The Stack

Despite knowing the inevitable outcome, I did bring my entire party from Wizardry II straight into Wizardry III for a ceremonial slaughtering. It just seemed wrong to do otherwise. That done, I was in for a bit of a shock: I couldn’t import any more characters. There’s a limit of 20 characters on what Wizardry misleadingly terms a “scenario disk”, and every single slot was filled with someone lying dead in the dungeon. To bring in anyone new, I’d have to delete someone. But who? My new characters? The ones that have been waiting for rescue since 2010? In the end, I opted to just wipe the disk and start over.

Now, once you’ve done such a thing once, it becomes very tempting to do it again. You lose a few parties to the maze, along with their sweet equipment. You’re only going to get a fraction of that equipment back even if you manage to recover their corpses. The dead guys are probably not worth the cost of resurrecting, and no one else has any XP. So what good is anyone? Might as well wipe the disk and stop them cluttering up the roster!

I actually tried streamlining the whole process: instead of creating a party at a time, I created 20 characters, imported them all at once, used all their gold to buy better equipment from the very start, and sent six of them into the dungeon. This might actually be closer to what the designers intended: when you’re through with Wizardry I, why wouldn’t you bring all your characters to the next game? And it worked pretty well, for a while. When one of them died, there was a replacement ready, and all I had to do was hand him the armor he had already partially paid for. But this was subject to an all-your-eggs-in-one-basket problem: when I finally did suffer a TPK, I basically lost everything. Might as well wipe the disk!

Really, I think, the key is to spread things out. XP is the real treasure here; as I noted in my 2010 posts, once you have even one level-2 character, the whole thing becomes much more survivable. So what I really should be doing is hoarding XP. When a party returns from the dungeon with more XP than they started with, don’t just send them all out again. Send one of them out again, with a bunch of new recruits wearing only as much armor as they can buy with their own funds. If they come back alive, add them to your stockpile. We’re essentially gambling here. Creating a character is ante. Sending a character with nonzero XP is a raise. You don’t want to go all in on every hand.

Coming right off of the power fantasy of Wizardry II, it’s impressive just how hard Wizardry III throws you at the wall. As you may recall from my posts of twelve years ago, Wizardry III, unlike Wizardry II, reduces imported characters to level 1. The only things you keep are your stats and your class. While this does introduce the possibility of putting a Ninja in your party from the very beginning, it remains the case that any level-1 character stands a good chance of getting killed in their very first encounter, possibly in a surprise round before you even have a chance to run away. I spoke of Wizardry II‘s eagerness to kill characters without warning, but there, it was at least always because of something the player did, and consequently learned not to do. The thing Wizardry III is teaching us not to do is go into the dungeon at all.

The in-world justification for disempowering the characters is that you’re not actually playing the same characters that you imported, but their descendants. Hence the title “Legacy of Llylgamyn”. Characters aren’t automatically reduced on import, but have to be linked into their descendants through a menu command (L)EGATE, a word that provokes etymological thoughts. (Oddly, the manual incorrectly gives the command (R)ITE OF PASSAGE for this. Were the manual and the executable in this anthology taken from different versions of the game? That would make sense for the first two games, which we know to be ports to the Wiz3 engine, but not for Wiz3 itself.)

This notion of legacy creates a false expectation. When a character wins Wizardry I or II, they’re awarded with a special mark on their stats page, like a medal in the form of an ASCII character. (In the version of Wiz1 that I played as a child, the mark was affixed to their name, but that isn’t the case here.) And these marks are inherited. Once you have a party that’s killed Werdna and recovered the Staff of Gnilda, it’s natural to think “These guys are special. They’ve got honors. These are the ones who will save the kingdom of Llylgamyn from disaster.” But then of course they all just die, and you wind up making a new party that you don’t bother going through I and II with because they’re just going to get deleveled and probably killed anyway. I can imagine getting fully-blinged-out heroes by sending a new party through I and II once you already have some powerful characters in place in III to act as bodyguards and shepherd them through the early levels, but it wouldn’t be the same. They wouldn’t feel like the real legendary heroes. They’d be more like tourists on a Legendary Hero package tour.

