Review-in-Progress (Open Thread)

[tetley]

Bartering for military techs and rush-building a military is actually kinda fun. Just like the tactic to build a forward military starbase and fight your war from there. Very satisfying ways to beat a high-lvl AI w/ bonuses.

[Solver]

I always built a forward starbase before picking a fight, and of course upgrade it with weapon assists. Which is yet another thing that the AI will never do.

And Sirian, stop taking any possible metaphor, I also want to have a couple of my own invention someday .

[DaviddesJ]

Originally posted by Sirian

Yes, but isn't that kind of silly?

Wait, it gets worse. If you ARE the military leader, then you can go conquesting with your forces and simply STEAL techs during invasions. Thus no real need in that situation to trade for them anyway!

I don't think either of these are "silly". It seems perfectly reasonable to me that a civilization could emphasize diplomatic relations and trade with other civilizations, and use that to acquire technologies that it needs. Or that a civilization could emphasize military power and use that to acquire other technologies that it needs. One can see many examples of both of these strategies being used effectively in human history.

It's only a problem if it's unbalanced (i.e., the costs are too low and the benefits are too high). That seems to be the case here, but my conclusion is precisely the opposite of yours: the basic concept is fine, but there need to be some (very substantial) adjustments to the underlying calculations and parameters. I don't have confidence that Brad will do that, but it's entirely possible.

[DaviddesJ]

Originally posted by Solver
I always built a forward starbase before picking a fight, and of course upgrade it with weapon assists. Which is yet another thing that the AI will never do.

Why not? This seems like it's among the easier things to implement. Will you eat the box if they do?

[Solver]

No, there's a bigger problem. As Sirian says, you can't achieve balance by plugging holes. There are concepts that are inherently unbalanced. The situation is that you can ignore a branch of the tech tree (say, military techs) and then get them all in a heartbeat through trade.

Such a diplomacy system where the AI is very willing to trade techs, a system open to tech-whoring, can't be properly balanced. You can make the prices higher, but you'll be stuck with a broken system either way. You'll either make it impossible to trade or still leave massive tech trades open.

Remember, all the AI can do is work according to its formulas, it can't understand the situation. Well, you can let the AI put a very high price on weapon techs in any case. Then those trades won't work properly and end up being broken. The way it's now, it's a good idea, having the AI look at your military strength before selling you techs. But thus, if you're weak, you can easily persuade them to hand the techs over. And then use the sliders to quickly get an army.

OK, admittedly, that wouldn't be a problem even if the trade system stayed unchanged but the AI had more brains when declaring war. If the AI actually had a war prep stage, and positioned its fleet appropriately before declaring, then you would be unable to get away with pushing sliders to military and trying to get an army. The AI would then begin hitting you the turn it declares, which currently never happens, if only maybe by coincidence.

[Solver]

Originally posted by DaviddesJ


Why not? This seems like it's among the easier things to implement. Will you eat the box if they do?

I meant will never do in the current implementation - I concur it's an easy tactic to implement in a patch.

[Shadowlord]

This might work for a fix: Instead of tech trading giving techs outright, it could instead reduce the amount of research required on that tech by, say, 90%. You'd still have to do a little research to make the tech available to you, and if you bought a bunch of war techs at once you'd still have to spend a few turns to have your scientists figure them out.

Alternately, what if it took your scientists X turns after trading for a tech before they understood it enough for you to use it? X could be proportional to the distance your closest prerequisite of that tech is from it. (e.g. if you trade for a tech that you could already research, it'll take 3 turns for your scientists to make sense of it. If instead you trade for something that has 2 prereqs that you don't have, it'll take 9 turns instead. (If your closest tech to it is one your scientists are trying to figure out, add the time left on that one))

Either way, you (and the AI) wouldn't be able to trade for all the war techs and immediately start pumping out a navy.

(I like the first one better, myself)

[yin26]

Originally posted by Sirian Yes, but isn't that kind of silly?

My hope, of course, is that issues like this --now that they've met the light of an unforgiving public-- will be adjusted properly. I know you don't have much faith now that the game is released, but I imagine Brad must be thrilled with phenomenal sales and seize the moment. However, it can go the other way, too. Sales were great, criticism generally limited to some hardcore gamers, and I doubt he'd want to mess around with the core too much.

That said, an AI that stages attacks before declaring war would go a long way. Also, diplo bonuses can be toned done every bit as much as you can hide horses and have them appear later with the discovery of contact lenses.

