Curious as to how other folks deal with dungeon maps in game?

37 posts / 0 new
Last post
 Do you have the group draw their own? Do you display the map for them?
I feel like i lose so much time into "The corrador branches to the east and the north.  There is a door on the west wall."
 Do you have the group draw their own? Do you display the map for them?
I feel like i lose so much time into "The corrador branches to the east and the north.  There is a door on the west wall."

You are right to feel that, because that's what's happening. Then there are the follow-up questions: "Which corridor seems most recently used?" "Which one slopes up?" "Are they trapped?" and then the answers and checks that go along with those. Guh.

I usually don't bother with a map. If their travel through an area is important, I might narrate it or provide a pair of concurrent skill challenges, but I generally don't map it or require them to map it. If I do happen to use a dungeon, and the dimensions of the dungeon map are at all important, I'm happy for them to see it. Why not?

[N]o difference is less easily overcome than the difference of opinion about semi-abstract questions. - L. Tolstoy

so maybe displaying an overall layout (say similar to a buildings fire escape plan) but still prep the rooms and such?
so maybe displaying an overall layout (say similar to a buildings fire escape plan) but still prep the rooms and such?

I'm not sure what you mean by "prep the rooms" but yes, show the players the map. They can still roleplay the characters as not knowing the layout, if they want.

I also would recommend getting player input as to what the dungeon layout is like. They might have good ideas about the kinds of environment and locales they'd expect to find and enjoy dealing with. If nothing else, if a player says "Ooh, is there a lab/armory/library/observatory/etc. in here?" then yes, there most certainly is.

[N]o difference is less easily overcome than the difference of opinion about semi-abstract questions. - L. Tolstoy

I was toying with the idea of doing a map layout displayed and then an overlay with a 40ft (or whatever light source size) hole so that they can only see that amount.
I was toying with the idea of doing a map layout displayed and then an overlay with a 40ft (or whatever light source size) hole so that they can only see that amount.

Inadvisable if for no other reason than how cumbersome that would be.

Don't concern yourself with what they know about the dungeon. At the same time, don't set up your dungeon in such a way that if the players know the layout that they can use that information to ruin the game.

What exactly is your reason for wanting the secrecy?

[N]o difference is less easily overcome than the difference of opinion about semi-abstract questions. - L. Tolstoy

i've got a mix of levels of experience in my players.  I want to avoid any metaknowledge as much as I can, just to remove the temptation from them.
i've got a mix of levels of experience in my players.  I want to avoid any metaknowledge as much as I can, just to remove the temptation from them.

Ah.

What would happen if the players gained that "metaknowledge"? What would happen if an experienced player gained that knowledge? What exactly is the temptation?

[N]o difference is less easily overcome than the difference of opinion about semi-abstract questions. - L. Tolstoy

I've used dungeon tiles and laid them out only as players approached the edge of the last piece. Similar to the D&D boardgames. But it was for an irregularly shaped room, not the entire dungeon. I've never set out a whole dungeon map. I go from encounter to encounter, and gloss over the travel in between unless they run into traps or something. Give them a choice - right or left, door or no door, up or down - but then just narrate it. Don't worry about giving them a blueprint
I've used dungeon tiles and laid them out only as players approached the edge of the last piece. Similar to the D&D boardgames. But it was for an irregularly shaped room, not the entire dungeon. I've never set out a whole dungeon map. I go from encounter to encounter, and gloss over the travel in between unless they run into traps or something. Give them a choice - right or left, door or no door, up or down - but then just narrate it. Don't worry about giving them a blueprint

The meta knowledge comes from them figurinsecret fret corridors or rooms based on the layout, as well as other details that their pcs wouldn't otherwise be able to figure out in a dark corridor. Thaaway for for them to figure it out, but it takes away from the challenge as well as depreciates skill sets of certain party members in the process.

what I do to combat this is give inaccurate information. I don't mean lie, but the information I give, such as a length of a corridor and it's facing, is not exact, but rather based on a character's ability judge the measurement. such as did they look down and guess the corridors length, or count out the paces, if they have applicable skills, etc.

as long as the DM is fair, and not being punitive or deliberately deceptive or misleading, it prevents any metagame
 Do you have the group draw their own? Do you display the map for them?
I feel like i lose so much time into "The corrador branches to the east and the north.  There is a door on the west wall."


