Well, this suppose that the AI works the same way if with FOW lifted for the player or not. If there is a bug, it could be that without CTRL+U, the map is revealed to the AI and with CTRL+U, the FOW is enable for the AI. That would explain different behaviors experienced by various posters

Eh, highly doubt it. I saw the same behaviour with and without cheat. With pumped up scanner and engines I was able to get into possitions within 4 turns to watch everything the AI was doing.

Well, this suppose that the AI works the same way if with FOW lifted for the player or not. If there is a bug, it could be that without CTRL+U, the map is revealed to the AI and with CTRL+U, the FOW is enable for the AI. That would explain different behaviors experienced by various posters

Now you are being rediculous... If that was the case then I would say that the coders of this game have absolutely no idea what they are doing, and I think there is plenty of evidence to the contrary (For starters, the game is good and works)
The only possibility is that is the FOW is lifted, then averyone knows everything, and that is definately not the case so we can assume that everything is working properly

But still it appears the computer seems to know more about anomalies than it should or else we don't because I can't see any way short of cheating or sensors of spotting anomalies. And my incedent with the enemies survey ship changing course was probably due to another race took the anamoly it was after

This I do think is a bug (or atleast a "feature" that doesn't make sense).. but If you set your ship to auto survey, you can do it to! so atleast it isn't a one way cheat

Oh ok, must be with patch. Before even if another ship had spotted one my survey ship wouldn't go.

AI planet knowledge cheat: Myth Busted

Now you are being rediculous

Nice. I was just pointing possible comparaisons problems between someone who experiences an omniscient AI with no cheat code and someone who experience a non omniscient AI with the FOW revealed for the player

If that was the case then I would say that the coders of this game have absolutely no idea what they are doing, and I think there is plenty of evidence to the contrary

Sometimes, mistakes can be made very easily and and are very hard to spot. You know, it is very easy to take the wrong value when using a boolean, and be sure to have the correct behavior when looking at the code

BTW, when enabling cheat code, it could be nice to take control of the AI to see what is its FOW. It would be also nice to use the same savegame and see how the AI behave with FOW lifted for the player or without any cheats enabled.

[post deleted]

@ jocmeister

not sure if this answers your issue, but i have reproduced ai (normal) behaviour on a small map. on turn1 reset your research spending to 50%. Research xeno research. Purchase a scout. Use survey ship to scout systems (with planets, if i have stellar cartography). Settle poor in-system planet, purchase whatever level of factory is available, set it to build a starport, then a lab. Set home world to build (not purchase) a factory, then a lab. end turn. launch scout and set it to explore a different (nearer, because its slower) system (with planets). Set homeworld to build (not purchase) a colony ship. When a habitable planet is found, purchase the colony ship and next turn send it to the planet. rinse and repeat. depending on planet location, i have matched/surpassed the ai in colonisation withthis approach.

i have not yet managed to match ai in military strength early game. i suspect i need to focus research on getting a combat tech, design a really poor fighter and set Mars-equivalent planet to build/purchase it.

I always build a sensor-15 drone first thing in my games to get an immediate look at my surroundings. EVERY time there's been an AI in range of me early on, I can see its scouts swarming all over the place before it ever leaves his home system with colony ships. Not surprising at all since Brad posted the code showing it must do exactly that.

The sentence "The AI isn't very smart" is relative. The AI can adapt very well from what I've seen while playing. There were a few dozen turns where the Torians had a ton of colony ships pressed up against their maximum range like kids pressed against a plate glass window, not coming or going, just trying to hope their way through. I didn't take that as a good sign, but later in the same game when the Arceans attacked the same spots over and over as a decoy to throw the largest fleet I've seen in-game yet at my rear guard where there had been no conflict... I smiled. That's when I felt respect for the computer like I was playing a worthy adversary.

Remember, it is a game. If you've seen the TNG episode where Moriarty gains consciousness, remember, they're trying to tweak an AI that can beat Holmes, not one that can beat Data.

Just chucking in some empirical obersation from in my most recent game.

The Torians were on a course for a planet they had definitely seen (they had a colony in that system) I took it the turn before.... that same turn, my scout picked up a PQ 26 planet literally a few tiles south of my new planet, in another system. The Torians, although I fully expected them to b-line to that planet, actually turned tail and went back a few sectors (about 8 game turns) and took a PQ 6 equally far from their capital....... in simple terms, they hadn't seen the PQ 26 planet or they would have taken it as it was only 2 or maybe 3 turns from their colony ship.

