Thursday, January 15, 2009

Designing a casual MMO based on Zork

This week's topic of "hey, look at that" among IF fans is Legends of Zork -- a browser-based "casual MMO" being developed by Jolt. (Licensed from Activision, of course. For those of you who missed the 90s, Activision owns all the old Infocom titles; it was Activision who published the three Zork graphical adventures after Infocom dissolved.)

The Great Underground Empire has recently fallen and the land is in disarray. The Royal Treasury has been sacked. The stock market has collapsed, leading even mighty FrobozzCo International to fire employees from throughout its subsidiaries. A craze of treasure-hunting has swept through the remnants of the Great Underground Empire. The New Zork Times reports that trolls, kobolds and other dangerous creatures are venturing far from their lairs. Adventurers and monsters are increasingly coming into conflict over areas rich with loot. It's a dangerous time to be a newly-unemployed traveling salesman, but it's also a great time to try a bit of adventuring.

(-- from Jolt's press release.)

A lot of people -- both active IF fans and long-ago IF fans who remember Zork fondly -- immediately started talking about this LoZ thing as "multiplayer IF", as a game that would be "like Zork" in some sense.

Yeah, no. Let's look at the first post on the official LoZ blog:

Gain experience and wealth as you battle creatures, dodge traps and solve puzzles. The game is designed to be played at your own pace, so you can log in and do some exploring whenever you feel like it. Achieve fame by challenging other players in the arena or form a group to take on some of the more difficult quests.

The card game Double Fanucci also makes an appearance, in the form of a full deck of 174 Fanucci cards that you can collect and use to improve your skills. [...]

(-- from the Legends of Zork blog.)

Experience points, money, combat, skills, buffs. This is an online CRPG. That's what they're announcing, that's what it is. I immediately said "Oh, Kingdom of Loathing with Zork monsters," and I wasn't the first one to say it, either.

So that's fine; I played a lot of KoL for a year or two. The question which I wish to tromp on today is, what kind of CRPG should a Zork CRPG be?

I am, of course, being arrogant and probably irrelevant here. LoZ is in beta-testing now; Jolt has done their design work. It's too late for me to be making suggestions, even if they had a mind to pay attention to suggestions from random IF amateurs out on the Web.

But it's such a cool question.

The default CRPG used to be rat-clubbing for gold pieces; you could be a fighter, a cleric, a mage, or a thief. That's what "CRPG" meant. D&D did it, so Ultima did it and Wizardry did it. Then there were variations (you need bards for the Bard's Tale) but that was the setup.

(Kids these days will tell you that the classes are tank, damage-dealer, buffer, healer, and controller... or something like that... I'm not a kid these days, so I'll leave it to them to explain.)

It's hard to argue that the Zork tradition is unrelated to the fighter-cleric-mage-thief quadrangle. I wrote a whole post about Gary Gygax and his fundamental interconnectedness to all things, including Colossal Cave and Zork.

But D&D style combat has never meshed well with IF. Zork 1 starts off with a swordfight, straight-up dice rolling and hit points -- and then that mechanic essentially vanishes from the Infocom tradition for seven years. The other combat in Zork 1 is so heavily plot-biased that it's essentially a deterministic puzzle: nearly impossible if you plunge straight into it, but easily winnable at the right place and time.

And that's how Infocom set up their subsequent games -- up until Beyond Zork, which had several interludes of typical RPG combat. I, like nearly everybody, "solved" those scenes by saving and restoring the game. It wasn't fun, it wasn't immersive, and it didn't fit in with the rest of the game. I don't think I'm far off the concensus if I say that those elements of BZ were a failed experiment.

So am I saying that a Zork CRPG should eschew combat entirely? Heck no. There are plenty of battles in Infocom games. You defeat enemies. But they're not D&D, CRPG, wear-down-stat-X-using-stat-Y battles. (I have another whole design screed about a combatless RPG, but I feel like I should implement it rather than blogging about it... maybe next year.)

But yet, we're talking about a CRPG here! I'd love to go off and say "Jolt should have implemented true multiplayer narrative IF," but they didn't -- as far as they've announced.

