Economy & Costs: Numerical Breakdown

I don't intend to do an exhaustive analysis of all the different costs and such of all the units currently in the game. Someone else might do that if they were so inclined. What I do intend to raise is the issue of the current economy's design.

I am aware that the game is currently unfinished, and obviously pretty much all the numbers will be changed and so on. That said, I am going to break down a few areas where I think the way the cost and game features are currently designed are significantly out of whack with how they should be.


The Mex

Obviously the starting point for any discussion of a TA style economy is the metal extractor. PA's metal extractor is different from TA's metal extractor in two respects. Firstly, PA's metal extractor yields a flat 6 metal per tick instead of approximately 1.5 to 2.3 metal per tick (let's just say 2 for ease of calculation). And secondly, it costs 300 metal to construct instead of 50 metal. As a result, the mex behaves in basically the same fashion as the TA mex- it is cheap to construct and easily destroyed. This is good.

I'm not entirely sure why Uber decided to go with 6 metal per tick on the mex instead of, say, 1, and divide all costs by 6, but it turns out to be basically the same (except for higher instability within a fixed amount of storage).

Obviously if you multiply all the income and expenses in any particular resource by the same quantity, the dynamics of the game remain the same. To simplify for this factor, I will discuss unit costs in terms of "mex ticks" where 1 mex tick equals one tick of metal extractor operation, rather than a metal cost which cannot necessarily be directly compared between games. In TA, a mex costs 50 metal and yields about 2 (often a bit less than 2, sometimes a bit more than 2), so a mex costs about 25 mex ticks to pay for. In PA, a mex yields 6 and costs 300, or about 50 mex ticks.

I would argue that maybe PA should reduce its mex cost by about half, possibly a bit less. However a lot has changed from TA to PA, most notably that planets are now spherical, so we will see.


The Energy Generator

The energy generator is the first area that requires immediate attention. A generator costs 450 metal to construct and yields 600 energy per tick. The issue here is that an energy generator only requires 75 mex ticks to construct

It seems to me that the current design of the game is intended to get players into the thick of blowing up units quickly. This means making energy as a limitation essentially 'go away'. From a testing standpoint this makes perfect sense. But from a gameplay perspective, your energy economy should actually be more of a resourcing issue than your metal income.

Metal requires territory control, which is expensive because of all the military hardware that entails. But actually building the mex is fast and cheap. Energy should be the opposite. You don't have to control territory to build it, but the resource and time investment required to significantly grow your energy economy should be quite substantial. This can be done with a large number of cheap, low-yield buildings (like solar collectors), or with a few very expensive high-yield structures (like fusions). However a cheap, high-yield energy generator like we have now is simply not good for gameplay. (I say high yield because a constructor consumes 1000 constantly; therefore 10 energy generators [only 750 mex ticks] equals 6 constructors on continuous construction permanently)

Again, I am unsure why Uber selected 600 instead of a smaller, more conceptually simple number; 6 generated and 10 consumed by a constructor would be identical, for example. But as long as the costs are comparably large I suppose it makes no functional difference whether energy is counted in ones or in thousands.


Constructors

Constructors are especially critical units, and in my opinion they are far too cheap at the moment. An Arm construction Kbot in TA cost 60 mex ticks to pay for. By contrast, a construction bot in PA costs only 180 metal. A paltry 30 mex ticks using PA's supermexes that yield 6.

I am open to argument on this issue, but it seems to me that mobile constructors are one of the very few types of units where gameplay benefits from a buildable object being more expensive, within reason. For practically everything else, smaller and cheaper is virtually always better. But mobile constructors make available freely assignable build power (unlike factories) and you should be paying for that privilege.

Constructors' cost-efficiency should be poor enough to encourage factories for dedicated production, but viable if you want to pay extra for freely assignable buildpower using many cons instead. Once the cost-buildpower ratio is set, then you need to determine how much cost and how much corresponding buildpower one con should have. And the cost of the constructor is determined by how early game expansion should play out- with raider units trying to pick off valuable constructors. This means they need to be expensive enough to be harassable targets (and not cheap, expendable, guaranteed-to-pay-for-themselves by sending them far afield) but also cheap enough that losing a few early on does not inevitably doom the player to defeat.

