I think what he is saying is they go directly from were they're built to a hapitable planet. If that is the case then, yes the AI has an advantage. However if the ship deviates in anyway from its course then it doesn't cheat, it just has a good idea due to stellar cartography as the guy before me said. I mean I can send a ship to were a star system is on the map and by the time I get close it could see were the planet is.
-Try your test again with Rare habital planets, and see if the AI sends ships to star systems with out habital planets. If it doesn't then yes, It does cheat.

-Try your test again with Rare habital planets, and see if the AI sends ships to star systems with out habital planets. If it doesn't then yes, It does cheat.

I think you meant, try a test with a galaxy with rare habitable planets. If the AI sends colony ships to stars WITHOUT habitable planets, then it proves your point that it's just sending out colony ships to stars on the chance that they will have good planets.

The Auto-Survey code to go to unseen anomalies was added in late gamma as a gameplay improvement. It in no way implies that the AI is given unfair knowledge about the map.

Your computer knows EVERYTHING about the map.

Only some of that knowledge is passed to you and the AI's

My experience has been slightly different. I suspect that the AIs can see what I can see. In my most recent game, twice, towards the middle of the game, I found a resource on the other side of the galaxy (huge galaxy). In each case, there was a mass rush of construction vehicles to the resource directly after that (the first one I was beat, the second one was closer to my colonies). I doubt multiple AIs simultaneously discovered the resource when I did, so something odd happened.

In GC1, the AIs knew the entire board (by my understanding). This was supposed to be changed in GC2 (again, by my understanding), but I am suspicious that some aspect was missed (like maybe how the AI reacts to what it knows you can see). Not knowing the internals, however, who knows?

You appear to presume that the observed behavior is a result of laziness in AI design. I think you underestimate the difficulty of creating a good AI, but, at least in this case, I suspect it is rather a bug.

Sigh.

Look, the code isn't very complicated. Here's a copy/paste:

for(ulIndex=0;ulIndexulTotalPlanets;ulIndex++)

{

pPlanet = g_pGalaxy->pPlanetInGalaxy[ulIndex];

if((pPlanet)

&& pPlanet->IsOnMap()

&& (pPlanet->IsKnownBy(QueryID()))

&&(pPlanet->GetQuality() > ulMinimumQuality)

&&(!pPlanet->IsColonyColonized())

&&( (pPlanet->GetOwner()== CIV_INVALID_PLAYER_ID)||(pPlanet->GetOwner() == pShip->GetOwner()) )

&&(pShip->IsDestinationInRange(pPlanet->GetXPosition(),pPlanet->GetYPosition()) == TRUE) )

{

//evaluate to colonize

}

See the "pPlanet->IsKnownBy(QueryID()))"?  The only way it can know a planet is if a ship flies near enough to spot it.

It's really that simple.

 

How does espionage work? I mean, once the levels are high enough, can you see (to a degree) what your target sees? In that case, shouldn't the AI be able to see what the player is seeing with enough money put into esponiage, so that its scout ships don't necessarily have to fly by the habitable planet? After all, if you map out the galaxy and it can see what you see, it only makes sense the AI would beeline its way to the good planets.

Of course, if esponiage doesn't show anything on the map at all, then something must be up if their scout ships indeed didn't fly by.

Ok well... I redid my test for a 6th and 7th time.

Now, I know I mentioned my "testing procedures" previously and unless you want to just blatently ignore them, well... so be it.

Now, I made a medium galaxy with RARE habitable planets and ABUNDANT stars. I had only 1 opponent and that was the Terran Federation. I recreated the game each time until I knew that we were on opposite sides of the map from each other.

All I did was create exploring ships, colony ships (for distance), star bases (cheap ones) and sensors. In both games this was the position of each player...

                    -

             -             -

       -                         -

-         ME           AI            -

       -                         -

             -             -

                    -

What I did was create constructors and colonized habitable planets so that I could "watch" the AI and what it would do like so:

                    -

             -  |          -

       -        |                -

-         ME   |          AI          -

       -        |                 -

             -  |          -

                    -

I also made sure that there were plenty of stars behind my lines.

Now, by the time the AI had enough distance to go past my lines, I could pretty much watch their every move.

