|
|
|
|
|
|
|
|
|
|
Writing
The Importance of Absurd Detail
Secret and Public Lore
Beyond Good and Evil and Stuff
Open Ended Storylines and Player Subplots
Content Beyond Powerlevelling (Shadwolf)
The Merits of Mindless Slaughter
Repeating Quests
Multiple World Plot Development and Contingencies
Dungeon Design
The Importance of Absurd Detail
Detail is what makes or breaks immersiveness. Adding little things to a game world (pen or computer based) that make it seem a little more "real" make it more pleasurable to the player. The less detail there is, the more you break the illusion, and the more you remind the player that he is playing a stupid game.
Sometimes, the level of detail in the pages seems completely excessive and unnecessary. Does it really matter to the average player of your game how good your basic grain yield is per acre? Does he care about the details of your ficticious history about the rise of the merchant class? Or the differences in a sword forged by standard techniques as opposed to pattern-welding? Probably not. He just wants to know what the best weapon is, and where the monsters are. However, they are important to some players, and it gives you a stronger basis for logical story evolution and lore.
It also helps you deal with questions from the players and unanticipated actions. When players start asking you about why they should care about defending the outlying farmlands, you can tell them that they will starve if they don't and why. If they want to know why their orc loot isn't netting them much (if any) coin from resale, you can explain that orcs use bronze and stone for weaponry, and why they don't have readily available high-quality iron and steel. And of course, you can provide basic information that explains to the players exactly why they have to go kill them in the first place, without shrugging and saying, "Duh, it's a role playing game, of course you go kill the orcs! That's what they're there for!"
There is also a certain "cool" factor associated with mad levels of detail. The first time one of your players discovers that the monster spawns actually migrate in a logical fashion, and react to massive defeats at the hands of players by relocating, they can think, "Wow, that's cool." If they can actually watch and supervise the construction of their house by hired carpenters instead of just double clicking a deed, that's cool too. Attention to the most minute of details helps to engross your players into the environment, maintaining their interest. Minute detail sets your game world apart from other game worlds with comparable fun factors and interfaces, and you can enjoy the corollary benefit of players who appreciate this using it as ammunition to flame your competitors, an indirect act of aggressive consumer-based marketing.
One note about extensive detail: Extensive detail is especially important when constructing the lore of your world, making sure that it all makes sense, it follows a logical progression, and most importantly is well-written. However, the general lore of your game setting is only initially released in small, broad amounts, like a basic timeline and common knowledge. The detail exists in-game as layers to discover and ultimately disseminate through the internet, but that's okay. The fact that your history makes sense, and that lore-minded players can research it, is another avenue of play, something that is very important.
Some of your game world's story, or lore, is going to be hidden from the players, as it should be. This gives them something to discover and be interested in, if such is their bent. However, one can go too far in hiding the lore of the land, and ultimately overzealousness in concealing the story makes players give up on it altogether, especially if your lore is particularly stupid.
In order to maintain player interest in your lore, it must be available to them to some extent, in an easily accesible public form. The best way to do this is to incorporate the lore into the game itself. Posting the lore of your game on a web site or some other external medium as a sole form of information only serves to remove immersiveness from the game. Most players won't even bother to do this, as the lore has no direct impact on the play of the game. For those who are actually interested in lore, it begs the question of why they would want to subscribe anyway, since they can simply read what they are interested in for free. This point is particularly valid if the player has no direct impact on the story. (See "Open Ended Storylines and Player Subplots" and "Multiple World Plot Development and Contingencies" for more about player impact.)