I just ran my group thru like 11+ sessions of Dragon Mountain. I bought 4 rolls of gaming paper and drew most the maps in advance. Then we used placemats, books, or whatever was on hand to simulate "fog of war". Worked quite well, though it meant more prep work for me, I felt it was worth it in how it sped up the game and lended to the feeling of the enormity of the dungeon.
It depends a lot on the type of campaign you want to run. If you want exploration to take up an important part then letting them know the map beforehand might not be a good idea. The last time I run such a campaign though, I did draw the map for them though while they were exploring. Trying the players to make maps from my descriptions always was too much of a pain ;) The current campaign I am running, exploration is not overly important. I rarely even bother detailing (big) dungeons. If exploration is dangerous, it is a descriptive skill challange. If finding ones way is not particularly dangerous or time limited I keep it to a short description.
narration tends to lead to confusion and players asking over and over what is where.

So, I have a battle mat and I tend to draw out dungeon maps quick and dirty - if a corridor is a little wider on the battle mat than I have on my (hand held) drawing, no big deal.  I also ask the players to help, especially when the drawing exceeds my reach.

In open spaces/wilderness, I only draw a map when there is going to be an encounter.  For most obstacles I drop coins or poker chips on the map to designate trees, rocks, etc.  This is especially helpful when I really do not care where each obstacle is located just that they exist.

In short, perfection is overrated.  If you spend a lot of time drawing out a map perfectly, it will probably take more time than you want.  Do not worry so much about the minute details unless they are important.

IMAGE(http://www.nodiatis.com/pub/19.jpg)

RedSiegfried wrote:
The cool thing is, you don't even NEED a reason to say yes.  Just stop looking for a reason to say no.
When I run games online, I have access to a Fog of War feature that I can turn on, revealing only what the PCs can see as they go. In this case, I have drawn up an entire dungeon architecture.

In other cases, it's usually better to either abstract the dungeon or draw and reveal the whole thing. By abstraction, I mean you develop only certain pockets of activity in the dungeon, encounter areas or the like, and connect them by "mazes of corridors," "maddening tunnels," or "freezing underground streams." Make it cool, and be as creative as you like or ask your players how they're connected as a collaborative exercise.

As far as just drawing up the whole dungeon on a battlemat or the like, you can trust your players to compartmentalize what their characters know and what the players know or you can just say the PCs already have a crude map of the structure. Draw the rooms and corridors and then fill in room details as they explore.

No amount of tips, tricks, or gimmicks will ever be better than simply talking directly to your fellow players to resolve your issues.
DMs: Don't Prep the Plot | Structure First, Story Last | Prep Tips | Spoilers Don't Spoil Anything | No Myth Roleplaying
Players: 11 Ways to Be a Better Roleplayer | You Are Not Your Character     Hilarious D&D Actual Play Podcast: Crit Juice!

Here, Have Some Free Material From Me: Encounters With Alternate Goals  |  Dark Sun Full-Contact Futbol   |   Pre-Generated D&D 5e PCs

Follow me on Twitter: @is3rith

I really think it depends on the game. If mapping and navigating and dealing with the dungeon as a hazard to be overcome by the players are part of what you are wanting, then it makes perfect sense. There are old-skool versions of D&D that really emphasize that, and they are fun if that is what you are wanting to play. There is a new indie game called Torchbearer (based on MouseGuard RPG and Burning Wheel) which really emphasizes resource management and navigating a dungeon.

In my 4E game, mapping the dungeon would feel un-heroic in contrast to the characters, their abilities and the overall story the players and I are creating. There actually isn't a lot of traditional dungeon crawling in my game, so if I suddenly started detailing 10x10 rooms and 100ft long corridors and handing out graph paper and pencils, I'd get a lot of weird looks.

We go with set pieces. If there is something interesting to encounter, that is when a map gets placed on the table and everyone knows it is an encounter
good stuff - its good to see that there is a lot of variation about what folks are doing
 Do you have the group draw their own? Do you display the map for them?
I feel like i lose so much time into "The corrador branches to the east and the north.  There is a door on the west wall."



I start by using GIMP 2.8 to make a map. On Civfanatics.com, in the Civ5 Mods, there is a "HexesTiles" mesh you can download and "slap over" your map and resize as needed. I like to have them start at a nice town. On the way out of town, there are always 2 events. One is usually an "on the way out of town..." and another is the result of your survival check. I print my "dungeon" as "sectors", if you will. Then I place each "Sector" down one at a time, declaring "Fog of War" as why you can't see the rest.

