmids, which is the better deal?" with "Without knowing what widgets, gadgets or zorkmids are worth, I can't tell you the answer.".*

*Warning to pedants: Yes, technically you have to know that they all have positive value, but I doubt Uber are going to give any units negative build costs.

This is going nowhere :roll:

You're making some sweeping generalizations about how things work, for example staging is not a worse factory - staging repairs units free of cost (beyond the initial investment). Moreover it seems like you're more committed to defending yourself than actually discussing the topic.

I was more looking for reasons why this may or may not be a superfluous addition, for example asserting that unit repair does not add any depth of strategy and why, or that the additional functionality of factories would ease production spikes and why.

You're absolutely entitled to your opinion but wild conjecture doesn't support it.

The important thing to note, is that repairing slows down the pace of the game, we don't want to fight a battle, and pull back to the nearest staging area or Engineer swarm, wait to repair, and then go back out. Laaaaaaaaaaaaaame.

Put more work on Reclaim instead! You know that battle you just fought? Well no the battlefield is a key area because of all the shiny wrecks you should be hoovering up! Take all those shiny new units your factories are continuously spewing and get them into the fight! You need that Metal to make more units for bigger fights! Chop Chop!

Mike

mhmm i initially had aircraft logistics in mind (fuel) but it occurred to me that factory staging would logically translate to land and sea and so I revised the OP to consider that. Staging for units without logistic limitations isn't necessarily desirable though, especially if it slows pace.

Oh look, it's that underhanded debating technique. If you say "your argument is full of sweeping generalisations such as" what you want to imply is "your argument has many flaws of which one is" but what you really mean is "I've only got this one point to make but I'll try and make it look like I've got several". Of course, by "free" you actually mean "20% cheaper than ctrl-k-ing and reclaiming and considerably less flexible". And if you want a game where you have to program complex unit AI behaviors for a 20% cost reduction, you can have that when you're the one who has to program everything. I hear Spring is open source, so nothing's stopping you.

You know what already does this, is already in the game, and doesn't require advanced unit AI and several UI options to function? Metal storage.

Okay, let's do some maths since you're calling my position "wild conjecture". You have 10 tanks, I have 10 tanks. They each have 10 dps and 100 hp. You, being in love with repair, have your units set to retreat at 10%. Now, that means they're retreating one second from death, so in practice most will not be able to retreat at all, but let's ignore that. Let's be generous and say it's a 30 second trip to the nearest factory, and it takes 10 seconds to repair the unit to full, and another 30 second trip back, which is an unrealistically favourable situation. Now, neither of us is microing and targeting priorities ignore unit hp (I see you never did figure that out) but facing and distance means that usually damage will be unevenly distributed onto one unit, so let's say 15 dps gets put on five random units and the remaining 5 all take 5 dps on average. This is also a favourable assumption for you, since even distribution of damage means all your units retreat at the same time and get chased down and killed, and loading all 100dps onto one unit means your unit only has a tenth of a second in which to escape, which means it almost certainly won't.

So, after 6 seconds, five of your tanks go below 10% and fall back. Five of my tanks are also below 10%, and all other units have taken 30 damage. But now I have 100 dps and you only have 50, and that 75 dps ends up being focussed on some other units, so you have five tanks that are taking 20 dps while I have ten tanks which are taking about 5 each. Two seconds later, five of my tanks die, but not before your remaining five have taken 40 damage, for a total of 70. My five remaining tanks have taken 40 damage in total, and are now taking 10 dps each, the same as yours. Two seconds after that, your five remaining tanks try to disengage, but fail, because they're the only ones left and are pursued and are killed one second after that. I have five tanks on 30 hp, you have five tanks on 10 hp. Meanwhile, the five tanks that disengaged have made it five seconds along their 30 second trip. Twenty-five seconds later the first one arrives at the factory and starts being repaired. Five seconds after that, you have one tank in the factory and four on 10 hp when my tanks arrive. Your tanks in the queue all die in under a second while four of mine take a mere 10 damage. I have four tanks on 20 hp and one on 30, you have one tank in the factory. I get three seconds worth of damage on your factory before your unit is done. It comes out, and proceeds to get trashed, as it takes a eleven seconds to dispatch my tanks, while my tanks destroy your last one in two, possibly without even taking any more casualties.