Therefore, the lore must have a very visible and obvious presence in the game world. If a ruined city-type dungeon environment is named after a historical warlord who once inhabited it, there should be traces of his presence inside the dungeon. There should be a nearby village full of people who can tell the players the story of the dungeon. There should be a book about it in a library in the central city, among many other books dealing with similar tidbits. If a town was founded/defended/whatever by a historical figure, place a statue of him in town with a little plaque explaining his deeds. You might also do this for player characters who are particularly heroic on a per-server basis, bringing more of a sense of identity to that server. Lore can also be important to questing, mage-oriented questing in particular. Remember, the literary mage was typically more a repository of bizarre arcane information than a fireball factory. All of these systems may seem a little heavy-handed, but sometimes you have to smack someone over the head with a book of lore to get them to notice, especially if they've been numbed by countless RPG's where lore has no importance.
Ongoing lore, such as the plot of Asheron's Call, also needs to get heavy publicity. Any company-mandated event is probably something of such magnitude that most of the world should be aware of it, even if it doesn't directly affect everyone. Town criers, gossiping NPC bar patrons, and such can help ram the event down the throats of your sleepwalking player base, without requiring them to go to a web site for information. The oral tradition can become part of your game world's color and propogate lore, in the form of NPC storytellers who spin yarns about what happened several months ago. You can even tailor these stories to the individual world, allowing player actions to be immortalized, giving them additional impact. A compilation of current and recent lore can be compiled into a journal accessible in-game. This can also be presented via a web site if you like, for the casual player who wants to check up on things while he's goofing off at work. Again, it does you no good if your lore for a past event consists of two horribly general paragraphs about what monsters came in for some mysterious purpose, buried deep inside an inscrutable and unnavigable maze of your publisher's site. If you want to use the web as a tool for game publicity and lore, try to get your own domain.
In regards to the use of secret lore, it's really important to make sure you have it mapped out in as much detail as possible, and that it all makes sense. This means a minimum of heretofore unexplained black boxes to validate events. If you think of lore in levels of detail and availability, you want to make sure the clever player can infer the nature of the next level of detail by examining the level above it. For instance, a general piece of knowledge is that the ancient empire of Sutannica was destroyed in a great battle. If players find information in a library stating that Sutannica's Field Marshal Brox was both brilliant and a champion of peasant rights, and another piece of information indicating that Sutannica's nobility was especially harsh on the peasants, they might infer that Brox was a key figure in a rebellion that toppled the empire. This may or may not be true, but when the players find the next level of detail that affirms or denies this, there should be enough supporting evidence in the preceding layer to make the truth make sense. If it turns out Sutannica was destroyed because a random wizard who nobody knew about opened a gate and let 40 demons out who killed everyone, this is a cheesy black box. It's akin to having the Lone Ranger survive a train crash from the end of the previous episode by jumping off a freight car at the last second, revealed only at the beginning of the next episode. It's cheap and stupid, and player-detectives will throw up their hands in disgust and never care about lore from that point forward.
Beyond Good and Evil and Stuff
One of the classic misconceptions about games, role-playing games in particular, is the definition of Good and Evil. Evil tends to be misused horrifically, applied to everything from a brainless NPC monster to a guy who lives to kill players for no reason. The scope of a true definition of "evil" is beyond the scope of this document, but suffice to say that real evil does not exist in a commercial game. If it did, the game would be too disturbing to support a viable subscriber base, and would probably violate a lot of laws. In game terms, as well as in real life, one can define "good" as being in accord with your own interests, and "evil" as being opposed to them. This is a inaccurate use of good and evil, but it's the way these terms were used to exhort children to march across the desert to take back the holy land (before they were sold into slavery), so it works just as well as any other.
So disregarding good vs. evil as a possible source of conflict, you have some realistic and perfectly viable choices for player (and NPC) motivation:
From the player standpoint, it also helps to define players' roles in society. A paladin who goes out to drive back the hordes of monsters that threaten the local farms is doing it for reasons better than "being good"; he is doing it to defend his homeland (Nationalism), to insure that his people get enough to eat (Economic Interests), and because the church has decreed that he must (Religion). A player who aspires to a noble title with lands and holdings does so to become rich (Economic Interests), status (Social Power/Fame), and just to say that he's the Earl or whatever (Personal Achievement/Fame). Understanding the motives of your players and their characters (hoping against hope that the characters are being roleplayed to the degree that they have motives of their own) is key when designing content, goals, and quests that you hope they will be undertaking, and making goals appropriate to each of these motivations attractive and fun enough for players to want to pursue them.