I used to use whole dungeons, from the classic battlemats, to my own designs. However, I got sick of players anticipating what lies ahead, and messing with different schemes to hide the road. I have tried placing paper over the map and adjusting it as they move, i have tried having a print out map behind the screen and drawing on a battlemat. There is no perfect answer!

On the way there, your first "dungeon" will consist of 2 pieces of paper, one in the plains and one in the forest, you will make a few rolls, and your random encounter will take you to a "blow up minimap" akin to a classic JRPG. Then, you arrive at the cave and I use the "5 room model" to present 5 rooms that offer a challenge.  Lets math this out.  Your first encoutner was in town, a second in the plains, a third in the forest, 5 more in the dungeon. Yup, its called a "shortie" because I can have you get 3 more encounters on the way back to town, one more in town, and usually have time for a second and third dungeon depending on "travel time".

As Merb101 says, I don't often use maps anymore, just "blow up" sections. Most of my players are kind of at the point where they can't be contained on such a map, so no reason to bother.  My games these days tend to have 1 - 3 battles of bigger scale and more damage with very few, if any, "dungeoncrawls".

To each their own!

Within; Without.

 Do you have the group draw their own? Do you display the map for them?

I feel like i lose so much time into "The corrador branches to the east and the north.  There is a door on the west wall."



I have a relatively detailed map in my notes, but we do a narrative at the table. I only draw something when it matters for something. But that "the corrador branchers to the east and north"  stuff is just done verbally and not mapped, if even done at all. Usually I just say something like "You follow a bunch of twists and turns and then something interesting happens". If I include a fork, I'll say "You come to a fork eventually, one way has the sound of the ocean and salty air, another looks extra dark and forboding" or something. Providing "north" if someone asks for it, I guess.

"In a way, you are worse than Krusk"                               " As usual, Krusk comments with assuredness, but lacks the clarity and awareness of what he's talking about"

"Can't say enough how much I agree with Krusk"        "Wow, thank you very much"

"Your advice is the worst"

I have tried it all--drawn maps on sheets of cardboard, laid them out using design programs, printed on wideformat rolls, and laminated textures onto cardboard construction. It's been a lot of fun experimenting. I am currently using hot foam knives and extruded polystyrene. However, not everyone is as crazy as I am for this stuff.

I do not like spending game time describing rooms and have found that providing printed maps gives us all more time for roleplaying and takes the guesswork out of combat. My players really really don't want to see spoilers. If the whole map is on the table, they will pretend not to know what's around the corner, but that's not as much fun for them.

If you want to put it out on the table as you go, you can either use a dry-erase roll out map, draw on a big sheet of paper, or print out room sections. I also have used that--adhere the map to sheets of matboard or cardboard, cut them out, and place them room-by-room as you go.

It really comes down to how important scale maps are to your style of DMing, what will contribute most to your players' enjoyment, and how much time you want to invest in preparation.
My players really really don't want to see spoilers. If the whole map is on the table, they will pretend not to know what's around the corner, but that's not as much fun for them.

What if they simply knew? Keeping things secret isn't the only way to make a game suspenseful, and it's not even a particularly reliable way.

[N]o difference is less easily overcome than the difference of opinion about semi-abstract questions. - L. Tolstoy

There's lots of different ways to run the exploration part of the game, but while I agree that keeping secrets isn't the only way to make a game suspenseful, I would say that it is reliable. There's no getting around the fact that, were you and I actually exploring a dungeon, we would likely have no idea what's around that corner. Could be a dead end. Could be a huge treasure room. Of course, a dead end on a map might be a treasure room and the opposite holds true.

But if players see the layout, they know some things their characters do not, whether they keep it from their roleplaying or not. My players prefer to maintain that barrier and I enjoy revealing things once they've decided their characters will take the risk. 

From a recent session...players are exploring mine which has been taken over by Orcs and Gnolls. Main goal is to chase down Orc Priest. They poked their head into one cave, spotted a couple of Gnolls sleeping and decided to sneak away. Had the map been on the table, they would have seen that this area was a relative dead end. A bunch of Gnolls to fight, maybe a prisoner to talk to. But they didn't and it's still bugging them. Maybe that way led to an easier way past their current difficulties? They chose not to risk it and won't know until they go back. At which point the Gnolls will have been the only creatures the players left alive in the mine and the defenses and traps will have turned from Orc-like to Gnoll-like. As an aside, I really enjoy that aspect of the game--the players made an odd choice and have changed the world in a big way, especially for 3rd level characters.
I usually add something to the story or an NPC to the story who can either provide a basic map (they've been there before), or describe it to the characters so that when I give them a map (very basic, no major details), I can say this is what the NPC described.
There's lots of different ways to run the exploration part of the game, but while I agree that keeping secrets isn't the only way to make a game suspenseful, I would say that it is reliable. There's no getting around the fact that, were you and I actually exploring a dungeon, we would likely have no idea what's around that corner.

