The Colors of Magic - Available Now

March 28, 2014

Plans are Worthless but Planning is Everything

As a GM I love to prep.  This isn't because I like to "tell a story" - I like to build toys and see how my players play with them.  I'm 75% toymaker / 25% storyteller.  I've talked about my prep process a little before.  I build a session in detail, but I never plan in significant detail more than two steps ahead.  That lets me throw everything out if my players go in an unexpected direction.

The title is a quote from Dwight D. Eisenhower, from the 50s.  He knew that no plan survived contact with the enemy, but he also knew the value of planning.  Planning creates movable pieces that can be rearranged to respond to unexpected events.  It fills in rough areas so that you have details thought through when they come up.  While you may still need to improvise in many places despite your plan, it will help there, too.  It makes sense:  If I spend an hour learning about carrots, I know a lot about carrots, but a little about root vegetables, gardening, vegetable farming, nutrition, supermarket supply chains, cooking, cell biology, etc.

WARNING!  SPOILER ALERT!  My Eberron players should not read any further!

What's in the file at the bottom of this post?

Today's post is my prep notes for an upcoming Eberron game.  This is for a long (12 hour) session, and may resolve in less than the full time, but I have other material prepared (not included) in case that happens.

This is Not Advice!

Most of the time on this blog, I give general advice that's good for any GM running the type of game I'm talking about that week.  This week, I'm not.  Not every GM is a "prep GM" like me.  I wouldn't say "there is no wrong way" to run a game.  Otherwise, I wouldn't have this blog, right?  I know there are wrong ways to run a game, because I've done almost all of them.  But there are a lot of right ways.

I'm also not saying "if you prep a lot, you'll do fine."  You can prep well, and you can prep poorly.  If you write a scenario that says stuff like "the PCs will do this and that," you may be railroading, or at least making too many assumptions.

Not prepping, or prepping very little, can also be done well or poorly.  There are lazy dungeon masters and then there are lazy dungeon masters. Sly Flourish doesn't advocate being lackadasical, he just wants to cut out the prep that you don't technically have to do, if you don't want to, so you can focus on the parts you like best.  If you like to prep, like me, Sly Flourish's advice is still good at helping you more efficient.

My prep notes attached here make the assumption that the PCs will follow the hook.  But this is dozens of hours into the campaign, and it's a hook they've set themselves:  They're the ones who sought an artificer to have Aric Blacktree's doomsday device analyzed, so presumably they will follow up and investigate the supposedly-sealed lab where the artificer says it must have been made.  I don't make assumptions about how they will get to the lab.  Encounter 1 is just a framework for a skill challenge, not the actual details.  It contains the DCs and success/failure check boxes, and some skill choices if the PCs choose stealth.  But they could try other tactics, and I can use that page to crib a skill challenge for any strategy they choose, from bribery to slipping in during a party to digging up from the Cogs.

More non-assumptions:  These notes have 6 combat encounters, but 3 have alternate non-combat resolution options, and 3 are totally avoidable -- After the first encounter (unavoidable action scene to get the pace up) the PCs can choose to go to LABORATORY or STORAGE.  They know the plot is in the lab, but being PCs, they will probably want to go loot the storage area.  I just have a feeling...

Some Ideas You'll See

Also, I like having ways to end a combat other than just killing everyone.  This is part of why I try to make around half of my encounters have built-in options for avoiding them with skill checks, and I try to mix roleplay and skill scenes in for variety.

