I'd start at the most recent post (bottom of page). Interesting theory. What do you all think?

 

"The days of the single-player game are numbered. "

 

http://groups.google.com/group/comp.sys.ib...0700963145fdbc#

I think he's dead wrong. AI is only going to get better as computers improve, and good strategy games take TOO LONG to play - there will always be a lot of trouble finding a live opponent who can give you enough time on your schedule. Also, far too many people in online games behave like complete jerks. As the AIs get better, that's going to be a serious issue, if it isn't already ...

I think he's dead wrong. AI is only going to get better as computers improve, and good strategy games take TOO LONG to play - there will always be a lot of trouble finding a live opponent who can give you enough time on your schedule. Also, far too many people in online games behave like complete jerks. As the AIs get better, that's going to be a serious issue, if it isn't already ...

I think the same as your, Warhorse.

I think he's dead wrong. AI is only going to get better as computers improve, and good strategy games take TOO LONG to play - there will always be a lot of trouble finding a live opponent who can give you enough time on your schedule. Also, far too many people in online games behave like complete jerks. As the AIs get better, that's going to be a serious issue, if it isn't already ...

 

Yep, I tend to agree. The argument might have some merit for the mainstream games, FPS and the like, particularly the ones that might appeal to the younger segment of the population. But I don't think strategy based wargaming ever fell into that category.

I like this discussion!

 

There are some truly capable AI designers out there. Most breakthroughs in AI seem to come in the form of doctoral thesis projects and the engines they produce are inevitably limited in scope and left to wither on the vine once the new doctor heads out into the world.

 

I may just have to agree with Eddy with the modification that 'telepresence' may be a better choice for the next 5-10 years while capable generic AI engines continue to mature. Small developers making wargames don't seem to have the budget and in-house talent to craft an AI appropriate for a game like Harpoon. Games with less variability (a few set battles) have AIs that are extremely capable, even chains of command, but an AI generic yet capable enough for Harpoon seems out of reach. Then there is the whole battle of making that AI useable by scenario designers (i.e. which dials can I present to the scenario designer such that he isn't overwhelmed yet still has the control he wants).

 

Let's all hope I'm proven wrong and a whole slew of grad students crafts a complete human replacement and easy interfaces!