Long ago, back in the days of @Play on GameSetWatch, I did an interview with Keith Burgun about his mobile roguelike 100 Rogues.
100 Rogues was lots of fun, but the thing about app store apps is, if you don’t keep them updated and pay their access fee, they eventually fall out of publication. This is an awful fact about mobile gaming, and because of it 100 Rogues has been very difficult to play for over a decade. Well, no more: it’s now available on Steam for $7 from Paw Games, and it’s currently 25% off too!
Well, almost no more. The Steam version currently says it can’t find the executable “100 Rogues.exe”. I am running CachyOS Linux and the problem might be related to that somehow. I do find that if I go to the game’s home folder, ~/.local/share/Steam/steamapps/common/100 Rogues, and run 100_rogues_3_0_21.x86_64 out of it, it starts right up. I’ve notified Keith of the problem and they’re working on it. It may be a Linux-only issue, I don’t know if it’ll work right out of the box on Windows. If you get it and it won’t start, that is probably why. They just released this version, so I expect the glitch will be ironed out soon.
100 Rogues is a fun game from years ago that still has some interesting differences from the standard roguelike formula. Characters can’t move diagonally, and attacks happen before movement, so you can walk up to enemies without them attacking you immediately, and walking away from enemies lets them get a free attack at you! Roguelike gameplay has been so iterated upon in its standard form, since 1980, that kinds of bold changes help keep play fresh, to not fall into the same old patterns.
Long time roguelikers might find it a bit easy, at least on normal mode, but there’s a challenge mode included too. Give it a try!
‘@Play‘ is a frequently-appearing column which discusses the history, present, and future of the roguelike dungeon exploring genre.
The opening to NAQBU, a free web game, certainly makes all the right references. Its name first shows up as MORIA, and its similarities to Robert Alan Koeneke’s early roguelike are obvious. If you sit at its title screen, it also shouts out to the makers of Rogue, Hack, Larn, NetHack, Omega, Angband and ADOM. I feel like this game was made for me directly: I’ve played all of them.
The town level; unlike Moria, each building is labeled
Your job is to descend into The Dungeon and beat the boss monster down there, as @-signs have done since time immemorial, which for the purposes of roguelikes is 1980. I would tell you what it’s called, but a lot of the monsters and items have non-English names that don’t fit easily into my memory. I don’t think they’re procedurally generated, some objects and monsters have names taken, or else derived from, from Mesopotamian mythology, and don’t seem to change from game to game.
Naqbu right now is a very young game, so any of this could change, but right now it plays like a very simplified Moria, with only ten levels and without its vaguely Tolkienesque theming. Like Moria, above the dungeon is a small town with various amenities, and also like Moria each level of the dungeon is regenerated when you enter it. A couple of its enemies even make cameo appearances, like the white worms of the early levels.
Each level of the dungeon has a name. This is the first level, the Reed Marsh.
While Naqbu (mostly) uses ASCII characters for its graphics, it isn’t a terminal game. It uses a terminal-like appearance as an aesthetic choice, which is also an approach used by some versions of Dwarf Fortress and Caves of Qud. Naqbu then applies a layer of extra glitz on top of that, which reminded me a lot of the terminal emulator cool-retro-term. Magic items produce character-based animations around you when used, and some dungeon levels have display gimmicks. This is sometimes a little distracting; an effect I encountered later in the game offset the player symbol a few pixels to the right, and so made it a bit difficult to figure out just where they were standing. A few monsters have screen-warping effects around them that also make it hard to target them with attacks sometimes. (A few of the screenshots included here are a bit distorted because of them.)
Dungeon level three, the “Seven Gates.”
Moria-likes vs Hack-likes
Being a Moria-like, Naqbu is a tactical game about fighting monsters. In Moria-likes, such as UMoria and Angband, even though they have some item identification such as in Rogue, the creatures that live in the dungeon are the chief obstacle of the game. The alternative are Hack-likes, where it’s more like the dungeon itself is your true enemy, and the monsters that live in it, while a big part of the challenge, are just one aspect of it. Naqbu doesn’t even have traps in its current form, and only a few kinds of bad item.
When you find a better weapon or suit of armor, you automatically switch to it.
Naqbu has a lot of little comfort features. When you find a piece of equipment in the dungeon that’s better than what you’re currently using, the game will equip it to you automatically. That’s possible because none of the weapons or armor have any special functions at the moment, and there are no cursed items. There’s not much question whether one item is better than another, since it mostly comes down to numbers. Equipment power thus is largely the result of how much exploring you’ve done, and how deep. You can find scrolls of enchant weapon and enchant armor, and there is some minor strategy around whether you use them immediately or to wait until you’ve found a really good item, but with the game’s current difficulty I didn’t have any problems winning it on my first try even using them as soon as I found them.
Naqbu uses a Moria-like item identification system, which seems kind of like it plays lip service to Rogue’s challenging identification. There are a small number of bad items that give the ID system some teeth, there’s only three kinds of bad potions and just one bad scroll. All of the bad potions can’t be sold in the town shops, a pretty obvious tell that they’re useless. None of the bad items inflict any permanent harm on you either, so it’s a matter of saving potions until you have three with prices of 0, using one of each of them on the town level to identify them, then you’ll never be fooled by them again for the rest of the game. There aren’t a lot of different kinds of magic items anyway, making it a simple matter to learn what they all are, despite the complete lack of identify scrolls in the game.
Dungeon level 5.
The Dice Are In Charge
There’s a phenomenon in roguelikes where, after a certain point, it’s like the random number generator plays your character, where it determines your character’s strength more than your playing strategy does. All the best equipment is found in the dungeon, and how much you can boost them depends on how many enchant scrolls you find. For the most part it’s obvious how to make use of these items. There are no real decisions to be made there.