  • Encounter 1 is a very open skill challenge.  My players like planning and executing "runs" in the old Cyberpunk 2020 / Shadowrun style, and this will probably entertain them for an hour.
  • Encounter 2 must end with the PCs destroying the sentinel automata
  • Encounter 3 can be avoided or prevented by bypassing the traps; but if it is triggered, it must end with the PCs destroying the guardian golems and escaping the traps
  • Encounter 4 can be avoided or prevented with a mystical challenge; but if it is triggered, it must end with the PCs destroying the wraith figments
  • Encounter 5 can be avoided; but if the PCs go in that room, they must fight to the death because the trap haunts will try to possess them and make them kill each other.
  • Encounter 7 can be prevented with a tense social challenge; but if triggered, the PCs can fight the warforged to a rout and chase them off or force them to surrender; or they can knock them all out and tie them up.
  • Encounter 8's opponents have different goals:  Hammerhand is trying to defeat the PCs, and he and the two berserk warforged must be killed or knocked out; while Beryl d'Cannith is trying to flee.
  • Encounter 9 is actually a live-action roleplay scene.  The rules don't get involved at all, and it uses an unusual GM technique.  

Encounter 9 is not something they teach you in the Dungeon Master's Guide.  I've said it before:  LARP is a crucible of GM skills.  Encounter 9 is an example of a technique I picked up from a LARP, brought into other LARPs I ran or played in (by talking to their GMs), and imported into tabletop gaming a long time ago.  It's also a great example of a situation where heavy prep can be superior to improv (the opposite is also true: Improv can be superior to heavy prep in some situations).

The idea is that if one PC is getting a roleplay scene (or a few of the PCs), but the others aren't present, and you can predict this, you can design the scene so that there are exactly as many NPCs as players whose characters are not in the scene.  Then you can write NPC bluesheets to give the players of absent characters.  In LARP - especially adventure style LARP - this is a great technique because it involves more players in a scene, and if you are running a scene with more NPCs than GMs, you need people to play them anyway (otherwise you're running a tabletop game at a LARP, which I generally oppose).  NPC bluesheets are hard to write.  In one side of one page of well organized, large font text with lots of white space, NPC bluesheets have to instruct the players who get them who they are, what they should try to accomplish in the scene, and how they should make it resolve.  The bluesheets need to be continuity checked -- that means you write them, then make sure they cross-reference each other correctly, their goals are in just enough conflict, the PC's role is clear, and they know how the scene should start and end.

So showing my prep process is a very self-conscious task.  Attached are my notes for this session.  I claim no ownership of trademarks or copyrights of stat blocks or rules or images or anything (I sometimes use art in here, and I pull stats from the compendium).  This is just for educational purposes, to show you what I do to prep.

Click here to see it. (PDF, served from

Self-conscious disclaimer:  This is NOT meant to be published or printed.  It hasn't been copy edited.  There are probably tons of typos and sentence fragments, awkward sentences, and occasional inconsistencies.  It hasn't been playtested or even run once.  

March 22, 2014

Distributed Processing

The rules in an RPG define what players can do, and can be strict or loose, depending on the system and the group.

A group that adheres well to rules tends to enjoy the structure they provide, the predictability they lend, and the power they give the players.  They may define what you can do with hard-as-steel gates, but they also give you the key to those gates.  Some RPGs require the players to do a good deal of system processing.  Many systems give the GM one set of rules (typically limited antagonist abilities, plus campaign and story design rules) and the players another (special magic or unique combat and skill abilities).  I've been running 4e D&D and Night's Black Agents lately, and both involve a good deal of "off-loading" system mastery onto the players.  3rd edition D&D and Pathfinder do this as well; as do the World of Darkness games.  A major benefit of these games is that they can give the players a very mechanical, predictable, world; provide those characters a lot of unique abilities; and not burden the GM with having to learn all the rules for those abilities.  I call that "distributed processing" because these games tend to have a lot of system options that only the players need to learn how to use.  For players who like it, system and system mastery are empowering and enjoyable.