Of course, this is just empirical evidence as seen in game, I couldn't possibly compete with the factually decisive statements made above by the proponents of AI cheating!

And, just for the record, there are a couple of real twats on these boards who seem to troll for the express purpose of jumping down everyone's throat who actually likes and enjoys the game and is prepared to stand up for it. If people post erroneous statements that criticise components, arguing against it does not mean you are "a fanboie" but rather that you are correcting the aforementioned erroneous statements. Yes, there are plenty of bugs in the game, but these "AI is teh ch33ter" threads are based upon supposition or badly flawed "tests".

Playing a Large Map as the Yor in the beginning on normal difficulty versus 7 AIs. As I am still sending my Colonyship to the first planets I found when all AI civs smack down on their second planet. About 10 turns later all AIs settle on their 3rd planets. This is when I am still buildning my second colonyship. My Flagship hasn´t even found a third system yet!

Ever thought of buying your colony ships like the AI does?Some will take loans out to get enough colony ships.

I've definitely seen the AI miss better planets that it hadn't seen and colonize inferior planets instead, even when the better planet was closer. There might be a bug here that causes the AI to see through the fog of war in some games but I doubt it's a common event.

There were problems with the FOW in the early beta, especially regarding games generated with Ctrl-N, you'd often see what you had revealed both in your current game and the previous game.

As a side note, discard any testing you do with the Torians. Their obnoxious sensor bonus means that if you're playing without the latest patch, they will see much more than you do, even with sensor barges.

I don't know about planets and colony ships but I *do* know that my ship on auto-survey always goes for anomalies in unexplored parts of the map. Just start and set your flag ship to auto-survey and it'll start racing around the galaxy without you ever scouting anything.

Okay, so I tested it. Here's a summary.

I started a new game. Gigantic map, frequent stars, rare planets, rare habitable planets. I set the scenario to Accellerated Start so as to get a leg up - you get 10,000 credits to start and a bunch of techs, including Warp Drive, which I needed for my strategy.

So first turn through, I created two classes of ships: The Sensor Bouy, which has 15 sensors and moves at 4. And the Interdictor, a small craft which has a single laser and two warp drives to move at 7. I bought a sensor bouy and set my spending to max out military, completely tossing research to the wind.

The second turn, I launched it and encountered a nearby minor civilization. I sold them all my techs, got lots more money, and proceeded to use the money to buy alternating Interdictors and Sensor Bouys, using them to set up a perimeter immediately around my empire, guarding my borders with the vigiliance of hawks. I also found a second star nearby with a habitable planet, and noted that none of the AIs were anywhere near my civilization. The sensor bouy set up station around the habitable star to the southeast, and I held off on colonizing anything.

Every scout ship that my sensors picked up, my interdictors flew out and destroyed. All of the AI scout ships had 2 sensors, and never got close enough to my home system or the second system to detect the second habitable planet. I made dearly sure of that, even though it meant ending up at war with every single civilization in the game. I encounter four other minor civilizations, each time selling all my techs to plunder their treasury, then destroying their scout ship they'd send to say 'hi'.

So now I'm sitting back, guarding my borders, making sure nothing gets through... and lo and behold, an Arcean colony ship pops up on my sensors, heading RIGHT FOR the single habitable planet in my area that NO scout EVER got close enough to see.

To sum up: It certainly appears like the AI is cheating.

Questions, comments? I have a save-game available if people are interested.

crickel

Edit: As a side note, the 'fog of war' being reversed is both: A. a plausible explaination for what is happening, and B. a plausible mistake. All it would take is accidentally leaving off a single 'not' statement in the code in some IF statement somewhere. If that were true, I would hardly call the programmers, 'the worst coders in the universe ever' or something like that. I've seen mistakes in code like that in place for years of use without being noticed.

Okay, so I tested it. Here's a summary.

Look like a valid test. You may want to send the zipped savegame to [email protected], with the above commentary.

Have you frequent savegames? reloading previous savegame wit hcheat code enable may be useful to spot when the Arceans decided to launch the colony ship, and if its destination have changed. You may also want to see if the behavior is repetable, by reloading savegame before the launch of the colony ship and see when the problem appears. It could be nice if you are able to get a savegame few turns before the annoying bevhaviors. Anything than can speed up the precise identification of the bug location is useful.

heading RIGHT FOR the single habitable planet in my area that NO scout EVER got close enough to see

on minor thing: how are you sure it is heading right to the planet and not toward the star?