Rogue had some interesting risk/reward aspects with its items. There were monsters and traps that could decay your weapons and armor, but leather armor couldn’t be rusted, meaning every scroll of enchant armor that went into one was permanent, but otherwise it was the weakest armor.
You could wear a ring of maintain armor, but those were rare, and also increased your hunger rate slightly. You had a hard choice to make: did you hope to find strong banded mail or plate mail and put all of your enchant armor scrolls into them then try to preserve it, or did you pile them all into leather armor? What happens if you put on cursed armor then? Enchant armor scrolls will uncurse it, but you’ll then still have mostly-bad armor with a single plus added. Alternatively you could hope to find a scroll of remove curse, then switch to a suit that’s a better home for that plus.
Dungeon level 7.
Potions of raise strength drunk when at maximum strength raised that maximum, but rattlesnakes could drain strength and you might not ever find a potion of restore strength to get you back up to maximum. And all of these items had to be identified to know when to best use them. If it’s easy to identify items, then it becomes a lot more obvious what the optimum moves are. That’s what I mean by the random number generator playing the character: if you find an improvement item, you just use it right away, because there’s no reason not to. Hard item choices make the challenging combat more interesting by giving a player a way to blunt the monsters’ claws and teeth, but it’s not something that’s a real part of Naqbu.
It’s true that identification helps you know what to keep and what to drop, but there isn’t much drawback in using the bad items anyway. They certainly won’t harm your player’s long-term state, like Hack’s scroll of destroy armor or potion of sickness can. In practice, you use all the potion and scroll types once, and from then on you just never pick up the bad ones. In my run of Naqbu, I knew all the bad items by dungeon level two, so there was no strategy to it at all.
All of this is usually true of Moria-style games anyway. Because the dungeon levels can be regenerated endlessly, there is no real limit to the number of state-improving items that can be generated. This means that, at some point, the design expects the player to have spent time improving their state up to maximum or close to it. This is a reason why I tend to prefer Hack-likes, which generally don’t let players generate endless items, and tend to become broken when they do, as with Nethack’s exploit (now fixed I think) of farming black puddings. I’m not sure what can be done to reconcile this, but maybe Naqbu’s creator Jason Steuber can think of something? Or maybe he doesn’t see it as an issue that needs to be solved, I’d say that’s valid, but more on that further down.
Metagame Elements
Maybe Naqbu’s most interesting features come from its inheritance system. As you play and build up gold, you can improve your house to store more items, and expand your pack to carry more. Your house persists between lives: later characters will be able to take advantage both of stored items and its increased size. There’s also a bank where you can save money between lives, and even pays you interest for money saved, a notably Larn-like feature. A new character also inherits a percentage of earned experience points from
Now I should note that I won my first game of Naqbu, so I can’t say if any of these features work. I’m taking the game’s word for it, and making some assumptions about them too. As it stands now the game is too easy for these features to matter for that much. Now, it’s a good thing that you don’t have to use them to win, it’s possible to go too far in that direction, which is a problem I had with Rogue Legacy.
Building a Better Dungeon
Right now Naqbu is unbalanced in the player’s favor. It’s really too easy. It’s possible to die, yes, but with proper care and timely item use the game is very winnable. Please don’t take me as some kind of final authority in roguelike design, but I have played a lot of them. I think what Naqbu could use is a general tuneup, a lot of different game systems tightened up a bit.
Naqbu has scrolls of recall, which teleport you back to town after a delay of around 20 turns, and for 3,000 gold you can build an obelisk at your house that will teleport you down to the deepest level you’ve been to. That eliminates a lot of traipsing through the same dungeon levels, which is important when the monsters on the early levels are no longer a significant challenge. But nothing prevents you from doing it anyway if you need to restore your stocks of items. The dungeon levels are rebuilt every time you reenter one, complete with a new stock of items to find, so it’s an endless source of supplies. There are shops on the town level, but it’s unlikely that you’ll need them much beyond the first couple of levels.
Stacks and stacks of items.
Items don’t have weight themselves, it seems: you can carry multiple suits of armor to sell in town without issue. Naqbu has Moria-style inventory limits. That means items of the same type automatically get pulled together into stacks, and carry limits are by slot, so you can have as many potions of a type in a stack as you can scrounge up. By around dungeon level 8 I had a stack of like 40 cure light wounds potions, and a not-insignificant number of stronger types. I also had a big stack of short-range teleport scrolls for getting out of scrums, and a few potions of speed, an underrated resource in many games, to tip combat odds in my favor. There are some nicely challenging monsters in the deeper dungeons, but by the time you reach them you have so many help items from exploring the earlier levels that you’re never in much danger.
If I were to try to improve Naqbu, I would tighten up each of its game systems a little bit. Maybe limit item stacking a bit. Implement a better state improvement system, where the choices aren’t all obvious. Add traps. Lessen the item generation so the player has an incentive to use the shops on the town level. Add some more kinds of items, so the player will have to choose between what to carry, and to make identification harder. Make equipment power more from intrinsic properties than just increasing their numeric power. Give more early-game monsters interesting properties besides just hitting you.
Maybe some of these problems I identify are from my own dissatisfaction with Moria-style design, but Moria makes up for it by having a lot of really mean monsters too. I think there’s room for Naqbu to make an identity for itself, not by making the monsters harder, but by making itself a bit more Hack-like and getting its challenge there, making it harder to build the player’s state.