Open Ended Storylines and Player Subplots
The best use of the RPG setting, in my opinion, is in the flexibility of an open ended storyline. The referee makes a detailed description of the world, its inhabitants, social dynamics, physical features, etc., gives the players the seed of the plot, and then roughly sketches out a number of different paths they might follow to "solve" the adventure, knowing full well that the players will miss most, if not all, of them. This requires an especially good gamemaster with a quick mind and improvisatory ability, to be able to cope with whatver harebrained scheme the party might throw at him. It is the most rewarding for the players, as they no longer have the feeling of being railroaded through a sequence of predetermined events, their only contribution being the rolling of dice to see whether or not they kill the enemy in front of them so they can proceed tot he next staged encounter. The players and the referee are effectively collaborating to create a work of fiction, and the satisfaction of doing so is shared by all involved.
Computers are notoriously horrible at this style of refereeing. To do so, the programmer must try to conceive of every possible approach to a problem a player might take, and code appropriately, using true-false values in combination with advanced fuzzy logic to deal with all the grey areas. Even then, it is almost impossible to determine what a player will do, and exactly how the environment should react. Worse still, a player doing something in an unconventional way logically has an impact on the further development of the story, requiring the computer to rewrite entire sections of the adventure on the fly in the perfect model, in accord with player choices and consequences. Multiply this with every possible quest in your world, and every player who will approach that quest, and you can understand why it's more practical to fall back on the "go through the dungeon this way and fight stuff" model. This model, however, is woefully inadequate for an ideal persistent world in which players feel they actually make a difference.
A compromise may be reached between open-ended stories and a linear plot, however. Since the determining factor in an open-ended plot is the actions of players, you must therefore (1) support advanced player interaction via trade, socio-political systems, and PvP, and (2) manage the logical consequences of such interactions on a case-by-case basis. The great advantage for the content team in this case is the fact that the actions of players are determining the various subplots, not development manhours. The disadvantage is that someone should be scrutinizing these subplots occasionally to ensure that they don't fly too far outside of realistic world dynamic concerns, and most of all, these will be happening on a per-server basis in a mirrored world environment (like any of the big three).
This last concern has certain implications for the concurrent linear plot, i.e. the "official" story of your game world implemented by the development team. Unless you want to effective run as many official plots and events as you have servers, increasing your content overhead by the same amount, your overall official stories cannot be significantly impacted by the player-driven subplots. This means that clear restrictions have to be placed on the players who wish to advance their own plots in the form of restricted building zones, maximum amount of political power wielded, etc. If Joe the Monarch builds a shell keep and cultivates 40 acres of land somewhere in the wilderness, that land is now no longer usable as a staging point for a massive invasion. However, if the keep is right on the border, and the invasion happens to start nearby, they will inevitably cause problems for Joe the Monarch and his holdings. In this way, player subplots and official storylines can coincide on a per-server scale, drawing the players into the official storyline, giving them new opportunities to participate and actually make a difference within the context of their version of the game world.
Content Beyond Powerleveling (Shadwolf)
One of the biggest issues in games is always powerleveling. Certain players will powerlevel and break the game and this becomes unfair to the other players. There are some fixes to mitigate this, such as "no numbers," but these will not eliminate the problem. The problem is more basic than this; it is about motivation.
Everyone has a favorite story or movie. They watch this movie or read this book and think "gosh, Aragorn is so cool. I wish I could be him." Fortunately, you have roleplaying games to allow you to play out this fantasy. You whip out some dice, roll some stats, assign them in the appropriate order and list all your skills. Next you buy your equipment. Now the best part: you write the background. Noble heritage... dispossesed...family heirloom sword...mysterious demeanor... etc. Now you are ready to play. What do you do?
