I'm coming back to this after a weekend off to experiment. It appears to me that the AI follows a certain set of rules. It has to send a scout to a planet it wants to colonize, even though it might know it's there already (when it shouldn't). Let me ask you guys one thing; Is it possible that the AI knows where the good planets are (maybe it *can* see what the player sees?), but sends a scout anyways to follow coding rules...and regardless of whether or not the scout makes it to that planet, the AI creates a colony ship after the scout (scout either gets within sensor range or is destroyed en route) and beelines for the known planet?

German print magazines have pretty much panned Galactic Civilizations 2. The best test was 75%, the worst, by the biggest German print magazine Gamestar, 62%. In that particular test, the tester Michael Graf opined that the KI does cheat, because it sends ships straight in a beeline at the best planets, according to him. Also, he says, it produces ( on a non-production intensive planet ) every turn a new ship, without having the funds to buy them and besides produces new technologies at an incredible rate.

While the last two things can be explained by different things ( full military or research production, for example ), the first item produced a lot of discussion on the forums of the Gamestar.

Several people seem to have independently found the same thing happen. I´ll summarize one of the examples they gave ( translating the whole thing is a bit much work. )

"I just tested this - KI on suicidal, middle sized map, middle probability for stars and habitable planets and a few opponents. I revealed the map by pressing CTRL - U and walled in the opponents flagship and scouts with battleships ( CTRL - N, CTRL - T ). Nonwithstanding, after a short time colony ships were on their way to several of the best planets in the vicinity - without scouting and not into "empty" systems. That leaves only one conclusion to me:

The KI cheats.

Frogboy either doesn´t tell the truth - or I found a KI Bug ( Because in other tests I did, without the ship walling, I also noticed how a few systems were scouted and then colonized. ). "

So, I am honestly a bit bamboozled here. I haven´t noticed that happening, but of course I didn´t probe as deep as that particular poster.

An opinion from Frogboy would be highly appreciated... we GalCiv2 fans want to bludgeon the responsible writer at Gamestar into admitting he made a mistake.

In that particular test, the tester Michael Graf opined that the KI does cheat, because it sends ships straight in a beeline at the best planets, according to him. Also, he says, it produces ( on a non-production intensive planet ) every turn a new ship, without having the funds to buy them and besides produces new technologies at an incredible rate.

What is the version of the game? As noted in thread (post 122), there was a glitch in the AI code that may allow him to know planets it hasn't seen.

Well, since they tested it just after it came out, I´d assume the unpatched version? I cannot say for sure, I am just one of the people discussing with him.

I didn't read any of the german reviews... do they only attack the AI? Or do they have other issues?

Cuz if they are only harping on the AI "cheating", well.. nearly every single-player game on the planet has cheating AI. Why knock a game for that unfairly?

I can only speak for the Gamestar review ( Gamestar is the most sold magazine in Germany for PC Game testing ). They panned the AI and the starting positions. It was only a one page review, so it wasn´t all that detailed. I´ll try to translate the relevant bits later on, as I need to go out now.

Ok, here goes the negative part of the review... I don´t really have the time at the moment to translate it completely.

"The bad part is that the AI seldom does play fairly. The opponents cheat themselves techologies and always know where are habitable planets, while you have to explore the galaxy. Also, the starting points are most of the time unfair. Sometimes you start directly beside a mighty enemy, sometimes in-between a bunch of unhabitable planets. That a lot of times nothing happens and the only thing you can do is to click the Next Round button makes GalCiv 2 only for professionals."

Gamestar uses a detailed system to determine the total score of a game. Here are their notes for GC2:

Graphics: 5/10 ; Sound 7 / 10 ; Balance 3/10 ; Athmosphere 7 / 10 ; Interface 8 / 10 ; Size 9 / 10 ; Starting Positions 3 / 10 ; AI 5 / 10 ; Units 8 / 10 ; Campaign 7 / 10 ( that would be the normal game they are talking about, not the actual campaign ).

Funnily, Civilization IV got 90% from Gamestar. Starting positions there got a 10 / 10.

Fans of GC2 went on to complain on the forum. Answered Michael Graf:

"I know what I experienced while playing. It´s not true that the AI knows where the planets are but not which are unhabitable. If not it wouldn´t send every SINGLE colony ship across the map to select the ONE inhabitable planet sectors away, without ever having sent a single scout ship that way.

But okay, you won´t believe me. And I guess also not that I try to give every single game a fair review. I am not prejudiced ( especially not in regards to GC 2, which I ---- as I had said before --- happily anticipated. ). And I repeat what I said in my review: I LIKE this game. I like it despite its weaknesses! But I am sure that there are round-based strategy gamers who won´t be able to enjoy it. Because of its weaknesses. And that angers me. There are design errors which can´t be overlooked.

It may be that you didn´t see the things I mentioned. That´s okay. But when the developers of GC 2 say out of PR reasons, that their AI is the best there is ( which, btw, they already did with the first part ), then it is not true."

Well, that´s basically what he said. I just finished two tests I ran on a small map... cheating away the fog of war, getting myself a Battleship, blowing away the scout ships of the opponent and colonizing all planets but one... I simply did not see the AI send colony ships straight to the last empty planet. But, oh well, I guess my perception is tinier than the one of a mighty game reviewer.

Well it would appear GameStar's opinion isn't worth taking very seriously.

Perhaps they should review games for more than an hour.  

I mean, look at their ratings. Sorry but that's a total joke.  Balance 3/10? AI 5/10? Whatever. Sounds like sour grapes. Graphics 5/10?  Was this a review or just a trolling represented as a review?