Which, come to think of it, fits the theme of legacy pretty well. We’re talking about creating systems to give the children of the rich and powerful a free pass and make sure they receive rewards they haven’t earned.

Level 6 of Knight of Diamonds is basically a big power fantasy, sending hordes of absurdly overpowered monsters at you so you can squash them with your even more absurdly overpowered heroes. That can be fun, but I did finally decide to get on with it and win the game. You don’t even really need to visit level 6 at all to do this. The ending takes place back on level 1, where you complete your quest by handing in the outfit you collected in the boss fights of levels 1 through 5.

There’s one way that level 6 connects to the ending, though: the riddle. When you fork over your loot to the goddess Gnilda, before she bestows upon you the final macguffin, she asks you “What is the answer to the sphynx’s riddle?”. The “sphynx” in question is on level 6, guarding nothing more than a shortcut back to level 1. So you don’t even really have to deliver the answer to the being who poses the riddle in the first place, but it’s a useful way to try out guesses without consequence. If you give Gnilda the wrong answer, you still lose your stuff and get nothing in exchange.

And now this post is going to descend into griping. According to one second-hand source I’ve found online, Sir-Tech customer service fielded more complaints about that riddle than about anything else, which would explain why they left it out of the Nintendo version.

First of all, I’m not at all sure that I actually ever found the bulk of level 6 the first time I won this game. That’s because it breaks a promise. We’re told that the use of a light spell like MILWA or (preferably) LOMILWA will reveal any secret doors — and that means that secret doors basically aren’t a factor for the bulk of the game; the only times you ever don’t have a LOMILWA going are (A) when you’re too low-level to cast it (which also means you’re too low-level to be even considering transferring over to Knight of Diamonds) and (B) when you’re in a darkness field of the sort I described in my last post. I honestly have no idea which of the doors on my maps are secret ones. All doors look the same to me. Well, except for a few in the lower levels of Knight of Diamonds, which are apparently extra-secret doors, invisible even under magical light. It would have been very easy for me to not know they were there.

And that would cause some problems for the riddle. See, the sphynx tells you to find three clues scattered around the level. Recall that you can’t teleport on level 6. If you trust that walls are walls, and you’ve thoroughly looked around the small space that’s visible available, it’s hard to see where they could be scattered. What do you do then? What the game expects you to do is walk into walls just in case the rules have changed.

Still, it hardly matters, because the three clues are actually useless. Here they are, see for yourself:

That king, the king who worships gold, will no more see his treasure room

That king, the king who worships power, will have none within his tomb

That king, the king who worships these, that king, he finds doom

Now, the brief bit of lore in the single page of documentation does talk about an evil usurper and an attempt to overthrow him, using the legendary Knight of Diamonds armor, that brought down the castle and turned it into the dungeon you’ve been exploring. So you might think to try the usurper’s name, Davalpus. When that doesn’t work, you might try the only other two named characters in the story, Margda and Alavik. But in fact the answer is “THE KNIGHT OF DIAMONDS”, and I don’t see how that fits at all. It is, on the other hand, a fairly easy thing to guess if you ignore the clues. The clues actually make the riddle harder.

If I squint, I can kind of see it justified as a twist revelation, that the “king who worships gold” etc. is actually Alavik, the actually-evil ostensible hero of the backstory, or possibly the original Knight of Diamonds that the armor was made for was actually evil, or maybe it’s even describing the player characters, accumulating treasure and levels for their own sake. Revelations in the form of riddles can be highly effective — Andrew Plotkin’s System’s Twilight has some smashing puzzles of this sort, where it tells you a story and then, some time later, asks you a question about things just outside of what was narrated. There, the very posing of the question provokes reevaluation, makes you come to new realizations just by making you think about things from a new angle. If that’s what they were going for in Knight of Diamonds, all I can say is that System’s Twilight did a much better job of it. There, when I realized what the answer had to be, I knew what it meant.

