AI behaviour and issue tracker

[LevelX]

Over the weekend I played many matches against the three AI opponents.
It's clear to me, that they can't be as good as a good human magic player.
But they are doing so many stupid things that I wonder if it's really useful to add my observations to the issue tracker.

I think the developers are aware of this big weaknesses and it would be wasted time to document them all to the issue system. Maybe there is already planned a big redesign, that solves many of the problems.

Or is it appreciated / useful to add the different observations to the issue tracking system?

Some examples of my observations:

The AIs use equipments a lot. They waste mostly all there mana to re-equip equipments to the creature the equipments is already assigned to. They do this with a full hand of other useful cards.

The AIs like to cast the Cage of Hand enchantment and to take it back to hand directly to recast it again and so on. Again, the hand is full of other useful cards.

They attack with a weak creature also if there is a much bigger creature that can block it. No effects exist, the AI doesn't cast anything. They do this very often.

They don't block an attacking creature they can block although they die if the creature won't be blocked.

They cast the nasty enchantments on there own creatures as also the good enchantments on the opponent creatures. And there was no genius move from the AI after to explain the behaviour.

Damage prevention effects will be used randomly and mostly on the opponent creatures.

The minimax AI tends to need very long time to decide what to to as the game lasts longer.

The new Monte Carlo AI doesn't show the "waiting for computer" message and nothing happens if you confirm the "Done" button as long as the AI works in the background.

[nantuko84]

I'd like to say that I created test framework for Mage a month ago, so I started created tests not only for cards and effects, but for AI behavior as well. This way it is much easier to fix these issues, at least because of them being reproducible.

So I would say - yes, we need to write these bugs as a part of the application, (please add AI tag while creating issue on google code issue page). If we don't add them as issues, then no one would fix them.