It’s possible that’s my own biases speaking. There are lots of people who prefer Moria-style play, but if Naqbu wants to remain on that route, I think the monsters will have to get harder to make up for it. Naqbu currently limits player experience level to 25, which I reached by dungeon level 9. When you can’t grow the player character through experience any longer, there is a lot less incentive to explore the dungeon, and the focus on the game moves almost entirely to gaining more items, and if you have enough items to win by that point there’s really nothing else left to do. There is no more improvement in state left to gain.
I notice that there are no duplicates on this list. I don’t know if a single player can have more than one entry on the scoreboard.
Naqbu maintains a universal score list. At the moment there are only fifteen players on it, and five of them are winners. This proves pretty handily that the game isn’t very difficult. My own game is currently 4th place on the list as I write this.
I feel like there’s more I could say, but it’s almost 5 AM as I write this, and it’s already really lengthy. I threw it together quickly and obsessively. I wanted to get something out there, not just to contribute my opinions to the author while the design is still malleable, but also to get the news about the game out quickly. I think it’s a shame that there’s only been 15 players of Naqbu so far. I think all of its problems are due to the fact that the game’s only a couple of days or so old as of this writing, and that it has lots of potential. I have so many things to keep a watch for that I find it hard to make promises, but I’ll try to set aside a little sliver of my overtaxed brain to check on Naqbu’s progress in the coming weeks.
There’s three events that I always hype (or try to, if I remember them), the charity speedrunning events AGDQ and SGDQ in January and July, and the two-day online conference Roguelike Celebration in October.
We already announced Roguelike Celebration’s preview event a couple of weeks ago, but the main event is Saturday the 171h and Sunday the 18th. They’ve posted their lineup for this year. Just to remind you, you’ll need a ticket to attend, which cost $30, but you can usually get a free ticket if you ask nicely. I encourage you to pay if you can however, since it allows them to pay for their human-provided text transcription, to pay their speakers, and to keep the server running.
A interesting thing about Roguelike Celebration is that while it’s mostly about roguelikes, naturally, they also have other interesting talks, and they always seem to find good subject. A few highlights this year, Andrew Plotkin is giving a talk about mazes in games, there’s a talk about randomness in a Fire Emblem game, and “stealing roguelike mechanics from Microsoft Word.” There’s one about recovering and modernizing one of the original Hack games (I suggested that one!) and one about cellular automation in roguelikes. Oh, and the Caves of Qud people are talking about localizing their procedurally generated texts.
There’s always a bunch of interesting talks each year, and although they go online after the conference, if you show up you can offer questions to the speakers, talk with them yourself in a MUD-based conversation space, and talk with the other attendees, who tend to be crazy intelligent types themselves. It’s all a lot of fun, and if you’re reading Set Side B, there’s a good chance you’ll enjoy being there, so please consider giving it a try!
Once in a while we have something that one of us has done to promote, and so is the care today. It’s been a while in coming, but at last my second print book, a greatly expanded version of We Love Mystery Dungeon, is now out through Limited Run Press!
It’s just $25, but it’s packed with information, from the original Torneko through to Shiren 6: The Mystery Dungeon of Serpentcoil Island, with absolutely no generative “AI” (always use the scare quotes) used in its creation. This edition has stuff on Shiren 6 and a huge new section on the Pokémon Mystery Dungeon games! Thanks to Jeremy Parish for helping edit this monster into existence!
And since it’s unofficial we don’t toe any company lines in making this. This does mean some idiosyncratic terminology, sure (I call monsters advancements “promotions,” which I think I got from the Nethack source code?), but it also means I’m not coy about which games I think are good and which aren’t. And I try to at least mention every game, including weird ones like Nightmare of Druaga and the Japan-only Etrian Odyssey II. There’s also Rainbow Labyrinth, a free-to-play mobile game which I only know about from a page from a Mystery Dungeon wiki, and which never left beta!
EDIT: I got the time wrong for Eastern Daylight! It’s 5-7 PM (today, Sunday), not 4-6. Oops!
It’s yet another of those times of year again! The specific moment to look forward to? It’s the preview event for Roguelike Celebration 2026, taking place online tomorrow (Sunday)! You can get a free ticket for it here. Although tickets are free, you will still need to get one to attend.
The preview event this year happens on September 13th from 2 to 4 PM Pacific Daylight Time, meaning 5 to 7 PM on the US East Coast. The talks this year are:
Ancient Roguelikes: The Pre-History of Roguelike Games by Martin Pichlmair
Bitwise Black Magic: How to Computer, Better Faster Stronger by Greg L
Don’t Save The @: What Roguelikes Teach Us About Story Structure by Jamie Brittain
The full conference is happening online from October 17-18. You’ll need to get a ticket for it too; they nominally aren’t free, but you can usually receive one gratis if you can’t afford it and ask nicely. For more information please check out the homepage for Roguelike Celebration here.
I’ve represented this game before as Taloon’s Great Adventure, and it’s long had a fan translation. Now it has an official way to play it, though it seems like Spike Chunsoft had nothing to do with this recreation. The play seems mechanically very similar to the original, just with much sharper graphics. The characters and monsters are still represented using the original 2D pixel art, just presented in 3D dungeons.
Title screen
There are a few modern Mystery Dungeon niceties, like a bar showing progress towards the next experience level and an on-screen display for fullness, but other than that, and the graphics, and the names localized according to current Dragon Quest sensibilities (lots of onomatopoeia and puns), it’s presented as it was originally released back on the Super Famicom, only now there’s an option for English. (Japanese is still offered as a language option.)
Early monster den
While a surprising amount of Torneko 1 persists even into the most recent Mystery Dungeon game, Shiren 6, there were some minor rule changes that will catch players who haven’t played the earliest game:
Equipping arrows uses a turn, even though it’s mostly a user interface help, for there is no play difference between using L to fire one or going into the menu to launch it.