[DaviddesJ]

Originally posted by Solver

No, there's a bigger problem. As Sirian says, you can't achieve balance by plugging holes. There are concepts that are inherently unbalanced. The situation is that you can ignore a branch of the tech tree (say, military techs) and then get them all in a heartbeat through trade.

Such a diplomacy system where the AI is very willing to trade techs, a system open to tech-whoring, can't be properly balanced. You can make the prices higher, but you'll be stuck with a broken system either way. You'll either make it impossible to trade or still leave massive tech trades open.

If it's too easy to take advantage of the AI in tech trading, then that is the problem, not the fact that one can get military technology in other ways than researching it oneself. You (and Sirian) portray this as a fundamental weakness of GC2, but I think it's only a superficial weakness, which is fairly easy to solve by tweaking the system. (I don't think Brad is likely to want to make the necessary tweaks, but that's a different issue, too.)

Civ4 illustrates some relatively straightforward approaches to this problem. There are others, too. There can be tweaks to the underlying game system, such as charging empires a cost, even when they trade for a technology, to integrate that new technology into their economy. (And this cost should include the integration cost for predecessor technologies as well, to avoid the loophole of trading for one advanced tech when you don't have its predecessors.) And there can be changes to the AI attitude toward trading. (E.g., AIs should realize that the "fair market value" of a non-monopoly tech is much, much less than a monopoly tech: if two different races can both sell me a particular technology, then, all else being equal, one of them should always be willing to undercut the other's offer, because at least they are getting something instead of nothing.) And there can be changes to the AI attitude toward research. (Sirian has correctly emphasized that it's moronic for the AIs all to emphasize research on the same techs, and ignore others---but still be willing to pay well for the latter!)

Is this stuff hard to do? It's not trivial, but it's much less hard than building the game in the first place. The issue is more one of will than of capability.

Unfortunately, I agree with Yin26 that the relatively high popularity and good reception for the game makes it less likely that there will be significant fixes.

[yin26]

A post from the official forum had a few things to say on this. Guess we're not alone:

If I understand correctly, the Tough difficulty is supposed to be when all algortihims are in place, correct? For the past few games I have been playing through with 9 races on Intelligent and everytime it seems that it's too easy. Of course the answer to this is to increase the difficulty, but I wondered what a game on this difficulty would be like if I did not build any military ships. So I did.

The result? A cultural victory to me. Now, The AI loved me from the beginning and I have no idea what I did. I was researching my techs and then selling them to everyone else and I ended up researching everything up to cultural domination. My influenece spread like no other and while everyone else was fighting their own wars, their planets were swiftly rebelling to my name. Maybe I was doing something right, but never once did any AI relations with me go below Neutral, even with a zero military rating.

I guess once the next patch comes out this won't happen as often, but until then, I guess I'll be playing Crippling or something.

Hooray.

Now if I could only get through the proxies at my University to get on Metaverse... I'm scoring about 25,000-50,000 on all of my games.

#1 by Nadia Surma
Tuesday, March 21, 2006 9:51 PM

That's the best strategy to win, hands down -- trade techs for BC to fuel your nonexistant economy to research your heart out. It actually it destroys them on Intelligent. Larger maps require a different initial strategy than smaller maps, but hey, you're human, you can adapt to anything.

I've tried other strategies, and none work nearly as well (or at all, on Suicidal).

The fact that you added more races (to 9, as you stated) just makes that strategy that much more efficient (more people to trade with).

Try Suicidal with 4 opponents (including one Terran)... that seems about the right challenge level for me (I still win, but I have to work for it).

On the higher difficulty levels the AI is able to throw more monkey wrenches into your plans -- mainly because they get bonuses you don't.

I'd like to believe that Brad will make strides in this area. Maybe he won't plug the holes completely, but he'll do good work, for sure.

[bonscott]

If you guys have been following the dev journals on galciv2.com you'll see that they are really working on a lot of things for 1.1 and they may but a beta up for it this week sometime.

The two biggest things I see are Brad is really working over the AI and doing some tweaks to the economic system. One thing he notes though is there is a balance to changing things too much and tweaking. His hope is once 1.1 is out that things like the economy will be set and not change much if at all after that. The AI however will continue to get improved.