You´d have to ask the reviewer. Here is his e-mail: [email protected]

I have to add that the AI of Civ4 got a 8/10 in the same mag. Totally ridiculous if you ask me.

Brad, you won't believe me the reviewer says that he liked GC1 and played it for months. But I have to admit I don't believe anything he says, so I don't know if it's true

But anyway, the reviewer of the PC Games -another german print mag - for GalCiv2 (GalCiv2 got 64%) atleast admitted that he tested the game only for about 10 hours, so it becomes quite obvious why the review is shit, though this won't change the damage, which is already done

It's going to be a few days before I'll get a chance to try this out... going out of town. Still if true, reminds me of the days back in school when we didn't call FM Frequency Modulation for a reason. A "Jurassic Park" reference would be, "life finds a way." Sounds interesting and it's something I'll test when we get back.

Well I guess that settles the argument on whether to do German versions of future games. 

PLEASE LOCK THIS THREAD

Thank you.



"Gentleman,you can't fight in here! This is the war room!"

PLEASE LOCK THIS THREAD

Why?

Well I guess that settles the argument on whether to do German versions of future games.

But don't forget there are people in Germany, who really like your game and even some of them are defending your game against this "reviewer". I can understand that you are sour, but the german online mags gave the game positive reviews, so it would be quite a punishment for those who like your game here in Germany and are not capable of English. The best would be to not send them (the german print mags) a version for reviewing. I don't know why Paradox is doing this anyway

Yeah, my opinion. While I am not affected by no German versions, the players who would love to play your game and don´t speak English would suffer. No shooting of the messengers, please.

I wouldn't give up on Germany. They are one of the biggest buyers of strategy games out there.

Damn, this thread never wants to die.

Well, It is definately the case that the German reviewer was using an un-patched version of the game.

Back when this thread was new and fresh I took an interest in this topic, as so I did some testing and came up with Link, which I posted way back when.

Then Frogboy came up with this:

Post #122

I did put in some code to make sure that the bIsKnownBy variable is intiailized to 0. Because it IS possible that under some circumstances that you could end up with garbage in that variable and so the AI thinks it knows about it. So try it out with 1.0X and let me know.

And the test is pretty simple. Grab a copy of galciv2 from before 1.0X and try out the procedure described by the German article, or the one I described above (which is basically the same), and you will see what he means. After 1.0X, it no longer works. Bug fixed.

Does this mean that people running pre-1.0X have a cheating AI? No. Not really. As it happens it is a rare case that causes the AI's colony ship to B-line to the best planet - it just so happens that that rare occasion is the one he simulated when he cheated (hey, if the player cheats then the AI cheats... I don't see what is wrong with that).

As I have mentioned in the previous link:
"What does the colony ship do if it has no where to go? ... Most of the time a colony ship gets to where it is going. If it doesn't, it'll find somewhere else it knows about, and if it can't do that... straight to any habitable planet regardless of what they know."

What are the chances of that happening? Highly unlikely. Most races have a second planet in their home galaxy. The ones that don't, the scout ship or flag ship tend to find one rather quickly. That's the first colony ship dealt with. After that, new colony ships are only created when there are planets for them to go to. The only chance of this bug coming through is if two races are racing for a planet, one doesn't make it AND the race that doesn't make it doesn't know of any other planet from it's scouting efforts and there is still one it doesn't know about elsewhere.... phew.

My guess is that for the writer of the article, this never happened. The AI just out-smarted him. And when he did the test, he stumbled across an old bug and came to the wrong conclusion.

But of course, if you have the patch none of this matters.

I don't trust the origional poster, did you notice how he changed his story when someone brought up occassional planets or rare planets. Then all of a sudden, yeah that's how I played it, I meant to mention that. He's clearly just trying to destroy the game for "multiplayers sake". Go back to Age of Wonders:Shadow Magic where they cater to mp's over single players. You lost any credibility after your 2nd post and certainly after Froggie showed how the code works. Plain and simply....you're lying.

OK. I found some time to test the whole idea. I'm running Beta 2a, tiny map, intelligence set one step above normal, common planets. Colonized all but one planet (out of sensor range, but in ship range). Killed their Flag ship and any scouts that appeared.

Result: The colony ship went into orbit over the home planet every single time. The times there was a planet in their home system, the colony ship would make 1 move towards it, realize it was now colonized and turn around and go into orbit.

I tried it on a large galaxy... same result. I even left a planet out of ship range; low and behold... nothing. The AI went to ground while under attack and was trying to save its colony ship.... every single time. There was no mysterious constructor making an appearance to provide a base and therefore range to the planet. Without instructions from a scout, the colony ship did nothing.

Sorry, I have to find the other results suspect.

I'm running Beta 2a

Sorry, I have to find the other results suspect.

Well, you have just demonstrated that the problem isn't in the latest version of the game, as anyone would expect for any version greater or equal to 1.0X after seeing the post #122

I very much agree. I started in the corner of the galaxy 5 times in a row. While those sons of BITCHES Altarians got the middle. Now I'm pissed. Then the Torians have a freaking ring planet that I want and they're bigger, but I have 70 soldering, they have 25. I've invaded that bastardly planet 10 times and with no good results, even when its 25 to 3.

They also know where the good planets are. I sometimes think the computer does this to smite me.

Those goody two shoes bastards also shove colony ships out like Santa Clause shoves out presents.

Plus I don't know how to cheat...yet.