A group that likes to play loose with rules either enjoys the GM-as-Storyteller dynamic, or they distribute the director/author roles more equally.  Perhaps they have short attention spans or no interest in system mastery.  Perhaps they play RPGs to sit around a table with friends, drink beer, and come up with wacky stories.  They may prefer story games like Fiasco, or games with greater narrative control in the players' hands, such as Fate; or any number of indie games.  Personally, I enjoy Unsung when I want a distributed-authorship game.  For players who hate system mastery, a mechanical, predictable world that makes them do mental labor they dislike to navigate it is a stahlhartes Gehäuse.  They will not engage in system mastery, and when placed in a game that requires distributed processing, they will become bored or frustrated, because their dislike of the system will keep them from engaging with it, which will prevent them from effectively interacting with the game world.  I call these "system-avoidant" players, which is not a derogatory term.  They avoid games that put more system in between their imagination and their ability to exert their ideas in the shared imagination space.  They also tend to dislike the "game" aspect of RPGs, since the game aspect usually gets mediated by rules.

The takeaway here:

If you're running games for a group that doesn't want to take on system mastery, choose a system with fewer player options.  Otherwise, the players won't learn their character's abilities, and you will be forced to run both your side of the screen and theirs.  They will quickly become frustrated with the situation, and so will you.

If you are running a game for a group that contains some players that don't want to take on system mastery, and some that do, you have a few choices:

1) Run games in systems with fewer player options that depend on system mastery.  If the group is at least 40% system-avoidant players, you may need to consider avoiding games that use distributed processing, or at least avoiding D&D.  This probably includes the 5th edition coming out later this year, despite its efforts to significantly dial back the system complexity.

2) Guide the system-avoidant player to a character that has broad ability to interact with the world, without any special rules (this may be hard if not impossible in D&D, where magic is always very complicated, and not having magic is limiting).  In Night's Black Agents, just ignore special attacks and cherries, and it should be OK.  In the World of Darkness, there is usually a character option that focuses on sneaking - this tends to be an easy system to interact with, with minimal mastery required.

Finally, if you're running a LARP, system-avoidant players can enjoy both elysium style and adventure style games.  They will dominate elysium style games (if they do) through politics and conspiracy, rather than systems and combat.  As for adventure style games, LARP systems are rarely so complex and important that a player who ignores them is substantially hindered by doing so.

March 14, 2014

The Hex Crawl

So you've heard about this thing called a Hex Crawl.  What is it?  Why should it interest you if you're not an "OSR" fan?

A hex crawl is an open-ended sandbox-style adventure that originated in old school D&D games.  In a hex crawl, the GM produces a map of interesting Points of Interest (POIs) for the players to explore within a fairly large geographic area.  The map is given to the players, usually as an in-character item.  They usually have a reason for exploring the area, but it's not clear how to achieve their objective without exploring the map more first.

The map is divided into hexagons, hence the name.  These help the players and GM determine how far apart things are, and whether the PCs will need to camp in the wilderness along the way.  They also tell the GM what route the players will take as they travel, which may lead to the players stumbling upon hidden POIs.  As they travel, the different areas of wilderness have "random" encounters, which are actually based on the terrain type and nearby monster lairs.

The player experience of a hex crawl is that of uncovering a lot of secrets about an area and exploring it however they want.  It feels like free exploration, though there are usually good story hooks for why they're here in the first place.  Then of course, the PCs will find more hooks along the way.

The POIs on the hex map are the meat and drink of the hex crawl.  At the POIs, the players encounter monsters, humans, or undead living in ancient ruins.  These creatures can be friendly, neutral, or hostile.  Monsters lair in ruins.  In large ruins, often multiple kinds of monsters form a symbiotic existence.  Humans hide out in ruins or make new villages inside or atop them.  Undead are unique because they might connect the history of the ruin with the present (or not -- they could be recent undead; say a vampire that moved in, or a necromancer making trouble, or a recently hanged criminal's ghost).

The current monsters represent the "A plot" of the POIs.  Because you have a map with several POIs on it -- known and secret -- the different POIs exist in a milieu that should generate relationships, conflicts, alliances, and plots.

