Why Include Randomness: Part 2

This is a thematic continuation of last week's post. To catch up, go read Part One.

Last week, I described one specific case in which using randomness adds some unpredictability to the outcome of your choices. This unpredictability can be a good thing; it forces the player to weigh their priorities against the possible outcomes before making a choice. This is interesting because it's what we do every day. We constantly, in real life, make choices, every hour of every day, and those choices affect the future. We make those choices by weighing probabilities, predictions, and priorities, and to do that in a game is fun.

That said, using randomness the way I specifically described it that time isn't necessarily that common. The most common types of randomness we see in gamebooks have to do with testing skills, if those skills are recorded numerically, or in combat--which is effectively another version of testing character skills. In this mechanic, the additive element comes in the preparation--picking which skills your character will excel in--instead of in the event of the challenge.  By the time you're testing to leap across a chasm, it's too late to make any choices that affect the outcome (other than perhaps finding another way to go...) But the existence of that chasm leap late in the game makes the choice meaningful which you made earlier about where to allocate your skills. Without the existence of the chasm leap scene, the number of points you put into Chasm Leaping becomes irrelevant.

Let's take a look at Destiny Quest, because it's a great example of this type of mechanic. In Destiny Quest, you run around, completing quests by testing (successfully) against your character's skills. As a reward, you acquire items which increase those skills, thus making you capable of completing the next quest.

In Destiny Quest, by the time you get to combat, you've already done all you can do. You've laid your bet, you've played your hand, and now all you can do is sit there and roll and hope.

Despite this, it is not a system without choice. It's simply a system where the choice is all front-loaded into the character development stage of the game. In Destiny Quest, the game remains fun because character development is an ongoing process. Every time you complete a quest, you get a reward. Often, this forces you to choose between two items which modify your character's abilities in different ways. Right there--that is the moment of player interaction. Which item do you keep, and which do you destroy?

The scenes of combat themselves are essentially very boring. For those of us who don't find rolling dice to be inherently exciting, there is nothing to do. You simply must carry through the busywork to find out whether the character build choices you made earlier will pay off or not.

But without that interaction-less moment later (combat) which tests your character build, then the choices you made during the build phase would be meaningless.

All that said, there still remains two possibilities, a skill test that involves randomness vs. a skill test that doesn't. Instead of having a Chasm Leaping skill rated 1-10 and a Killing Goblins skill rated 1-10, you could simply have each of those be a binary trait. You either have it or you don't.

In that model, when presenting the reader with a chasm, the game would ask, "Do you have the Chasm Leaping skill? y/n" and the player would either be able to jump it or not. None of this, "roll and add your skill and try to be difficuly number x, y or z."

This model has the advantage of being simpler, and in a gamebook, that often makes it worth it to go that route. But it doesn't allow for as much real choice. If you know for a fact you will be able to jump any chasm, there's no reason not to go for it every time. If you've got about a 20% chance of horrible painful death if you jump a chasm, there's now an interesting choice. Is it worth the risk? If you're running for your life, probably. But if you're out for your morning constitutional, then probably not.  Each time you run into a chasm, your mind will instantly start evaluating whether it's worth it or not. Sometimes it'll be easy, but there's still a real choice, and when it's not easy, it becomes a good game experience. And as with last week's example, this choice retains its replay value compared to a situation where there's no luck involved.

Why include Randomness?

A few weeks ago, Dave Morris raised an interesting question on his blog, "Does interactive fiction need randomness?"

One of the things I like about Dave is that he's always at the forefront of gamebooks (a term I'm using generically for our type of interactive fiction, as the term "interactive fiction" has been claimed by a slightly different genre). Dave pushes the limits of the field, questioning assumptions and imagining the genre not as it is, but as it could be. I admire this a lot, but it doesn't mean I always agree with him on each of the particulars ;)

In this case, I would like to make a defense of randomness in interactive fiction. I won't argue that every gamebook should have a random element. Like ice cream, sometimes you want one flavor and sometimes another. Sometimes one person has the flavor they prefer, and they aren't interested in any other. What I do want to argue is that randomness can be additive, and here's why:  unpredictable results create interesting choices.

A while back, I wrote about the types of choices that one finds in gamebooks. To sum up the main points... A) The core game mechanic in a gamebook (regardless of the presence or absence of other rules) is the sequence of narrative choices the player makes. B) It is very important that those choices be interesting in order for a gamebook to be fun. C) For a choice to be interesting, the player has to have enough information to make an educated guess, but not so much information that the best option is obvious.

In short, to give the reader an interesting choice, the author must reveal some information, but also conceal some information.

Think of the classic "Which Door" choice: the reader is presented with something along the lines of, "at the end of the tunnel are two identical doors. Do you open the one on the left, or the one on the right?" What the reader doesn't know is that one of these doors leads to instant death, while the other leads to tasty cake! This sort of crap may have been what dudes were into in the 70's and 80's, but by today's standards, no part of that is a good game experience. If you lose, it feels arbitrary and abrupt. But even if you make the right choice, it's not like you can take any credit for it. Success isn't satisfying if you didn't earn it.

On the other extreme, what if you knew right off the bat what was behind each door? This is, literally, the classic, "Cake or Death" choice. This is also a complete flop as a game experience, because the answer is obvious (unless you're some sort of death-seeking masochist, in which case I think you have better things to worry about than the quality of a gamebook.)

The point is, any situation in which the player has complete knowledge of the results of their choice will eliminate the existence of the choice. A choice has to do with predicting results. If you know the results with 100% certainty, the preferable choice will always be apparent. Even if all options lead equally to success or failure, then there is still no choice to be made, because the direction you choose is irrelevant to the results.

This is where randomness comes in. Randomness introduces a fog of war to the scene. Even if the player could read ahead and see all the future paragraph sections, if there is an element of randomness, the outcome is no longer certain.

Of course, information can also be concealed narratively. This is an excellent way of solving the problem, which gamebooks should always use. The only caveat is that it has no replay value.

If you take those two doors and draw a symbol of cake on one and a symbol of death on the other, then the player has an interesting choice. Is the dungeon being honest, or lying to him? Perhaps something has come up earlier to lead him to trust, or mistrust, the signs in this dungeon. The player has to think about it, predict a result, and make a choice. But once the player has read through that scene once, the next time he comes down that hallway with Character McHammenheimer the Second, the player immediately knows what's behind the door--there's no choice any longer.

Contrast this to taking the cake and death, and putting them both behind one door, but there's only a 10% chance of death, and a 90% chance of cake. Or you can go through the other door and continue on your day without taking the risk.

Whoa... now you have a choice. Just how tasty does that cake look? Is it worth a 10% chance of immediate death? Mmm... it looks pretty tasty, but immediate death?

Best of all, this remains an interesting choice, no matter how many times the player reads through the game.


Afterword: I do not think that cake is a very good reward, nor death a very good punishment. Obviously these should both be replaced with positive results that do not rely on virtual taste buds, and negative results that do not instantly kick the player out of your game. You do want people to be playing your game, right? So why did you just boot them out again? How to design positive consequences that are meaningful without being gamebreaking, and negative consequences that are meaningful without being crippling is a different topic entirely.

New Fighting Fantazine is out!

I just want to make a quick post to point out that the 11th edition of the gamebook e-magazine "Fighting Fantazine" has recently come out.

What's especially cool about this one is that it includes an all-new, original 219 section gamebook by Stuart Lloyd called Ascent to Darkness.

Go check it out! I know I will :)