We might be fully informed about the dungeon, and the exploration of it might not be the point.

But if players see the layout, they know some things their characters do not, whether they keep it from their roleplaying or not. My players prefer to maintain that barrier

I wonder why.

and I enjoy revealing things once they've decided their characters will take the risk.

Why wouldn't they take the risk?

From a recent session...players are exploring mine which has been taken over by Orcs and Gnolls. Main goal is to chase down Orc Priest. They poked their head into one cave, spotted a couple of Gnolls sleeping and decided to sneak away. Had the map been on the table, they would have seen that this area was a relative dead end.

And, it sounds like, not worth anyone's time to bother with. They could have saved lots of time at the table, which I guess is my main concern: getting right to what's fun, rather than hoping boring stuff can be made fun.

A bunch of Gnolls to fight, maybe a prisoner to talk to. But they didn't and it's still bugging them. Maybe that way led to an easier way past their current difficulties? They chose not to risk it and won't know until they go back.

What if they never go back?

[N]o difference is less easily overcome than the difference of opinion about semi-abstract questions. - L. Tolstoy

There's lots of different ways to run the exploration part of the game, but while I agree that keeping secrets isn't the only way to make a game suspenseful, I would say that it is reliable. There's no getting around the fact that, were you and I actually exploring a dungeon, we would likely have no idea what's around that corner. Could be a dead end. Could be a huge treasure room. Of course, a dead end on a map might be a treasure room and the opposite holds true.

But if players see the layout, they know some things their characters do not, whether they keep it from their roleplaying or not. My players prefer to maintain that barrier and I enjoy revealing things once they've decided their characters will take the risk. 

From a recent session...players are exploring mine which has been taken over by Orcs and Gnolls. Main goal is to chase down Orc Priest. They poked their head into one cave, spotted a couple of Gnolls sleeping and decided to sneak away. Had the map been on the table, they would have seen that this area was a relative dead end. A bunch of Gnolls to fight, maybe a prisoner to talk to. But they didn't and it's still bugging them. Maybe that way led to an easier way past their current difficulties? They chose not to risk it and won't know until they go back. At which point the Gnolls will have been the only creatures the players left alive in the mine and the defenses and traps will have turned from Orc-like to Gnoll-like. As an aside, I really enjoy that aspect of the game--the players made an odd choice and have changed the world in a big way, especially for 3rd level characters.




I think you've hit on something important. If the game is about exploration and braving traps and hazards, absolutely keeping things secret and surprising is important. If  you are running a heroic game, it's not about the dungeon, its about the wizard at the center of the dungeon trying to summon a monster. Traps and blind turns and 10x10 rooms with secret doors just slow you down getting to that wizard. It's two different games.

Take for instance the Lord of the Rings movie. Huge scene in a huge dungeon. Not a dungeon crawl by any stretch of the imagination. There was the door, the tomb and the bridge. Three scenes. The rest of it was just getting them there. Not once did they look for a secret door, worry about torches or rations or potable water. Not once did they fall into a spiked pit because someone forgot to check for traps. None of that was the point.

So it absolutely depends on the kind of game you are playing.

1. In this case the players were not informed about the dungeon ahead of time. They could have followed up on a chance in town to get more info and maybe a map, but they prefered not spend the gold. So exploration is part of the point.

2. I think they like to keep the barrier between player knowledge and character knowledge in respects like the map because it increases their immersion. The more they know about what's coming, the less suspense. Their characters know about their world, some of the threats, and what their mission is. A significant part of the fun is now knowing the rest until they get there.

3. In this case they chose not to take the risk because they were injured and they had a time element to deal with. It was a dangerous place to get into a pitched battle (multiple unexplored entrances) and they chose to sneak on in hopes that the Gnolls wouldn't wake up and follow them.

4. The party knew that prisoners were being kept in the mines. They had already rescued some and gotten valuable info and one ally. If they had driven off the Gnolls, or snuck past them, or killed them, they might have freed another couple of slaves, one of whom knew something very valuable (though she didn't know it was valuable). The players chose to forgo the risk, even if it meant possibly rescuing more prisoners. They made a choice about their abilities and position and it made sense at the time. 

