I'm planning a mini mystery campaign (hoping for it to last 3-6 sessions), but I'm having some trouble deciding how to go about conveying information to my players. The plot is basically that four people have died, and there are three possible suspects (it doesn't really work with fewer victims or suspects, I won't go into why because I know that at least one of my players reads this forum). There are three main investigative avenues which my players can follow - one is to go to the crime scenes, where certain clues will point towards the assassin. The second is to interview NPCs who knew the victims, and the third is to interview NPCs who know the suspects. These will either provide direct clues, or give the players leads to follow.

Its the NPC interviews which are giving me issues. My initial plan was to have 2-3 NPCs with information for each suspect and victim, but that's already 14-21 possible NPCs. So then I narrowed it down to 10, which seems about right - it means that information is widely enough dispersed that they can't get a complete picture immediately, but not so sparse that its too easy. However, I don't know how to make my players feel engaged during these NPC conversations. I certainly don't want to lock information behind die rolls, but on the other hand I don't want to just dump information on them, and make them feel like they haven't actually discovered this information themselves. I don't want them to feel like they're basically going through a shopping-list of NPC conversations, particularly if they decide to gather as much information as possible before they start to do any sleuthing.

I'm perfectly comfortable acting the part in those NPC conversations and trying to make the conversations themselves engaging, but I worry that the players won't feel like they're actually accomplishing anything. So I guess my first question is: am I going about creating a mystery in entirely the wrong way? Is relying on NPC conversations a bad way to go about it? And my second question would be, has anyone else had a similar situation, and did you find a way to resolve it?

So your mystery depends on the successful collection and assembly of at least 10 pieces of information obtained by interview and several more obtained by examining the physical evidence at the crime scene.  Without at least 11 distinct data - none of which come from the suspects, or victims (speak with dead spells are a landmine) or eyewitnesses - we can't logically determine which of three people murdered which of four other people.  This is a pretty complex investigation.  That worries me, off the bat.  I mean clearly YOU know what actually happened, but whatever that was, it must have been a f*****g goat rodeo.

This all probably depends on your players.  If you can go over your methodology with them without revealing too many plot points, I would.  Just come right out and say: "There are 10 witnesses and each of them have one important clue as well as some basic info they all share."  Do not worry for a moment about making it too easy, just make sure the RP is fun.

I'm increasingly convinced that D&D mysteries are an urban legend.  I had similar situations long ago and never resolved the problems.  I'm better at this than I was then, but by virtue of that same experience, I just doubt you can get a decent puzzle-box mystery when at least half of the writers of the story are improvising.

If I had the guts to try it, I'd honestly just do it like those improv mystery plays from the 90's.  Just decide when you want to do the reveal and, at that time, take a vote from your players about who they WANT it to be, and have three different endings prepared.

I think there's a lot of fun to be had, RP wise in riffing on those cop-show interview scenes, but of course, those are scripted.  The cop always comes up with a great walk-off line even though you and I never do.  When, instead of a script, you have five people all racing each other to the bottom to see who gets to be "bad cop," and "the psycho," things can become a fiasco - a really mesmerizing Marx Brothers trainwreck.  Then you end up with at best a comedy session which usually wastes time.

At the other extreme, you may have one or two people at your table who mentally check out until someone asks them a question or they hear "roll initiative!"  For those people, spending two hours sitting there letting your two talkers do all the talking is going to quickly lead to a shopping trip session.  I don't know why, but what they always want to do, at that exact moment, is go buy a bear trap, or nickel-plate their eagle companion, or whatever.  And this also, usually wastes time.  The crucial thing to remember is that there is no such thing as running a solo errand while your group pursues the main quest.  It is an absolute illusion and a misconception of reality.  There is only one DM and you cannot "save time" by asking that one DM to run two scenes at once.

So, what have I babbled myself into saying?  1) You know what actually happened, but give your scenario another look and see if this can't be simplified; 2) You know your players and if you're worried that they won't enjoy doing interview duty then you're probably right; 3). If you do have a mix of different kinds of players, though, maybe pair them off with each other in teams.  One talker and one dice-roller on a team.  It doesn't take any longer or shorter to do ten interviews with four investigators at each one than to do ten interviews with two at each...again - only one DM.  What it does do is maybe avoid the talkers talking over one another and give the dice-rollers a bit more of a chance to get a word in edgewise; 4)  If you need to fill time, the players' main suspect can get a knife in the back and the killer can just be someone else, if you need to wrap things up, the ongoing investigation can always panic the criminal into making a fatal error if things start getting too far off course.  Or a surprise witness can be uncovered.  