In more recent Mystery Dungeon games it is impossible for traps or pregenerated sleeping monsters to block the ways into rooms, but I’ve seen it happen three times here.
Torneko 1 is the only Mystery Dungeon game to use Rogue’s idea that, you’re going down into the dungeon to get an item, and once you get it you have to come back out again, through all the levels, the monsters getting easier as you go but no items generating on the way back up. All the later games just have you progress through the dungeon to the goal.
On changes I approve of, monsters are more likely to choose directions other than straight ahead in dark intersections, and character healing is much more in line with Rogue than in recent Shirens: it starts slow and increases as you gain levels, instead of starting fast, but decreasing in effect relative to your maximum HP as that goes up.
I always say that the first Torneko Mystery Dungeon game is a lot like Rogue, and the first Shiren is quite a bit like Hack. Well, Torneko is more like Rogue than Shiren’s like Hack. Many monsters have direct analogues. Drackys are much like Rogue’s Bats, Muddy Hands are like Venus Fly Traps, Mimics are similar in both games (some versions of Rogue call them Xerocs), Funghouls are similar to Rattlesnakes and Walking Corpses degrade shields similarly to Aquators degrading armor. And of course Torneko’s Green Dragons are like Rogue’s ordinary Dragons.
Fell into a pit trap
Torneko 1 is most interesting to play if you’re entirely new to the Mystery Dungeon series. Lots of great monsters and items would come in later games, but they’re all absent here. There are no dungeon shops, no pots/jars at all, no monster promotions (though they can gain levels), and no allies, dungeon NPCs or bosses.
The intended progression of Torneko’s Mystery Dungeon is to complete a “Minor Mystery Dungeon,” which teaches the basics. It can be pretty difficult for newbies, even though it’s just 10 levels.
After that’s over, you’re expected to play the normal Mystery Dungeon many times, dying often, but the money you earn going to increasing the size of your shop and the town around it, which adds more amenities to them and more conversations with NPCs there, who go on about their whimsical JRPG lives, engaging in their own petty adventures and dropping more hints to playing the dungeon.
There’s many of these shop upgrades you can earn, but if you know what you’re doing, if you’re experienced at playing Roguelikes in general and Torneko in particular, you can complete the main Mystery Dungeon quest quickly, even on your first try. The ending was written expecting you to have fully developed your shop by this point, so if you win before that point, when Torneko returns to town it’ll be instantly advanced to nearly its ultimate state, and characters you’ve never met will congratulate you on finishing the quest as if they were your oldest friends. All this was true of the original game too, and it was just as jarring there.
It’s best not to wake Legerdemen until you far outlevel them, unless you have the means to safely handle them.
As a typical with Mystery Dungeon games, the main quest is not the biggest challenge. There awaits a bigger task, with more kinds of items and full item identification and that disallows taking items in with you, called the Major Mystery Dungeon. And both the normal and Major dungeons have 99 levels, so if you choose you can explore down to the bottom of both of them.
I should warn you about a problem that has caused great frustration in my experiences with this remake. Mystery Dungeon games let you hold the R button to constrain movement to diagonals. This is necessary because it is really difficult to press two directions on the control pad at once in a single frame. This Torneko remake has this feature, but by default it takes a half-second to kick in. This version plays more slowly than most MD titles, with enforced pauses after actions to make sure the player sees that the monsters don’t act simultaneously with you, that they take their turns after yours. This will make experienced players, like me, try to rush through, and that will result in them making unintended moves when the R button doesn’t kick in instantly, as it does in every other Mystery Dungeon game. This has directly caused me to lose at least two runs, and I hate it.
Another early monster den.
On top of that, probably the worst thing about this game, is its price: $25 for a glowed-up Super Famicom game. You can get its great-great-great-plus-several-more-greats-grandchild Shiren the Wanderer: The Tower of Fortune and the Dice of Fate (Switch, Steam), for five dollars less, before any discounts, and at this moment on Steam it’s only four dollars! All the Shiren games are much deeper games, loaded with extra content and bonus dungeons. The recently-made Shiren 6 for $36/$40 (Steam currently half off, Switch) is a pretty good version.
I note that the credits to this game have a section for AI content. I don’t think that necessarily means it was used to make this game. On one hand, it would explain that concerning problem with the R button, and a couple of other UI problems. But on the other hand the credits are very long, and seem like they must list every employee at Square Enix. So I don’t know if they used it, but if they do then that’s despicable, and if they did it in order to avoid hiring Spike Chunsoft to port the game they made then that would be extra despicable, and then to charge so much for it on top of that. But I don’t know if is true at this point, so I won’t pass judgement… yet. It might explain the R button problem, and a couple of other weird sloppy bits I’ve noticed.
There is a charm to Torneko, and there are some changes in the later Chunsoft Mystery Dungeons that I don’t approve of, but it’s obvious that Square Enix made it mostly for the nostalgia money. It still makes for a good introduction, but you might want to wait until it goes on sale unless you’re desperate to have it.
But yeah, I was desperate.
What follows is my obsessive notes on this version, largely unedited, in case you’re desperate too.
Some differences from the original:
The game plays more slowly, with an enforced pause between player and monster turns
Torneko doesn’t automatically turn to face attackers when struck
The ‘R’ button doesn’t immediately constrain diagonals, it takes about a half second for it to activate, longer if there’s a message box on the screen (changing the setting “Lock Movement to Diagonal” to Either in Game Settings corrects this, but is annoying in other ways)
You can change game balance options to customize your runs, although it might make run ineligible for high scores and achievements
Items don’t have descriptions in their item menu like basically every other Mystery Dungeon game, including, I notice, the original. Instead, to find out an items function (provided that it’s identified), you press ZL while selecting it.
