AVault's GalCiv review - AI cheat claim

I have just read the Adrenaline Vault review of GalCiv and this jumped out at me -

As a reviewer, it is seldom prudent to accuse the AI of a game of cheating; I will say, however, that the alien races of Galactic Civilizations seem to be fairly adroit at making something out of nothing — even at the easiest of difficulties.

In a fairly mundane game at the beginner setting, I was the fortunate recipient of a corvette class ship, found in deep space by a survey craft, within the first few turns. Finding the home world of the Yor nearby, I destroyed their survey craft and their initial colony ship, effectively limiting them to their single planet with my early blockade. Within a few turns, they asked for peace, ceding all of their financial and technological assets to the humans.

Immediately after the ceasefire, not one — but two — constructor units set out from the planet. It would have taken approximately fifty turns for one of my planets to build one at that point in the game, and even if the first was built by going into debt and arranging payment plans going out hundreds of turns, it would have precluded the launching of the second ship.

Because of these seemingly free constructor units, AI races are better able to exploit the resources of the galaxy, even if players can use starbases more effectively when it comes to combat.

This jumped out at me

First, the addition of HyperThreading technology to Intel?s latest line of processors, allowing a single CPU to effectively emulate a system of multiple processors working together, has put Galactic Civilizations into the limelight for its deeply multi-threaded programming ? something that has been a part of the series since its initial offering in the days of OS/2.

And made me, a programmer, go who cares? why is this in review? why is this relevant.

I haven’t seen the AI have a large build advantage.

“Galactic Civilizations is one of the few 4X releases in which the player can always win the expansion war”

Maybe he should have taken it off the fucking Beginner setting he said he was playing at.

and god knows how it takes one of his planets 50 turns to build a Constructor. Even at the beginning of the game with a single colony you can pump them out every 7 turns if you increase spend rate and military spending to 100%.

Indeed. It wasn’t a very good review. It doesn’t help that most of the AVault staff prefer to play FPS.

As far as I can tell, the GalCiv AI cheats in two ways (both of which are documented):

  1. The AI always knows which planets have class 15+ planets and makes a beeline for them at game start (this is actually mentioned and “explained” in the backstory). You can counteract this pretty easily by exploring like a madman and getting a few engine techs to start, thus outrunning the AI. Or get the Stellar Cartography tech which gives you the same “cheat”.

  2. The AI is handicapped based on the difficulty level - at lower levels it gets less income from planets, at higher levels it gets more. I think at “Intelligent” the AI gets income equal to a human.

Sounds like he played approximately two games.

I must confess I am extremely disappointed with the Avault review.

I can deal with a 3.5 star review if the review was fair. For instance, let’s say that the reviewer said “While some may find the game fun, the graphics in it are so distracting as to ruin the experience for me.”

I would still disagree with it but that’s his opinion and he’s entitled to it.

But the Avault review is flat inaccurate. To be honest, I cannot believe that a reviewer would stoop to the age old “The game beat me, it must be cheating…”

The AI knows the game as well as I know it which means it knows how to do a lot of things that a new player may not. For instance, the Yor and Arceans (remember, each computer player has its own AI engine and plays quite differently) will increase their taxes at the start of the game, lower spending to basically 0 and then buy on lease with the minimum down to crank out constructors and colony ships. It will also aggressively sell things to minor races (just as expert human players do) to get more money to keep doing it.

That is why the Yor and Arceans regularly lose planets to the I-league – if their strategy fails, they go into debt and their people rebel. But when it works, it works really well. It’s a gambit and it’s one of the advantages of having different AI engines.

To be blunt, saying GalCiv’s AI is bad or unintelligent is like saying DOOM 3 has bad graphics. GalCiv has its shortcomings but bad AI isn’t one of them.

I’m just curious: why do people read Avault reviews? I can understand a developer reading the one for his game, but does anyone else really expect anything from someone whose chief job qualification is “willing to work for free”?

Just to clear this up a little, apparently they know where every Yellow-class star is, but they do not know what the planet ratings are until they explore the region. So the AI will make a beeline to the yellow stars at the start, but it won’t know whether it’s heading to a sweet 22 class planet of a marginal class 15 planet.

I don’t think many here put much stock into Avault reviews and have not for a while. I think they jam them with word count to make up for time they did not spend actually playing the game like a high school term paper you did little to no research on.

Unfortunately, the nasty little unwashed masses who do not have giant sized brains brimming with gaming knowledge like all of us here, may read that and get a completely wrong idea.

Does anyone other than a game’s fanboy population send these reviewers email with complaints about their reviews? Like if I sent an intelligent, constructive one and said, “Hey fuckwad, did you know this, this and this is completely wrong in your idiotic review? Did you even play the goddam game!!!” Would he take that as valid criticism? I, of course, would include several more exclamation points in my email.

