Back in 1983, when computer games were still constructed of mud and straw, people much like myself spent more time than is healthy playing The Hobbit. This ZX Spectrum text adventure had, like all half-decent games, bits that required smart thinking. Or the more likely approach of trying everything. These choke-points, and there were perhaps two in The Hobbit, are burned into my memory for one reason. “Thorin sits down and starts singing about gold.” That bloody line must have cropped up a thousand times, and to this day if, say, I’m buying a medallion, I’m on a hair trigger in case the jeweller turns out to be a lazy but tuneful dwarf.
Since those heady days a never-ending stream of games have flowed, and what I recall from so many is rarely the great dialogue but the repeated, irritating lines and the jarring banter.
The trouble is that when a game is going well, and unfolding its engrossing plot with lifelike and well-acted characters, you get carried up in it and don’t notice how highly acceptable it is. Great dialogue and acting doesn’t draw attention to itself. That’s why it’s great – you simply absorb it, believe it and care about what happens next. But games are flexible. Which is another way of saying games must cater for idiots being idiots. What’s the solution? Have the sidekick you’re stupidly battering say the same “Get off me!” line over and over? That’s not very realistic. In reality, when you repeat a cycle of bad behaviour, the victim usually modifies their response over time. Anyone who’s married knows that. But is providing a long and varied list of responses, possibly even increasing in annoyed intensity the longer the abuse goes on, simply rewarding the abuser? Doing dumb things in games just to get differing replies is pretty much the same as online trolling.
The rock: the player needs to be told the same thing many times. The hard place: you don’t want to use up time, memory and effort saying the same thing. The even harder place: players getting stuff wrong or being counter-productive would annoy those who have to endure it. It would soon alter the relationship between the characters, probably souring it permanently.
The answer, according to Leach’s Fifth Law of Dialogue (random laws that we’re not going through in order), is to minimise the responses to mistakes or willful oafishness. What’s needed is a line to state the problem – for example: “You must construct additional pylons.” This gets played every time you make the error. If you hear this more than once, it’s your fault. Of course, that StarCraft line has the advantage of being spoken by a computer character who will trot it out all day long if needs be. It’s the nearest useful equivalent to a one-note error beep.
And in fact a similar approach works with people being asshats in a game. Even if you can opt to frag your platoon or hack up your guru, if the audible ‘reward’ isn’t worth hearing, it’s a sport that will soon die out. And without entering into dialogue about what’s happened, there’s no problem with the ongoing relationships in the game. The Sarge from Halo’s line “A favour. Don’t kill my men” is as far as you need to go.
So, here I am, Mr Dialogue Man, talking myself out of a job. Well, no. The time for massive banks of responses, or buckets of barks as we hilariously call them, is in the meat of the gameplay. Again, Halo is a good example, with over 5,000 lines. As you play you’re repeating similar actions, but hearing new lines all the time. So you’re getting the variety when the good times are rolling, and that’s when you really need it.
There can be an interesting effect at work too. Creating the buckets of barks is tough, not-much-fun toil, and if loads are required a writer would have to be more than human not to resist straying a tiny bit from the workaday. For example, you need nine ‘into battle’ lines. You go through “Charge!”, “Attack!” and so on, and by eight or nine you sneak in “To certain death!” Once in a while, a player will hear this immediately prior to seeing his troops obliterated. This player, although clearly lacking in the tactics department, may announce on the Internet that your game ‘knew’ the odds and thus the certain death remark was triggered specially. Much like seeing faces in clouds, the gamer’s brain has imputed a degree of brilliance to your game which isn’t there. This effect works well for richly dialogued RPGs, as players expect to find clues everywhere, and every action may have long-term consequences.
So the truth is that repetition, for so long the horror of game writers and audio teams, is no bad thing when it indicates that the player is at fault. Basically, game-buying public, if you don’t want to hear it, play the game properly.
The other truth is that massive variation in barks and banter is always good. Oddities or funny asides work too, but they have to be mighty rare. And when the writer completes his 5,000 lines of battle cries and smiles, knowing that there’s one perfect, impossible-to-find Inglourious Basterds quote in there, he’ll be told a week later that it doesn’t work in French, Italian, German or Spanish, and could he replace it with another anguished howl for someone getting shot?