Wednesday, November 18, 2009

Movies: Why you should always split up.

Recently, in a series of posts about surviving the traps of the Saw franchise, I tossed out the theory that the best way to survive a slasher attack would be, despite the injunctions of many a horror blogger, to split up.

Here's the idea: If you are being targeted by a slasher, you and the rest of your party should book it in opposite directions as fast as you can and continue running until you are either killed, escape the slasher's sphere of influence, or find genuine help (like, say, you reach an Air Force base willing to scramble Apache helicopters to aid you). I hypothesized that, with such a strategy, the killer would easily get one or two victims; but the longer the game went on, the less likely he would be to snag another victim. As time progresses, the distance between targets increases and the knowledge the killer possesses about the location of future victims decreases. At some point, the killer is dealing with steadily expanding distances and he's essentially guessing about where he needs to go.

It sounds good on paper, but the persistence of received wisdom - "Don't split up!" - made me doubt. So, I decided to simulate the problem with a pencil and paper thought experiment.

Though we are all throughly aware that "assumptions" make an ass out of "you" and "umptions," we need to make a few before we can test any theory. Here's what I assumed.

Assumption 1: The slasher is reasonably human-like.
First, strategy of any sort only matters against somebody who exists in time and space. Call it the Freddy Axiom. What's the point of running away from Freddy? He'll just pop out of the wall ahead of you or turn the floor into a giant sweater-wearing snake or something. There's really only one way to beat Freddy and that's to be a person the screenwriters have decided should live. For the purposes our experiment, the killer must move through space and eat up time with his actions.

Assumption 2: The slasher isn't psychic, but neither are the victims.
The slasher is limited in its knowledge of the its surroundings and won't magically know exactly where it's victims are. Slashers, like any predator, have to find their victims. When they can see or hear them, this isn't a problem; but there's an operational limit involved here and, after that, they're essentially on random. This also means that the slasher won't get inexplicably delayed fighting some co-ed who has suddenly discovered that she can throw sofas at a dude using just the power of her mind.

Assumption 3: All things being equal . . .
Because I'm doing this simulation with paper and pencil, we've got to keep it stupid enough for me to actually handle. We're going to assume average speeds. We'll admit that there would be differences in routes and whatnot that would introduce constant variations in speed. But I'm not smart enough to simulate that. We're also going to assume an average kill time. Sure, sometimes Jason just buries an axe in a girl's skull and calls it a day. But, when he's feeling his oats, Jason sometimes ties somebody shut into their sleeping bag and hoists the whole shebang over a tree branch in order to position the victim over campfire. Instead of selecting all the varying kill times, we're going to say that it all comes out in the wash and standardize his kill delay.

Assumption 4: There's splitting up and then there's splitting up.
For our purposes, splitting up means putting an ever increasing distance between you and the other members of your party. Dividing into three groups to explore a creepy farmhouse isn't splitting up.

To test the theory, me and my cubical-mate created a little game. It works like this. There are 8 potential victims and 1 slasher. Everybody starts at a central point. The game goes in turns. Victims always move first. In a turn, each character can move their full movement allowance - measured by an arbitrary unit of length we very scientifically called "a unit." Potential victims always just book it the hell out of there, moving in a straight line away from the central point. They fan out at 45 degree angles. The slasher is free to move in any direction. If the slasher contacts any potential victim, they are removed from play - but killing them takes up the slasher's next turn. Play continues until all players are dead or they have moved 10 units away from the slasher. At that point, the slasher "loses sight" of the players and the slasher's motions become randomized. In theory, the slasher could still catch somebody, but in practice the chances of that happening are minimal (it never happened in any of our games).

We played three variations of the game: Game 1 featured a killer and victims of equal speed, Game 2 featured a killer that was slightly faster than the victims, and Game 3 featured a killer who was twice as fast as his victims.

Despite the popularity of the plodding slasher, we played no games in which the killer was slower than the victims because it became very clear that, according to the game model, a slow and steady killer catches neither diddly nor squat. So, we've learned our first lesson. Wannabe slashers take note: Go fast or go home.

Game results:

Game 1 - The killer takes a victim in Turn 1. And that's it. The killer takes off after his next victim, which he always stays a little more than one unit behind. The other victims begin to disappear off his radar by Turn 6. Curiously, the game - according the the strict construction of the rules - never officially ends because the killer never loses sight of that poor potential second victim. Like Sidney Prescott, Not-Victim-2 is cursed to be pursued for the rest of the franchise.