You have a really cool character. You know he is good at smiting evil and doing heroic deeds. Hmmm... I guess you go look for some evil to smite.
Alternately, you are a game master or game designer. You have a fantastic idea for a game world. The people of Shangrila have lived in peace for centuries. The have known limitless prosperity and developed amazing new sciences. All this was shattered when the evil Necromancer Urghblech unleashed his mighty army of hideous orcs upon these unsuspecting people. Can anyone save them in time? Great - now what do the players do? You need a plot. Wait.... you don't need a plot - you have an army of orcs for the players to kill! When they get too tough for orcs the Necromancer will summon Ogres for the players to kill. These will later be augemented by Ogre mages, Dragons, Red Dragons, Blue Dragons, Copper Dragons, Curium Dragons, Radioactive Polka dot Dragons, Elder Radioactive polkadot Dragons and ultimately by something called a K'tl'aa'draclqw that looks like a multitentacled jello cube and fires thousand point damage area affect lightning bolts and is only affected by +329 or better cheese based weapons.
With players like this in worlds like this, the goal is to kill monsters as efficiently as possible. What is the best way to accomplish such a goal? Powerleveling. So how do you eliminate powerleveling? Take away the motivation. Maybe the event of the month involves finding the spy who has infiltrated the imperial court. Perhaps the queen is dying of a strange illness and the king has promised great riches to the first person to find the plant that is needed to make the cure. Perhaps a terrible artifact(s) is found and anyone wielding it/them gains great power but is slowly drawn to evil. Some players want to obtain the artifact for their own use - others wish to destroy it. These are simple plot ideas which are taken from common fantasy genre stories. There are thousands of starving authors out there writing this kind of stuff. I bet they would be willing to sell such a storyline for far less than it costs to hire a programmer to come up with "There are these infinitely respawning monsters everywhere and the players have to kill them."
Okay, so now you have a plot. What does this accomplish? The players become immersed in the game. They work to accomplish these goals. Some players with more limited imagination will still waste their time powerleveling, but most will be more interested in seeing how the plot develops. The best gamers will be rewarded by becoming integral parts of the storyline. Never underestimate the power of recognizing a player. When the web page anounces that Lord Ragan has been appointed as Field Marshall of the army of Blackmarsh, you have made that player's life complete and most of the other players are going to wish they were him. These goals must be meaningful, however. The anouncement that Chef_Dewd34 has won the royal cooking contest because he was the first one to make rice krispy treats does not add anything to the game.
(Rice Krispies is a trademark of Kelloggs corporation. Useless cooking contests to enhance "roleplaying" is a trademark of Turbine games)
The Merits of Mindless Slaughter
Most of the content in this section is concerned with providing for a realistic, deep, immersive environment that allows roleplaying to flourish, and gets beyond the prototypical dungeon crawl model for fantasy gaming. This is because no MMORPG on the market, or any that look promising in the near future, provide this very well. However, this is not to say that hack and slash isn't appropriate in the MMORPG. In fact, combat must be a significant portion of the overall content. This is because:
Assuming you can't just have the monsters stay dead once killed, which punishes everyone after the first guy to find them, you need to find some sort of middle ground. The basic elements of realistic and intelligent hack and slash presentation are:
Consequences of player-monster conflict are harder to deal with from a practical standpoint. An advanced strategic-level system would be required to "keep score" of monster-human conflicts and adjust spawns accordingly. If the newbies keep trashing the local swamp goblins, it's unlikely the latter would hang around and take it. They would leave, get reinforcements, or possibly all kill themselves in a spectacular but doomed suicide rush on the town. If they vacate the area for any of these reasons, they need to pop up somewhere else, preferably somewhere equally accessible to the characters who would most likely benefit from hunting them. The location of the goblins would be fluid, as they migrate hither and yon, looking futilely for a spot where they can establish themselves. Note that unintelligent enemies like animals wouldn't be as likely to make such an organized effort, though they might be pressed into new territories, and certain monsters that are tied to a location for whatever reason would stay in the same general vicinity.
