Rant (beware, long post)

[vonduus]

So I just lost another game. Or rather, I didn't loose it, I gave up. Not that loosing or giving up is a problem, I have lost more games than I have won. When you win, you don't learn anything much, it is from loosing that you get better, learning from your mistakes. The day I win every time the game is dead for me. But that day never comes I am afraid.

The situation was that I wanted to try out a golem. I tried one out before, but that was a disaster, before I knew what happened the golem was dead and the aip went crazy. So this time I was going to do it right.

I planned to take one or two human colonies early on to get a ressource boost, then a Zenith power plant or two, and then the golem. And that was basically what I did. I had some problems in the early game, because the ai kept building a lot of heavy turrets, and they are a real pain if all the ships you have got is mkI's and mkII's. So I complained, and Keith promised to make a cap on heavies.

After quite some time and some heavy fighting I broke through into the universe. I had found a nice human colony on a planet where there also was a broken Black Widow. I nerfed all adjacent planets before I took this planet, to make sure the ai wouldn't take the golem away from me before it was repaired, like it did last time. I had also isolated a planet with a Zenith power plant, and I bought one from the trader on my home world, which is also a planet with no hostile neighbours. So I repaired my golem, and meanwhile took out a few more planets. The aip was around 150, and after the golem was repaired around 200. But then things went bad.

The ai got pretty angry because of the golem, but I knew that. Somehow I had the upper hand, he attacked again and again with raiders and space planes, no problem. The real problem was when I had to go through a certain planet to get to an advanced factory, both factories were placed on adjacent planets in the far end of the universe, together with all the data centers. The only other way to go to the factories was litterally through an ai home world. Said planet contained some 2500 ships, mkIII and mkIV, plus fifteen heavy turrets mkII. At this time i had a fleet of 8-900 ships, mkIII's and a few mkII's, reinforced by some Core Grenade Launchers (a real nice little ship!). I made sure I had a full stock of supplies, brought up the Black Widow, and attacked.

To make a long story short: Three hours later I was in possession of this planet. During that time I had lost my entire fleet three times, except a few Dreadnoughts. I had no more supplies, so I had cracked open a Zenith Reserve, and they did the job. Unfortunately they were core ships, so my power supply went into the negative, until I had turned off half of the ships.

So why did I give up? Because during this truly epic fight, where I took out a total of perhaps 4-5000 ships and thirty or forty heavy turrets (the ai built replacements faster than I could take them out, there were fifteen heavies to begin with, and after I had taken out maybe ten I noticed that there was still twenty to go) - during this fight the ai managed to sneak into the system with the human colony and take it out. And to make matters worse, he also managed to take out the conquered Zenith Power Plant. And the plant that produced Grenadiers. This, combined with the fact that the ai don't like that you repair your golem (what I had to do repeatedly during the fight) meant that at the end of the battle the aip was on 540. And in the meantime the next planet before the factory was alerted, and had built twenty heavy turrets and some 3000 mkII and mkIV ships, plus the usual 15 heavies. I had some one hundred ships, no resources and a negative power supply. At this point I found it wise to go for an armistice. Like Ludendorf in 1918.

So what have I learned from this?

Heavy turrets in the hands of the wrong intelligence is dangerous, they make the game feel like the trenches of World War I. But this is acknowledged as a balancing issue, and will be taken care of, have I been told. Apart from this, the penalties for having a golem does not match the benefits, imo. Likewise with the human colonies. You get a decent ship/a decent ressource boost, but it is way too expensive aip-wise. To me, this whole experience boils down to the following rules:

1) Don't ever take a human colony. It is not worth it, as it is always placed very near a wormhole, and there is no way you can protect it without keeping your main fleet on top of it at all times. 

2) Don't ever take a golem, it is not worth it. It will cost you a lot of aip just to get it functioning, and it will cost you a lot of aip to keep it maintained.

3) If you really want a Zenith Power Plant, build it yourself at your safe home planet. Never take one, unless you are prepared to keep your main fleet on top of it at all times.

4) Always stick with your conventional troops, you don't need golems to loose a game, and they actually make it harder for you to win.

I like the ideas of golems and human colonies, and I would really like to play with them. So someone please tell me that I am wrong, and why.
 

[x4000]

The thing about golems is that they are super risky to use.  For players playing well below their skill level (or using cheats), who just want to have a fun romp and want to see a lot of massive destruction, golems are an obvious choice.  For serious players playing at or slightly above their skill level, golems are dangerous; enough so that a lot of the advanced players avoid them.

To be perfectly honest, this is by design: if golems were too easy to use at the higher levels of play, then the game would devolve into a race to get them, and everything else would lose value by comparison.  Look at a late game of SupCom: in my games, the experimentals were so powerful that a single flying saucer could win an otherwise unwinnable scenario -- by itself.