Example:  The PCs are looking for an ancient artifact called Sehanine's Tear in the ruins of a river valley.  They have a map showing a village and a wizard's tower.  When they get to the village, they learn about a band of pirates operating out of the ruined temple on the other bank of the wide river.  The pirate band is allied with a local lizardfolk tribe who live in an abandoned gold mine.  The lizardfolk disapprove of the pirates' predation, but depend on the purloined meat the pirates bribe them with.  The pirates are enemies of the fishing village across the river that the PCs find themselves in.  The village used to be protected by the wizard in the nearby tower, but the wizard recently disappeared after saying he was going off to explore a mysterious palace that appeared in the mountains above the hills across the river.  The tribe of lizardfolk cannot get meat for themselves because displacer beasts lair in an old coliseum, in the hills and prey on the game that the lizardfolk used to hunt, killing deer, goats and lizardfolk hunters equally.  The pirates are having their own problems:  Their leader recently became "ill" and cannot tolerate sunlight.  He spends his time in the crypt below the temple, plotting a "new strategy" for dealing with "those pesky villagers."

This milieu gives us several POIs with solid A plots.  Get some hex paper and put a different terrain type in each hex.  Place the POIs in hexes.  If you are using random encounters, use the dangers in your POIs to decide what encounters are likely to occur in what geographical areas.

For non-D&D games, this is all you really need.  For non-D&D games, each faction or threat is relevant because of its relationship to the local milieu.

But if you're running D&D, you have another step!  You need to install proper dungeons in each location.  In D&D, each POI contains a plot point (e.g. "the ruined coliseum needs to be cleared of displacer beasts before the lizardfolk will stop protecting the pirates") and each plot point is an excuse to have a dungeon.   The next step is to dream up the B plots for each one.  The B plots are the lingering remnants of the histories of the ruins in the area.  See last week's post for advice on traditional dungeon design.

Quick aside:  Let me explain the difference between dungeon design and typical RPG design:

  • Typical Design:  Each plot point in a typical RPG usually involves some challenge that the PCs must overcome, a consequential decision they must make, or a cost they must pay.  The displacer beasts in the hills would represent a discrete challenge, that may be broken into sub-challenges: 1. Find out why they came to the hills.  2. Undo the cause of their migration.  3. Chase them back to where they came from.
  • Dungeon Design:  The discrete challenge of the displacer beasts would be resolved within an isolated geographical area full of discrete zones where encounters can happen in a bounded way.  Instead of 3 challenges, the GM has the opportunity to expand the number of encounters by adding a B plot, which generates additional encounter hooks.  In this case, the coliseum is more than just a site for overcoming the 3 basic challenges of ousting the displacer beasts.  Perhaps the displacer beasts come from a portal to the feywild in the ruins, which opened when the fey palace appeared in the mountains above it.  The PCs must find out how to close the portal, and that leads to more encounters.  The B plot can also connect to other POIs:  The PCs may need to venture to the fey palace to end the overlap of the two planes of existence.

In any case, you need to place the ultimate objective -- the location of the thing the PCs came to the mapped area to find or accomplish -- somewhere in the area.  Select a location for it that -- based on the web of plots in the area -- the PCs cannot find without interacting with a few other POIs.

So in our example we have 6 POIs...

  • Village
  • Wizard's Tower
  • Ruined Riverfront Temple
  • Abandoned Gold Mine
  • Ruined Coliseum
  • Mysterious Palace

Two of these are known to the PCs initially: The village and tower.  They will learn about the palace in the village.  They will also learn about the pirates, and if they search for the pirates' base, they will find the temple.  The temple will lead them to the lizardfolk in the mine, which will lead them to the coliseum.  The coliseum is on the way to the mysterious palace, as well. They could ignore the pirate problem and chase after the missing wizard, and that lets them skip the temple and mine; but they will still encounter the coliseum and palace in that case.  If we want to place the ultimate objective in this setting, Sehanine's Tear, we would put it in a location in the coliseum that can only be accessed by something they find in the palace.  That way the PCs will have to explore most of the map before getting their objective.  So we will put the artifact in the hands of an ancient treant living in a fey grove on the other side of the portal to the feywild.  The PCs can't interact with that portal without getting the Rowan Staff that was stolen from the friendly wizard by the evil Eladrin in the palace.