Long story short, I think PA should consider doubling, perhaps tripling the cost of constructors. Perhaps more.


Combat Units & Turrets

Obviously combat units are the area most subject to change in the immediate future, so I won't comment on them at length. But they are too expensive, especially the higher-tech units, which are vastly more expensive than their lower-tech counterparts. I would like to see higher-tech units that are of more comparable cost to lower tech units, but with much more strongly emphasized strengths and weaknesses.

Defensive turrets are very cost-effective, and that is a good thing. But the cheapest ones (especially the missile turret) are actually somewhat expensive and hard-hitting. Not every turret needs to be useful in an assault; sometimes all you want is one Defender/Pulverizer (35 mex ticks) to shoo away scouts and to keep an eye on an area. If you wanted to make a huge defensive line, you just make a lot of them, but the decision to build zero or one is much less binary due to the reduced upfront cost.


Nukes

Nukes are the area of the game that I consider to be the most egregiously silly in basic design. The numbers really do speak for themselves.

Nuke Silo: 5410
Nuke: 32,400

Antinuke Silo: 4500
Antinuke: 17,280

I don't think I need to explain that something is seriously amiss here. Why is an antinuke over half the price of a nuke? For that matter, why must the defender even pay for an antinuke at all when the antinuke silo costs nearly as much as the nuke silo does? And why on earth are nukes an eye-popping 32 thousand metal each?

Furthermore, consider that a Halley costs 55,000. The total cost of a silo and a single nuke is 37,810. Now I don't know a reasonable cost for a Halley yet. But I guarantee you that a nuke should not be in the same cost ballpark as anything that can be used as a planet killer. Firstly, nukes are not as destructive, and secondly, antinuke exists as a hard counter.

Therefore, I propose that antinukes cost about 2000 metal (maybe 1000, or even less if it works out), cost nothing to fire, and have a tremendous range increase (2x? 4x? More?). And I propose that nuke silos cost about 3000 metal, and a nuke cost about 6,000 metal. And perhaps allow a nuke silo to build, carry, and fire multiple nukes simultaneously, say hold up to four at a time. If you want more nukes, build more silos (not necessarily in the same place).

To effect this change, I would actually propose adding a new model for antinuke. Perhaps a strategic anti-air structure/unit could double as anti-nuke. Then, use the current antinuke structure as the normal nuke silo which builds and fires nukes within a certain range on the same planet. And then use the current Nuke Silo structure to construct very expensive "ICBM" style nukes that can be fired at unlimited range, or can attack other planets, but only carries a similar payload to a local nuke (which might be intercepted by anti-orbital).

The idea being that nukes would no longer be game enders in the way they were in TA; the Halley, or certain types of orbital units fill that ultra-high-cost game-ender role. Antinuke is very available. But firing variable numbers of nukes is now economically feasible, if expensive. But again, multiple antinukes is also feasible. And don't forget the most effective defense against nukes; spreading out so that there isn't 6000 metal worth of targets clustered together. Using a nuke to destroy a factory, an energy generator, and a couple turrets is a huge loss for the nuking player, and having several nukes get intercepted is worse.

Unlike super-high-cost nukes and antinuke, this arrangement creates tension for both the nuke user (even if it hits, does it pay for itself? How bold/risky should this next nuke be?) and for the defender (where would I nuke if I were him? How many nukes would I use? How many antinukes should I build or move to cover that area at the moment?) The game ender role is redundant for nukes, and available and effective antinuke creates dynamic counterplay even with "overpowered" cheap nukes.


Conclusion

I am aware that it is too early to start fine-tuning numbers. But I think it is fairly obvious that some very critical sectors of the economy and costs are not just in need of fine-tuning, but are in fact orders of magnitude massively off-base. The difference in numbers is so significant that it isn't mere 'fine-tuning' that is required, but a complete overhaul of the basic concept, as is necessary for energy generators, constructors, and nukes/antinukes.

good analysis ledarsi, i don't check the forums much anymore but your posts are always worth the time to read

Love the write up,

About nukes, the reason why it costs 32,000 metal is because metal is also the direct cost of time

When my econ is bombing in my last game, I slapped down a nuke silo, then threw 5 adv fab bots on it to assist in building a nuke..... I didn't time them, but damn it reloaded pretty quick, 1 every minute?