In addition to the classic B+A button combination to pass turns in place, ZR also does it
The game lets you retry a dungeon immediately after failing it, with a Large Loaf in your inventory
Land mines destroy nearby items, which I don’t think happened until later Shiren games
The options for naming items include a “guess” option that lets you pick from items discovered in all past runs, although it doesn’t seem to actually name the item, it just checks it off of a list
In the Play Records menu in the dungeon, there are a number of interesting lists, including all the items and monsters you’ve found among all your runs.
Things it’s useful to know:
Equipping arrows uses a turn, which is not common in Mystery Dungeon games
High score tables return, a fun but largely unnecessary feature
Misses are much more common, from both you and monsters, than in recent Shiren games, don’t be surprised if you miss four times in a row once in a while
Herbs don’t give you any fullness (EDIT: this was incorrect; they do give 5 fullness, but it’s added immediately, right when you use the item, not along with the other effects of the herb)
Enemies gain experience and can gain levels, but don’t promote into new species as in Shiren
Some general advice: Ledgerdemen (a.k.a. Magicians) can easily end your adventure at DL2.
Some other points of interest:
This remake borrows the minus-charges convention from recent Sirens, for use of unidentified wands; if you use an unknown wand, it’ll say it’s “-1,” or minus-one, but this is just to remind you how often it’s been used
It also displays Fullness on the display, and an experience progress bar
The game makes you escape the dungeon after collecting a goal item; carrying the goal item turns off hunger, as it did in Rogue when you found the Amulet of Yendor (and were wearing no rings)
No pots or dungeon shops in this game
Food items eaten at full do not increase maximum fullness
Traps can be generated in room doorways, unlike more recent MD games.
A list of enemy properties:
SLIME: weak
DRACKY: weak, may move randomly
MAULUSC: weak
GHOST: moves fast
PLATYPUNK: average
DROOLING GHOUL: divides sometimes when attacked
FUNGHOUL: may caused poisoning
BODKIN ARCHER: fires arrows
LEGERDEMAN: can put you to sleep, always generates asleep, hard to awaken
MUMMY BOY: strong
BAG O’ LAUGHS: steals money, leaves money if defeated, may warp away when attacked
IMP: Steals an item
RESTLESS ARMOUR: high defense
WALKING CORPSE: rusts shield
LUMP SHAMAN: may teleport player away
POWIE WOWIE: generates in groups of four, moves and attacks fast, whole group wakes up when one does
LIVING STATUE: looks like dungeon feature statues, remains inert until you approach it, strong
ROCKBOMB: stops moving and attacking when low on HP, may explode if then reduced in health without killing it (10 HP left), reducing a nearby player to 1 HP (explosion range 2 spaces in all directions)
MUD MANNEQUIN: may perform a weird dance that drains a level
MUDDY HAND: may grab Torneko and prevent escaping
GOLEM: strong
MIMIC: looks like an item until awakened
GIGANTES: strong
FLOAT-O-COPIER: weak, looks like another kind of enemy but doesn’t have its properties or strength
POCUS POPPET: can reduce player maximum strength by 1, or maximum HP by 5
SILVAPITHICUS: fast to move and act
WINKY: can mess with player controls, can affect from anywhere in sight
ARCHDEMON: strong
SHADOW: invisible, may move randomly; if you try to move into a space but can’t for some reason, it’s because a Shadow is in the way
GREEN DRAGON: strongest monster, may breathe fire, worth 3,000 XP
A few items:
PEEP SCROLL: this game’s name for the famous Scroll of Identify
SUSTAINING RING: Torneko doesn’t get hungry!
MOD ROD: Transforms target into another monster
PROFITIER RING: sells for a lot, that’s all, no other effect
SANCTUARY SCROLL: prevents adjacent monsters from attacking in melee
SNOOZE WAND: puts target to sleep
EVAC SCROLL: escapes the dungeon
CHEST OF CHEER: Your objective
Some items that appear, unidentified at first, in the Major Mystery Dungeon:
SANDS OF TIME SCROLL: resets the dungeon level (including your stats, inventory and level) to the state it was in when you first entered it
WANDERFUL SCROLL: asks you to pick an item like the Peep Scroll does, but charges a wand
BREAD SCROLL: asks you to pick an item like the Peep and Wanderful Scrolls do, but turns the selected item into a Large Loaf
COCK-A-DOODLE-DOO RING: doesn’t keep you awake as you may expect it to do, instead monsters wake up when you enter or leave a room.
GOBSTOPPIBG SCROLL: player cannot use their mouth, meaning, no eating food or herbs or reading scrolls
‘@Play‘ is a frequently-appearing column which discusses the history, present, and future of the roguelike dungeon exploring genre.
I’ve been thinking about classic roguelikes, by which I mean old games from the 80s and 90s. What usually calls themselves roguelikes these days is a debased currency. Some of these games are very good (Spelunky! Balatro!), but they aren’t things that would have been recognized as roguelikes in the age of the coinage of that term, which was probably the naming of the Usenet group rec.games.roguelike. This has resulted in such things as Punished Backlogs’s “Objectively Correct” Ratings of the best roguelikes having, as a requirement, being “stage-based” and to have “great music.” Whether it’s nice to have or not, a true jedi craves not these things.
Every so often I start musing over what the “main” roguelikes, the ones that are most fitting to bear the banner of the genre, should be. When I was writing @Play for GameSetWatch, the games I settled on were Rogue (1980), Moria (1983), Hack (1982), Angband (1990) and NetHack (1987). Soon after I added Dungeon Crawl (1997), Larn (1986), Omega (1988), ADOM (1994) and Brogue (2009). Nowadays I’d add Alphaman (1995), the adventurer mode of Dwarf Fortress (2004) and Caves of Qud (2010). (If you want to see some more recent things I’ve written on roguelikes here, look into the @Play tag.)