In a typical RPG, each can be the location of a particular challenge.  In D&D, because of the unique feature of dungeons, each should contain at least a small dungeon (2-4 encounters) if not a large one (5+ encounters).  Each encounter should relate to the A plot, B plot, or arc plot of the hex crawl.  If your players like random encounters, you might still use them if they camp in the wilderness.  You can also make "random" encounters have plot relevance by having them reveal secrets about the nearest POI.

The example hex crawl would have the following encounters:

  • Village (2-4 encounters):  There would not be a dungeon per se.  There would be several skill encounters in the village, as the PCs piece together where the wizard went.  They may also find a pirate spy - a villager who has been coerced into putting out signals to the pirates to warn them if there are soldiers or adventurers in the city.  He would represent a combat encounter and a skill challenge opportunity to learn the location of the pirate base before the next pirate raid.  On the South bank of the river, the PCs are not likely to have random encounters.  This area is civilized.
  • Wizard's Tower (5 encounters):  This is definitely a dungeon.  The wizard has died, so the PCs may eventually come here to look after his affairs or loot his treasure.  It contains constructs and cunning magical traps.  On the South bank of the river, the PCs are not likely to have random encounters.  This area is civilized.
  • Riverfront temple (6 encounters):  This is a ruin with a dungeon.  There would be encounters with pirates but then also with vampires, as the leader of the band has been turned.  Deep in the crypt is the leader's vampiric master.  The master vampire slumbered in a sarcophagus for centuries until the Eladrin Lord led the pirate leader to the secret "fount of eternal youth" which turned out to be the vampire's sarcophagus.  He did this in an attempt to draw the wizard into his trap, for the wizard was the only thing standing in the way of the Eladrin taking over the valley.  Random Encounters:  On the North riverbank, the PCs are likely to encounter vampire spawn or pirates, depending on the time of the encounter.
  • Abandoned Gold Mine (4 or 6 encounters):  The old gold mine has a bunch of lizardfolk, but it also has a B plot with hidden rooms and traps devised by the paranoid, greedy dwarves that dug in here.  D&D's unique subterranean threats would be useful monsters for the sealed dwarven areas that the ignorant lizard people couldn't get into.  If the PCs are friendly to the lizardfolk, they will talk about the dwarf vaults they can't enter, and offer to let them past peacefully to explore the old vaults, in exchange for a share of the treasure (how much is another skill challenge).  Random Encounters:  The PCs could encounter lizardfolk patrols or basilisks, which are friendly only to lizardfolk, and therefore tolerated by the tribe.
  • Ruined Coliseum (5 encounters):  This crumbling structure stretches hundreds of feet into the sky in places, and its crumbling terraces house agile and deadly displacer beasts.  It also has a subterranean area that was used to house slaves for arena fights in ancient times.  That area has secret passages dug by rebellious slaves centuries ago, traps devised by the jailers to keep the slaves from escaping, and fey creatures (in addition to displacer beasts) that came in through the magical portal.  Random Encounters:  Displacer beasts!
  • Mysterious Palace (4 encounters):  A magical palace appeared earlier this year.  It is an Unseelie Eladrin palace that crossed over from the feywild.  The local wizard crafted a Rowan Staff to banish it back to the feywild before the Unseelie fey could cause trouble, but they tricked him and killed him.  His cat familiar survived and is hiding near the palace to give the PCs cryptic hints in charades about the dangers within.  Random Encounters:  Unseelie fey creatures, such as eladrin, spriggans, etc.
I hope this description of a hex crawl was useful for the GMs that read this blog.  My goal was to describe what a hex crawl is and how to design one.  My example hex crawl should actually be enough for a mini-campaign, with a few dozen encounters worth of material - enough for two levels in Pathfinder or three in 4e.  If you ran this without dungeons, it would go a lot faster.

