Notes on Half-Life 2

But first, some notes on Half-Life.

Half-Life was not a revolution.  Some people talk about the importance of innovation, how games like GTA and The Sims where so succesful because they came out of nowhere and were entirely new, but Half-Life showed that it is possible to create quite a splash simply by taking a popular genre to the next level.

Half-Life came out less than a year after Quake 2, and was very similar to it:  a linear funhouse walkthrough in a science fiction setting with monsters and the like.  Quake 2 was no slouch of a game, but somehow Half-Life ended up making history and Quake 2 is only dimly remembered. What made Half-Life such a classic?  Both Half-Life and Quake 2 had continuous worlds;  both Half-Life and Quake 2 had story;  both had the best enemy AI to date.

One thing Half-Life is that Quake 2 is not is a triumph of simulationism.  Half-Life tries its darndest to answer the question, "If we accidentally opened a dimension for alien creatures, what would it really be like?"  Many aspects of the game support the simulation:  the guns are real-world guns;  powerups always have a reason for being where they are;  the enemy AI is incredible - as if you were facing a real team of coordinated soldiers instead of zombielike monsters that just rush you en masse;  and there's a storyline that's internally consistent.

About the only concession Half-life makes to gamism over simulationism is Gordon Freeman is very hard to kill.  Over the course of the game he probably takes around a thousand bullets.

Another thing Half-Life did was immerse you in the story instead of alternating story and gameplay, Ms. Pac-Man style.  Not everyone has played Half-Life all the way through;  but everyone who played it at all remembers the opening sequence - the trainride that gives you an overview of Black Mesa, followed by the bang-up, James Bond beginning with the exploding equipment and the invasion of the creatures, with you in control the whole time.  When a scientist is snatched by creatures right in front of your eyes, you have to ask yourself:  "What if I got here later?  Would I have missed this event?"  The answer, I'm pretty sure, is no:  the game triggers the cutscene when you're close enough.  Still, the illusion is convincing.

We've experimented some at Treyarch with our third-person games, trying to eschew cutscenes in the same way that Half-Life does, and it is an incredibly difficult task.  It seems that the player is always facing the wrong way, or the critical events are too small on the screen, or the player will be able to interfere with the events of the in-game-cutscene in some unexpected way, and we always give up and revert to the tried-and-true cutscene.  The Half-Life aesthetic requires painstaking work.

Obviously, if having a story was all it took to make something good, then all movies would be the same.  All stories are not created equal.  Half-Life has a story that is more interesting and believable than Quake 2 without becoming the didactic overwrought melodrama that is Metal Gear Solid.  There are multiple factions:  the aliens, the man in black, the Black Mesa scientists and guards, the opposing force.  How they interact is interesting to unravel.  The story is internally consistent.  The story gives us a character that is a little easier for videogame geeks to identify with, I think:  Gordon Freeman, the scientist who, in a pinch, can grab a gun and kick ass.  A nice change from the musclebound thugs typical of video games.  And the story purposely leaves things unexplained, left out:  you feel like there's a rich backstory, a world, which you only see a slice of.  The tip of the iceberg.  So Half-Life is a triumph of narrativism as well as simulationism.  (http://www.indie-rpgs.com/articles/1/ for gamism-narrativism-simulationism)

And, of course, it's a fun game as well:  the enemy AI provides not only realism but also interesting choices, and the level design, which has a habit of showing you a goal and then putting some obstacle in your way that you have to navigate around, is addictive.

So Half-Life represents a no-compromise triumph;  no matter what style of game you prefer, Half-Life shouldn't alienate you.  Perhaps their consensus-building cabal approach helped acheive this.

So what does Half-Life 2 bring to the table, beyond better graphics and the best facial animation system ever seen in a computer game? 