If you cut that cost in 1/3, I could spew out a nuke every 25seconds or so..... pretty nasty.....

I agree T1 Fabs should cost a bit more along with T2 fabbers, not sure about double, they still are key targets


and to the cost of T2 tanks vs T1, well, they are much much better, at least 3:1 if used the way they should....the extra range on T2 tanks compared to say, the Ant is more than enough to make it 10-1 to the ant depending on your targets defenses as the T2 can out range laser turets ontop of dealing more damage and much improved HP

You haven't kept up to date with the changes.

Basic/adv. turrets have the same range as a Leveler. Ants are a hell of a lot weaker now. A basic laser turret kills an Ant in one shot and shoots twice a second.

If you want to charge defenses you can only use bots or Levelers. Bots still die in one hit, but there cheaper, do more damage, and the time between entering the turrets range and the turret entering the bots range is less then an Ant.


As for fabricator cost, in TA you would use only 1 fabricator for basic stuff. In PA you want to use at least 2 or 3.

Agree on all counts except Nukes. Especially agree on Fabbers/factory cost realignment. Spamming air fabbers is fine but at the moment rather too easy. Same problem that FA had with engy drones...

Personally feel that the sheer AoE KIll-ALL of the nuke should be very expensive. Make nukes cheap and mid game armies become pointless.

That said, you have hit on a problem that PA currently has: there is no mid-late game in terms of mobile (non orbital) units. We go from tanks to nukes/orbital to planet killers. And that's a steep slope.

Now if somebody were to throw in a giant spider with a laser on its head...that'd be interesting.

*cough*

I prefer cannon hands and laser head with a missles backpack but to each his own.

Engineers effectively cost 800 mass. They're half as energy efficient as factories, and so each one needs an extra energy plant to run. They've already quadrupled the cost of constructors if you actually look at the game as a whole.

Again, you should look at the actual costs of killing a planet. Doing so requires a minimum of three Halleys, one orbital launcher, a lander, a dozen advanced construction vehicles and an advanced vehicle factory, and that's if you're doing it completely uncontested. Total mass bill: 201,410. You have to count the cost of getting to the asteroid and all the units you'll leave on the asteroid when it crashes into the planet, as you're not getting them back and it's not economical to recover them at the moment. 200k metal gets you five nuke silos with a missile each and change, which is basically enough to kill everything on a planet, but is far more flexible.

Firstly air fabbers are too weak at the moment. They use twice as much energy compared to other fabricators (200 energy for every 1 metal) and they also build at 0.6 time the speed. They also cost more metal to build and are much easier to kill.

As for land units, they only become useless when they wipe everything off the face of the planet. Building a super expensive satellite doesn't stop units from flattening your ground base.

Paying for a constructor and more energy is very different from spending more on the constructor alone. First of all, that energy can be used for other purposes, whereas resources invested in a more expensive constructor will only get you that constructor. Expensive constructors means you are actually investing in mobile buildpower, not spending a little bit on mobile BP and the rest on more energy.

However the even more important point here is that the constructor becomes much cheaper and more expendable, while its other "piece" is safely generating energy behind your defenses. The risk of sending a constructor out into the field to build mexes depends directly on the cost of the constructor- not the total cost of its effective build power. If that constructor gets killed, you only have to pay 180 metal again to get another one. To further emphasize how ridiculous this is, remember that a single mex costs 300 metal in PA.

Giant stacks of ridiculous numbers of aircons were a pathology in TA, in my opinion. PA should try and avoid having the same thing happen. I'm not actually sure that making them cost more metal to construct, and more energy per buildpower is enough to actually stop it, but it's worth a try.


