End Errors: Game, Postscript
For the last month or so, I’ve been talking about how terrible endings are in video games. Over the course of two articles, one focusing on a decent variety of problems, the other exploring a single issue in depth, I think I managed to at least identify what the five biggest issues are. Unfortunately I don’t feel either article explains that most important of all questions, the “why” of it, adequately. Additionally, as someone who prefers being proactive it bothers me that neither offers a remedy to such common maladies.
Personally, I find this dissatisfying. Sure it can be fun to commit to some deep analysis, but if you aren’t offering some sort of solution you might as well be a latter day Thomas Aquinas counting angels on a pinhead. Or in this case, I suppose its extra lives on a pixel?
So . . . what IS the reason why most games have terrible endings? Why do developers act like it’s OK to deny denouement, or end at all? Why don’t they think investing time into making a satisfying climax is important?
Well if you ask the developers, they might just point out that the reason lies in gamers themselves.
Or, to quote the digital Lee Carvallo . . .
“Would you like to continue?
You have Selected ‘No’.”
Folks in the gaming industry were, early on, seemingly blind to why any given trend in design was good or bad. The whole dynamic was about as basic as it got. Some people made games (developers), and some people bought games (gamers), and nary was there a question by the makers as to why the buyers bought.
It didn’t seem like much of a problem for gamers. People bought whatever seemed the most fun (nebulous as that may be), had the best graphics, or simply appealed to a particular interest of theirs. Or if they were younger like I was, bought into the corporate hype that led to the console wars.
From the development side of things, there was conjecture about what people bought or made a given game objectively “better”, but it was mostly opinion without consensus. Studios had their own internal methods and numbers but were also in heated competition, so it was difficult to separate blustering self promotion from the truth in order to determine real quality. In 1992, was Mortal Kombat a better game than Street Fighter 2? Today, most would probably agree it wasn’t, but if you were looking at what put more kids in arcades (at least in the States), the market might beg to differ.
Without any objective data available to everyone, it’s pretty hard to go on anything other than sales figures after all. But then, that’s not really accurate. Data was available, it’s just that there wasn’t effort put into obtaining it.
It was only with the slowly earned acceptance of gaming through the nineties that serious academic study of game design began to blossom post millennium. Nowadays, analyses of player behavior and psychology are becoming increasingly common. Even better, while data and the derived metrics (good band name?) from it are often collected individually, much more is shared at places like GDC to more developers; everyone gets access to information that can make gaming better as a whole!
As with your average H.P. Lovecraft protagonist however, knowledge comes at a price. In this case, additional information that has a grim bearing, for all signs continue to unfortunately indicate that most gamers – the core group of people paying for games and making the industry function at all – don’t finish games. We’re talking the vast majority here, somewhere in the realm of 80-90 percent.
Unfortunately, going into why gamers don’t finish their games is its own overly convoluted issue that I’m probably not even qualified to tackle – I finish games after all (you kind of have to when you review them). Though that might make an interesting topic, I’m trying to stick with just the development side of endings for now. As far as I’m concerned, the fact that this key piece of information is common knowledge amongst game developers is critical all by its lonesome.
The road to Development Hell is paved with Practical Considerations
Simply from a business perspective, this statistic removes incentive for good endings with the folks at the top. Or so it can be deduced. Producers are more obligated to get the game finished, and above them are people who probably care more about return on investment than profundity. How can it be justified to either party that more time or money is needed for a quality conclusion that most won’t see? Especially if it means missing an important release date?
Prime examples of this phenomena include Knights of the Old Republic 2: The Sith Lords, where developer Obsidian Entertainment was forced to wrap every wayward plot thread in a single overbearing conversation, before just sort of stopping. Why? To hit the Christmas shopping season. Dark Void is another: the plot jumps ahead halfway through the 2nd act and into an underwhelming boss battle from nowhere, leading to a haphazard and dissatisfying ending. Why? Well, the game had been in development a LONG time, and it had to get finished and sold at some point, right?