In a way, the hex crawl offers the PCs more flexibility and freedom.  They can explore the area in any way they want.  They can skip areas if they want, or try to delve for as much treasure as they can get.  Because the area contains enough encounters for multiple levels, the order that the PCs explore in will define how easy the area is.  Some encounters may be too hard, and they will have to flee and come back later.  Some may come out to be too easy, which is fine.

In the example, if we presume that a Pathfinder GM runs this hex crawl across levels 4-6, we might have encounters ranging in CR from 4 through 8.  The level 8 encounters should be reserved for the Treant and the Eladrin Lord.  The Master Vampire encounter might be CR 7, as would the hardest encounters in the Palace, wizard's tower, and the sealed area under the gold mine. A half dozen CR4 encounters would make up the easy fights and "random encounters" if used.  Most  of the other encounters would be level 5 or 6.

If you have any questions, ask them here or find me on twitter @RunAGame

Note!  I added to this idea with more on Hex Crawling with Icons.

March 7, 2014


It's been a while since I posted.  Things have been very busy in my life!  Here's the Friday post for this week.  I'll be back on the weekly schedule for a while now.


The word "dungeons" is so common in D&D that it's almost comic.  Yet it's important to look at what it really means and what makes a "dungeon" a unique play element in the most iconic of tabletop RPGs.

A dungeon is isolated so that it can have monsters and treasure.

The dungeon is isolated by some factor of its design.  Either it is geographically remote ("two days North of here") or hidden ("somewhere beneath the city sewers") or alien ("through the portal to the Shadowfell") or sealed up ("in the vaults beneath the Great Library").  Its isolation causes it to be mysterious, and allows it to be full of treasure and monsters.  If it were not remote, the treasure would have been looted by now, and the monsters would either be killed or escape and destroy the nearby civilized lands... and that would mean it's remote and isolated.

The ogres that are raiding our village seem to be coming from the old cliffside temple, but we thought the Paladins of the Silver Flame sealed that up generations ago...

A dungeon is full of discrete areas so that the encounters are distinct.

Every encounter in a dungeon is distinct from the one before.  The roleplaying encounters, exploration challenges, and combat encounters are all isolated from one another.  For whatever reason, the enemies in the dungeon don't all band together to mob the protagonists.  It's not a mystery -- the dungeon's layout determines this fact.  It's full of remote areas and sealed rooms, secret tunnels and long passageways, locked doors and cave-ins.  If the areas in the dungeon were close together, all of the monsters and foes inside would hear the sounds of combat and come rushing to one another's aid.

The ancient temple is carved into the cliffside, with long passages connecting distant rooms that were once used for some dark purpose...

A dungeon is old so that it can have an A plot and a B plot.

In TV terms, the "A plot" is the main story of the episode, which resolves at the end of the episode, but may be connected to an "arc plot."  The "B plot" is a subplot that also resolves at the end of the episode and also may be connected to an arc plot.  Most of the on screen action revolves around the A plot.  In a dungeon, the A plot is usually "what's happening lately" and the B plot is usually "what happened centuries ago when this ruin was first built."  

Ogres have moved into an ancient hidden temple that was made by cannibalistic werewolf cultists of The Fury, but 90 years ago, Paladins of the Silver Flame came and killed them all and sealed the temple.  But a few years ago, there was an earthquake and the bricked-over entrance collapsed inward, and a few months ago, ogres moved in and used the temple to stage raids on the local village.

The B plot has mysterious effects on the A plot

The B plot colors the dungeon.  Most dungeons are re-purposed structures in remote areas being used as a base or home by criminals, cultists, or monsters.  They have moved in and taken the space as their own, but the space doesn't fit them perfectly.  It leaves them too spaced out to come to each other's aid if they get raided (by the PCs!) and has some lingering shadows of the past.  Those lingering shadows color the treasure found there (magic items and works of art from the B-plot time), monsters themselves (some lingering monsters from the B-plot time) and secrets (passages, puzzles, traps, etc.).

