Traitors

>Have a plot, where somebody is working against the party
>It's a shapeshifting demon
>One of the NPCs assures the party that it must be someone in a position of influence
>There are eight possible candidates
>Party starts investigating
>Drop hints that a few of them might be the demon they're looking for
>But the thing is
>I haven't decided which one of them it will turn out to be in the end
>Or it might be a totally unrelated, ninth person they've also met
>I probably won't decide in quite a while
>Possibly until I flat out reveal the traitor

Is this okay? As a player, would you be pissed?

Doing the Agatha Christie, i see.
Honestly, as long as it makes sense and is consistent it'd be A-OK with with me.
(But don't tell your Players that you didn't have it planed out all along, ever)

If I find out, yet. Just don't let me find out. If you can keep the clues interesting and the twists engaging I wont suspect anything, but if everything is a grey blob of unclear clues the investigation is obviously not going to be very fun.

Yes, I would be enormously pissed if I found out. Depending on how good you are at perpetrating subterfuge and how perceptive your players are, they might not realize of course, but at least speaking anecdotally, that kind of Heisenberg reality where the GM casually alters the world to provide drama for the PCs is abhorrent.

> That NPC that isn't a traitor, just an asshole.

Totally forgot to mention - while I totally intended that the party will try to find out who is working against them, the game as a whole is not a detective, and I will not let the investigation be the main focus. I intend the traitor to reveal herself eventually on her own.

Honestly if the party has decent reasoning for why they settle on a particular character i dont see why it shouldnt just be that character, see what their interested in maybe have the first character they start investigating be murdered by the demon for whatever reason and have a few suspects have alibis they can discover to narrow it down and throw them off at the same time. So long as the partys reasoning is sound enough and theres no definite answer this is a good way to make them feel like they solved it without there really being a set soloution. You would just never tell them this fact. If theres no reason it couldnt be whoever they pick and you decide its someone else with no setup theyre gonna think youre just being a dickass

it can work, creating drama and exciting situations its the GM's work after all
but yeah this kind of thing sounds shitty...
why are the players even making an effort on finding the demon if no matter what they do they won't catch him until the GM decides its time?
taking away players agency just doesn't work

Sounds like every NPC my DM ever came up with

it is a shitty thing to do
your players are thinking and taking actions based on the evidence you give them
you are pretty much treating them like idiots and lying to them. I personally would be pissed.

Even with Christie it's at least always possible to deduce who the culprit is before he's unmasked.

You would be pissed. I would be pissed too.
But only if we found out

>It's not a problem to say, cheat on your wife/gf, as long as you don't get caught.

Depend.
If you do it to "trick" your players by switching the traitor identity at the worst moment then fuck you.
If you do it because you want to follow the flow of the game, why not.

A shapeshifting demon could successively impersonate different people though.

So all the work they put into this investigation will be pointless? Even if it's not a mystery game, as long as they're enjoying the investigation why not let that be the focus for now? Putting in effort to find who the traitor is just for them to reveal themselves no matter what seems really shitty. I for one wouldn't be satisfied with that result.

they will find out, mostly because I think the reason OP is doing that its because he cant write a proper traitor plot so he keeps swaping it around. anyone can notice when the plot its bullshit behind the scenes, and the longer the adventure gets the harder it will be to get away with bullshit.

You basically have two approaches here:
The Christie: decide who the culprit is and lay out clues so the party can follow and reveal them (or massively misinterpret and finger the wrong guy).
The Stout: drop clues that can feasibly be connected to every suspect and decide on the culprit later. is a classic twist, too.

I'd say, start with a Stout. If your players don't seem to enjoy the investigation. proceed. If they do, retool at the earliest opportunity.

Semi-related: if I moved a quest objective around inside a non-linear dungeon to be in the last place the party looks, would you be similarly agitated?

Yes.

not that guy but I would.
the GM its not the players enemy and dungeons are challenges, if you keep moving goalposts you are treating your players like idiots.

>>Or it might be a totally unrelated, ninth person they've also met
No matter what you decide but really don't do that.

I myself would do it like this:
Have one obvious suspect to begin with.
He/she is shifty, secretive and kind of an asshole.

Use this how much your players are actually into it. If they really aren't then just abandon the thread there, have him be the bad guy and move on.
But if they are into it he get's killed by the actual traitor, and all his actions get brought into a new light:
He was secretive and an asshole because he didn't trust the PCs, he was at the scene of the last incident because he was on the actual traitors heels, etc.
By this time you should have established enough things about the other characters that you yourself know who the real traitor actually is.

...

Wouldn't the objective always be in the last place the party looked?

Because after that they stop looking.