Of course, then there’s everyone’s favorite punching bag in this regard, Halo 2.
Pushing out games before they’re fully cooked is naturally a problem, but it’s compounded by the resultant effect of another common datum: players need to be hooked quickly or they usually won’t buy a game at all. Which makes perfect sense really. Why would anyone stick with any story if it doesn’t capture interest in a reasonable length of time?
Films have a “10 Minute Rule” for this reason, and game developers do try to learn from their Hollywood cousins in this regard. One of the few areas of games that gets focus tested to hell and back is the opening, and many a game has spent excess time and effort making sure the opening will keep the player invested for the rest of it, or at least enough to make a purchase. This is something easily justified to the folks holding the purse strings it seems.
By itself, focusing on a solid start is a good thing; you can’t expect your audience to necessarily wait for your game to grow its beard. “Don’t worry, it gets better in Season 3″ is an excuse that requires a lot of trust, and in gaming the equivalent of “At least get through the tutorial level” can go to some extremes that defy logic. It’s an uncertain proposition for the player, and until a beard fills in all you’re left with is a bad teenage mustache.
This only becomes a problem when there’s an obvious lack elsewhere in the game. When the opening is hammered on by the team to be made great, but subsequent levels or features obviously can’t live up to the standard it sets. It’s essentially the logic why many fans get angry by the inclusion of Multiplayer in otherwise single player games: it draws resources away from the intended focus.
Prioritizing resources (time and staff) for openings is exacerbated by the process of making games, which is heavily based on iteration and often as sloppy as Augustus Gloop. Game development has roots in software development after all, and is prone to one of its main problems: creeping. For the uninitiated, feature creep (or scope creep) occurs because games are made in a series of of iterative “builds” – software drafts essentially – and developers, clients, and producers keep adding new features (such as a level, or new attacks) beyond the initial plan during production. It “creeps” due to the nature of the work; as games need several iterations anyway, why not toss in a new idea or two each time?
This approach is a bit of a double edged sword. On the one hand, it can lead to enjoyable improvements and polish and there are many games made better for it. On the other, it can cause priorities to shift mid-project and tasks to go unfinished, extra delays (see Duke Nukem Forever), and even cancellations (too many examples to list).
Even with projects that aren’t mortally wounded in this process it can still lead to a bloated, unwieldy game with too much stuff in it. While seemingly benign, this can lead to players not finishing their games and thus reinforce that problem. This particular issue even has a nickname, the “second refrigerator syndrome“.
Individually, each of these issues is a minor threat to a developer making a fabulous finish, but it’s when they start combining that they become some kind of terrible terminus Voltron.
Here’s a hypothetical situation that I’m betting happens all too often: a team spends most of their time making sure the fundamentals of the game work, then when that’s done spends even more time ensuring the opening is solid while a few designers with less to do get sidetracked with extraneous stuff they keep adding before running headlong into a deadline they can’t slide on. Uh oh! The team is forced to make a mad dash to get original priorities accomplished on one end and cut out the fat on the other and one thing (possibly of many) that isn’t up to the snuff is the ending!
It’s easy to see how any single aspect to a game, including an ending, can get devalued in these situations, especially if you then toss on the apathy that the persnickety little “no one finishes your game” factoid surely adds. The priority is finishing the project, not ensuring the project has a good finish.
What scares me though, is that the practical realities over the history of game development mean that this is accepted as the norm. At this point, I wonder if the average dev team wonders why they should we care about making a good ending at all.
Yes. Why indeed?
This might be a bit of a tangent . . . but I’ll try to make it brief.
The Journey only matters most if they play “Don’t stop believin”
“It’s the Journey that matters, not the destination.” and similar phrases were thrown out in defense of Mass Effect 3‘s lackluster finish. In fact I’d say this quote, and the basic philosophy behind it, is the primo defense for crummy video game conclusions over the years. To me, this sounds like the talk of an apologist.