The ghosts of the slain werewolves have possessed the Ogre Tyrant and his two brothers.  They have started acting strange.  They have ordered that the band capture and imprison villagers in the old cages beneath the temple, and they seem to be sleepy and quiescent as the full moon approaches, in three nights' time...  

But there is hope...  In an old situation report found in one of the first rooms, you learn that four of the paladins were lost and never accounted for.  They explored deep in the temple's East wing, and disappeared.  It was assumed they fell into a trap or were killed in a secret passage.  If you can find their bodies and appease their restless spirits, you may be able to take their silver weapons, which are the only way you can defeat the possessed Ogre Tyrant and his brothers.

More, some of the ogres are disturbed by the Tyrant's strange behavior.  Just after the PCs find the situation report, they encounter Rashga Grey-Beard, the band's elder, who will actually converse with the heroes instead of attacking them, asking them if they know any ghost stories about this temple.  If they are honest with him, he will take several of the ogres out "raiding" for a while, clearing the way for them to get to the inner sanctum, where the Tyrant's brothers guard the passageway into the Killing Rooms downstairs, where the Tyrant stands silent watch over the terrified soon-to-be-victims.

A dungeon is a discrete unit of arc plot.

An arc plot is the plot of a season of television - a story with a beginning, middle and end.  Serial TV shows have several arc plots over the seasons, and some arc plots span multiple seasons (that's why some things aren't resolved in the season finale).  The plot of the campaign, in D&D, is designed to link a string of dungeons together, so that it keeps giving the protagonists a reason to explore a new isolated area composed of discrete zones filled with challenges.  Each dungeon has a clue, item, or objective that unlocks the next step in the arc plot.  The next step inevitably leads to a new dungeon.  

In your plot outline, you may have "track down the surviving witness and find out what really happened to the Duke."  So you create a dungeon and put the surviving witness in it.  You can have the witness be a villain in the dungeon, or an innocent victim of the villains in there.  You can have the villains there connected with the arc plot, or not.

...a secret known only to the three witnesses who saw the Duke's murder.  Two of them were found dead within a day.  The third, a priestess of the Silver Flame, fled to the North.  She was last seen on the road to Morda, a small mining town along the mountains...

...Yes, the priestess was here.  But some ogres raided the town, and she was kidnapped along with the innkeeper and his family when they burned the inn to the ground.  Please, adventurers, will you go to the old cliffside temple, slay the ogres, and rescue the kidnapped villagers?  Your priestess may be among them.  If you do not, they will surely come back and kill the rest of us...

A dungeon is full of choices.

A linear dungeon is boring.  Give the PCs forks.  Remember that the decisions they have must have consequences.  Do we go left or right?  That decision is boring.  Give them some information about what lies each way, even if it's incomplete or confusing.  Give them tactical and strategic options.  Give them roleplay choices.

Do we go left, through the underwater section, where the ogres can't be standing guard, but other dangers may be present (like drowning)?  Or do we go right, where we know ogres patrol regularly?  Do we sneak past the patrols or fight them?  Do we disarm the trap or try to lure ogres into it?  Do we take the secret passage, leaving monsters behind us, or kill them first?  Rashga proposes to take some of the evil ogres out to raid the town to get them out of our way while we deal with the Tyrant.  Do we accept his offer and hope the town militia can handle a smaller ogre band?  Do we let him go, but plan to press on the rest of the way without rest, hoping to finish up here then race back to town in time to chase him off?  Or do we kill him, even though that means we will be forced to camp the night to recover from our wounds?  Does one of us swear an oath to convert to the religion of the Silver Flame to satisfy the paladins' ghosts, or do we destroy them as abominations, as their own religion would have us do?  The innkeeper has been possessed by a werewolf ghost as well.  The full moon is coming very soon.  Do we kill him out of mercy and compassion for his family as the priestess suggested, or do we chain him up and hope to find a cure?