And regarding nukes, it seems like nukes belong in the PA midgame, not endgame, and should be priced suitably. Nukes are vastly inferior weapons to planet killers, which are unconditional destruction, with unlimited range, that will wipe out such a massive amount of enemy infrastructure that the game will pretty much end immediately.

Nukes are limited in range, do not necessarily deal game-ending amounts of damage to an enemy spread over multiple planets, or even sufficiently spread out on a single planet. And they can be countered directly using antinuke. Therefore, they need to be a lot cheaper. This also makes nuke gameplay more dynamic and active, both with many more nukes being used by both sides, and with many more antinukes intercepting more nukes.

Large armies are not made irrelevant by the presence of nukes. Antinuke is currently a bit of an exotic item due to its insane cost, but very efficient and inexpensive antinuke would simply be mandatory for any sufficiently large base or army. And, naturally, smaller bases and forces not worth nuking remain useful in the midgame. But if you want to use a large army, just bring antinuke (static or mobile). Possibly several, so a concerted nuclear strike of a ton of missiles will fail, or at least doesn't pay for itself.

If a nuke costs 6000 metal, then an army of 60 units that cost 100 metal each should bring at least 1 antinuke, so that the enemy must fire two to three nukes in order to expect to wipe it out. But now that nuclear strike cost 12,000 or 18,000 metal, just by adding a 1,000 metal antinuke. And in practice I think discrete blobs of forces out on the map will be an early-game phenomenon that will give way to large fronts and battle lines with overlapping antinuke fields.

The real cost of a lost engineer isn't the price of rebuilding it, it's the time you lose sending out its replacement, and engineers being cheap doesn't change that. Their cost relative to a mex is more due to mexxes being abnormally expensive. A mex in SupCom paid for itself in 18 seconds, IIRC, but a PA mex needs more than three times as long to bring in its own cost in mass.

Less awful ground unit pathfinding and longer ground unit build ranges are likely to make that go away. The reason why construction aircraft were preferred was because large swarms of construction vehicles were simply too dumb to act efficiently.

Large swarms of constructors can be used to construct buildings, assist production of factories, reclaim resources, etc. A factory, on the other hand, can only build the units in its build profile. It cannot move, it cannot assist other constructions, and it cannot reclaim. It should be no surprise that players prefer large groups of cons instead of factories. And unless factories are very, very efficient compared to the flexibility of constructors, that is a rational decision to always prefer cons whenever possible.

You are correct that aircons were chosen in TA because they stacked without limit, unlike kbot and vehicle constructors. But a giant army of bot constructors is only marginally better because land cons are less universally mobile, and don't stack infinitely.

Zero-K had a good answer by placing significant emphasis on nanotowers, which are efficient, immobile build-power with a reasonably large range. They are useful for assisting factories and for nearby construction, but not for capping mexes. Factories being the most efficient (by a large margin), followed by nanotowers, and then relatively expensive and inefficient mobile cons.

The problem with nukes lies in their simple dichotomy. It's a deep fundamental issue for a game that needs to progress beyond nuke play, and doesn't have much to do with with the cost of nukes or anti nukes.

Nano towers are hardly a solution for lathing needs. ANYTHING can be made effective by giving it hueg numbers. A highly efficient tower will not change the fact that air units get to move FAST, and are capable of reaching places that ground units will never touch. What's really at stake is a symptom of ground constructors lacking any kind of unique advantage or trait, some kind of advantage that can't be synthesized through numbers alone.

I agree with you 100% about the issue being that nukes create an incredibly strong dichotomy (presence vs absence) in TA, SupCom, and currently in PA.

However you are incorrect about the costs of nuke and antinuke. The cost has everything to do with the dichotomy.

Suppose there was only one type of air unit that cost 30,000 metal to construct, and it can only be attacked by one anti-air unit that costs 20,000 to construct, and destroys it. The exact same kind of dichotomy is created because the discrete difference between zero and one of those units is so huge. Having one of the 30k planes must either win the game, or it is destroyed because the enemy has its hard counter, and you're out 30k. The discrete difference between success and failure of both is also incredibly severe.

Cheaper air units and cheaper anti-air create a spectrum of air presence and a spectrum of anti-air, which creates dynamic gameplay as air units move about in groups of variable size through varying density of anti-air.