One thing it doesn't do is mess with the Half-Life formula:  it must have been tempting to give Gordon Freeman a voice, but they resisted the temptation. The thing I noticed most about Half-Life 2 is that it's incredibly toyetic.  "Toyetic" was a word given to me by a friend who used to work at Mattel who doesn't like being mentioned in my blog.  It means, "like a toy."  An amusing sidenote is that the guys at Mattel are trying to make their toys more like computer games, while we're trying to make our computer games more like toys.  Or toy chests, anyway.  The collection of guns in your typical FPS are already toyetic;  a set of toy guns, each with their own kind of play.  Half-Life 2 gives us a bunch of new toys above and beyond the usual collection of weapons:  the air raft, the gravity gun, the dune buggy, the ant lions, portable gun turrets, squads of soldiers.  Each toy comes complete with a context to make it interesting, and makes Half-life 2 feel like a brand new game, not a rehash.

Once upon a time there was a game called Trespasser that boasted the most realistic physics yet seen in a computer game.  The game was universally reviled:  "a crate stacking simulation" it was called.  After its release, physics was avoided like the plague by game companies.  It wasn't until recently that Havok made physics desirable again.  Although Max Payne, Thief, and Deus Ex all have realistic physics, the physics in those games seems mostly like a cosmetic add-on.  Half-Life 2 is the first game to make physics a part of the gameplay, with the creation of (admittedly contrived) physics puzzles.  These puzzles are at the perfect level of difficulty for a mass-market computer game:  difficult enough to be nontrivial and make you feel somewhat clever for solving them, but not so difficult that many people are likely to shelf-level-event on them.  I never had to consult gamefaqs to complete it.

The story in Half-Life 2 is even better than the story in Half-Life.  A dystopian future in the aftermath of the Black Mesa incident, it fills in some of the details missing from Half-Life but still leaves us hungry for more information.  The way the Combine uses missiles loaded with headcrabs to pacify the populace of the city feels like something Larry Niven would write.

A rule they had in the making of Half-Life was they weren't allowed to develop technology unless it was going to be used in the game at least twice.  I'm not sure, but it seems like they increased that number to three for Half-Life 2, which does help make the game a little longer.  When you have quicksave, it does seem necessary that you reuse the same gameplay a few times.

The biggest complaint people have with Half-Life is the ending, where you go to Xen and do jumping puzzles until you get to the final boss fight.  It was frustrating, the rules of the game seemed to change, it wasn't the kind of gameplay people had been enjoying so far.  About all you could say for it was it made for a nice change after a lot of shooting. Half-Life 2 did not make the same mistake.  There's an unwritten rule that the final level and final boss have to be the most difficult part of your game.  Lately a school of thought has sprung up that this is not what people want.  They actually want an easy but spectacular ending.  Half-Life 2 tests the theory, by giving you an incredibly powerful gun that makes the final level and boss fight pretty much a cakewalk, but a spectacular cakewalk, as you're kicking so much ass.  So it's not one of the many games that I played all the way to the final boss and then quit.

Half-Life 2 almost never felt like a slog.  Almost every other FPS I've played have sequences or levels which seems like more of the same old thing, a drag.  In general, with Half-Life 2, when I got tired of a sequence, I'd get to the end, and a new kind of gameplay would be introduced.  I wonder if they cut sections that were deemed too boring, or if they've found the magic formula for how to keep a game interesting.

I feel like I need some paragraph to sum up...but since the Half-Life phenomena seems like a lot of subtle aspects that work together to form a whole greater than the sum of its parts, I'm not sure what to say.  And I'm hungry now.  Must eat.

Comments

[Factory]

"Half-Life tries its darndest to answer the question, "If we accidentally opened a dimension for alien creatures, what would it really be like?""
While I generally agree with this statement, I would also say:

"... the guns are real-world guns;..."
Well the guns are real-world-like, but only the pistols and the shotgun are copies of real world counterparts.

"... powerups always have a reason for being where they are..."
Erm, Half-life suffered from goodies-in-crate-o-philia as much as any game.

--

"The answer, I'm pretty sure, is no: the game triggers the cutscene when you're close enough. Still, the illusion is convincing."
To maintain that illusion it's vital that the player not see the scene more than once, the mechanisism gets revealed then. HL2 avoids this by putting lots of autosave points in. (prolly they put them in just after any scripted story event, I haven't checked)

--

"Once upon a time there was a game called Trespasser that boasted the most realistic physics yet seen in a computer game. The game was universally reviled: "a crate stacking simulation" it was called. After its release, physics was avoided like the plague by game companies. It wasn't until recently that Havok made physics desirable again."
Hmm are you sure that it was Havok that was pushing physics? I would have said that extending ragdoll physics to the rest of the game, was more of a driving force. Certainly all the interest in physics predates Havok (at least the usage thereof).