While there is merit in this philosophy as a way to look at life as it’s probably better not to dwell on that destination, I’m willing to shave my head bald and pull a Lex Luthor when regarding entertainment:
Entertainment isn’t life. It may be a reflection of life, but it isn’t life itself (at least not until MMOs gain sentience). In fact one of the key fundamental points of difference between entertainment and life is that that the destination matters as much as the journey.
To say that it doesn’t just reeks of an alarming acceptance of laziness and a blatant ignorance of prior precedent. I seriously doubt there would be soccer riots if endings (in that case, to championship games) didn’t matter. The same goes for storytelling in general: just think about the number of stories you’ve personally read, watched, or been told that hinge on a twist ending, a reveal that gives context, or a punchline that makes the joke work.
Consider for a moment, the Richard Matheson novel I Am Legend, a favorite of mine. A tense and harrowing tale of a man struggling to survive alone past a vampiric apocalypse, it’s smart and well written. A great read that I highly recommend, along with much of Matheson’s body of work (you can skip 7 Steps to Midnight though, blech).
However, if you quit reading before the conclusion you’ll miss out on an ending that isn’t so much a twist in terms of plot, but of perspective. It’s an end that not only gives the entire story its true meaning, but justifies the story’s title. Stopping before the end results in a loss of importance in general, and a severe lack of understanding about why the story works at all.
I could go on: The Lord of the Rings‘ ending gives the story much of its poignancy. Night of the Living Dead (the original) ends in a state of dark irony that turns it into social commentary more than at any other point. Ditto Memento, and obviously any mystery story, or story that has any sort of moral.
Especially any story with a moral.
For many stories to have value, meaning, or applicability, the finale needs to be as good as all that came before. Running out of gas at journey’s end might mean you cross the finish line, but rarely does it result in first place. I don’t think The Lord of the Rings would have caught on as it did without the sad facts of Frodo only being saved from his prior mercy and the Shire being rent asunder; these inured it to the zeitgeist of the time (hippies) in a way that earlier forays into fantasy simply didn’t, and thus propelled it forward to what it is today.
I put at the start of this trilogy of articles a little aphorism I made up (as far as I know), “Without art in the ending, a game ends up without art.” That is my response to why developers should care about their endings. Even if they have practical reasons why they don’t, I really hope that game developers don’t believe that endings don’t matter. Because they do. A lot.
Perhaps I’m being too alarmist. It’s difficult to believe that developers don’t realize such basic concepts. It’s just that the situation is so damn common in gaming that it seems like some grand apathy has taken hold, some large, collective blockage. To be honest, I’m grasping at straws a bit trying to figure out what it is, and perhaps it’s a fool’s errand to even try and peg this down so simply.
But if that’s the case, then I best make myself a fool all the way through, and lay bare a few possible solutions to this problem . . .
A preliminary plan for progress perhaps?
It’s easy to say that most problems that arise during development (beyond just poor endings) are due to bad management. I’m sure that’s actually true in many cases. A simple solution is to just as easily issue a command like, “Hey Devs! Learn to schedule!”
Another simple solution is to promote the auteur. To pitch creative heads like some sort of Nietzschean Übermensch who will wrangle their teams in with a dictatorial fiat, force producers to acquiesce to their delays and to make sure their games are so engaging that all of the players see them through to the end no matter what! But there’s the problematic case study of Peter Molyneux: an auteur in control whose games still feature ALL of the problems listed above!
However these don’t seem very “actionable” pieces of advice vague as they are. Besides, I promote the auteur all the damn time. Also, both of these suggestions only cover narrow aspects to an issue that needs solutions to cover a wider range of problems.
Keeping in mind that this is from someone with limited industry experience (which admittedly makes this rather presumptuous) I still hope that what I’ve deduced about might be useful. So I offer three ideas that might help. One on the part of developers, another on the part of gamers, and one that’s an alteration to an industry wide standard.
Developers: Don’t save the end for last! Put it as close to the start as possible.