Likewise, an ideal system would allow for the consequences of monster success. If the Broken Fang tribe of orcs moves in near a little fishing village and is unmolested, they will grow in power and influence, attracting more allies to their cause, and the hamlet will be overrun until the players take care of the problem. In this way we can start to approach the treatment of intelligent monsters as more than attack dummies, viewing them instead as rival kingdoms and peoples who are out for land control just as much as the humans.
A simple example of fluid monster distribution is the idea of the "frontier" area. This can be used in conjunction with the fluid spawn zone concepts described above, and it provides an ideal and logical way for players who just want to go kick some ass to do this. The frontier is a portion of the map (possibly but not necessarily the edge of where players can go) which is an entry point for new monsters. Monster groups enter the land at the frontier line and begin looking for places to settle. More powerful monsters, being more intimidating than their weaker cousins, will settle in convenient spots closest to the frontier, forcing weaker monsters to push further toward the players' territory in search of a home. Therefore, the enemies will get tougher as the players move closer to the source of the monsters, a comfortable convention that will satisfy the most unimaginative of Telengard afficionados. The actual outline of the danger zone of the frontier is fluid, changing as players destroy camps of monsters, forcing them to respawn at the frontier and begin looking for settling places all over again. The frontier can be used as a plot device, i.e. the prototypical threat of invasion from the vile barbarian humanoids that can only be stopped by the heroic adventurers.
One of the many problems with MMORPG quest simulation is the fact that quests are repeatable forever, barring artificial constraints like a reward that one can only get once. For some quests, this is appropriate (like the harvesting of a spell component, or bringing food to an area that needs it), but when you've gotten the Sword of Gulrag the Beheader for the 5th time, and had to stand on line to do it, suddenly the Sword of Gulrag the Beheader doesn't seem quite as epic as it used to. Eventually, it will be either listed on web sites as a "required" piece of equipment for a particular character type (if it is unreasonably good), or it will be relegated to the graveyard of quests nobody bothers to do (if it isn't). Unfortunately, the static repeatable quest is extremely easy to implement and maintain in an MMORPG with a low administrator:player ratio. It's a fire and forget sort of deal (until one of your static quests turns out to horribly unbalance the game).
One possible solution for the boredom associated with static quests is to simple make them non-static. There are two ways to do this, and both of them involve extra work for the administrator:
However, the unique superquest presents a problem that impacts many aspects of a mythically heroic game world: everyone thinks they should be the hero. When one person gets the Holy Lance of St. Augustin and ends the quest, that's one person out of all your subscribers, and you can bet that the other subscribers are going to bitch about it. This can be alleviated somewhat by making unique quest items balanced according to zero sum, so that the person with the Holy Lance of St. Augustin isn't perceived as getting a twinky device that allows him to dominate all the content in the game from that point forward, including (presumably) other superquest items. The superquest reward may be better on average than a standard weapon, but then again the guy with a saber from the local shop isn't being haunted by daemons who eat his gold and try to kill him, or afflicted with rotting sores or whatever.
The quest that goes away eventually is appropriate to most static quests as people think of them now. Examples might be bringing grain to a town gripped by famine, hunting an invading army of humanoids that have occupied a town, or recovering stolen diamonds for a well-to-do merchant. After a while, the town is back on its feet, the humanoids have been driven away, and the diamonds are all back. The quests end, and new ones take their places.
A quest of this sort that does not go away must be of low importance, or appropriate to lower power characters, like getting a type of ore out of a really dangerous mine for a master smith, who then uses the ore to make items, and thus requires more. Another might be delivering the mail through a dangerous area. This is the standard "Fedex" quest used in the only way that makes sense, and it is rightfully droll. Characters who undertake this quest will be relatively poor and/or low-powered, and the rewards for carrying out these mundane tasks should be petty enough to stop interesting them once they become better established, leaving the quests for other newbies to perform as they go looking for the fluid quests, which change from patch to patch.