I watched their colony ships go straight to stars systems with habitable planets and ignore the star systems with NO habitable planets. Not only that, their colony ships didnt make any "turns" or adjustments in their vectors (unless sommething got in the way) when heading towards habitable planets.

I also made sure I watched the "anomalies very carefully". You can tell when scouts do not have survey ability... they just randomly go all over ( I watched all their scouts and plotted where they could have gone). MANY anomalies behind my lines were in no way within their field of view. So once I noticed a ship come over my field of view boundary, I watched them for a turn, and plotted a direct course to see where they would go if they continued on that vector. And not to my surprise, 100% of the time, the AI ship was headed towards its goal without every having "seen" it.

Oh, and until you run the same tests, just try my simple little survey ship test with your own ship. Get it way out in the middle of nowhere (where the only path "seen" is behind it) and stick it on Auto Survey. You will then see the ship veer off into the unknown DIRECTLY to an anomaly.

Now, you can blow all that away and call me a fool. I hardly care. I am hoping for 3 things here.

1) For Stardock to take a look at their AI code a little better since this "is" a single player only game.
2) For non-fanbois to coherently read and understand my "testing" and either back it up with their own testing or debate me if they find different results.
3) For Stardock to see the value of putting in a multiplayer option.

Now, for Mr. Sunjah Kahn and any other fanboi's that like to bash those asking for a multiplayer option instead of debating, I have read the "need" for multiplayer in the forum and I have also read that if they get enough requests for multiplayer they will add it in. That is what I am hoping for.

You sir, seem to have no "gaming" friends and only get online games where you know NOBODY. Yes, its rough out there, but I would hope your parents instilled enough self confidence in you that you could enjoy a game that is fun and not be constantly insulted by little kiddies.

I have a very round group of friends who LOVE multiplayer games and that is all we do. We played the hell outta MOO2 (tried MOO3 but it was ghastly), played Alpha Centari, played all the Civ games etc etc. And no doubt about it, all of us like MOO2 the best. And ALL of use were ready to buy this game until we heard there was no multiplayer.

But, to look at this from a business point of view. They state that they would rather save a buyer 10% off the retail price because it does not have multiplayer. Wouldnt it be "smarter" business wise to open up your prospective customers and let them pick and choose if they want to buy? I know at least 20 people right now who will NOT buy this game w/out multiplayer =/ I am the only one that bought it and so far its very fun (cant believe they didnt put tactical battles in... but I guess I can see why) but I would much rather go to work tomorrow and discuss my "continuing" game with my coworkers and laugh and talk about the battles we have had, instead of just ending a game and starting a new one

PAR

Um, Par... They said that the framework is built in to the game for multiplayer. And that they may release a FREE upgrade to multiplayer in the future. So chill, kay?

Nice clean code is so refreshing. I've seen things that would make an experienced programmer cry....in terror at the idea of debugging it.

Even without the code example, it is obvious that the AI uses scouts to figure out its surroundings. It's actually more obvious on larger maps rather than smaller. when you see scouts in your area of space you can be certain the ai will build colony ships imediatly for any non colonized worlds in your area.

Hmmm...that's odd, because not only does my survey ship shut off "Auto-Survey" if it can't see any more anomalies, but I've seen the AI scout it's butt off with scout ships before it sends colony ships. Those scout ships seem to move around pretty fast. Isn't it possible you are missing the AI's ships? Or more likely, that the AI is sending a ship to every star system?

Frankly, I don't know why I should believe your rather strange test when the code clearly says that what you describe can't actually occur. In my experience, the AI has seemed to explore with scouts before colony ships.

I could be wrong though.

I haven't done any experiments and I suppose there could be an overlooked factor that somehow allows the AI to benefit from your experience as well as it's own but I would wonder why, if they know where the usuable planets are, do they spend so much of their treasure on fleets of scouts?

Maybe they scouted the area before you cleared the fog in your example. I have found that the AI rushes to build extended range scouts early. In a recent game I was beating the AI like a cheap rug in colonization but they had me in ship range by several sectors.

It is well known that the AI has some advantages at the high levels of the game but nothing in their behavior has given me any reason to think that they know where the planets are at normal. If they do, perhaps they need to, because I have no problem swamping them in the early going at normal difficulty. It is their lack of fatigue vs mine in the last third of the game that I have the most trouble with.