--

"Half-Life 2 is the first game to make physics a part of the gameplay, with the creation of (admittedly contrived) physics puzzles."
Hmm I'm not too sure that HL2 was the first game since Tresspasser to do physics in gameplay, although I cannot think of any off the top of my head. It's almost certainly the first FPS since Tresspasser to do so though.

--

"Half-Life 2 tests the theory, by giving you an incredibly powerful gun that makes the final level and boss fight pretty much a cakewalk,"
Hmm powerful, but it's a 'wierd' weapon, which I think detracts from the climacticness.

--

Oh and let us not forget the rather bad Team mate AI, which is ok as long as they do not try to go into buildings, where the flaws are revealed. IMHO dropping those parts of the city section might have been a good idea, or perhaps making them more open (limiting combat to the streets perhaps?).

[William van der Sterren]

Jamie,

(as an AI developer) I don't think your statement "both [Quake2 and Half-Life] had the best enemy AI to date" does justice to the Half-Life's AI, and to the big step forward in game AI it represents.
Compared to Quake2, Half-Life added:
- friendly NPC's who successfully followed the player around, added to the atmosphere with their comments, and provided some assistance
- (the illusion of) squad behavior, with grunts (seemingly) cooperating with another to suppress, flank or throw grenades
- "idle behavior": most NPC's are doing something fitting and believable in the game world when the player first sees them
- replayability, because much of the AI (individual and squad) behavior really responded to the player's actions
- a seemless mix of scripted and autonomous AI

This made the Half-Life AI a benchmark for FPS game AI, and in the following years many games have struggled just to match Half-Life's AI.

William

[Dave Johnston]

William,

Regarding the AI, it's probably important to note that since HL1 came out, there have been lots of games with far far more powerful AI. The thing was that in HL1, the AI that was there was good, and fun. It reacted and performed in a way that made playing the game fun. I think too many people picked up on how good the AI was and tried to 'beat' it by making 'cleverer' AI.

HL1's AI wasn't great because it was clever. It was great because it was clever enough, and worked both for and against the player. It could trick the player occasionally, but the user could 'trick' it. It was just right. Lots of games consider AI as something for the player to beat, when in fact, the AI should give the user a challenge and work with the user to make the challenge fun.

Good AI in games is more about the combination of good AI and good game logic than smartness and intelligence.

[quanta]

A great article! You're right, HL1 excelled because it tried to simulate reality as much as possible. I remember in the first level you could walk into a public washroom and use the sink, turn on the hand dryer, even flush toilets. Pickup items weren't animated, hovering icons. The puzzles were always logical (as opposed to hitting random switches, humping walls to find hidden rooms and grabbing keycards). And imagine, a rocket launcher that carries a limited amount of rockets and needs to be reloaded!

HL2 upheld the tradition of simulation, but in the process doesn't feel as revolutionary compared to today's games.

[Terpfen]

I agree generally, but I just had to say one thing:

Metal Gear Solid isn't overwrought drama because it doesn't take itself serously.

[Tom]

Interesting to see that you only mention good things about HL2; I believe that it has significant game design flaws as well and so it isn't The Game of the year, at least IMHO... a few examples:

Weapons are very limited in their numbers (the combine rifle is the only new gun, and all the alien/energy weapons are taken away); and they aren't properly balanced. The magnum is overpowered for both close range and as a sniper gun. The capacity of the rocket launcher is far too limited with 3 rockets and thus the player is forced to leave cover and search for ammo during the gunship/strider fights far too often. Forcing the player to use the gravity gun only in Ravenholm can be quite frustrating for those who don't like that weapon.

