Brad Wardell Interview, Part 1

[DanQ]

Brad Wardell is the Project Lead on Galactic Civilizations, set for release February 25th of this year. He is also the Founder and President of Stardock Systems, the company that has employed a dozen full-time staff for the past one-and-a-half years developing the follow-up to its original game release of the same name in 1994. This single-player, turn-based strategy game set in outer space is being published by Montreal-based Strategy First.

In the first of a three-part interview with Apolyton CS, Wardell talks about the exclusion of multiplayer components, artificial intelligence "cheating" and comparing the current GalCiv with its predecessor. A part of this one-on-one dialogue will be published every Monday up until and including February 10th. GalCiv becomes commercially available on February 25th.

Comments are welcome in this thread.

--------------
Dan; Apolyton CS

[FrustratedPoet]

Very interesting.

I hadn't been following GalCiv at all, but it interests me and I'll definitely keep an eye out for the next part of the interview.

This will please many people:

I think the key to a good game AI is for those players to play by the same rules you do. They don't get free money (except at the most extreme levels), they don't just get ships given to them. When I play a strategy game and I take out some crucial building, I want to be assured that I've really hurt that player. Often times I'll play a strategy game and take out a crucial unit or building and not be sure if it really mattered, that the AI won't just conjure up another. In GalCiv, they have to build everything the same way you do.

[moomin]

Sure will. I'm really looking forward to this game a lot more than to Moo3 at this point - because I know it was designed for delivering a superior SP experience.

Oh, and Dan! I'm pleased!

[Asmodean]

Yeah...the multithreading is definately one of GalCiv's strenghts. This AI can give you a run for your money.

Asmodean

[haz]

no multiplayer

I just wanted to log in my absolute support for Mr. Wardell's emphasis on a superlative single-player experience, and specifically, a concentration on improving game AI. I think Play the World has demonstrated how difficult it is to create a satisfying and workable multiplayer experience in complex TBS games like Civ and GalCiv.

I always go through the same process with new TBS games -- I love them at first, find them relatively challenging, and indulge in the strategic options. However, without exception, after a couple months of gameplay, the AI seems pathetic and predictable. I then consider turning to online opponents (if the game supports multiplay), and after a brief horrendous experience with some a-hole who cheats, or (more often) someone who just throws you the cheeseball strategy they picked up from Apolyton (:0), I give up.

The greatest, and most difficult task, facing the next generation of game developers will be figuring out how to utilize the increasing processor speeds of PC's to create more "human-like" AI. I can appreciate the difficulty of this, since how humans think, weigh variables, and decide upon strategies in their everyday behavior (much less, PC game behavior) is something that psychologists and neuroscientists probably won't completely understand for another couple 100 years. In terms of games, the more variables you input, the more fun the game is -- but the more difficult it becomes for the AI to juggle them effectively and coordinate an overall strategy. GalCiv seems to be one of the first games to confront this dilemma head-on, and I have nothing but respect and admiration for the designers and programmers. I look forward to seeing how it turns out.

[darcy]

Good interview.

I don't get Brad Wardell's point about "cheese tactics" though. He makes it sound as if such unbalanced maneuvers are alright to have in single player.

Sure, multiplayer exposes them faster, but if the AIs are really so smart, shouldn't they exploit "cheese tactics" too if they can? Players sure will, even single players.

[Draginol]

There's a big difference. In on-line games, cheese tactics get the network effect. In single player play, most people won't experience cheese tactics. At worst, they'll find some fan site that will list cheese tactics but the player themselves won't see them.

TA, for instance, had com-napping. Warcraft III has hero rushing. Sure, the people who like cheese tactics think this is fine but the rest of us just want to play the game in the way it was designed.

[Asmodean]

One of the best cheese tactics I can remember was the "Railroad across ocean" exploit in Civ1.

That particular game didn't have MP, of course, but it was still a crazy tactic.

Asmodean

[Harry Seldon]

I always enjoyed the 'auto-reload' in MOO, where you'd get chased away after tying or losing a battle and immediately send them back into orbit around the planet so they'll battle the next round with full weapons.

This game looks awesome. I can't wait to try the AI, and if the metaverse works as promised, this should have an amazing replay value.

[moomin]

As an old OS/2 GalCiv veteran, I'm a bit saddened to read about how the new version deviates from that immortal classic, but I have faith in Stardock's vision.