Let's stipulate that the design problem is "a CRPG with the flavor of Zork". We will have stats, treasure, and skills. The game will be about burning time on repetitive actions that crank up some numbers until you can succeed at tougher actions.

But -- it doesn't have to be arranged the same way as Ultima 1 and Wizardry. I want to see how far I can break down the traditional concepts.

Treasure. When you win fights, you get treasure. Treasure buys stuff. Okay, that's good. But what is treasure? Nearly all games have currency -- a simple scalar of how rich you are. D&D had gold pieces (and a table of other coins, roundly ignored). CRPGs followed suit, although some call them "credits" or "meat".

Zork doesn't have money. Treasure, yes. Coins, sure. Barter, sometimes. Money, no. You collect and use items, distinct and distinctive. (Even the stack of zorkmid bills is a unique item -- you never cash it in for face value.)

What if we built our hypothetical Zork game (let me stress that I'm not talking about Legends of Zork here, I'm making stuff up) on a "monetary" basis of unique treasures? We could randomly generate their names and descriptions -- that's not hard. Succeed in a quest, find an antique Dwarvish black opal. Do it again, discover a handful of silver-inlaid knucklebones. Or a rare blue faience brooch. These don't auto-convert into gold; they retain their identity.

The point is to trade these in for useful tools and items -- I'm not throwing that idea away. So there will still be an underlying monetary value. Maybe it'll say "...rank-3 treasure" on the back of the brooch. But you would still have the experience of finding something, something new and interesting. (Even if it's really generated from a template algorithm.) The game designers could throw new adjectives and templates into the database occasionally. Find a treasure that takes your fancy? Don't sell it -- put it in your trophy case to display!

And you can open a market for players to swap for treasures they want.

There are design consequences, of course. You can't pile up treasures as a reward, the way you can with gold pieces. Eighty treasures are not eighty times as cool as one; they blur together and spoil the point. So rewards become much more granular -- you might get one or two per hour. That affects the rhythm of buying (or bartering) tools for further play. That affects the way you design those tools, and the challenges that the tools resolve.

Doesn't this start to sound more interesting than yet another Wizardry knockoff?

Algorithmically generating instances of things is a good trick. Let's run with it. How about locations? Kingdom of Loathing has lots of locations, but you visit each one over and over again, and the descriptions never change. (Well, rarely.) Let's set up our game so that you enter the forest, and find a unique, freshly-generated forest clearing.

(Again, the template and random-table work for this is pretty easy. You aren't trying to fool the player into thinking that there's a human being writing this stuff. It just has to read well. As long as the game doesn't lead you past a hundred of them in quick succession, players will buy into the descriptions. My own experiments with this tool are in Hunter in Darkness, and the cave section of my web site.)

If you explore out from the clearing, you find more forest locations. New room names, new descriptions. These are just "instances", in the usual MMO sense -- but you're going to visit them frequently, first to "solve" a location, and then passing through to more distant ones. You'll remember the descriptions; they'll feel like a real environment. And if you can invite other players into your instances, they'll find your part of the forest to be different from theirs.

(Maybe draw all this out on the traditional IF box-and-line grid... Oh, I'm not against graphics here. But if you have a generic forest illustration that appears above varying textual descriptions, I guarantee that the players will read and be interested in the text.) (If you can procedurally generate interesting forest illustrations too, you're really in clover...)

I did promise to get back around to combat. The burning-stats-against-enemy-stats model is a rich and well-explored mechanic, and I'm not going to try to discard it. (Today...) But who says that a "combat" must be a blow-by-blow struggle against a monster?

The fundamental act of Zork is exploration. What if the basic quest of our Zork CRPG was exploring a dungeon? (Or forest copse, or temple...) An "attack" would be the entire act of entering a room and facing its challenge -- by stealth, or trickery, or courage, or willpower. You'd still find monsters in the dungeon -- but the rhythm of the game would not be fighting blow-by-blow-by-blow, but rather exploring trap-by-monster-by-maze.