*Sigh* This is the exact mentality that makes me agree with the argument that playing online is no fun... There are so many people out there that immediately assume "CHEATING" whenever something happens they don't expect/can't do themselves. And it is precisely the same here.

See, when I hear these stories the first thing I think is not "OMG, the AI is cheating", but "The AI is smarter than the player"... Which is usually the casemor of the time. online or against AI.

There are so many more plausible reasons for how the AI knew to go there; They saw it before you managed to setup your border, They have better scanners than you think... maybe the noticed you killing all their scouts and figured there must be something good there. Honestly, I don't know.. I didn't write the thing...

But what is certain, whatever there system is, it obviously works pretty well.

As for 'reversible fog of war', I don't think you fully understand how the process of getting a game out works... They have put countless days and months testing the game, and this would be one hell of an obvious error for a programmer to notice (considering they have a better idea of how the AI should be thinking, and that the AI would be doing the complete opposite to what they expect).

Secondly, it is highly unlikely that changing the visibility for the user is going to make any difference at all to the visibility of the AI, especially if that change is "make it work" I don't think you understand how totally unlikely that would be, taking into account the way these games are programmed, and that it would make obvious changes in behaviour on the AI that would easily be noticed by the programmers.

I think you all should give the programmers the benefit of the doubt that they have, indeed, created an AI that is smarter than you are (no offense intended).

But, so I don't look like I am sticking up for the programmers without any proof - Crickel, yes, I would like to see the save game.. Maybe I can have an idea where you guys are coming from.

Secondly, it is highly unlikely that changing the visibility for the user is going to make any difference at all to the visibility of the AI, especially if that change is "make it work"

Are you telling us that if you remove the FOW for you through cheat code, it won't be removed for the AI? Sure, I can understand for debugging puprose, that if the FOW is lifted for the player, the AI must still have it to test the validity of the exploration process. BUt I don't know what exactl the CTRL+U cheat code do. I know it remove the FOW. It gives also the universal translator. But what does it exactly do to the AI?

As for 'reversible fog of war', I don't think you fully understand how the process of getting a game out works

Perharps. I know that nobody can do full regression tests on the AI. And if you aren't aware about strange behaviors, you may not investigate enough time to see it correctly. The test protocol made by Crickel mean to invest time in testing with a specific way to play the game. Not sure that this kind of test is made on all builds

BTW, my point of view in this topic is very simple: consider that everyone tells the truth about what they have experienced.

There are so many more plausible reasons for how the AI knew to go there; They saw it before you managed to setup your border, They have better scanners than you think... maybe the noticed you killing all their scouts and figured there must be something good there. Honestly, I don't know.. I didn't write the thing...

Sure, but the test made specifically said:

The Sensor Bouy, which has 15 sensors and moves at 4.

I also found a second star nearby with a habitable planet, and noted that none of the AIs were anywhere near my civilization.

The sensor bouy set up station around the habitable star to the southeast, and I held off on colonizing anything.

Every scout ship that my sensors picked up, my interdictors flew out and destroyed. All of the AI scout ships had 2 sensors, and never got close enough to my home system or the second system to detect the second habitable planet.

So thess should invalidate some your proposals. Andthe arceans haven't important sensors bonus. But there could other explanation:
- the Arceans flagsip enters a wormhole and ended unnotice in the player space. Then it manage to escape unnoticed through another hole
- the AI had send a colony ship blindly. The main problem is how to exactly know the precise destination set for the ship

I think you all should give the programmers the benefit of the doubt that they have, indeed, created an AI that is smarter than you are (no offense intended).

Strangely, it is my position. But if some people report stange behavior, then tests and investigations are needed in order to have a proof of a buggy behavior if there is one.

Strangely, it is my position. But if some people report stange behavior, then tests and investigations are needed in order to have a proof of a buggy behavior if there is one.

And I am pretty sure that exactly this is what they did during the testing phases of the game. And also partly why I want to see the save game files...

But if you go through these replies, either two things show up, a) scouts are always seen before colony ships, or colony are seen, but we don't have any evidence whether or not other ships have been before.