Both of these solutions pose the same fundamental problem for the development team, though. Work is thrown away, and new material must be created. Unfortunately there is no great solution to this, outside of a quest generation engine (which has never worked well), so new material has to be cooked up on a regular basis, even if much of the basic material is recycled in some form or another. A dungeon for a unique or ending quest may be a big pain in the ass to design, and you don't want to throw it away, so it may become a home for itinerant humanoids (or a new city!) once the boss bad guy is evicted.
The reward for all of this additional overhead is immersiveness. Long term players will be keeping track of, and maintaining interest in, the lore of the land without having it thrust down their throats. "Remember a few months ago when Gular and his band sieged Thistlewood?" Players will be actively looking for current news and rumors to find out where things are happening. Those that find this annoying can always go to the frontier and kill stuff until they get tired of it. If content is dynamic, though, there will be something for them to check out once they do get bored of killing.
Multiple World Plot Development and Contingencies
Of the three major MMORPG's on the market right now (AC, EQ and UO), only AC makes any attempt at an ongoing storyline. UO has some "lore" from the previous Ultima series games, which makes no sense and doesn't matter anyway: any plot that happens in UO is usually by accident, and created by the players themselves. EQ has some sort of weak background lore based on a combination of 15th generation Tolkien ripoffs and DikuMUD, but it never ever changes; the world of EQ is even more static than that of UO. AC has an ongoing storyline, with events and such, despite the fact that the vast majority of players don't care about it past the concerns of what new loot is available with the event, but at least there's the attempt. This section concerns itself with AC-like models incorporating an ongoing plot as part of its design.
In an MMORPG model, one has to resign oneself to the fact that a large portion of your plot is going to static, and therefore linear. What this means is that if next month you are planning to have the Rovers of the Frost Barrens start raiding border towns, they're going to be doing it on every iteration (server/shard) of your game world, even if those towns happen to be very well garrisoned in some worlds. This also means that if in January you drop a bunch of clues and lore and such about an upcoming catastrophe into the world, but only one of the shards manages to find your clues, that catastrophe is still going to happen on all worlds in February. If the world that managed to figure out the lore ahead of time deals particularly well with the catastrophe as a result, while all the other worlds get bombed into the stone age, the following month's events will not reflect this divergence between worlds. If you start diverging like this, eventually you are running as many games as you have servers, with that much more content development overhead. This is similar to the problems outlined with open-ended storylines above.
Again, the situation relates to development power in the hands of players, but as it impacts the company storyline, it rests more on the technology you develop to deal with player power than the players themselves. If, for example, you allow players to build houses through a dynamic construction engine that allows for building damage, repair, and the like, then there are definitely variable circumstances that will work themselves out. Depending on what sort of frontier system you are using, the Rovers of the Frost Barrens might run right into a popular player built town at the edge of civilization, kill the NPC guards, and start burning buildings. This sort of thing definitely changes the dynamics of one quest from world to world.
Occasionally, though, you still run into problems. The business with the Herald Crystal in Asheron's Call is a perfect example. On nearly every server, eventually somebody killed the uber-godlike Herald Crystal despite other players trying to defend it, which was expected, and necessary for the next monthly event to make sense. Every server, that is, except Thistledown. Eventually, Turbine had to send in an invulnerable admin/player to kill the thing to further the plot. Everyone felt cheated by this, and they had reason to. However, it had to happen somehow. Could it have been handled in a better way? Possibly. The point is that it had to happen. If Thistledown had successfully foiled the plot int his manner, implementation would have had to scramble their asses off to give them a different event from every other server. Not only is this horribly impractical, but you would get even more whining from all servers about being "left out" of the events happening on Thstledown, or on any other server besides Thistledown.