At this point, I could probably just go for the ending. The characters I imported were overpowered for the scenario, and even if they hadn’t been, the game is as eager to help you advance quickly as it is to kill you without warning. The dagger that instantly transforms a thief into a ninja, available only as a rare drop in the previous scenario, is available for purchase in the town. There are similar items findable in the dungeon for attaining the other advanced classes, and if you (justifiably) don’t trust that kind of magic and want to make Lords and Ninjas the hard way, there are items that facilitate that too, giving permanent boosts to various stats.

But even as it accelerates advancement, it slows exploration. And that’s why I haven’t ended things yet. I still want to map out the dungeon as thoroughly as I can. (That’s why I’m replaying this, remember?)

How does it impede exploration? First, there are places you can’t teleport to. Teleporting blind is a foolish gamble anyway, betting that the spot you’ve picked isn’t occupied by solid rock. But when you can’t even teleport to places you know, you lose a valuable shortcut, as well as a way to be sure where you are.

Speaking of which: There is a low-level spell called DUMAPIC that’s extremely valuable for exploring confusing areas with teleporters and spinners and indistinguishable rooms. All it does is tell you your current coordinates and heading. When the game really wants an area to be disorienting, it just disables this. There are anti-magic fields where no one, player or enemies, can cast any spells at all, but there are also specific anti-DUMAPIC fields. Such a field covers the entirety of level 6.

Even worse are the darkness fields. By default, you can see one space around you. With a spell providing light, you can see two more spaces beyond that, and that is the state in which any sane person plays most of the game. When in magical darkness, you can’t see at all. The first-person view is simply filled with a grey dither pattern. The only way to tell where the walls are is by bumping into them, making mapping slow and laborious, and at the same time absolutely necessary. Darkness is so discouraging that the first elevator back in Wizardry I is protected by nothing else, secured by just the fact that the player’s first instinct on entering darkness is just to try to get out of it as quickly as possible.

Level 6 of Knight of Diamonds combines all of this. You can’t teleport to it. Reaching it for the first time means sliding down a chute into darkness where you can’t use DUMAPIC. The only way I know to find the coordinates where the chute comes out is to teleport straight up, and cast DUMAPIC from there. There are stairs leading down from level 5, but they’re behind a one-way door; if you’re being sensible and not teleporting blind, the only way to find them is from the underside, approaching the level 5 boss backward.

The boss fight against the shield on level 2 of Knight of Diamonds is immediately preceded by something new for the engine: a riddle. Wizardry I and II have both had riddle-like things going on before this — notably, on level 1, you get a mysterious hint about how to reach the boss, but the solution there is “use a teleport spell to get past the teleport square that creates the endless-corridor effect”. But here on level 2, for the first time, the game wants you to answer in text, typing in the answer.

It’s not a difficult riddle. It goes on for so long, you’re pretty much guaranteed to figure it out before it’s over, and yet it keeps on going. The thing that impresses me is that it fits so naturally into its environment that I’m not sure I even noticed it was a new mechanic the first time I played through this episode. It just seems like an extension of two things that already existed: spots that give you a bunch of text when you step on them, and spots that block your way unless you’re carrying the right key item.

And that makes me think back to the riddles in Wonderquest. I was quite critical of those, in my most recent post on the game earlier this year. They didn’t seem to fit in at all well there, even though there were in some ways better-integrated with their environment: they had game-mechanical effects beyond letting you do something you could do if they weren’t there, and sometimes the content of the riddles concerned game content. So what gives?

I think what’s really going on is this: Wonderquest is primarily a puzzle game. Wizardry is not. It has puzzlish aspects, like confusing mazes and figuring out how to best take advantage of the mechanics, but it’s not focused on puzzles in the same way that Wonderquest is. So when we see a weak puzzle like a text riddle in Wizardry, it feels like just another manifestation of what it’s been doing all along, just extending its general design philosophy along a different branch. Whereas in Wonderquest, it feels like a downgrade, temporarily abandoning rich and complex mechanical puzzles for something relatively trivial.