Quote from TimCurtin >>

So your mystery depends on the successful collection and assembly of at least 10 pieces of information obtained by interview and several more obtained by examining the physical evidence at the crime scene.  Without at least 11 distinct data - none of which come from the suspects, or victims (speak with dead spells are a landmine) or eyewitnesses - we can't logically determine which of three people murdered which of four other people.   

Ah no, sorry that's my fault for not being clear. They don't necessarily need to speak to all those NPCs to come to the correct conclusion, its more the case that each of those interviews will give them a clue which they can use, but they could certainly figure it out with fewer. I've been doing some reading around the topic, and the advice I saw a lot was to make sure you had several clues pointing in the same direction, in case the players miss one.

But yes, on top of my worry about the inconsequentiality of the RP not being engaging enough (since they're always going to get the information they need regardless), I do worry that it might be too complex. Its quite hard to gauge, since it all makes perfect sense in my head!

1) Set up a revelation list.  This is a list of conclusions that the players need to come to in order to figure out who the killers are/where they are.

• Revelation 1: The murders are occurring in a warehouse by the docks.
• Revelation 2: The butcher is killing people and feeding their bodies to his pigs.
• Revelation 3: A crime boss is paying the butcher to provide this service and keep it quiet.

2) Once you have the revelations, create 3 clues for each revelation that the players might find, each that can help guide the players to the revelations.  You might have townspeople be able to provide information about the warehouse.  In the warehouse, there may be a butchers bib, or cleavers that can be identified as those of a butcher, or maybe some paperwork with the butchers name on it that was hurriedly dropped.  Maybe there is a note from the crime boss or his minions giving instructions or a list of victims to the butcher.  Maybe there is clothing of the victims at the pig sty where the pigs are.  Maybe townspeople can tell the players that they saw someone carrying a large back to the pig sty in the dead of night.

3) Assign clues to nodes.  The nodes could be people, places, or things.  Sometimes they'll be tied closely to clues.  At this point you assign the clues to these nodes so that when the players investigate them, they can find information about the other nodes.

• Node A: The Warehouse: Butchers Cleavers, Pigs fur, bloodstained note with deliveries to the butchers shop
• Node B: The Butchers Shop: A Matchbox to the crime bosses hideout, bloody clothes and identification of the victims, a manifest of incoming/outgoing shipments between the pig farm
• Node C: The Pig Farm:  Body parts buried in the muck that the pigs missed, A personal item that can be traced back to the butcher, a deed of ownership that the farm is owned by the crime boss.

Now you've got some locations for your party to visit, clues for them to accrue, and they can start putting some things together.  They may find the butcher at the butcher shop, and in that situation he can try to lie, or if the players really pin him down he could be another source of clues.  Notice that some of these clues point towards the next higher level, which would be a confrontation with the crime boss who wants these people to disappear.

In this example, the NPCs that the players can talk to really serve the purpose of just trickling them small bits of information about where they might be able to find clues.  I don't often pre-script what information they might be able to give, but often rather just use them as a vector to either help facilitate the revelations or clue finding.

You can break it down into Who, Why, and How. How more or less covers When & Where as well.

So the trick is to give several NPCs a motive - one hated the victim, another inherited their fortune, another inherited leading shares in a business, etc.

Then give some of those NPCs an opportunity, the How, in which they could have murdered the victim. These plots work better in a closed system, such as "someone died at a party" and no-one has left ,so someone there was the murderer. This is also the most familiar sort of mystery from TV shows.

Then, give all but one of the NPCs an alibi - the victim was seen after their opportunity, so they didn't do it, or their opportunity was interrupted by the butler, that sort of thing.

Then assign everyone the facts which, when unraveled, will point to the murderer. Make sure the Murderers facts are duplicated by other NPCs (EG, the murderer has the fact that the victim was seen at the bar at 8pm, and so does another NPC. The Murderer might then lie about seeing the victim at the bar, saying instead it was at 7pm, making it look like someone else had the chance to commit the murder). Then, the players will establish that some of the facts don't line up, and every one was learnt from George the Butler, and if they discard his information then all signs point to him, so George Dunnit.

You can add layers by giving the NPCs their own secrets, EG one says she didn't talk to the victim, but someone says they saw her talk, and it turns out she was talking to them because of some secret which could be a motive, but then the victim was seen later so she can't have done it. 

You'll probably have to make yourself a timeline with each NPC on it for your own reference, showing their interaction with the victim, and any twists you're throwing in.