How he missed the major complaint that this Fred Waddell character has given piss poor support for supposedly “his” game, I have no idea. :wink:

I’ve always seen a lot of people that absolutely love the Avault reviews. Just swing through the Gone Gold forums sometime or USENET or whatever. People often talk about their reviews and how great they are. Mostly it revolves around the fact that the review itself is very long.

Most of us here have a lot of trouble with them because we’re simply more informed in general. But so many folks believe in the Avault stuff…which is why they’re still cranking this shit out daily.


Here’s a link to Avault’s HOMM IV review:


“The AI in HOMM IV is remarkable…”

The issue I have with Avault’s review is that it reeks of lack of having played the game and secondly seems to make issue of something that is irrelevant.

The AI in GalCiv at lower levels doesn’t cheat. But let’s say it did. So what? That would make it like virtually every game out there. The question should be, “is the game fun”?

As a buyer, that’s the bottom line question, “Is the game fun?”

But the thing that really gets me is the inconsistency in reviews. If I ran a game site, I would try to have a system in which reivewers, before they submitte their scores, could see what other games the site has reviewed in the past and see what they were scored. That way they would have a benchmark.

HOMM IV is a pretty fun game I think, but how can one give it a 4.5 for its AI while GalCiv gets a 2?

Or how about this:

Avault ranked MOO 3’s AI higher than GalCiv’s.

As a fan of all HoMM games, at first blush, it seemed like a 4.5 star game. I liked the innovations, the heroes on the battlefield, the new overland map graphics, the new creatures, the new choices you had to make when cultivating your towns. That was after a couple nights of playing. When you dug into it, the glaring deficiencies could not be ignored no matter how hard I tried.

My guess is, as you feared Brad, the guy just didn’t spend enough time playing GalCiv and neither do some of the others for their respective reviews. That Heroes review, along with UGO’s, are a joke. Highly anticipated, AAA titles must have to cause your computer to self destruct in order to get a bad review. Everything else gets a cursory glance, but the same fat-ass word count.

As a buyer, that’s the bottom line question, “Is the game fun?”

Holy Jesus, Brad!! Run!! Hide!! Before Tom gets in here and finds out someone was throwing the “F” word around. He’s like the Spanish Inquisition!

LOL! You should keep that quote and post it on your website, right next to that “cheating AI” claim in the GalCiv review.

For the record, I’m hyper-sensitive to cheating AIs and I didn’t notice any incidence of cheating in my far too numerous games vs easy and normal AIs. It may be true that the AI makes a beeline for yellow stars but I’m not much bothered by that. Otherwise, the AI is certainly the best I’ve seen in any strategy game so far, especially when it comes to gauging it chances at war and diplomacy. Finally a game where the computer players aren’t all disoriented psychos!


I’m trying to not be too harsh on Avault. I’ve been writing columns for them for 2 years. But right now I’m so ticked off that the veins in my neck are pulsating.

I mean, I wrote the AI. Accusing it of being bad or cheating at the low levels is like accusing me of cheating. I take it personally (too personally I admit). GalCiv is the work of a lot of people but the AI I wrote 100% of. I know what it can and can’t do.

At the higher levels, we give the AI lots of money. But at the lower levels, I had to code all sorts of strategy tricks to get it to squeak through things.

Like the constructors, how is the AI able to build so many constructors early on? Because it lowers its spending to practically zilch and increases its tax rate to around 50% (this is only 2 of the AI personalities that try this, others don’t because it’s a risky gambit) and then they lease stuff like crazy while going for easy to get techs and selling them to minor races.

New users don’t realize that the minor races are such cash cows. Expert player worry that the minor races are an “Easy exploit” because you can get so much money out of them. But that’s why they’re there.

Well it appears that we now know why he thought the AI was cheating: Leases.

The reviewer wasn’t using leases, didn’t apparently know about them.

Hence the whole pacing and AI arguments are, unfortunately, pretty meanignless.

Not utilizing the leases in GalCiv would be like not going after the “Creeps” in Warcraft and wondering why you lose. They are a crucial part of the game’s strategy.

But Avault reviews are well-researched! He must have played the game or all of… two days!



No doubt a bad review, but I`m surprised that a reviewer would not a least contact the developer with questions before going to print with such a negative review of a major release by a established and respected publisher/developer. ( Corporate Machine, et al )

It`s not like you are impossible to get to… :wink:

I do know that both Bob Mayer and Scott Udell of CG - CGOL tend to do that, and Steve certainly did while writing his very accurate and unflattering review of Sim City 4.

Having said that, more easly accessible ( in one place ) and detailed Documentation in perhaps PDF format, as well as a Tutorial Scenario would go a long way to ease people into what is certainly a complex interface.

On the other hand, one could argue that the reviewer should not need to contact the developer to ask questions or clarify points on the game. After all, the average user doesn’t have that sort of access (usually… I guess Brad is sort of the exception, there, since he makes himself publically available on a number of forums).

That said, the reviewer should also play the game enough to make accurate criticisms, which the Avault reviewer clearly did not do.