I am sure that some of the advanced players will shortly tell you that golems are not strategically worthwhile. This is not so (although, for a few like the Cursed Golem that might be true).  But for a black widow, I've personally used those to clean out multiple mark IV planets before it died (without any support).

There are a variety of strategies for successfully using a golem, but they are far from easy to use (again, by design).  Used incautiously, they can make a bad situation worse, or can even lead straight to death.

Here are some general tips:

1. Golems really piss off the AI, so make sure all the adjacent warp gates are killed before you start rebuilding one.

2. They will generally raise the AI Progress by around 100 just for repairing them once, and further repairs will cost another 100 AIP for every 75% damage you repair to the golem.  And if they die, there's another 100 AIP.  You must have a goal, and a strategy capable of achieving this goal, worthy of this cost before you even think of repairing one of these monstrosities.

3. Offensively: the longer you stay in one place with a golem, the worse it will do -- it causes the AI to reinforce more strongly, and in greater numbers each time it does reinforce where the golem is.  It is best to have the golem go in, do its dirty work, and then move on.  If that means leaving even a few hundred straggling AI ships loose as threat, so be it.  Make sure you have turrets and/or a large fleet waiting to mop up after the golem; if the golem tries to do its own cleanup work, it's likely to cause more of a problem by giving the AI longer to react.

4. Offensively: again, golems are all about the blitzkrieg, preferably against unsuspecting targets.  The absolute best way to use them is to strike a target that is unprepared.  Generally, if a planet has been on alert for hours and is hugely reinforced, the golem is going to take more losses than it should at that planet, simply due to the amount of firepower that is concentrated on it.  You can use a golem in this way, but the fast-strike-followed-by-other-ships-mopping-up strategy is again the way to go.  That's a lot of AIP cost for one planet, though; it's cheaper just to nuke the thing unless the planet is filled with Mark V ships (and even then, there are cheaper things to do).  Most golems are better used to cut a huge, immeasurably-damaging swath through a bunch of semi-defended AI planets.  Nothing else can sweep through so fast and so effectively (a few of the more specialized golems are even more challenging to use, though).

5. Defensively: keep your golems away from any planet with warp gates, and you'll be able to use them as a pretty incredible defensive structure for no permanent cost aside from the ~100 AIP cost of initially rebuilding them, plus whatever you spend on repairing them further.  In most games this may not be worth the cost.  But in maps where your home planet is continually at risk and otherwise pretty impossible to defend due to a plethora of inbound wormholes or whatever other factors, this can be a gamewinner.  Also: it can free up the entire rest of your fleet to concentrate itself on offensive actions which can also be of immense value.  Also: a single golem can often eat a CPA for lunch, depending on the golem, the size of the CPA, and how the CPA approaches your planets.

6. Do note that these are cheapest in terms of energy, AIP, and other factors if you use them offensively to do as much damage as you can before they die, and then let them die (or, if you are very quick about it, get them near the point of death and then withdraw them to somewhere safe behind your lines to just sit in storage).  Continuously repairing them over the course of a long game is a sure path to a loss in most cases, as it makes the AI take way too much interest in your goings on.

In general, the two suggested uses for most golems in advanced play are:
A) Offensive blitzkriegs that are over and done before the AI knows what hit them.
B) Defensive bulwark away from warp gates, guarding a key target so that the rest of your fleet isn't trapped on guard duty.

Hope that helps!

[vonduus]

Great, thanks a lot for the answer, especially the part where you state that golems are a pain by design. Being the way I am, I knew when I wrote my rant that tomorrow I will break my own rules and try one out again.

And it is a very wise design decision, too many games suffer from a super-unit that always wins the game for you - you mention the exp ufo in SC, I could mention the howitzers in Codename Panzers, or the nebelwerfers in Close Combat. These units are the main reason why I stopped playing those games. I guess something like this was what I expected from the golems, but I am glad that I was wrong.

Perhaps I am playing a bit over my true level, but this is the only way I can get better. I mean, I don't believe in 'true levels' at this stage, I remember a very short time ago when I had a hard time against two ai's level 5. Now I have a hard time against two level 7's on a grid map. But in every game I learn something new. I have been playing for three months now, and I don't even know all the units yet! The Grenade Launcher was new to me, I really like that unit.

I bet that the placement of the human colonies right next to a wormhole also is by design - a very evil design, that is!   

Blitzing seems to be the way to go. And as you say, I must know exactly what I want from my golem. This is great, I love making plans, it is half the fun. I really like your idea of using up the golem, and then run away and not repairing it any more. This will be the order of the day - tomorrow.

