October 13, 2008

What Happened

Imagine a game that begins strongly, propelled by ambition and purpose, and in the final act totally deteriorates. That's not interesting to me anymore, even as a process story. It happens with such frequency now, it's way too familiar. The entire project, probably, was inspired by a single concept, but as a game it needed levels and plot, and as the designers strayed from that initial flash of brilliance to build the basic structure necessary to support it, they lost interest and it turned out workmanlike. Or maybe they worked on it for too long, got tired, stopped trying. These games run out of steam, time or money and bashfully crawl over the finish line or don't finish at all.

Games try and open with their best foot forward and shove the mechanically weaker stuff at the end. Bioshock's failing is a fundamental disagreement between story and game, but it still ends on its worst two levels: an escort mission and an overly conventional boss fight. Crysis dramatically narrowed its last-act focus to forbid the freeform strategy the game previously encouraged. Far Cry and the original Half-Life are classic examples: aliens showed up and abducted the quality.

I wonder if the opposite is possible: bad start, outstanding finish. Not like a slow burn, but like the developers went through the same declining creative trajectory all the above games did, but happened to build the game back-to-front. As if the designer envisioned a wonderful ending set piece, executed it flawlessly, but couldn't conceive of an interesting game leading up to it. Letdowns, broken promises and missed opportunities but in reverse. From trainwreck to transcendence. I want to know what that game would be like.

First impressions would be very different. They might improve because we wouldn't have the game's best moments for comparison. It would certainly exit on a perfect note: an alternate-universe Bioshock would conclude on a sequence of astonishing and sustained theatricality and Crysis would expand its possibilities instead of restricting them. The bad levels would not disappoint since the game wouldn't have built up our expectations. Nobody cares about an awful conclusion to an awful game, it's when the weaknesses are preceded by greatness that they really sting. Invisible War was only reviled because it was Deus Ex: Invisible War.

More likely, first impressions would be fatal. If the beginning is abjectly horrible, that will characterise reviews and player opinions. No one credible has anyone ever advocated suffering through ten hours of video game misery on the basis that later the sun starts shining a little bit. The game would hook nobody and its wonderful ending would go largely unseen. This is almost surely why all the worst parts of a game find themselves at the end. The levels are too expensive not to use, but the game's too expensive to risk opening with them. When Obsidian had to cut content from the second Knights of the Old Republic, they took it from the ending. They couldn't get away with a similarly unfinished beginning. They'd be called on it instantly, and fortunately for them reviewers are less likely to mention a disappointing ending because they think it's spoiling.

Assuming players could stomach the whole thing, would it be like the game only got better? Or would it be like watching the game fall apart backwards. Imagine you got through the uninspired and undetermined opening, then saw the game hitting its stride, and you know at that point in their development they've got lightning in a bottle. But you've seen how it turns out for them, you've seen the inadequate levels they make next. You know they end up losing it. Like flipping through a photo album from the back; every part of the game is a snapshot of the development team, and the beginning, the last thing they did, is beaten down and marked by failure. Somewhere near the sequential end of the game, though, they're on fire and they know it. Basically, imagine experiencing Bioshock or Crysis or the discography of the Rolling Stones or Elvis Costello in reverse.

That's what it is to play Bioshock or Crysis for the second time, when you know how they end. You're not watching them escalate anymore; this time you're not along for the ride. You're an observer, not a passenger, and you know all the goals and grandeur don't end well. For all your agency, it's something you can't prevent. They're going full throttle but you know they're about to crash. It's going to break your heart and theirs.

And take that replay experience and translate it to the first play. I don't think that's ever happened. Bring that on.


Spencer Greenwood said...

It's a shame that endings seem to be the first thing to go when developers feel the need to shave parts off their game.

It seems to me, though, that having a good beginning and a bad ending might be a situation which is symptomatic of an inherently flawed game. Often, if interesting themes, attractive aesthetics and some design innovation are present from the beginning, we begin to see the game as 'good', but compare this to the novel. Many novels are spurred by gripping themes, and are written coherently, but fail to provide a satisfying ending because the author didn't really have anything to say - they were just flapping their gums and trying to be really cool for a few hundred pages.

Perhaps we need to change what criteria we have for judging a game as good: perhaps games need to be seen as a whole, and, in that case, reviews should contain spoilers, if for no other reason than that it is difficult to give them a fair treatment otherwise.

Often, a game can be pretty and contain fantasy settings, emotional relationships and references to euthanasia, but not be a coherent whole. We wouldn't forgive this trait in a movie or a novel. Why do games recieve this treatment?

Valentin Galea said...

A corollary to this is when the ending ups the difficulty to 11 just for the sake of being the end.

I recently played Psychonauts and loved every second of it, until i got to the last level, that Meat Circus, when it got so frustratingly hard it almost erased all my good memories of the game.

