TEAM FORTRESS 2 AFTERTHOUGHTS


GFW:Are you ready to talk about Team Fortress 2's two false starts at long last?

ROBIN WALKER, COCREATOR: We're not really ready to talk about those because we want to release a Raising the Bar.style book.

GFW: Once upon a time, Team Fortress 2 looked like what Battlefield became. You later went the other way, opting for simplicity over the supercomplexity of today's top team-based shooters.

RW: What we found as we worked on TF2 over the years and went through all these iterations is that depth doesn't demand increased complexity. TF2 is a sucker punch -- it only looks simple. We're already seeing large-scale skill differences. Right now, we know the nuances. When the Shack group, who are incredibly hardcore FPS gamers, challenged our team, we totally crushed the living crap out of them. Strategy is about balancing decisions against constraints. Your team has 12 guys, and you ask, "What are the best classes we can choose for these conditions?"

We started to see what the true core of our game was and stripped out stuff that wasn't needed. You can always point to any feature and say it's fun. The question is, what are the negatives associated with it? Grenades are the obvious example.

When we decided to try Team Fortress without these things, the team assumed, "Oh, god -- there's no way we can get on without these!" Then we playtested, and all agreed: The game is actually better. And there's certainly a threshold you can cross over where too much complexity prevents people from appreciating the depth. Look at chess, rock-paper-scissors...



GFW: The Sims creator Will Wright loves Go [an ancient and strategically complex game with simple rules].

RW: Yeah, an insanely simple mechanic. You each take turns putting a single piece onto this table, but people spend centuries writing strategy analyses. It literally took us years to get to the point where we were able to do that as designers, to even realize it was right.

GFW: This is a radically different perspective than that of the modmakers you once were -- where with modders, the idea is frequently to find as many weird and new approaches as possible. Times have changed. Can you imagine making Team Fortress Classic for consoles?

RW: The console thing's interesting because internally ... well, we didn't even think about it. Not as much as much as we should have. The sort of streamlining and reductionist design had nothing to do with the game going out on consoles. We don't think that there's any real difference between what console and PC players find fun. The interfaces differ, though not the gamers themselves, not enough to where we should be thinking about that when we're designing. And the other thing ... it's a really interesting observation to say that TF2 is counter to the modmaker philosophy. Modmakers have a set of advantages: They can aim for a niche audience in that if you're going to make a mod for Half-Life 2, your audience is already defined for you.

GFW: Did you design critical hits to provide a degree of access to people who don't traditionally play shooters?

RW: A lot of people assume that they're there to help new players. Some of it's pacing. We looked at Counter-Strike, where you happen to get lucky at times and get five consecutive headshots. And then you talk about it for another three weeks.

GFW: Is that luck or skill?

RW: It's skill affecting luck. You can't control the Gaussian distribution of fire when you pull the trigger, but skill is involved in centering that Gaussian distribution on an opponent's head. The idea that there shouldn't be any luck in an FPS is silly.

We created Team Fortress 2's critical-hit system so that critical-hit chances increase over time based on performance. There's the skill. If you're a good player, you'll always have a higher chance of hitting criticals than novices. On top of that, there's the flat-out fact that crit chance is determined when you fire, right? You can miss crits.

Adding randomness increases the range of possible tactical skill options. Sometimes, the crazy thing works. If a game is 100 percent deterministic, then only one tactic works. The more randomness you add, the wider the range of options becomes. When your game is 100 percent random, every option is equally viable, right, which we don't want either. So the critical-hit system was us saying we want to specifically tune this stuff, whereas it's just fallen into place in the past.



GFW: One side effect I see is similar to gambler's logic. I call them old nerds' tales. For example, someone insists the likelihood of hitting criticals increases when he's outnumbered. Or that broken Demoman bottles do more damage.

RW: Yeah, people are generally terrible at establishing cause and effect where chance is involved. I've seen someone say, "I shot three crits after taunting. Taunting must cause crits."

GFW: How to play slot machines....

RW: Take your lucky coin and turn three times. We couldn't curb this if we tried. And it's a way of coping in cases where an opponent wins.

GFW: And that's the other side effect: blaming losses on luck.

RW: That was intentional. One way hardcore players salvage their egos is by saying that the other guys got lucky. Their egos are very much tied into how much fun they're having, and so the system explicitly lets them feel like they weren't beaten fairly. And the reality is, if not crits, they'd find other reasons, right?

GFW: At one point, you planned to sync players' VOIP [voice over IP] input to the characters' faces...

RW: Yeah, we talked about it a bunch. We got more mileage out of facial states -- like characters looking afraid when their health is low.

GFW: Is this an area where you can continue to add to the game?

RW: Funny you should mention that. We're actually bringing all the voice actors back. You're not just going to get more maps from us; our plan is to keep pushing everything. More dialogue -- you'll hear really, really rare lines sometimes -- and more reactions, like characters encouraging teammates to come stand on the damn capture points already.

GFW: Or yelling at the Medic who only has eyes for Heavies. Are you happy with the predominantly Medic-Heavy pairings you see in public servers?

RW: Heavies have a few significant weaknesses, specifically to Snipers. Plus, Spies are pretty dangerous to Medic-Heavy combos. Internally here, Snipers shut down Heavy pushes. We started seeing more invulnerable Demos and Soldiers. In general, you want to threaten them into using their invulnerability. And then you pull back [and] let them push. When their invuln is over, you kick the s*** out of them. One problem we see is people still shooting invulnerable [enemies].

GFW: And that's actually another old nerds' tale, that you can decrease the duration of their invulnerability with concentrated fire.

RW: Really! I hadn't seen that one. I mean, as a Soldier or a Demo, I'll sometimes still keep shooting the invuln, but I'm trying to bounce them because if you separate them the Medic's beam breaks. If Medics aren't starting to realize the value of assisting other classes, though, we'll think about tuning something. Right now, übercharge buildup is based on the health of the target. In other words, if you're healing a Heavy and there are three other guys around you who all have full health, you'll get no bonus for switching. We'll probably start rewarding you specifically for switching targets -- encourage you to spread that love around.