AS USUAL: AI cheats on combat results

[Lt. Killer M]

Now I am really pissed: for 5 turns, a Conquistador unit stands till next to a town of mine. On turn 6(!) it attacks, chances are around 11.7:14 for me. Surprise, surprise, it wins.


Can anyone tell me why it waited for 6 turns????? Why not attack ASAP? Why, when there was neiter reinforcement nor units moving away?



I can only conclude that the AI knows combat results will be 'good' or 'bad' and attakcs only when 'good'.

[mad]

btw, this is not the first time I encountered 'weird' behaviour that benefits the AI.

[col]

There is a very good article on combat posted by Arathorn at CFC.

http://forums.civfanatics.com/showthread.php?t=137615

Maybe youre being a bit paranoid - or maybe the AI couldnt afford the loss until that turn, then got lucky.

[Lt. Killer M]

maybe - still, I have seen it too often. 3 times already since I have the game! [eek]

but, maybe I am paranoid, maybe I was fucking correct last time I said the AI cheats on this! Maybe in Civ3 the Ai DID know the map and the unit positions.

Amybe it is a bit weird that in CIV, the Ai is also capable to *know' where your troops are and never places troops in an 'oops, that road is blocked, I didn't know that' position - even when I make DAMN sure they can't have known! Maybe Firaxis likes to boast how good the AIs are and then tries to make them get there by using the usual shortcuts. Maybe it's time they learnt hat a truthfull 'we had to give them that or they would have sucked' is better than boasting and delivering shit.

[Kemal]

The AI also sometimes waits to attack until it feels it can mount a decent assault at once, not spread out over several turns. Maybe this is what happened in your case too?

[Pastorius]

Weird. Last post by Killer, active topics says, but Kemal has the last post. [hmm]

[Melifluous]

Kemals post is 3 mins after Killers.

I'm guessing that the Kllier went back, edited it to deliberately mispell somethnig.

Within the 5 mins = NOT an EDIT.

[meli]

[Matrix]

Was it healing?

[ProPain]

quote:Originally posted by Lt. Killer M



Amybe it is a bit weird

That is a bit weird for sure!

[Pastorius]

[rotfl]

[Beam]

quote:Originally posted by Kemal

The AI also sometimes waits to attack until it feels it can mount a decent assault at once, not spread out over several turns. Maybe this is what happened in your case too?


Garee with that observation. In one of the test starts I played there was a surprise attack by a AI SoD on a reasonably well defended city which I lost. Cause I found the sequence by which defenders were picked a bit weird decided to some reloads and watch what happened with some units swapped etc. At some point in the reloads the city was so well defended that the AI did not attack the city at all, instead went for assorted pillaging attempts. In other words: said Knight might be waiting for the right time to attack? Why did you not kill it btw? [lol]

What I learned during the process is that the comp always picks the best defense vs. the attacking unit, i.e. if a Melee unit is attacking a unit with very good abilities vs Melee can be choosen over a defending unit with higher overall strength.

Although arty type of units cause coll. damage I still firmly believe that a SoD type of diplo mission is the most effective in 4. ;)

[Pastorius]

Amybe so, but difference being less of a 10horsemen + 2 spears SOD (just a general example), and more of a mix, no?

Mix being a more complex promotion based one.

[Lt. Killer M]

quote:Originally posted by Kemal

The AI also sometimes waits to attack until it feels it can mount a decent assault at once, not spread out over several turns. Maybe this is what happened in your case too?


nope, there was just this one unit.

[Lt. Killer M]

quote:Originally posted by Matrix

Was it healing?


nope. healthy all the time.

[Lt. Killer M]

quote:Originally posted by Beam

In other words: said Knight might be waiting for the right time to attack?

but nothing changes - but, it seems, the RNG rolls.

[grahamiam]

quote:I can only conclude that the AI knows combat results will be 'good' or 'bad' and attakcs only when 'good'.

hmm, I had an AI Knight (2 promotions) attack an Elephant on a hill (Combat 1 (+10%)) last night and lost. So, I guess the AI doesn't cheat on every attack :) What was really weird, though, was he had 4 LB's and 3 pikes in a nearby town. These units never left the city, allowing my 2 Elephants to pillage the snot around all the nearby tiles and cutoff his horses (no more knights :) ), cows, and a free pass to his silks. Very wierd behaviour, as he had enough units to wipe out my 2 elephants. This was in the SG LoTR-17 last night.

[Pastorius]

Seing as this is mr Jsut's thread, and that he, back in the days where he used his proper nick over at the other forum, complained about the tank vs spearman, I thought this quote would be interesting.

Not that people here whine though [mischief]

quote:Posted by Sirian about the gunship in Spearman vs tank thread (http://forums.civfanatics.com/showthread.php?t=141050):
Players need to get past the graphics and "realism" and deal with the gameplay. The Gunship is doing exactly what it is meant to do. It's not good at attacking fortified cities, so I have no sympathy for complaints of "Longbow killed my Gunship!" Yeah? So what. A single human doesn't last for 6000 years. America was not founded in 4000BC. It doesn't take hundreds of years for even ancient sailing ships to sail around to the other side of a continent. IT'S A GAME, PEOPLE. Gameplay first, realism second.

[Lt. Killer M]

Paal, you're entirely correct, that's gameplay. But I measure Firaxis by their own claims, and it appears that - again - they fail.

Last time, at first nobody believed me that the AI knew the positions of my troops in the fog. Nobody believed me that it could be impossible to generate a leader with certain elites. And Firaxis certainly kept mum.

Guess what?
I was right on both counts.