At it’s core, this idea is simple: a team moves the implementation of the ending of their game up in the timetable, and spends the same amount of time polishing it as they would polish the beginning of the game. Basically, get the beginning and ending of your game done first, then take care of the middle parts. The idea is that if you have to start cutting stuff out and scale back, at the very least the ending is there and it’s something you’re happy with.
Primarily this is meant to prevent KotOR 2 situations from happening, but it also has a bonus of limiting feature creep. Adding a new ideas or piece of content either can’t affect the ending at all, or it means re-doing the ending to accommodate the new feature(s) which automatically makes implementing them carry more tangible weight.
Oh, and there’s another reason it could be useful too. If done early enough, it gives the team a chance to consider the ending more fully to help prevent the nasty problem of “if only we had gone a different path.”
Now I realize that this is still a very obvious concept, but I’m wording it very generally on top of it all. But it kind of has to be if it’s going to fit into anyone’s style.
There are as many methodologies and management styles when it comes to game development as there are freckles on an island of gingers. A brief tour of Gamasutra, a website mostly focused on the development and business sides of game production reveals this within minutes. If you counted every story that boils down to “Our game proved successful and my team uses X practice, therefore it is THE BEST practice!”, you’d run out of digits within as many pages.
Despite such diversity though, I am left with an impression that most if not all the biggest obstacles developers face mid project tend to result in the same problem: losing their focus. Feature creep alone is that problem writ large. Unfortunately any solution I offer is going to have to be a bit vague for maximum applicability, and it’s going to have to try and take focus loss into account.
I think this idea might do just that. Tackling the ending early when the team is still in full stride and hasn’t entered the death march many dev teams describe the final crunch of a project to be, might help to avoid the (sometimes inevitable) loss of focus to come. Still, that’s a problem that probably comes from poor management or scheduling, so perhaps this is just a naive hope and something everyone already knows (or does) anyway. Any developers who end up reading this, I’d love to hear if the idea has any merit or is as obvious as I fear it might be.
But of course, the problem with endings doesn’t just lie in the domain of the developers who make them, for if anything that nasty statistic reveals that it’s as much the players fault as anyone . . .
Gamers: Don’t give up! Finish your games!
Seriously, I can think of no worse statistic for a creator than one that reveals that the vast majority of your audience simply won’t get to the end of the game, even when they like the rest of it! Why try to make something meaningful if the attempt will be futile? Such a question is starting to haunt developers who would like to make something meaningful, and it’s hard to say their medium is capable of allowing for such profundity if no one’s willing to commit to it.
But the other real reason it’s important that gamers finish there games comes down to feedback. Apart from whatever a game developer can collect themselves, what players think about a game is rather limited, especially regarding endings. Either because players don’t finish them or are jaded from past examples, no one really complains about crummy endings specifically. Or at least, they didn’t.
I’ve said it before, but for this reason alone the Mass Effect 3 controversy is important. Not only does it show that people care about the narrative to a game, but it’s generating information for BioWare (and everyone else) about what does and doesn’t work for a game’s ending in a way that silence never could. While most may be obvious – don’t introduce major plot elements during the resolution – some of it isn’t or is wholly relegated to interactive storytelling – player choice is mostly illusory, but an ending that breaks the illusion without making it a key to the narrative (such as BioShock) comes off dissatisfying.
My theory is that the relationship between players not finishing games and games having bad endings is directly proportional. If you consider the fan reaction to ME3 when in light of data on player behavior from ME2 that says about half the players finished the game (far above the average ten percent) I’d say the theory holds some water. If it does, then the one thing players can do to make endings better is to simply see their games through to the end.
But then, gamers not finishing games is generally a response to the game themselves. I said earlier that delving into this why might be beyond my scope, and so it remains . . . unless I become grossly reductive with the facts. So let’s try that!
I’m going to make some big assumptions and boil ending impotence down to one core problem in relation to two different factors: gamer uncertainty over difficulty (being too hard for them) and how how long games are (no knowing how much time needs to be invested).