[RCIX]

That seems like a good cost for golems, but look at it this way: You bill them as this games Experimentals, yet they cant win an unwinnable situation by themselves without causing a ton more pain then they are worth. I honestly don't like them for that fact (i kinda feel like they were put in as a fun toy but not really useful in situations where you would think it would be sensible to apply them). If that's how you want it then that's fine, but i am just saying i know i don't like it and i bet a bunch of other people don't.

[x4000]

RCIX -- I hear where you're coming from, but there's not an alternative from where I am sitting.  The usual mode of doing things is game-breaking for serious play with most other games, and since AI War is foremost a game about serious play (given its target audience), that's not a good thing.  Golems in their current form are difficult to use but sometimes useful, which is better than the logical alternatives:

Alternative 1: Golems are very easy to use, but not really all that super of a superweapon in order to keep them in balance (many strategy games do this).

Alternative 2: Golems are very easy to use, and a superweapon in more than just name.  Thus everyone has no choice but to use them (or kill their enemy before their enemy gets a chance to get them), which shifts the focus of the game to just be a race to the superweapons (or a rush fest in general).  A ton of strategy games do this, I didn't feel the need to do the same.

Alternative 3: Not having any player superweapons at all.  There are plenty of serious strategy games that do this, too, and until the Zenith expansion AI War was one of them.  But I think they do add something positive to the game, just not what people expect on the surface.

I could go on and on and on about this, but I guess I'll leave it there.

Or, well, one last thought: despite the proliferation of nuclear weapons in the world, how many have been actually used in war?  Just two, so far as I know, and those were done as part of a blitz of sorts against an opponent that was not expecting nor prepared for them.  But other than that, many countries have these giant weapons, but none use them because the reprisals would simply be too great.  Too often games portray superweapons as something that can be used with impunity, but history and fiction has shown us time and time again that having a big weapon often makes you the first target (the post of a machinegunner in Vietnam was a horrible one to have, because you were a priority target for the enemy to take out; the empire builds a Death Star in Star Wars, and suddenly the rebels are mobilized more than they ever had been in the past; and on and on).

Anyway, I'm done now, for real.

[I-KP]

For what it’s worth I don’t use Golems either: they tend to make a good situation worse (i.e., if you’re doing that well then you don’t need them, or you have excess strategic capacity to spend on posing) and a poor situation desperate (i.e., if you’re in the sh_t then Golems are only going to attract even more heat well beyond the reach of the machine itself).  For me Golems are the equivalent of Nukes: if I find myself looking at them with covetous eyes then I know I’ve already done it wrong and sense begins to scream in my head that making the already hot kitchen radioactive isn’t going to help much in the long run.

I happen to agree that they have probably been positioned at about the right spot in the game and I wouldn’t dare tell anyone not to use them if they wished to; however, for me they will essentially remain as eccentric quirks that can be ignored quite easily, which in some ways does feel like a wee bit of a shame.  This isn’t a dig at AI War, it’s just more my style of wargame play.  In my grognarding days I was known never to be much of an exponent of the big and expensive war machines (cf. the Sturmtigers, IS3s and Pershings of WW2) and always tended to favour the bread and butter assets instead (cf. the PzIVs, T34s and Fireflys of WW2).  I rarely lost any battles so there must be something in it.  It won Russia the war (once they got organised) after all!

[Doddler]

I understand the risk reward structure, and really I like it, but the whole game revolves around managing AI progress.  Golems are non-essential and huge progress sinks.  The way I understand it, the entire game's design is supposed to provoke a 'how can I do this with as little progress as possible' train of thought, the game revolves around making sacrifices and risks in order to discourage the AI behemoth from suddenly rolling over you.  

Right now the Golem or No Golem decision is a no brainer, the game design is in place to discourage it's use.  I can't think of a situation where using a golem would warrant its cost.  You pay 100 progress for repairing it, you pay every time to repair it, you even pay 100 progress if it dies (which doesn't make much sense).  It also costs a huge amount of resources to repair, causes bigger waves and reinforcements, and costs an immense amount of energy.  Most games I play end with progress around 600, using a golem could easily increase progress by 50%.  And for what? None of the golems have a usefulness that match the reward.

The problem is that there's so much better.  The Dyson Sphere, for example, once liberated literally controls the flow of the game, and trashes dozens of AI worlds.  It doesn't cost anything once liberated, and changes the course of the game way more than a golem would.  A superfortress is probably a better defensive tool than most golems, and doesn't cost you an arm and a leg.  Taking 5-10 planets is a low risk and arguably better reward.  It's just not really a tempting thing.

[RCIX]

X, you make a very good argument, and given that line of thinking, i do still feel that a change is needed for them but not necessarily one that makes them easier to use. I'll fire up the ideas machine, and given what ideas i've cranked out in the past i'm sure i can come up with something!

