Jolly Joker wrote:The biggest problem is that people want an AI that they can be proud of of having it beaten because it is so darn smart.
That would be nice but I agree with you that it would cost way too much in development time, development resources, and computing power.
I agree with JJ that in order to provide a REAL challenge to an experienced [not necessarily expert] player then the AI is going to need some cheats. What those cheats are, how they are structured, etc. makes all the difference in the world though ....
Seriously, let's say you want a really tight game: if the AI plays under the same rules, how can the game be tight? First thing is, you need different amount of starting money and resources.
I disagree. The first thing that you need is you have to let the AI see the entire map. This will [somewhat] counter the effect of players who already know where everything is because they've played the map before.
That, however, is not nearly enough: if you get through the first, well, 5 or so weeks, you'll STILL beat it easily.
I don't imagine this is in any way an admission that the way the AI bonuses are structured implies that it is grossely unfair to human players?
Back to the discussion I imagine that knowing the entire layout of the map isn't enough but it's a good start. The AI theoretically should know enough of the area to plan how to take initial mines better than humans can. This should give the AI a boost in resources because it can plan for exactly how many troops to take to which mines in what order to maximize castle builds, creature growth, etc.
So what? Do you want other production numbers for the AI? Like we had in H 2 and H 3? What?
What other production numbers are you referring to? Are you indicating that on blanced MP maps that the AI received automatic creature bonuses in H2 or H3? I don't recall this. If you could point out a source I would appreciate it.
Either way giving the AI bonus creature production would be the sort of thing that I could NEVER support as a general rule whether it's in H2 or H5. [It's ok if the map maker wants to do it but the game itself should NEVER do this by default]
Actually I find the current situation not bad: the AI seems to have as much money as it needs to build and buy.
I think that the structure of how it gets bonuses is terrible. Currently the AI has far too much of an advantage in the first 4-7 weeks of a game because it can afford to build up it's castle and buy out troops without any resource restrictions.
Taking away the initial bonus and giving the AI daily bonuses of either +X% of what it is able to make itself by resource & money gathering OR making sure the AI has Y% more resources than the human players would cut down on this inequity dramatically.
However, attacking needs going to the one attacked, so you are on the defense first.
Each and every game shouldn't have to put you, as a human player, in such a defensive mode. It would be nice if an offensive philosophy would be a viable strategy but alass it's not.
If the AI doesn't beat you in the time of material superiority for whatever the reason it goes on the defense. It will have the material it gets into town and you will be forced to fight a - sometimes - very tough end battle.
And no matter what you do there really isn't a way to "lay siege" to an AI because the resource bonuses always kick in allowing it to fully buy out everything in its castle. It would be nice if there was something that the humans could do [at this point of the game] to prevent this.
If you have more than one opponent the game may develop different based on whether a far off AI player can conquer other AI players. If that's the case the AI will ruthlessly attack sending you one army after another.
If this happens, esp. if the victorius AI doesn't lose tons of its troops then it's almost a "game over" right there. The AI can afford to buy out everything in multiple castles even though you have no hope of doing anything similar to this.
Reading this here you'd think here are two different opinions (except my own). One part is saying, well the AI is no challenge at all (due to Nival's philosophy of "losing graceful"), while the other part is saying, it's no fun because the AI has to massive advantages and you don't have something of a regular play in the beginning because of the overwhelming advantage.
In a very odd sort of way that is exactly correct. The AI itself seems pretty feeble and must be propped up by the massive bonuses that it gets. Because the AI doesn't stand much of a chance otherwise it must act like a school bully and [while it's bigger than you] make sure that it takes your lunch money.