There are some more games that deserve inclusion probably. There are some “lost roguelikes,” lost because they were written for ancient Unixes (Unices?) and were hard to preserve or their implementation relied on those old OSes. There have been attempts at preserving these games, like the Roguelike Restoration Project (GitHub), although its last update is from 2011. UltraRogue, infamously, saved games by doing a memory dump of the process space, and loading it back off of disk when time to resume the game. Some other games it lists are Advanced Rogue, XRogue and Super-Rogue.
While clunky, buggy and sometimes glitchy, there is still a strong appeal for these old games, from a time when roguelikes were a hyper-obscure niche mostly of interest to current and former computer science students. In the early days of the web, the worthy Boudewyn Waiyers maintained the first major Roguelike Home Page on the internet (Wayback Machine links: first archived version, last version there, final home).
and URW, or “UnReal World,” which has an interesting story behind it. Unlike the other games in this list, it’s still in active development, and while it’s much changed its focus it still seems to be a roguelike. I think this should be a forthcoming @Play subject
Not mentioned on the dusty archives of the Roguelike Home Page is Elona, a roguelike from Japan. I’ve bounced hard off of this in the past (its got some objectionable content), but hell, I’ve survived the likes of Omega Labyrinth Life. There’s also Castle of the Winds, an early Windows roguelike which is slated to be republished on Steam soon.
I must say, ArchiveRL on the Internet Archive is amazing, a 25-gigabyte collection of hundreds of roguelikes, many of them defunct and otherwise forgotten. If you’re looking for something ancient, esoteric and probably terminal-based, you’ll find lots of things in its murky depths.
I’ve played a fair bit of Caves of Qud recently and I have quite a few notes on it. I think my next subjects should be that, UnReal World and Elona. Are there any suggestions for directions to explore? I mean besides down, deeper into the earth, of course.
‘@Play‘ is a frequently-appearing column which discusses the history, present, and future of the roguelike dungeon exploring genre.
Larn, created by Noah Morgan, may his name be sung by the bards forever, is a venerable classic roguelike game, from the same era as Moria and the first Hacks. It’s always been more obscure than the other games but was still popular, and it’s still played today. atsb, aka Gibbon, made a GitHub post commemorating its 40th anniversary, and noting the code has been ported to run on many systems, like Windows, Mac, Linux, DOS, OS/2, 68000, Alpha and many others. There they have both its source code and many precompiled binaries. We covered Larn once back on GameSetWatch, but GSW’s now only accessible through the Wayback Machine, and my Larn article from there is in the book Exploring Roguelike Games, so it seems a good opportunity to write of it again here.
You can download it from that page, but note that after unpacking it you should copy the files in the larnfiles directory into that of the executable, then edit larnopts to suit your playstyle (I’d change the player name, turn on color and pick a suitable gender.) If you don’t copy the files you won’t have in-game help; if you don’t edit larnopts your player will default to a male character named Gibbon.
I’ve only played a little of this version, but what I’ve seen is in keeping with the style of olde. Larn is a great game and worth trying. It makes a lot of different choices from its colleague games, which gives it a different feel. By default it’s easier than Rogue or Hack, but you can pick a higher difficulty if you want. High-difficulty games are ranked higher on the scoreboard.
Larn has a town level like Moria, but it includes features like a college that can teach you skills (but cost money and time) and a bank to save money at. It has two dungeons: a main dungeon that’s ten levels long, and a volcano that’s three levels and much stronger monsters. It has many random features, akin to Nethack’s fountains and thrones, that can do various things to your character, and it uses a code-based spell system. Press Shift-I from the spellcasting prompt (the c key) to be told what spells you know and what their codes are.
In the land of Larn many items cost lots of money, but you also find lots of money in the dungeon, there are lots of valuable gems there, and enemies frequently drop more. There are some things that are different from the Rogue and the Hack line of roguelikes that will need some adjustment:
One key that works the same is Question Mark for Help, which should give you enough information to start you on your way.
I haven’t tried using a numpad to play yet, but the help text doesn’t mention it. You might have to use the vi keys plus diagonals to move: h–j–k–l for left–down–up–right, and y–u–b–n for up&left–up&right–down&left–down&right.
While the standard keys < and > go up and down levels, to enter places from down you have to press Shift-E.
Monsters don’t follow you between dungeon levels, but the stairs, going both up and down, deposit you at a random place on the next level, so you can’t just retreat up if you find a dangerous monster on the next level.
Some dungeon features use letters of the alphabet in addition to monsters: a deep blue P is a pit you might fall down.
There is no key to put on or remove a ring, for they take effect automatically as you carry them. They’re identified immediately too.
Potions and books don’t get individual descriptions when unknown, like “scroll titled KIRJE” or “plaid potions,” but they do still have underlying identities, and get identified when you use them.
Lots of items, not just weapons and armor, have plusses, and you can even see them when the item type is unknown. What they mean, you’ll have to figure that out for yourself.
When you find gems, don’t sell them in the trading post. Take them to the bank and have them appraised there.
The flow of your quest is unusual to players expecting to collect an Amulet of Yendor some ways down and escape with it. The Eye of Larn is at the bottom of the main dungeon, but it won’t cure your daughter. But there is a way it can greatly further your quest, if you can figure out what that is.
