Any need on inventing an Impossible AI?

[Stratego (dev)]

I had feedback saying that the AI is not hard enough.
In the dev. version the AI got better with the ability to build stable/barrack/arhery ...

But i have here a few more ideas to get the AI better than the humans (or at least catch up to the human):
Idea 1: Impossible AI would have not only the default techs but the "blacksmith" and "adv. center" techs also
Idea 2: Impossible AI would have a good chance to build unit faster than the human player, like:
a) every turn 30% chance to decrease the build time not by 1 but 2turns!
b) OR, every turn a 20% chance to build up the unit immediately (even if that is a catapult, if the AI has the luck than it can produce it in 1 turn!)

what do you think? or have other suggestion?

[Longinus]

Interesting... it would be good for late-game players.. especially Idea2-b... I would suggest that to make it more impossible all HP of enemy units and structures are added by 2-5% more..

[kon]

hmm... catapult the bots are already rapidly being built

[Pepe]

Any idea on how to make it so hard without "tricks" or "bonuses". How the bot "mindset" works? The best would be to make it learn from human or to be powerful enough without the help of these bonuses. I am not a programmer so can't help making the bot hard to defeat! how can you change the bot behavior?

[Stratego (dev)]

Currently it is a simple logic, based on this article: http://www.gamasutra.com/view/feature/1 ... hp?print=1

And it is almost hitting the resources of the phones (i am profiling it already - hard to improve its speed)

So any learning algorithm direction is not the good way, and also takes years to implement .

But i am open to any suggestion that fit into the current logic.

[Stratego (dev)]

any thoughts?

[anvil]

here's a few thoughts.

the first that pops to mind is have 3 or so difficulty levels incorporating your ideas.

second would be to buy with gems your ideas either as a per game purchase, or select to use per game once purchased.

third would be to have simple ( ) "AI" builder where you have sliders to increase the % you list above.

I would really like the choice to up the AI or not.

[Pepe]

This won't make the AI impossible to win as it is just a small tip. Missionary and priest should be trying to convert the strongest unit. I see in many games they try to convert the weakest or any aleatory unit instead of the most powerful (and of course, "convertible").

[Stratego (dev)]

actually priests are not convert intentionally - so they do not "attack" or even approach the enemy in order to convert, but if any enemy accidentally in range, they will try to convert.