If I had to make a series of movies based on the Wizardry games, Knight of Diamonds would start with the heroes from the previous film. Flush from their success against Werdna, they descend cheerfully into the dungeon under Llylgamyn, encouraged by how relatively easy (but not trivial!) the opposition is. There are a few close-calls, followed by laughter and back-slapping camaraderie. Then they all suddenly get killed. The rest of the movie concerns a completely different team, inexperienced and nervous about attempting the quest that even the famous slayers of Werdna couldn’t complete.

It’s not just that the game is eager to kill your characters. That’s been a factor from the beginning. The impressive thing is how it springs it on you suddenly and unexpectedly. I found an item that, when used, grants the user a bunch of experience — not enough to gain a whole level, not at this point, but enough to get you a substantial amount of the way there. Afterwards, I noticed it was still in my inventory. Well, that’s actually fairly normal: magic items in this game don’t have charges, exactly, but they have a percent chance of breaking after each use. It looked like the idea was to keep milking this thing for XP until it broke. It turns out that when it breaks, it kills the user.

And not just kills, but renders them “Lost”. Recall that death isn’t immediately final in this game: you get a couple of chances at undoing it, either at the Temple of Cant (which costs money) or with your party’s own cleric (free, but has a greater chance of failure). If resurrection fails, the deceased is reduced to ash, which takes a more powerful/expensive resurrection spell to undo. If that fails, the character is Lost, and cannot be recovered. Here in Knight of Diamonds, the kid gloves are off and things can just send you directly to Lost, skipping the intermediary steps.

And bear in mind that this is a dungeon designed for characters who beat Wizardry I. You can import characters who haven’t finished the quest and killed Werdna, but they still need to be fairly high level to survive the encounters in the dungeon here, especially the boss fights against clothing. So every senseless death represents the loss of a significant time investment. The game just wants to periodically deliver a big setback, to pad out the play time.

When the game was new, there was at least a little recourse: when something bad happened, if you were quick enough, you could eject the floppy before anything got written to it. I kind of get the impression that by Knight of Diamonds the designers were expecting the players to do this, and compensating for it. But it’s not an option under emulation.

Anyway, in addition to losing a character to greed and hubris, I managed to instantly Lose the rest of my party by carelessly teleporting into solid rock. One moment a promising party that I expect to last the rest of the adventure with only occasional substitutions, then poof. The shock of it is an experience you’d be hard-pressed to find in modern games — on purpose, at any rate.

As I’ve said before, Wizardry II: Knight of Diamonds would be treated as an expansion rather than a sequel today. You could easily imagine it as DLC. It’s not self-contained — it has no character creation system, so characters have to be imported from Wizardry I. It adds six new dungeon levels, and a whole bunch of new items and monster types (including wereameobas), but operates under exactly the same rules and magic system as the original. Where Wiz1 has a sizeable manual, the documentation for Wiz2 is a single page giving some new lore, a backstory concerning conflicts in the kingdom of Llylgamyn.

This lore is all but irrelevant. All you really need to know about your quest is this: You’re looking for the pieces of a very powerful outfit, and before you claim each piece, you have to subdue it. At the end of level 1, you have a boss fight against a suit of armor, and once you’ve beat it, one of your fighter-types can put it on. It’s extraordinarily good armor, and grants regeneration. Further levels will, if I recall correctly, involve fighting the helm, shield, and so forth. In other words, it’s the same basic idea as Kevin Wilson’s The Underoos That Ate New York!, but taken more seriously. Not entirely seriously, mind. Wizardry has always had a silly streak.

Unlike in Wizardry III, characters aren’t disempowered on import. You lose all your gear, which is a bit of a blow, but you keep all your experience levels and money — which means you can immediately go and buy the very best things available in the shop. Which aren’t the best things available in the game, of course. You can find better stuff in the dungeon. But still, it means my party is reasonably well-equipped and still just as overpowered as when they fought Werdna. So my progress through the dungeon is mainly limited not by the need to grind, but my desire to map the whole thing out. The dungeon levels are a great deal more satisfying than in Wiz1: more intricate and elaborate, and more pleasing to the eye when mapped out. It’s the usual effect in these older series: the first episode is made by people with no level design experience, the second by people with one released game under their belt.