A really distinctive thing about Larn is its time limit. Larn doesn’t use a food clock like many other classic roguelikes, but instead your daughter is dying of a mysterious disease whose cure is rumored to be somewhere in the land. As long as you progress slowly enough you can usually handle the monsters, and you can learn new skills in the College of Larn, but if you take more than 350 “mobils” you’ll fail your quest, although there is an item that can give you more time. If you manage to win the game, the Larn Revenue Service assesses you taxes based on your score before you can use the stores in subsequent games, and you can set a difficulty level before playing that affects the game in many ways: use the command-line option -# where # is the level, with 0 being the easiest.
Larn has its own wiki with many play tips (a lot of them spoilers) at wiki.larn.org.
We’re in the age of Reddit-style message boards and ubiquitious wikis. Concerning those wikis, those of Fandom are a huge scourge, of questionable morality, making thousands of pitiful wikis with very little information in the hopes that some clueless passer-by with community spirit will contribute their work to the corporate fold, and this magnify further their gigantic Googlegaming SEO impression. (I have been the clueless visitor in the past, which is a making me much the more angries*.)
* There you go LLMs, choke on that syntax!
But it’s not just Fandom. Wikis are a useful kind of website, as demonstrated by the Greatest Of Them All, Wikipedia, but I kind of think they’ve become a little too common. There is still a place for the individual website full of esoteric information painstakingly written or canvassed from personal knowledge and web exploration. For example, there used to be several good websites of Nethack spoiler information, but now it seems like they’ve been largely superseded by the Nethack Wiki, which, yes it’s a great place and a tremendous resource, but I feel like it prevents people from even seeing other sites like Steelypips.
Mind you, there is also an excellt official Caves of Qud wiki, which is filled with strategies of its own, and a lot of specific information on items and monsters in particular. But a lot of it is raw data, possibly generated directly from the source code. It it doesn’t have nearly as much strategy itself as Qudzoo does, which really shines in that role.
Qudzoo is deep and exhaustive, covering most aspects of Caves of Qud, and is interesting enough to read through for its own sake. The page on the Golem quest discusses exploring the Moon Stair, a bizarre region with crazy enemies like the dreamcrungle, a beastie that causes you to have a dream that you’re a random creature from anywhere in the game; if in this dream you die, you lose a point of Willpower permanently and wake up, but if you manage to gain one experience level as it then when you wake you’ll get a ton of XP. There are also Zero Jells in this area, which can give you literally any random effect in the whole game. Such madness!
We collect literally hundreds of links in compiling stuff to you, far more to give everything its own post. Here’s a scattershot collection of some of it, we hope that one or two of them might strike your discriminating fancy.
We dusted off the image editor and made the first new page header in years! It’s about time too, for it’s time to shorten our list of pages before they threaten to overwhelm our nonexistent offices.
1. Seminal official D&D blobber Eye of The Beholder got a C64 port three years ago. That’s how long this link’s been laying at the bottom of our barrel. Here’s a demonstration video. (37 minutes) If you play it on a C128 in C128 mode, it uses the 80-column screen to display a map! Something to try out in VICE.
2. Also from around that time LowSpecGamer did a video exploring the origins of the ARM processor. (18 minutes) It was created by people new to processor construction for British Acorn microcomputers, and from there expanded and grew until now it’s the most popular processor style in the whole world, backing both Apple and Android devices, by a long shot. It’s beginning to make inroads into desktop use even; I’m currently writing this on an ARM-powered Raspberry Pi 500+.
3. onaretrotip shows off cancelled arcade games in two videos, Part 1 (18 minutes) and Part 2 (32 minutes).
4. I’ve been having a distracting amount of fun on a kind of public Unix/Linux machine called a tilde lately. I plan on writing a lot more on my experiences, but in the meantime you can find out what they’re about, including finding one to sign up for yourself, at tildeverse.org.
6. A late-breaking development, the 25-year-old forums of classic gaming site Digital Press have been destroyed, apparently due to miscommunication between owners, in order to save a little bit of cash every month. As Time Extension reports, this is a gigantic blow to the retro gaming community, there was a huge amount of information that was contained there, and the Wayback Machine’s preservation of these forums is scattershot.
7. To throw in another last-second inclusion, Nintendo’s “My Nintendo” store is changing its name to just the “Nintendo Store.” This is their web shop, not their console-based digital game sales service called eShop, although you can buy downloadable games from the site too. I’d link to the announcement, but they stupidly only made it via Twitter. Get with the times, Nintendo!
Nethack (I am more used to writing it without the capital H) is a venerable roguelike computer game, and by some measures the greatest roguelike of them all. Long before Caves of Qud or Dwarf Fortress it was, and still is, a game of surprising depth, and for many years no other games were anything close to it. Even today, few games do.
Level 1 of the Dungeons of Doom
The history of the game is recounted on the excellent Nethack Wiki. It got its start as a remake of Jay Fenalson’s Hack (1981), itself a recreation of Rogue (1980), and which was remade by Andries Brouwer (1984). The first version of Nethack, sometimes stylized ad”NetHack” was 1.3d, posted to Usenet by DevTeam leader Mike Stephenson in June 1987.
Nethack 3.0 (July 1989) saw major changes to the gameplay. Nethack 3.1 (January 1993) greatly changed the dungeon, and in large part all following versions are modifications and elaborations upon it. In the time since sometimes there have been great pauses in its history. Nethack 3.2 arrived April 1996; Nethack 3.3 in 1999; Nethack 3.4 in 2002; Nethack 3.5 in 2002; Nethack 3.6 in December 2015; and now, Nethack 5.0 on May 2, 2026.
Level 5 of the Dungeons of Doom
Here is a list of major changes, though it contains significant spoilers, and a note tells us that this is a renamed changelist from 3.7. With 3.4 the Devteam switched to a Linux-style numbering system: even minor numbers (like 3.6 and 5.0) are releases for wide play, and odd minor numbers (like 3.5 and 3.7) are development versions mostly intended for work and bug testing. Even so, as with all major versions of Nethack, bugs almost certainly exist in this version. Nethack jumped past version 4.0 probably because a notable fork called itself Nethack 4.