5. If they never return to the mine, that's fine. The Gnolls will become a threat in the area, having basically been given a whole mine full of dead Orcs and orc weapons. The players will eventually learn of this and may decide to return. I don't know, a lot can happen before then. 
Take for instance the Lord of the Rings movie. Huge scene in a huge dungeon. Not a dungeon crawl by any stretch of the imagination. There was the door, the tomb and the bridge. Three scenes. The rest of it was just getting them there. Not once did they look for a secret door, worry about torches or rations or potable water. Not once did they fall into a spiked pit because someone forgot to check for traps. None of that was the point.


Very true. Every time I listen to Chris Perkins I am amazed by the stuff he leaves out. I try to cull dead weight whenever possible.
I've tended to find that players not having information might lead to more suspense. But it almost certainly leads to less action than players just having information and being able to act on it. The suspense in this case comes from not knowing outcomes of their actions versus not knowing what's around the corner and, in some cases, talking the issue to death instead of acting. I'm sure we've all experienced the players, being faced with the unknown, taking up valuable session time planning for every eventuality and mitigating every possible failure, even when they don't know what might actually be threatening them. We try to minimize that in our games in favor of consistent forward movement.

No amount of tips, tricks, or gimmicks will ever be better than simply talking directly to your fellow players to resolve your issues.
DMs: Don't Prep the Plot | Structure First, Story Last | Prep Tips | Spoilers Don't Spoil Anything | No Myth Roleplaying
Players: 11 Ways to Be a Better Roleplayer | You Are Not Your Character     Hilarious D&D Actual Play Podcast: Crit Juice!

Here, Have Some Free Material From Me: Encounters With Alternate Goals  |  Dark Sun Full-Contact Futbol   |   Pre-Generated D&D 5e PCs

Follow me on Twitter: @is3rith

I've tended to find that players not having information might lead to more suspense. But it almost certainly leads to less action than players just having information and being able to act on it. The suspense in this case comes from not knowing outcomes of their actions versus not knowing what's around the corner and, in some cases, talking the issue to death instead of acting. I'm sure we've all experienced the players, being faced with the unknown, taking up valuable session time planning for every eventuality and mitigating every possible failure, even when they don't know what might actually be threatening them. We try to minimize that in our games in favor of consistent forward movement.


Agreed. Most devious trap I ever devised was an untrapped set of stairs. The practically built a suspension bridge over it, quite brilliantly, I might add. 

When I feel my players are taking unrealistic amount of debate time, that's when guards come by. Or a rockslide. I had one set piece two sessions ago where they knew this whole spiralling downward tunnel was unstable. So I rolled 2d4, set the timer on my phone for as many minutes and as soon as it went off, I dropped d8s all over the map. If you got hit, you took damage. Did that all the way down. Kept things moving at a nice pace, though I lost a d8.

I do think you point out a pitfall when the players have no idea at all what's around the bend. Indecision can be paralyzing.
I do think you point out a pitfall when the players have no idea at all what's around the bend. Indecision can be paralyzing.

That's one time when acting on some metagame knowledge can come to the rescue and help the players realize that they're wasting their own time.

[N]o difference is less easily overcome than the difference of opinion about semi-abstract questions. - L. Tolstoy

That's one time when acting on some metagame knowledge can come to the rescue and help the players realize that they're wasting their own time.



Right. There is a fear of some players and DMs that metagame information will hurt their immersion. In my experience, this comes from people who are more focused on maintaining character immersion primarily when, in my view, the best way to achieve that is to focus on being immersed in the scene and, by extension, your character. In order to do this, you have to be able to use metagame information to build the scene and tension in a way that immerses you. That's almost impossible to do if the control of information is entirely on the DM's side of the table. Focus on character immersion in this regard is actually a hindrance in my opinion. It's the scene that matters!

No amount of tips, tricks, or gimmicks will ever be better than simply talking directly to your fellow players to resolve your issues.
DMs: Don't Prep the Plot | Structure First, Story Last | Prep Tips | Spoilers Don't Spoil Anything | No Myth Roleplaying
Players: 11 Ways to Be a Better Roleplayer | You Are Not Your Character     Hilarious D&D Actual Play Podcast: Crit Juice!