The 'goodies in crates' sickness is very disturbing by now. Many sequences are almost impossible to complete without health loss, and thus the player will always come upon a large supply of medpacks/shields after getting out of these areas. This feels very artifical and takes away from the immersion of the game, especially because the player also has to break the crates somehow to get the goodies.

I have to mention Halo as a game that managed to invent new and better solutions for these typical FPS game problems. Valve should've copied the recharging energy shield concept, and the ammo consumption/resupply balance (especially for the rocket launcher). And I consider Halo's AI (both enemy and friendly) to be better than what I've seen in HL2. The zombies and other monsters were OK to be dumb, but the soldiers seemed to be stupid, too - probably even worse than their HL1 predecessors.

I also didn't like the total linearity, complete with invisible walls and invulnerable characters, but I can accept it as a possible direction for gameplay, considering the increasing demands of asset creation. But I'd still rather have more games of the Thief/Deus Ex/GTA kind, where you aren't traveling on rails...

[SpiderMonkey]

"The capacity of the rocket launcher is far too limited with 3 rockets and thus the player is forced to leave cover and search for ammo during the gunship/strider fights far too often."

Um, isn't that the point? You get a lot more exciting moments out of the desperate dash to a rocket crate than you would if you let players just sit in a corner and hide. I really liked the alternation between fight and flight - a theme that persisted throughout the game.

The other weapons also work well because they are orthogonally differentiated and so each have their own uses. In how many other games would you still be using the SMG right up until the end?

Ravenholm exists to teach players to use the gravity gun, so that seems fine to me. Goodie-holding crates were always very clearly marked - a big step up on the original.

I had no problem with the health-kit-abundance. At the end of the day, Valve is aiming to provide an experience, not a challenge as such, if you get what I mean. The objective isn't to stop the player progressing but to make them feel what you want to feel as they move through the game.

I think I sound like a Valve-fanboy having written that. Don't get me wrong, there are things I don't like about the game, but I don't think there's anything bad about those issues.

[Hutz]

I disagree about Half-Life 2's final boss being good. In fact, I also disliked the first game's final boss greatly.

I'm going to generalize even further. I haven't yet played a PC first-person shooter that had an enjoyable final boss. Wolfenstein, Doom I and II, Quake I, and Half-Life 1 have bosses that are either mundane puzzles or just bigger, stronger regular enemies. And Half-Life 2 had a final "challenge" that was far too easy, then they gave you an ending that was a total cop-out.

My favourite FPS final boss? 006 from the N64 Goldeneye 007. It was frantic, challenging, and not a puzzle; not too mindless (it did require skill) yet not a total mind-boggler. More FPS's need a final boss that has that much excitment and satisfaction.

[SpiderMonkey]

The boss (queen alien) at the end of the Marine campaign in AvP one was quite entertaining.

Anyone who didn't play it:
You can't kill her. She stomps and charges at you and also hurls boxes at you from around the environment. You have to complete a sequence of events around the cargo bay it's set in, in order to flush her out into space, but not get sucked out yourself.

[Raphael]

Jamie...I think your very last comment in the article sums up why HL2 (and HL) were such a success. It feels like Valve was able to strip away all of the extraneous shite that typically keeps most good games from being great. I think this is akin to the editing process in films or writing. I don't know anything about the actual development of HL2, but I'd guess that Valve edited a lot of content to end up with the polished, shiny gem they ended up with.

Great article!

[Kevin Mandeville]

Great site. I really enjoyed reading your comments about Half-life and HL 2. Although, I got a bit peeved when you started discussing the ending of the game. I quickly stopped reading so I wouldn't spoil it for myself. I haven't completed the game yet! Maybe include something in the beginning of your posts that says you include spoiler information. Otherwise, I like the site, and I'll be adding it to my blogroll, and my site feed list. I'll be back.

[jim]

the best final boss i ever met, ever, was at the end of 'contra' for the super nintendo, seriously, get one, and the game (costs less than the gold package of hl2, lol) and see what i mean.

[kp]

Yea this game was great, alot easier than Half Life, but far more better graphics and story line, i really wish there was a better boss but the end does have a good lead into a possible half life 3, altogether i love both of em cant wait till the next one