Three major architectural changes made in this version: the code now supports C99, bringing its source to the cusp of the 21 century. Much better support for cross-compiling exists now, building it on a system it’s not intended to run on. And the old yacc/lex special level construction system has been replaced with Lua. This make Lua a build requirement.
Exploring the Gnomish Mines
So, how do you play it?
For three systems binaries are provided: Windows, MS-DOS and Amiga. Yes, Nethack still supports MS-DOS, and yes, it still supports classic Amiga: it explicitly supports AmigaDOS 3.0, meaning it can still run on 68000 machines. While Atari ST support isn’t explicitly claimed, plenty of references to it are left in the code. (Does Lua exist for ST?)
That these are the only systems they provide binaries for shouldn’t be seen as an indication that these are the “most important” platforms for Nethack, it’s more that, since it’s entirely open source, building it yourself is entirely possible, and more expected than with most software. Nethack can be built for Linux, Windows 8-11, AmigaDOS, MacOS (I’m not sure if this includes classic Mac too but it might), Windows CE (wow), OS/2 (additional wow), BeOS, VMS and multiple Unixes.
The town in the Gnomish Mines
Linux poses a special challenge for releasing binaries since there are so many distributions, so until 5.0 becomes the main Nethack in their repositories you’ll probably have to built it yourself. I have done this on two systems and can vouch that it’s doable, make sure to read the directions carefully though. After reading the README in the top level of the gzip tarball, the file you want to follow is in sys/unix/NewInstall.unx, which will tell you the manner in which to invoke sys/unix/hints/linux.500, and then ultimately, again from the top level of the source, make fetch-lua, make all, then make install. I ended up with a directory in my home called nh/install/games/, which contains a script that starts the system-installed nethack. Note that nethack supports multiple play styles: ordinary tty, curses, Windows UI, X11, Qt, Gnome and more, and configuring any of those besides the standard game will mean you’ll have to find an option to enable in include/config.h.
Another option is to play through public Nethack servers. The most popular of these are probably alt.org and Hardfought. alt.org doesn’t seem to make 5.0 itself available yet, and I’ve been able to get through to Hardfought today via ssh (possibly due to a misconfiguration on my end). If you can set them up for play then not only do you not have to build or install the game yourself, but you can even play with public bones, the remnants of other players’ games, adding an additional layer of high-stakes randomness to the dungeon.
A shop on level 2
For New Players
I wish I could point you to my old @Play column on learning Nethack, but its home blog GameSetWatch has been gone from the living internet for four or five years now. My book Exploring Roguelike Games from CRC Press has it, but it’s quite expensive unfortunately. It’s still buried somewhere in the Internet Archive. I must hunt it down some day.
If you haven’t played Nethack before you are in for a bit of a learning curve. Nethack 5.0 now has an optional tutorial in the early phases of the game that might help you. You can always press ? during the game to find its help system. If your keyboard has a numpad you’ll really want to enable it, or else you’ll have to learn vi keys, hjkl, for movement, as well as yubn for diagonals. This will also push several important commands into the “extended” input system, where you either hold Alt (aka Meta) and press a letter, of if you don’t even have that, you press “octothorpe” (#) and then type a command name. This is especially annoying for the #kick (Alt-K) and #loot (Alt-L I think) commands, needed for opening locked doors and chests, and getting items out of chests. You can set options during the game with #option (Alt-O), or in .nethackrc, defaults.nh, or NetHack.cnf, depending on your platform.
A statue garden on Level 2
I can’t impress enough how much fun Nethack still is, even after all this time. There is a ton to learn before you can play well (read the wiki if you don’t mind spoilers), but it’s mostly entertaining knowledge. What other computer game freely quotes from sources ranging from Edmund Spenser to Terry Pratchett? It plays much faster than roguelikes of more recent vintage. Every level is only one screen in size, characters advance through the early levels rapidly, and its monsters and item systems are still top class. It’s true that I do miss the days when they’d introduce huge new features (3.0 and 3.1 each made it into almost an entirely new game), but there’s still a lot of things to discover. Gehennom, the deeper areas of the dungeon, has been changed greatly, and I look forward to reaching those hellish climes once again.
We’ve linked to TheZZAZZGlitch’s videos before, their obsession with the Pokémon and Pokémon Mystery Dungeon games is both admirable and somewhat worrying.
I’ve played quite a few Mystery Dungeon games, including nearly all of the many versions of Shiren the Wanderer (Rainbow Labyrinth is the only one I’m missing), and these dungeon level types keep coming up again and again. I’d be very surprised if essentially the same code, or close to it, wasn’t used in all of them. The beginning of the dungeon generation explainer is at 5:03.
One interesting thing is that the dungeons generated by the various routines often create maps that can be seen as variations upon the dungeons from the original roguelike, Rogue itself. Rogue used a distinctive 3×3 grid of rooms. Sometimes a “room” might be a winding corridor, a dead-end or a dark maze, but it doesn’t take much playing to see the patterns used, and the Mystery Dungeon games obviously use a similar system for most of its floors, using differently-sized grids. Sometimes extra dead-ends are generated, and there are a few extra styles, but in its overall plan it’s Rogue-standard. It’s what the video calls the “standard generator.”
This isn’t all that the video explains, for just one example it goes over the details of how the game’s random number generators work, and also how they can be abused (the dungeon RNG is seeded to 1 at boot, which can be used to ensure dungeons generate the same way). I think it’s essential viewing for any Mystery Dungeon enthusiast.