In precisely the same fashion, cheaper nukes and cheaper antinuke will make a much more continuous spectrum, with a softer distinction between zero and one nuke, or zero and one antinuke. Instead, nuclear strikes of variable numbers of nukes have to contend with varying density of antinuke. And having nukes land or be intercepted is not necessarily immediately game-deciding. There can be back-and-forth with nukes being launched, being intercepted, and being successful, with both sides managing nuke launching and nuke defense as an additional dimension of gameplay on top of armies, aircraft, ships, economy, orbital, and so on.

The problem with less than 20 000 cost for nuke is the pace of game. Currently if you see enemy tanks/bots travelling toward your base, its possible to put all engineers to make one nuke and wipe them out. If the cost would be 6 000, we would need moving anti-nuke platform/unit to protect your armies. Otherwise nukes would be constantly used against huge armies and there is no way you could protect them.

Everything have to be synced with the speed of game, but I'd like to see a bit faster pace than currently. Tho its okay, but microing units is alot less important if everything moves very slowly. You wont make that much difference, sometimes you can but not that often.

Considering the current balance of the game, this has got to be the silliest suggestion I've read on these forums so far...

If anything, nukes might be too cheap/easy to build.

Your numbers seem way off. Why would you have nukes cost the same as a couple of levellers?

Also what's hard about countering flying engineers with fighters?

The missile defense turret is not hard hitting. It is easily destroyed by a few ground units.

Nukes are very tricky to balance. Reducing the cost of them is very dangerous. One of the best defenses against a large blob of units is a nuke. If a nuke's cost were reduced that much, then it would be impossible to attack your opponent with an army. You'd always get nuked.

I do agree that anti-nuke's cost should be reduced a little and possibly increase the range a little. However, increasing the range by 4x or more would mean nukes are now invalid in destroying a base because anti-nukes would be so incredibly powerful. So then nukes would only be valid in destroying attacking armies. So then the only way to play PA would be to turtle.

Also. It's incredibly important that anti-nuke launchers can build three missiles and nuke launchers can only build one, otherwise nukes would be too powerful.

There's an issue with making nukes cheaper. The target of a nuke doesn't have to be a well defended base, it can also be a large blob of units that is threatening your own base. Now let's say you make both anti-nukes and nukes cheaper, then it would mean that engaging the enemy with ground units becomes a very difficult thing to do.

You'd be essentially forced to go into a sort of trench war, where units can't move out without taking heavy losses if not total annihilation. Unless you somehow manage to creep the enemy with anti-nukes.

It goes without saying that there must be a mobile antinuke option available. If indeed antinuke is integrated with strategic anti-air, this becomes even less of a burden on a mobile force.

Another possible implementation would be to integrate mobile and stationary antinuke into a single asset that converts back and forth between the two. Primarily for convenience, so you don't accidentally move a strategically significant antinuke when box-selecting. But this also allows fixed antinuke to be redeployed instead of carpeting a large area with many antinukes, and allow players to move inefficiently positioned stationary antinukes. This is especially interesting if we make antinuke intercept nukes flying through its range; not just nukes aimed at a point within a fixed distance. Positioning your antinuke between the silo/submarine/launcher and its target would be an efficient way to defend against shooting at any targets behind the antinuke.

Basically, people are assuming that antinukes are stationary and claiming "oh, what about defending an army against nukes" which fails because mobile antinuke will almost certainly be available. Secondly, people are assuming nukes' value is independent of the cost of anitnuke. Cheaper antinuke directly reduces the effectiveness of nukes, and requires that their cost be reduced in order to remain equally viable.

Reducing the cost of antinuke by a factor of 10 with free interception changes everything about nukes. You can't just look at the cost of the nuke and how much stuff it *could* blow up. The probability of it working is a function of a lot of factors.

It is my position that a low cost and a low probability of success is superior to a very high cost and a high probability of success (because of the relatively high cost of antinuke). This creates more options for how to use nukes, and more counterplay against them as well.

That's currently how anti-nukes function.