I'm sure most have seen this but there will be a couple new options that many in this have asked for:

1) "Blind" influence at the start. In other words you'll only see another civs influence on the mini-map after you meet up with them or cross into their sectors. Just like Civ 4 or even GalCiv 1 for that matter. I like it as an option.
2) No tech trading. Personally I'd only choose this as a challange game, not a normal game for me as I really like tech trading. But others don't like it. Brad is also tweaking the AI to try to balance it somewhat with no tech trading. So that should make many people happy.

[yin26]

I'm turning off trading just to kill the numbing micro. The diplo abuse isn't really the worry for me...if we had tech auctions! I was never bothered by seeing the influence, though I understand why others are.

It will be interesting to see if AI tweak alone (or in the main) will be enough to close some of the gaps.

[bonscott]

Yea, I don't mind seeing the influence from the start as it kind of fits the backstory of that you have had contact with everyone already via the Star Gates and thus know where their homeworlds are in relation to the rest of the galaxy. But this will give those that don't like it the option to turn it off.

[DaviddesJ]

Originally posted by bonscott

2) No tech trading. Personally I'd only choose this as a challange game, not a normal game for me as I really like tech trading. But others don't like it. Brad is also tweaking the AI to try to balance it somewhat with no tech trading. So that should make many people happy.

I wonder if this will affect the gaining of techs when conquering worlds. That seems likely to be even more broken if there's no tech trading.

[MojoJojoUK]

Originally posted by Solver
I always built a forward starbase before picking a fight, and of course upgrade it with weapon assists. Which is yet another thing that the AI will never do.

I've never seen it, but Frogboy specifically mentioned this tactic as something the AI will do in GalCiv 2. If I remember rightly, this was something added to GalCiv 1 after observing top metaverse players doing it.

Hello all, been following this thread with interest the last few days. My main thoughts have been that a lot of comments are only relevant in the situation where tech trading gives you a massive lead. For example, Sirian's take down of the Altarians on page 4 of this thread. Yes, it's a bit silly that the AI gives you planets to make peace with no guarantees, but if you are in the situation where you can take 6 out of 14 planets in a single turn, what is the AI supposed to do? They're obviously screwed. The only sensible solution is to surrender to another side - but if the AI does that too easily everyone surrenders without putting up a fight. Which isn't much fun.

Anyone, my point is that criticising the AI for bad moves when it is in an unrecoverably bad situation is a a bit needless. I would like to see how you all find the game with no tech trading (which should also get rid of problem of everyone using the same weapon).

Anyway, observations I have made about the game, some inspired by this threas:
1) Being able to move really fast should be limited. 3 engines per ship maybe, or some exponential rise in cost for more engines. Making the AI use engines a bit more would help, but I think it's just too easy to abuse. e.g. planet hopping transports, moving from one system to the next with no opportunity to intercept.
2) The AI not preparing for war is a good point. I think what is supposed to happen is the AI moves units towards you as it descends through negative relation states - I've certainly seen the Dregnin have ships hover around my planets when tehy were "Wary". But they don't do it enough - think they get distracted by other sides.
3) The AI adapting its weapons - it's being praised as something that does work. But it is fairly simple. And it doesn't do it in preparation - if it knows it's going to attack me, it should research and build correct weapons in advance. Suspect this may be difficult - probably AIs end up in a constant arms race with each other.
4) The AI isn't very effective at fighting at all.
5)One of the reasons for the funny economy is to give the AI an advantage because it can micromanage it nicely.
6) The extra races are pointless. Thy are not distinctive, and muddy the mix, making the original races less distinct in turn. Something GalCiv 1 got right was the races managed to be distinct while having exactly the same units and abilities as everyone else; their behaviour made them distinct. This seems to have been lost under the budget allowing more races. Although I may find this changes after further playing.
7) The alignment system seems broken. The new choosing alignment idea was designed, but apparently not enough time was found to implement it properly; hence the AI never using it. Also the single payment to change alignment seems cheap - a better system would be payments over time, like espionage. You would only be allowed to start making these payments ("charity" and "terror fund" maybe) after you have researched the ethics tech - so changing to a different alignment will delay getting the super weapons, as well as cost lots of money.
8) Think some of the strengths of GalCiv 1 came from its limited budget and features - it had tight focus and got many things right because of it. Some of this has been lost - one thing that springs to mind is that games don't seem to end quickly.


Ug, sorry for he long, rambly post. I've been saving these points up before I registered, see. And I'm too tired to express myself well at the moment. I will try to do better in future.