End result: You: Lose all your tanks. Me: Lose 5 or 6 tanks which are all heavily damaged. Maybe I'll repair them with an engineer. Probably not because it's not worth the effort.

So, by retreating all you accomplish is turning an even fight into a terrible rout. Sure, you can claim that my numbers are made-up, but really, that's about as favourable a situation you're going to see. Ten seconds for two evenly-matched tanks to kill each other is pretty long, and that supply route is unrealistically short. If you want to attempt a more rigorous proof yourself, go ahead, but at the moment the only person who's actually tried to logically analyse the problem here is me.

I don't understand the question and I won't respond to it

To your point about storage, it was something I was explicitly hinting at with my language so it's no achievement you brought it up

My concern was that it could perhaps make storage redundant, then again storage is about absorbing income spikes, not production spikes. Engineers are better suited to deal with spending spikes by improving eco, thus turning a temporary spike into a long term boon, or improving position with static defense. At least that's how it worked in FA. However once your eco was fully saturated and your position secure you generally abused the variable mass drain between units by building experimentals instead of spiking unit production. It's interesting to contemplate and exemplifies the depth of this type of economy.

And why they are short-lived is discussed here http://forums.uberent.com/forums/viewtopic.php?f=61&t=46037&start=0, because of no unit ai script and overwhelmed players. Blackops got a t1 air stage and it is cool.
In real life repairs are vital part of mech warfare, because anyone trying to go all-out suicidal Supcom-like would quickly lose to attrition. Supcom is such a complex game that no man plays it properly, because of lack of Psylon brain and mind-machine interface to actually manage all goings proper way. Thats why Starcraft is so popular: if you accelerate yourself to be Korean, you can actually have full control of this simple game.

@jurgenvonjurgensen: wow thats some text wall

Have a go at my logic: we both have 10 tanks, i set mine to retreat at 50% because thats the standard (play Distant Worlds for practical examples) and focus fire yours. If you try to focus fire mine, you cause them to retreat, thus making you unable to focus fire all your tanks at one because of stretch. So you have to either dont pursue them or take heavy pounding by driving through my healthy tanks. Naturally as you do not retreat, your tanks are focused and die. So simultaneously with my 2 tanks being routed you lose at least one of yours, resulting in me having almost all my tanks being repaired (i may lose some if you pursue) and you losing half of yours. If you try to push, you will run either into my staging base or reinforcements. GG. Oh, and i can use transports and engies to speed up repairs. Or even have something like Mantices repairing each other after routing short distance.

Even more simple example: lets say we have identical tanks massed into identical waves, with identical trickle of reinforcements. I retreat, you push. What happens? I get more reinforcements than you, because your reinforcements have to travel increasingly longer distance, and mine are increasingly shorter. Inevitably your wave gets exhausted first.

It is a basic rule of war that to attack you should have superior forces. 'Significantly' superior. Take the Napoleonic war in Russia: Kutuzov simply routed all the way to Moskva while waging guerilla war of attrition. Who got owned?

Retreat and repair is an op strat, and only feeding loser n00bs dont use it. Period.

*edited by Uber*

About the reclaim, good point. Reclaim and repair would own only reclaim though.

Individual unit micro in PA? Looks like you're playing the wrong game. Which from your subsequent post is pretty obvious since you're trying to apply DotA analogies to an RTS game.

Irrelevant to my point. I was discussing factories being used as staging facilities. Those strategies do not require pulling units off the front.

No, here's what happens. You retreat. I push slightly. I get the wrecks from the battle and 'win' the engagement and stop pushing while my engineers give me better eco than you. Then it doesn't matter that you're getting reinforcements faster because I can afford to run more factories than you and simply have more reinforcements total. As long as I keep winning I keep getting the wrecks and I can keep up a higher rate of production. And the moment I stop winning, I stop pushing and still have a higher rate of production since your retreat gave me some mass points (unless the point you were defending was worthless, in which case it would have been a strategic mistake for me to attack there in the first place. It's to my advantage to allow players to waste mass defending things I have no intention of attacking).

You'll note that Napoleon didn't have the ability to order his soldiers to eat dead soldiers in order to produce more soldiers.