Game 2 - In the game we ran, the killer moved one and one half units to every unit his potential victims did. As in the slower game, a victim dies right-off in Turn 1. The killer gets another victim in Turn 5, but victims start falling off the radar in the same turn. The killer gets a third victim in turn 13. The killer takes a final victim in Turn 26. At that point the remaining four players have escaped.

Game 3 - In the last game, our killer hauled ass. He could move two units to every one unit the potential victims could move. As in Games 1 and 2, somebody bites it right off. The second victim falls in Turn 3. People start dropping off the killer's radar by the Turn 5. The third victim goes in Turn 7 and the fourth in Turn 13, but by that time everybody else has dropped of the radar.

Conclusions: Splitting up - defined as getting as far from one another as possible as quickly as possible - saves half the party even when the killer is considerably faster than the victims. In fact, after running the game three times, we estimate that even the fastest killer could get no more than five victims. The rest would inevitably fall off the radar and escape.

Admittedly, this is all pretty abstract. We didn't figure in the obligatory twisted ankle, the presence of a second killer, or late-franchise supernatural shark-jumping elements. Still, I think it points to the fact that the conventional wisdom is wrong. How do you survive a slasher flick? Split up!


boyhowdy said...

I wonder how the volunteer dilemma (or its inverse, maybe?) might work into this. What's the likelihood of survival if one (or more) of the folks separate and the rest stick together? If the slasher happens upon the group, it'd be happy days for the outliers. Does the group's chances increase if he/she/it encounters one or more of the individuals first?

Is your cube-mate up to the task?

The Igloo Keeper said...

I'd hate to be the guy who had shout out 'turn 7!' to the slasher though.

Madelon said...

Interesting stuff, as usual...

In response to boyhowdy's comment: splitting up into group plus one essentially constitutes of two 'players', which means that the chances of the slasher catching up with the group are quite high.

It might be interesting to take something into account regarding the ability of players to hurt the slasher, thus slowing him down or removing him altogether.

This would make an interesting boardgame, really... Reminds me of 'Last Night on Earth'

boyhowdy said...

@Madelon: Absolutely! And when you consider that he still has to take a turn each to off the members of the group, those outside the group are almost guaranteed to get away (still assuming that the slasher finds the group first).

I guess my conclusion is that the the best way to survive a slasher flick is to convince everyone else to stay together and then get the hell out of there.

Also, totally agreed on the boardgame. It would actually be pretty easy to mock up a software version...

CRwM said...


Interesting question. That's basically the model most common to the genre: the group mostly stays together and stray individuals go out.

We would have to add some new rules to the game, such as figuring the kill delay on a group-kill.

I haven't run the necessary games, but my "research partner" and I theorize that the death-rate would increase (just as it does in the movies).

In terms of the game as we designed it, the slasher is completely aware of all activity within 10 units. This means the slasher "player" can tell who is a decoy or sacrifice.

If he's the same speed as his victims, he would choose not to chase the loners, and instead take the "group" as his first kill. Since the strategy for escaping a same-speed killer gets you seven survivors, there's no profit in sacrificing two or more players to get the others safe.

If he's faster, then it becomes a slaughter. He can pick off the loners and still catch the main group. Everybody dies.

That's what we predict.

That said, there might be profit in grouping folks so that the dispersal angle is greater than 45 degrees. Forty-five degrees is still a pretty tight cluster in the beginning of the game. That would have to be tested.

CRwM said...


Though, if you apply other slasher rules, you could maybe socially engineer better chances. The kill patterns are always the same. Imagine that the first victim is at the 12:00 position of a clock face. The killer actually kills victims clockwise or counter-clockwise, but he can never skip a victim or change direction without losing more victims. If you could put, say the "tramp" at 12 and "the druggy loser" at the 2:00 position, perhaps you can lure the slasher to go clockwise and escape by being at the 7:00 position.

CRwM said...


It would be neat to add more detail and make it a genuine game. But we just wanted to figure out if geometry worked out.

Thanks for the heads up on LNoE. I was familiar with the Zombies!!! board game, but I've never seen Last Night.

zoe said...

ahhhh, you just made my day again.
hysterical post, thanks!

Sasquatchan said...

cube-mate ? so you're back to the realm of the employed and what not ? (ie no freelancing now? )

CRwM said...


My pleasure.

CRwM said...

Screamin' Sassy,

I'm still freelancing (in office), but not for much longer. I start a real grown-up full time gig next month.