The opposite is true as well, Bioshock as you said has a very very weak boss. Diablo 1 had a very weak 'Diablo':). The last fight in Half-Life 2 is a dud, etc etc.

An excelent ending is the one for Half-Life Episode 2 - the last fight with the Striders. It's non-linear, it's challenging but not to the extreme, and it can be done in a number of different ways.
It's actually the opposite of Crysis, the ending is an open-ended battle while the rest of the game is on rails!

Matthew said...

Sometimes publishers get mad if we developers spend too much effort on the ending of a game. “Why are you wasting our money?” they’ll say agrily, “Hardly anyone is going to see that part of the game anyway. You should be putting all of your resources into the first three levels.” So we spend an inordinate amount of time making the first three levels as interesting as possible. Those are the levels that get used as demos to the executives and at trade shows. At some point we realize we have to finish the rest of the game; if we’re lucky we’ve committed to a good final act and the middle sags, but if we aren’t we end abruptly, or cut the ending entirely.

Kylie Prymus said...

Serialized games have the potential to accelerate in greatness as the game goes on. The great thing about serials is that they have the opportunity to incorporate player feedback on both narrative and gameplay. Patches and DLC can only do so much because they assume an initial launching that is complete and stand-alone. If the developers are hitting on something right then there's no need to change it - until the chorus of gamers starts to feel like it is getting old. If something can be done better then subsequent installments can adjust.

The end product of this may be a complete game that just happened to be released in serialized form. Many famous novels did this - particularly during the 18th and 19th century novels were released in installments in magazines and often the author was just making their deadlines and responding to feedback from the readership. Today we look at those novels as whole entities and not as serials and we may get a sense that a novel that started weak got better over time. Games could do this too. But it's unlikely that a completely mediocre serial would even get subsequent installments.

But that raises the spectre of relativity. Are the last parts of, say, Bioshock independently mediocre, or do they just not live up to the earlier promise of the game? If the game had started with the quality of the end parts but been serialized, would players have liked it enough to warrant continuing the series and approaching the greatness we now see in the beginning?

All of this may be to say, perhaps we need to encourage more developers to take the serialized approach. And I mean truly serialized the way .hack tried to be, not simply divided into independently purchased chunks like Siren.

n0wak said...

A previous commenter mentioned Psychonauts, and that's a prime example. I never made it all the way through the game -- barely even started -- and it's my shame, but I lived through it vicariously on the 1up FM podcast. At the end of the "Backlog" they had an interview with Tim Schaffer where he admitted that the end wasn't up to snuff. All the effort and testing and quality assurance was focused on the earlier levels, for numerous reasons. All of which are valid but, really, do show the emphasis placed on having a strong start. It's what they make demos out of.

Duncan said...

Difficulty is a separate issue, I think, although no less valid. I've written about this before, I'm sure, but in the final act when a game is trying to accelerate the story and gives you some time-sensitive goal and the guitars come in and everything, that's when it should get easier. If the game sets up a premise like you have three minutes to defuse a bomb and save the world, the ideal result is that the player does it, barely, on their first try. A couple of games let you slam-dunk the ending like that, and the player feels really good about themselves and the game. But most games become hugely difficult and start demanding pinpoint precision, which stops dead all the momentum that they actually want to create. Valentin, you mention of Half-Life 2 as an example of a weak boss but I think that entire final Citadel sequence is great because it lets the player feel overpowered and tear through the level, instead of it being an exercise in quicksaving attrition.

Psychonauts is an offender, although apart from the difficulty I think the Meat Circus is a fundamentally okay design. I wouldn't necessarily say that about the last two levels of Bioshock. Kylie Prymus, I take your point about relativity, but I'd argue those last Bioshock levels are both independently mediocre and disappointing within the context of the game. The former is partly the reason for the latter: it was this ambitious, creative and theatrical game and it's going to end on the shooter cliches of escort mission and SNES-era boss fight? It doesn't even offer an interesting take on either of those.

I didn't get anywhere close to addressing everyone's points. I'll try and make it up later. Thank you everyone for leaving such great comments though.

L.B. Jeffries said...

Bioshock's chief failure was not having a fourth or fifth act. After you demonstrate to Ryan the quality of your golf swing, the plot pretty much fizzles out. Like you noted, the level design sinks with it. Levine himself admitted this which he defended by saying he didn't think anyone was going to care about the story. Given the fact that without the plot it's just another FPS with RPG mechanics, it isn't a totally irrational judgment call on his part.

Still...most developers are beholden to the Q&A and game design process. Some kind of method for keeping the story intact while all the development is occurring has to be created. Given the success of Half-life 2's episodic content, it looks like the best way for stories to grow are when the developer is familiar with the game engine, has a bunch of assets to rely on, and tons of feedback on what they want out of the story. That way, the writer has a very firm grip of what to write and what won't be cut as opposed to the constant revision that development currently creates.

Or as someone said above, think Victorian serial novels.