What’s great about such reductionism is that it has an equally reductive solution!
Industry: Games need Running Times and Skill Ratings!
Think about this for a second. When you go to pick up a book, you immediately know how long it might take you simply by looking at its thickness and weight. A glance at its pages tells you more: if the font size is small it will take more time than if large, scanning the first page will give a sense of the type of language the author uses. Within seconds, a potential reader can get a decent estimate of how much time the book is going to take them.
A similar bit of info is available for films: running times. It’s something you see on the back of the DVD and most theaters have the length listed at the box office. Similarly, if you go to a ski slope you’ll find skill rating given for how challenging a particular course can be. Tours through museums tell you how long it’s probably going to take to wander through with the group. The list goes on, but the message is the same: the measure of the commitment expected is itself expected.
At least in literally everything other than a video game, which have no declarations (definite or otherwise) in regards to finite time. A game like Xenoblade Chronicles could engulf literal weeks, Portal 2, a handful of hours, but both take up the same amount of shelf space at the store. Ideally, a game review covers this, but there are often inaccuracies, a lack of consensus, and unless a game is particularly long or short, the chance it isn’t mentioned at all because it doesn’t matter in the qualitative analysis of the game’s other aspects.
The same uncertainty lies with difficulty. What’s “normal” for Ninja Gaiden Sigma is sailor swearing hard for Halo 2, and then there are games without difficulty settings! Half-Life 2, Dark Souls and L.A. Noire all lack that option, and all are astoundingly different in terms of challenge. Again, reviews could help warn against notable examples, but difficulty is highly subjective measure to begin with, and the lack of consensus makes the issue moot.
From a consumer standpoint it makes no sense that such uncertainty exists for such basic information. Uncertainty is a commitment killer! I’ve seen friends lose interest in RPGs that were too long when nearly at the end, or give up on challenging games right after they’ve cleared the last major hurdle because they just didn’t know if they were going to have to put up with ten more minutes or twelve more hours. Everyone’s more willing to see things through when they know how long something takes!
This is an issue of product labeling if you ask me, and that is the only viable solution I see: better labeling. Just as there is content information in the form of ESRB Ratings (and its international equivalents), there should be estimate of average running times and perhaps simple notations on expected skill to finish a game. It needs to be from an objective source and on the danged box for it to be truly useful.
Of course I already foresee some problematic questions arising. “How does Multiplayer factor in since it can extend the length indefinitely?” and “How do you gauge the skill level required on a game that offers multiple difficulty selections?” spring to mind. Then there’s the arbitrary nature of rating anything, but at least it’s no worse than the systems in place for content.
It’s a better solution than listening to inconsistent reviewers or the bias from a publisher’s marketing department. They always have the same answer: the difficulty and length are “just right” (and infinitely replayable until the sequel). Not only do I think the benefits from such a system would outweigh the obstacles and the alternatives, but I honestly think it will help with the main issue of players not finishing their games. Eliminate some of the uncertainty and you promote completion rates overall, which in turn promotes better endings overall, and you don’t have to initiate some industry wide trend toward making all games as easy as the last Pajama Sam adventure.
Besides, this last idea must have at least some merit, one website has already jumped on board. It’s just something that should be official and universal as far as I’m concerned.
And now the conclusion!
So there you have it. A last look at a problem that perhaps I alone care about, and three ideas that could help to alleviate it. It sure took a while, didn’t it?
Yeah, I’m going to have to work more on the whole brevity thing. But if you did make it this far, why not respond in the comments below?
Are endings as problematic for you as they are for me? Does the fault lie more with the players, developers, or producers? Do any of these solutions have merit? Do you have any of your own? What is the airspeed velocity of a migrating swallow?
It’s a topic I’m definitely interested in (obviously) and would love to hear other opinions on, but for now, I think I’ve said all I can on it. Like the Master Sword at the end of Link to the Past, I intend to put this topic away . . .
Until next time!