[orzelek]

I would be content with Alternative 1 presented by X I think. Currently golems as they are are simply something that lies on the planet and you prefer not to touch (even not repaired one boosts waves). Using them requires first huge investment in energy and time for repairs and costs you so much progress that you could take ton of planets instead and have much better economy for example, or capture more goodies in the process etc.

[RCIX]

I wouldn't go that way. SC2 did, and now people are hammering it for that. I literally heard one player describe them "having a punch like an arthritic turtle". Just give me time

[vonduus]

I have now tried to get something out of the golems in a few other games, but unfortunately none of these games got off ground due to immense problems with something as profane as heavy turrets. I have made a complaint in the 3.088 thread, so I won't repeat this here.

I have given a lot of thought to x4k's arguments. What convinced me was the part about super-units winning the game for you: Too many games suffer from this flaw, most strategy games that I have loved and nonetheless play no more is suffering from either an unbeatable economy tactic or from a super-unit. If golems had been what I initially expected them to be, they would  probably have killed the game for me in the long run.

But perhaps the fear of creating a game-breaker has made x4k make the golems a bit "too balanced"?

Consider these two scenarios: (1) I am on my own, and want to have some fun. (2) I am with a friend, and we want to do some serious gaming (which is a lot more fun than just having 'some fun'). In the first scenario I might try out the golem, just to see what kind of thing it is. In the second scenario we will begin with analyzing the golem's ai  progress profile in detail, and as a result of this analysis we will never try out a golem, no matter the strategic/tactical circumstances. Because a pure analysis of the numbers alone will show us, that no matter how we handle our golems, we will end up with more aip than if we had just left it alone. And given that they are not super-weapons, why give ourselves this handicap?

So what has actually been introduced into this hard core strategy game, that per definition is not intended for casual players, is a casual element that no hard-core gamers will actually ever use. I for my part will definitely explore the golems by using them, but I am pretty sure, that after I have tried each of them out, I will probably never repair one again (unless one of them should happen to be a super-weapon after all).

I play every game with the intention to win. Sometimes I do 'rehearsals', but this is just basic training: I run through some series of motions, to learn some things, to get used to other things etc. My golem sessions are in this category: It is training, and I expect thereby to reach conclusions that I can apply as rules in upcoming, serious games, where there is only one criterion of success: I win. 

Therefore I think the balancing of golems have been overdone: The general analysis of numbers alone hints that golems are of no use if you want to win, and actual play with golems convince you, that your general analysis was right. Conclusion: Don't use golems.

Imo the penalties associated with golems does not match the benefits. Penalties are huge, benefits are tiny.

[Awod]

I think they're fine, in-general I don't use them but that's OK. They're situational, actually in a game I'm playing there's a black widow golem 1 hop away from my latest outpost. It's looking very fine right now since also 1 hop away is a string of IVs leading to a homeworld,(I suspect/hope)

It's looking like an even better idea since X said black widows are good.

[x4000]

There is one thing that has bugged me about golems for a while, I guess, though the recent posts are what have brought this feeling up from more than a subconscious level.  And that is this: golems cost you AI Progress when they die.  That's basically counter to the "use them then lose them" mentality of them, and basically doubles the AI Progress cost of even an initial use of them.  So that's something that I'm cutting out in the next version, which should make them all a lot more useful, without being too overpowered.  Thus the incentive is to pick them up, use them in a crushing rushing raid, and then let them die.  Or do something else that doesn't involve ongoing repairing of them, which would ratchet up your AIP costs.  Of course if you want to repair them more, you can, but the ongoing AIP cost of that makes it so that isn't an option that is exploitative.

This is a pretty major change, of course (effectively lowering AIP by 100 for all the golems except cursed, which is reduced by 10 instead).  I feel reasonably comfortable that this won't make them game-breakingly overpowered, though, and it should make them a lot more attractive for people to use more often than before.  My plan is to give this a while and see how it does, and if things still need more tweaking in a few weeks or a month based on aggregated data, then we can revisit.  These are such huge units that I'm not intending to make many massive changes to them too quickly, since that could get things way out of whack.

[Kjara]

I might suggest that if you scrap an broken golem, you incur whatever aiprogress it would have cost to repair it to full(that way if a golem doesn't fit in your strategy, you can't just take the planet and scrap it to get rid of the wave multiplier)?  Other than that, I love the change(since I was one of those who had used a golem once, and that was just for fun/ended up playing around with it, then loading back before I repaired it and finishing the game seriously).

[x4000]

Actually, even now you can scrap an incomplete golem at no cost.  Or, wait -- actually, I don't think you can scrap golems.  But at any rate, there was never an AIP cost for scrapping or losing an incomplete golem.