There is an interesting trick one can extrapolate from various hackneyed time travel novelists. One theory about "changing the past" is that time is self-healing, and that while stepping on an ant 20,000 years ago might have some repercussions, eventually things will smooth over, and Rome will still fall, Hitler will still come to power, and the atomic bomb will still be developed. If we reduce the scale of this example a little, we can apply the self-healing aspect of a multi-server plot to the overall chain of events, but allow things to diverge slightly on a server to server basis if we have the man-hours to implement it.
For instance, say your overall plot functions on a 2-month cycle. You are just now implementing an event that involves refugees from a neighboring kingdom swarming into the play area, followed closely by a tribe of marauders that has evicted them. You've already figured out that the next big event will involve an invasion by the players into giant territories to stop their impending threat. During the two-month refugee/pursuit event, many different things can happen. The players may, as expected, kill every giant they see, and at the end of the two month period, the local nobles rally and exhort the players to go kick some giant ass. However, maybe the players are particularly brutal, and kill the refugees, which might lead to a temporary truce with the giants who can assume the players are on their side. The players may fail to deal with the giants, and some towns and such may suffer as a result. The players may be trying to fight the giants, but the fluid "frontier border" of giant influence is at a standstill. Different servers can have very different versions of the event, but somehow they all have to lead into the next development.
The key here is to always have a contingency plan, and to be able
to make one up on the fly if something really strange happens. This
is not unlike the normal refereeing of an open-ended plot, except it's
not truly open-ended. You must have all game worlds coincide
at the end of the cycle. Looking at this as a graph, you start with
a point (the event), then branching lines into various possible continuations
of the event based on player action, then a reconvergence with rationales
in place to bring the players back into line. The preceding example
might look like this, if you happen to use bad web tables in lieu of graphs:
Refugees enter play area, followed by pursuing giants.
Players gain entry to the giants' homeland to attack them. |
If successfully implemented, a contingency web like this can give the players a real sense of consequence and accomplishment, rather than the feeling they're just being railroaded into the next chapter. You will also be far less likely to hear players complain when you have to pull a deus ex machina to kill the Herald Crystal because you didn't plan for it.
Good dungeon design is not unlike decent FPS level design. There are a number of technical and fun factors to be considered, like resting areas, hazards, range of visibility, camping spots, etc. However, to be believable, a dungeon or similar construct must also be logical and explainable to the most nitpicky of players (whom you will never really satisfy anyway, but you can do better than the competition).
First, realize that the dungeon is the most overused and hackneyed device in all of gaming. They can be lots of fun, but after a while the smart players will start to wonder, "Where did all these dungeons come from and why is this bizarre array of monsters living in it?" Once they figure this out, they will start to notice that the dungeon lacks basic necessities like storage, food preparation areas, sanitation, etc. This kills any immersiveness they might have felt on their first exploration, and reduces your construction to a running joke amongst fantasy gamers.
One thing to keep in mind is that you have to explain why all these dungeons are lying around everywhere. Excavation is a very labor-intensive and expensive process, and most intelligent creatures won't go to the trouble of building a dungeon unless they have a good reason. Nocturnal creatures may build them for the same reason dinural beings use artificial light sources: so they can keep being productive when they would otherwise have to be sleeping. Extremely rich and powerful beings may in fact excavate underground for storage, detention, emergency shelters/escapes, or to expand living areas in places that are difficult to build aboveground, like a castle on top of a craggy hill that's already gone as high as it can practically go. Other creatures may take up residence in abandoned excavations like tunnels, mines, or natural cave complexes. Keeping in mind the purpose and origin of complexes like dungeons can go a long way toward making them believable.
Another obvious thing to do is to just not have so many damn dungeons
everywhere. An aboveground complex like a fort, town, or castle can
be just as enjoyable as an underground complex, and be far more believable.
If your lore allows for a very old civilization that left a bunch of buildings
lying around, why should an itinerant group of intelligent monsters bother
to build something new when they could just move in? Why would they
go to all the trouble of mining through solid rock when it's easier to
build wattle huts with a palisade? Just use your head.