Here, Have Some Free Material From Me: Encounters With Alternate Goals  |  Dark Sun Full-Contact Futbol   |   Pre-Generated D&D 5e PCs

Follow me on Twitter: @is3rith

Whenever I hear Chris Perkins say "You feel that..." I wait for a little metagaming nudge. Players deserve it, particularly if they are intent on shoving their nose up the wrong tree. Or should I say, the less fun tree. 
I don't think it's the province of the DM to tell a character how he thinks or feels (though it's admittedly easy to make that statement during play). This suggests an agenda of some kind, something the DM wants or needs the players to do. Accordingly, that doesn't sit well with me. It also means the DM is still in full control of the game's information which isn't ideal in my view.

No amount of tips, tricks, or gimmicks will ever be better than simply talking directly to your fellow players to resolve your issues.
DMs: Don't Prep the Plot | Structure First, Story Last | Prep Tips | Spoilers Don't Spoil Anything | No Myth Roleplaying
Players: 11 Ways to Be a Better Roleplayer | You Are Not Your Character     Hilarious D&D Actual Play Podcast: Crit Juice!

Here, Have Some Free Material From Me: Encounters With Alternate Goals  |  Dark Sun Full-Contact Futbol   |   Pre-Generated D&D 5e PCs

Follow me on Twitter: @is3rith

Agreed, but what I'm really referring to are in-game observations designed to aid the player's comprehension. 

"You feel that spending another 30 minutes arguing over the price of map will get you nowhere."
"You feel that the silence echoing through the mine indicates that no Orcs remain and you could safely rest here."

That sort of thing. Since really telling the players what they feel is an outright violation of the first commandment of DMing: The only thing players get to do is play their characters and thou shall not take that from them.
I see. Yes, if you're telegraphing information to the players, I guess I'd just frame it differently. Or just ask the players to make that determination because they have control over information just as the DM does in our games. Or, in the case of your examples, leave those to the rolls (if success and failure are both potentially interesting as to outcome).

I can't agree with that commandment either, but topic for another time perhaps. 

No amount of tips, tricks, or gimmicks will ever be better than simply talking directly to your fellow players to resolve your issues.
DMs: Don't Prep the Plot | Structure First, Story Last | Prep Tips | Spoilers Don't Spoil Anything | No Myth Roleplaying
Players: 11 Ways to Be a Better Roleplayer | You Are Not Your Character     Hilarious D&D Actual Play Podcast: Crit Juice!

Here, Have Some Free Material From Me: Encounters With Alternate Goals  |  Dark Sun Full-Contact Futbol   |   Pre-Generated D&D 5e PCs

Follow me on Twitter: @is3rith

I usually add something to the story or an NPC to the story who can either provide a basic map (they've been there before), or describe it to the characters so that when I give them a map (very basic, no major details), I can say this is what the NPC described.



Along the same lines, OP try to take a look at the Scales of War adventure Rescue at Rivenroar. It involves the PCs having to talk to NPCs who have been taken hostage in a dungeon, and try to get some directions and clues from each one about where the others might be. Since they're all scared witless, and have been moved around by their captors, each NPC is an untrustworthy source who can't give perfect directions, only vague, and sometimes slightly incorrect ones. If your players know the premise, this could be a way to keep them on their toes.
I tend to just narrate the environment unless there is a specific reason for the players to know more exact dimensions. Straight out combat, distance sensitive traps in the area, meeting with an npc that has potential to turn ugly, so on. 

I find it is a time saver to avoid maps unless they seem necessary. The more time I spend setting up a map that is likely to be used for five minutes at the most, the less time the players have to explore the next area. That pretty much sums up my feelings on maps at the table.

Maps in general, I tend to keep an idea of what the players are heading into in mind, but I don't draw out specific maps unless I know for certain something will be going on that requires one. I had a very detailed map drawn out for the starting city of the campaign I am DMing (first attempt) and virtually every building of even minor importance in it. I thought this was a great idea and vital. Set the PCs loose and they entered one tavern before leaving the city. Now, my feeling is that maps are generally overrated. 
I use Maptool, hence the logo ;)

erdana, Arial, Helvetica, sans-serif">I work with computers all day so I once refused to have anything to do with them during my leisure time.  In fact, I am something of a Luddite   However I have found the tool saves me DM time (and time is my rarest commodity) and at the same time its FOW and light functions add to the general game experience.

erdana, Arial, Helvetica, sans-serif">As a small point, I prefer pre-built maps where possible since I think this leads to more natural feeling encounters.    If players explore the whole game environment via a map they move about more naturally, don't stick to "marching order" and can genuinely be surprised because the safe areas look just the same as the rest.