I’ve got a story for you. It starts with our hero in a state of inaction: the next step forward is to take on Werdna, the wizard at the end of dungeon level 10, but I don’t have the courage. “Am I ready? Am I sure I’m ready? I remember seeing my parties get torn apart by this guy and his vampire minions in my first pass at the game. Maybe I should do another few grinding passes, just until I have a ninja in the party, or at least more than one character who can cast TILTOWAIT (the most powerful direct-damage spell). Just to be sure.” And so I dilly-dallied for days.

So I’m at the point where I can do a complete circuit of level 9, hitting all the rooms, and still have enough health and magic left to go several rooms into level 10. Level 10 is a series of seven twisty but branchless corridors leading to seven guarded rooms, the seventh being Werdna’s lair, which is clearly marked with a message outside so you don’t stumble into it by accident. Each room comes with an opportunity to teleport back home, even if you’re out of teleport spells, so it’s not a bad place to hunt for rare items.

Now, the rare items I’m hunting for always come in chests, and nearly all chests have traps; this is why it’s important to have a thief in the party. But even high-level thieves have a chance to fumble and trigger the trap they’re trying to disarm. One of the most concerning traps is the Teleporter, which sends the party to a random location on the same level — mercifully, it never teleports you into solid rock, but if you’re on a level you haven’t fully mapped out, you can easily wind up teleporting to a place you have no idea how to get back from. Reader, I hit a teleporter on level 10.

A smarter, or more paranoid, player than me would have cast a teleport spell straight back to the surface at this point. But I figured it wasn’t so bad, right? You can’t really get lost on level 10. I was in a room, and I didn’t know which room of the chain it was, but it was bound to have the exit route that all the rooms have. I took a step forward. Suddenly I was in an encounter with Werdna. The trap had sent me straight into the last room, skipping over the warning message. I hadn’t even healed up from the encounter that gave me the chest with the teleporter.

Fortunately, all my worries meant that I had put off the encounter long enough to make it downright trivial. My priest and bishop successfully silenced Werdna before he got a single spell off, then dispelled vampires while my mage spammed TILTOWAIT. I barely suffered a scratch. And now I guess I’m moving on to Knight of Diamonds.

Now, there are many versions of Wizardry I, and the one I’m playing today is not the same one I played in my youth. They’re both PC versions, but the standalone box I had back then had a very different UI from the one in the Ultimate Wizardry Archives CD-ROM anthology, the former displaying the first-person view of the dungeon in one corner and filling the rest of the screen with stats and messages, the latter keeping the dungeon view full-screen and overlaying temporary windows on top of it. I understand the full-screen approach to have been originally created for Wizardry III. And neither is completely identical to the Apple II version that preceded it, or the console versions that followed it. Some versions change the content, too. There are entire maps that are completely different on Nintendo. I don’t know why. Names of items vary wildly by platform, as if they had been translated by different localization teams, except that they were originally written in English. (Maybe they got translated into Japanese and back? It’s not as wild a supposition as it sounds; somehow, the series has long been most popular in Japan.)

Still, the Ultimate Wizardry Archive port is completely faithful the original maps, even when they seem like they must be wrong. There are portions of the maze that are inaccessible without teleporting, including part of the tunnels forming Andrew Greenberg’s initials on level 9, as if they left out a door and didn’t notice. There’s a sort of quest chain on levels 1 and 2 where you find keys to access other keys, culminating in a gold key that has no use. Apparently in the Nintendo port, the gold key gives you access to the elevator from level 1 to 4, just as the blue ribbon lets you ride from 4 to 9. Maybe that was the intention all along. As a child, I always thought of Wizardry as a class act, with its sleek and elegant black boxes and its substance-over-style design. It tarnishes this impression somewhat to notice not just such slapdash QA, but that the apparent mistakes went without correction in a port made 17 years later.