Of course you'd have to have a rich set of "combat" (exploration) mechanics. You'd have options on each move; you'd have tools and resources to use up. Ink and map parchment? Bread crumbs? Arrows? (I smell a Wumpus...) Maybe willpower and courage and steath and cunning are your solvent stats, expended against the dungeon -- just as the traditional CRPG hero expends his hit points frugally, trying to reach the orc's last hit point before the orc reaches his. Reach the end of the dungeon, and you find a treasure.

Ooh, treasures! They come in lots of varieties, right? (Since we're randomly generating them.) So they can have properties. There's the depth we want for the exploration "combat". Treasures can bribe monsters, treasures can jam traps, treasures can be left behind to mark your path through a maze. (Very Zork, that idea.) I know, I said treasures would be rare -- but you're bartering some of them for tools, and tools can be randomly generated too. Ropes, spikes, torches. Oil and batteries, food and water, all usable as "hit points" against the dungeon.

Clues! Use up clues to solve puzzles. I'm not talking about actual IF-style puzzles. Tell the player "There is a mysterious altar here, covered with Gnomic runes." To pass, he has to expend some of the Gnomic lore or rune lore in his inventory. Instead of healing potions, you find more lore. Instead of strength potions, you find a book of Gnomish history, which enables you to understand all Gnomish puzzles better...

You see where I'm going? Starting to go? Sketching a path in the direction of going? It isn't Zork-the-text-adventure -- as stipulated. But it tells the same kind of story.


I see I completely forgot to talk about character classes, you know, the fighter and the mage...

Well, I don't know if I want them, in the traditional sense. The adventure-game experience rather assumes that you, the adventurer, go everywhere and do everything. You cast the spells and defeat the monsters and solve the puzzles. Then you go back and find all the alternate solutions too.

But there are different approaches, and maybe specialization is okay. (Sneaking, courage, etc, etc.) Or maybe those specialties should be per adventure? That might be fun. Gear up as a thief and go after that dungeon -- but if you're defeated by a surfeit of mazes, you'd try again from the riddle-master's point of view. Or the warrior hero, for monsters. "Healer" is not a concept that makes sense, given the model, but there just might be room for the dungeon master to make a few key changes for his own benefit...

Comments imported from Gameshelf

Jota (Jan 16, 2009 at 9:45 AM):

So there will still be an underlying monetary value. Maybe it'll say "...rank-3 treasure" on the back of the brooch.

Hmm. I'm inclined to think that rather than just having a band of gnomes running around and scribbling appraisals on the backs of all the treasures, it might be cute to make the value of a treasure directly proportionate to how complicated its name is. So a rank-1 "opal" would be a common newbie treasure, but an "antique Dwarvish black opal" would be rank-4 and necessarily much more rare.

(By the way, did you know that this page uses a captcha that requires javascript but does not appear to actually say so if you have scripting turned off?)

Andrew Plotkin (Jan 16, 2009 at 5:23 PM):

I like that idea, but I don't want it to be strictly proportionate. You're going to start out seeing a lot of low-level treasure, and I think it would give the wrong idea to be "opal", "ruby", "gold" all the time. There should be a nice mix of long and short names, even at low levels.

(Plus, "The Amulet of Zork" is a high-level treasure, even though it's short.)

Yes, that captcha setup is built into Movable Type. It bites me on a lot of blogs. I'm not thrilled, but I'm not about to rewrite the code.

Denis Moskowitz (Jan 20, 2009 at 7:33 AM):

It would also be an interesting mechanic to use the adjectives - to buy a sword you need 10 ranks of treasure, but to buy a gnomic scroll (increases Gnomic Lore stat by 50) you need 10 ranks of Clockwork treasure.

Andrew Plotkin (Jan 20, 2009 at 10:28 AM):

Unquestionably. The more this text ties into game mechanics, the more the player will want to see the items (and locations) as real.

I don't know if it can get up to the level I'm used to in text adventures, where visualizing the game as a real environment is the only way to make progress. But it's worth getting in even a sliver of that.

No comments:

Post a Comment