Ok, so b turns up more than a, but only because b people are louder (considering the topic of this thread that is to be expected)

If the people had hard proof that there was no scouts, or other ship that had seen if before (which they don't, since they are working within the limits of the game, and therefore could miss information) AND that sending colony ships to stars they haven't been to before isn't just one of their other strategies (ok, so no one has seen one go to a bare solar system, but it hasn't been ruled out yet), then you are just jumping to conclusions to say that there is "proof".

So yes, more tests are needed... if only there was a way to be 100% sure of what the AI is doing... oh wait there is, except people are going out on a limb and saying "oh, turning on that cheat might all of a sudden cause the AI to do exactly what they are supposed to"... thats sounds more like a plea for an excuse to be correct than anything else, and definately no proof that that could be the case.

So, assuming that everyones observations are correct, then the AI does things we don't expect them to do, except when we did... expect them...to

except people are going out on a limb and saying "oh, turning on that cheat might all of a sudden cause the AI to do exactly what they are supposed to"... thats sounds more like a plea for an excuse to be correct than anything else, and definately no proof that that could be the case.

Well, maybe I wasn't clear enough: I was trying to point out that enabling cheat mode can cause the AI to react differently. And that maybe, the tiggers associated with cheat code may help to spot the bug. Obviously tests are needed to confirm any hypothesis and I have never told otherwise.

So, assuming that everyones observations are correct, then the AI does things we don't expect them to do, except when we did... expect them...to

Well, the problem is to know if the things the AI do are coded that way or results of a bug. The only way we have to spot is to try to reproduce the strange behavior while carefully discarding all hypothesis.

Playing a Large Map as the Yor in the beginning on normal difficulty versus 7 AIs. As I am still sending my Colonyship to the first planets I found when all AI civs smack down on their second planet. About 10 turns later all AIs settle on their 3rd planets. This is when I am still buildning my second colonyship.

The AI isnt cheating, hes doing what most experienced player would do. He is buying colonoy ships the first few turns. Remember, the buy button is your friend.

As for the debate. Its clear that the intentions of the original poster was to point out flaws in the AI to support his case that there should be Multiplayer. His tests are most likely bias. So are the ones done by other trying to prove him wrong, most likely. In the end, if you like the game. Enjoy it. If you think the AI is cheating and the game to hard, then turn down the difficulty or find another game. If you want Multiplayer, then post your comments in one of the numerous threads already about it And say you prefer mutliplayer and that your friends do to.

If you want something added, you should be more positive. The developers are more likely to listen to "I want multiplayer because me and my buddies love to play multiplayer space sims and this one would be great with multiplayer" then "We need multiplayer because the AI is a cheating bastard".

If you want something added, you should be more positive. The developers are more likely to listen to "I want multiplayer because me and my buddies love to play multiplayer space sims and this one would be great with multiplayer" then "We need multiplayer because the AI is a cheating bastard".

Ok, well thanks for your objective reply. I got to the beginning of the second page of this thread and realized that the majority of the people have negative comments to say about me personally and my "supposed biased" tests. So I just skipped down to the last one and read yours.

I honestly never really meant to be negative. I guess the title really was negative and it should have been a little more PC... but never-the-less I DO like this game and have enjoyed 2 huge galaxy games already (one where I got slaughtered and another where I did the slaughtering:))

Anyway... just to point something out OBJECTIVELY.. here is a well respected game reviewer who just happens to notice the exact same thing that my original post was implying.

http://www.gamerankings.com/itemrankings/launchreview.asp?reviewid=670112

Final musings There were some problems in the original, the AI seemed to still magically know where the best planets were and ...

Although he goes on to say at the end of that sentence "...the point I make here is that while reviewing the game I haven’t experienced these problems..." I dont know if he meant the ORIGINAL Galciv had these problems and he was comparing this game to that game or he STILL see's that the AI magically knows where planets are.

Anyway... Read it yourself.. its a pretty good review of the game.

PAR

Uh, what a hot debate ! But in fact I don't care a dime if the AI knows or not where are the planets/goodies !
I've no real feeling of competition vs the AI, it's not a player ! It's just a game where I think I should be smart enough to beat it, else I lower the AI level .
Did anyone remember old MoO and MoM ? AI was just a cheat bag, they had stuff generated out of nowhere . Same in Civ series (1-3, dunno for 4).

when i dont know where the habitable planet is, I send my colony ship to first star system with more planets, and if i dont find it, i go to the next one, but i find it most of the time...