Valve Software's Doug Lombardi

Left 4 Dead and more under discussion

AI is a key element. How is your use of this different to other titles?

Because this is a mix of singleplayer and multiplayer, you sort of have to have this new way of authoring it. First-person games are usually scripted, 'this monster performs in this play so is placed here', 'when this room is triggered he does this', and there's a little bit of randomisation to make it feel less static. but in this case you've got four people playing; we want it to be highly re-playable.

The monsters have there own behaviour, then we have this thing called the AI Director. This is the interesting twist. We're taking what you'd called standard AI; monsters who act within certain rules, knowing where they can go, what they can do, and then there's the notion of the director. This will keep it dynamic, keep an eye on pacing (from Half-Life we learned of battle fatigue, to avoid sensory overload), so we schedule this in the game. We add humour, we scare you. Peaks and valleys. Finally, there's the idea of dynamic difficulty. The game goes up a ramp depending on the difficulty you've chosen in a normal singleplayer game. However, becuase this is multiplayer as well we're letting the AI alter how overwhelmed players become as the game progresses. So if you choose easy, the AI might start making the game harder anyway as you get better.

Given the co-op focus, is story important to the experience?

Only in so much as giving you context. Making you feel a part of the experience. For us the most interesting stories will be told after the game has been played. The re-call of what happened. What you were thinking.

We want players to govern their own experience, so we help them out with context. The basics, you know, 'get to the hospital where there's a helicopter, oh but there's some zombies in your way'.

Is the game mission-based?

We're calling it campaign-based, we're sort of stealing the RTS term there. We've strung together five really big maps and call it a campaign. Take the hospital level shown here as an example, you actually start in an apartment building, and you've got to make your way down and through the streets, to get to the roof of the hospital, that takes place over five huge maps.

Are the maps mainly urban?

No, earlier we were demonstrating a level outdoors, in the cornfields, where you have to make to a farm house at the end. And there are zombies rushing out of the corn, and its amazing how creepy that is, just naturally. There's two urban and two more rural.

What's your favourite instance in the game?

A lot of stuff is being done proceeduarlly, and there will also be a few scripted moments, so there's big crescendos, etc. For me, the best part is always when you've caught up and you think you're at the end of the level, and then something goes wrong and you're back in it again. That moment of excitement, when something swings from nowhere. That false sense of security. Just earlier there was a moment like that, and all four players went 'whoa, holy shit!'. That's the moment.

Whenever you play CoD 4, Counter-Strike or whatever, when you're playing with a good team, and its going well. When you've ground out that tough victory and everyone throws their hands up, that's the moments we're looking for. That's the big idea. When I hear that from play-testers I know we're doing something right.

Is polishing and attention to detail more important than an incredibly ambitious or complicated design?

I think sometimes people try to complicate. Portal is a good example of that. There was a whole bunch of other things we thought we should do with Portal. Art being one of them. They were high contrast, what have you. Then we tested it and people didn't know what they were doing. Don't confuse the player, that's the key, don't let them struggle to find out what to do - challenge them to do what they know they need to do.

A game shouldn't be about finding what needs to be done, it should be about accomplishing that. We try to be ambitious in the design, but through playtesting we try to be honest, strip-back as necessary. Lots of folks tell us Portal was short, but people were so glad they played to the end. That to us was another 'we did something right' moment. We're not creating half a game with the hope that people will make it to that point!

Monitoring difficulty is part of that, not getting too elabourate with the art is part of that, because people will get distracted. If everything is beautful in this room I don't know what to look at, give me a shiny object and four grey walls, I'm going to look at the shiny object.

The game is coming out on PC and Xbox 360. Are there any differences between the platforms?

Not really, no. The multiplayer options work slightly differently, obviously, but once you're inside the game, the only diferences relate to the controls.

Would it have been possible to make Left 4 Dead for the PC only?

Oh sure.

Some people seem concerned by the state of the PC platform as a cutting-edge format. Do you worry its becoming the preserve of MMOs?

Look at what the Battlefield guys are doing. People get too obsessed with NPD's US sales numbers. It says that 12% of the games sold last year in the USA were on the PC. Well, it says 12% at retail in the USA. Make the complete statement. Look around. The money being made on Steam, and World of Warcraft. Probably more tham 12%!

There's other places to make money - advertising, cybercafes. In Asia micro-transactions are huge. We're barely scratching the surface. Go to the Korean Game Show. Its all about that. They're making tonnes of money. That model might not work here, but its impossible to say there's no money in PC gaming.

So a new approach is needed?

Yes. Game makers need to get out of the 'make a 15 hour game and sell it in a box at the store' mentality. Then the PC doesn't suck any more. Of course, the platform holders spend millions on marketing. No one pushes the PC in this way.

Would Valve ever go micro-transactional?

If we ever get a game that fits that, we'd love to do that. I'd love to have a game that we gave away saying 'however much you want to invest in this, this is yours'.

Would Counter-Strike work with this?

We'll see how DICE do! We tend to take lot of risks on a lot of new ground... so that's one area where we're going to let someone else take the risk.

When is Left 4 Dead out?

We're aiming for the first week of November on PC and Xbox 360.

Thank you very much.

By Luke Guttridge