The version I’m playing now makes one change that I find really significant: in porting to the Wizardry III engine, it inherited Wizardry III‘s alignment mechanics. In the original version, you chose each character to be Good, Evil, or Neutral on creation, and that was pretty much it for them — I’ve seen claims that picking fights with friendly monsters provided a miniscule chance of turning Good characters Evil, but it never happened to me. A character’s alignment limits both what classes they can be and who they can adventure with, so this effectively meant you couldn’t have a party containing both a Lord (Good only) and a Ninja (Evil only). ¹Except by giving a Thief’s Dagger to a Neutral Thief, anyway. But Wizardry III made it so easy to switch alignments through your treatment of friendly monsters that the distinction became almost meaningless: anyone can adventure with anyone else, given enough time to arrange it. But at the same time, shifting alignment is randomized enough that it also became easy to split your party’s alignment, rendering them temporarily incapable of getting together again if they’re split up, either voluntarily, by quitting the game, or involuntarily, by getting someone killed and resurrecting them in town. This creates a pressure to keep acting your alignment all the time to avoid splitting the party. If you’re playing a Good party, you have reason to leave all friendly monsters alone, which is simple enough, but makes the Murphy-grinding slower; if Evil, you have reason to attack all friendlies, even the ones that are liable to cause problems. Wizardry I as originally released lacked such considerations.

I knew Wizardry I was grindy, but I had forgotten just how grindy. I’ve said before that the ideal balance in a CRPG is that if you just explore each area thoroughly, the encounters you have along the way give you enough experience (and/or loot) to keep going. The levels here fall miles short of that ideal. You run out of dungeon to explore long before you’re ready to face Werdna. I’ve been orbiting level 9 for a while now, waiting for rare equipment drops like it’s a gacha game. Those goodies are must-haves for the final encounter.

If it weren’t for that, I’m pretty sure that the Murphy’s Ghost back on dungeon level 1 is still a more efficient way to get XP, partly because there’s essentially no delay between encounters, but largely because of the complete lack of risk. The monsters in the lower levels of the dungeon have ways to undo hours of progress instantly: ninjas with decapitation attacks, level-draining undead. Your main counter to these things is to kill them before they can do anything, so high-level battles frequently come down to who goes first. And that’s affected by experience level. So it helps to grind Murphy for a while before you go loot hunting, but putting more time into your characters also raises the stakes.

I haven’t done this yet in the current go-round, but: Part of making an ultra-powerful party is changing everyone’s class at least once. This is an option you don’t see in a lot of CRPGs, even ones modeled on Wizardry. Changing class puts you back at level 1 and resets your stats to their racial defaults and reduces your spell slots to just enough to cast every spell you know once, but leaves your hit points alone. Mages have very low hit points for their level, making them a point of vulnerability for your party, but if you take a fighter with over 100 hit points and convert them into a mage, the result is a mage that can actually take some damage. Or, contrariwise, a high-level mage turned into a fighter becomes a front-row combatant that can do a few powerful spells in an emergency. There are two classes, the Lord and the Ninja, that can only be made by converting a character into their class.

Trying to make a Ninja the by-the-book way is an exercise in frustration, though. It requires a minimum of 17 in every stat, and your stats go up and down at random whenever you gain a level. What the manual doesn’t tell you is that there’s an easier way: with an item called a Thief’s Dagger, you can turn a thief directly into a ninja of the same level, circumventing the usual class-change process. That’s largely why I’m still cycling through dungeon level 9. I’m waiting for a Thief’s Dagger to show up.

Changing class the normal way has one other effect: it ages the character, presumably from the time spent training. Age is one of the basic character stats, albeit one that I don’t pay much attention to most of the time, and it has the peculiar property that different characters can age at different rates: send someone to rest off his injuries at the inn for a few weeks, and he’ll come back weeks older, while everyone else stays the same. (In practice, this never happens, because it’s both cheaper and more efficient to go into the dungeon where you can use healing magic.) Apparently characters will start spontaneously retiring once they hit their fifties, but this isn’t a serious limitation — characters age very slowly as long as you don’t waste time at that inn, and new characters can start as young as 14, which is a little horrible when you think about it. I mean, okay, that’s a pretty common age for a character JRPG, but JRPG characters basically never die. They just have their progress reset a little. Whereas first-level Wizardry characters are basically fed into a meat grinder by the mad overlord in large quantities while they’re still children.