My last session ended with a very serious crime being commited. The PC:s have retreated to their home but being famous the guards know where they are. 

So that's how the next session will start but im not sure what the players will do in response to a lot of guards trying to arrest them. How do I plan a session without knowing what is going to happen? Should I create a bunch of possible scenarios I can pull out at a certain time or just wing it and hope for the best?

My group is rather unpredictable and in a case like this they would 98% of the time fight it out. Make it a combat encounter they cannot win but dont kill them. have the guards grapple them and attempt to shackle them. Of coarse give the an opportunity to resist and whatnot using skill checks. Have the guards use nonlethal methods after a while to avoid death. And in the worst case if the players all drop due to taking damage dont kill then but rather have them wake up in a jail. Also you could do something like have them make a DC check to resist being knocked out when a guard tries to subdue them. You could even make it a chase and separate the party really leaving the the players overwhelming. Hell, you could even let one get away only to really be on the local shit list as he works to get his fellow party members freed. Or maybe you will get lucky and they will go quietly. Basically your simply going to have to railroad the players and there is nothing wrong with it. Sometimes you simply have to because they just dont pick up on any of your plot hooks and the like. My group does this all the time. I have the next session ready and all plained out then when we play they decide to go wondering off. In most cases I let them because I try to offer as much freedom in the game as I can in most cases. So I just roll with it and makes approbate changes to the current session an the next.  But when you in a way point out that they need to move on but dont. Its time to force them but in a way that feels natural.        

Depending on the crime and the type of game, I would weight the confrontation in favour of one outcome or another. If for example it is a somewhat standard heroic campaign and the players just murdered half the well loved local nobles for the fun of it, then your guards rocking up with overwhelming force and a high chance of a TPK through either death or imprisonment seems fair, but even if they do escape they are now going to have difficulties doing anything in that kingdom. If on the other hand it is an evil campaign then the players taking reasonable actions to defend themselves should probably be enough to keep the character safe-ish.

If this is a large settlement then forces that can deal with troublesome adventurers seem plausible. So depending on the game type, severity of crime, and resources available to the settlement you will want to figure out just how dangerous the guards elite forces are.Then you need to think about the immediate and long term consequences of the party surrendering, being defeated and captured, stealthily evading capture, fighting their way out...

I would not worry about planning it all out in detail, just make sure you have a few ideas for how things can advance in each scenario and that should be enough to fill in the gaps as you go.

Two points here:

• I've had this happen. Try and arrest them, if they resist, have overwhelming guard force take them out. The rules specifically state that you can opt to do "non-lethal damage" with a regular attack, so you can "TPK" the Party without actually slaying them. Then put them on trial, then throw them in jail. That's what I did with a Party. They lost 3 months of in-game time, and missed several important deadlines and rendezvous. Actions have consequences.

• Let me find my soapbox ... you actually never know what's going to happen. But dealing with a "free form plot" isn't hard - it just requires the DM to do some role playing. What's the current adventure conflict about? Who is involved? What do the people & organizations involved want? What do they know & believe? What is their personality like; what tactics will they employ as a matter of course; what tactics won't they use? If you can answer all that ( and IMHO if you're DM'ing an adventure you should be able to ), then figuring out what everyone will try and do next is fairly easy: the cult will try and assassinate the Party for desecrating their shrine; the cabal of Wizards will give the Party the aid they're asking for ( but will attach a mission to the Party: a favor for a favor ); the city guard will just try and conduct the prison transfer the way they're supposed to .... etc. .... Where their actions collide, you have events that the DM has to decide how they'll resolve; and where their actions collide with the Party , you have encounters for your adventurers. You can run whole adventure arcs that way - in fact that's how I run all my plots.

IMHO, trying to build branching plots, and account for all possible scenarios is a ton of work, where you are absolutely guaranteed that you'll have to throw most of it away - and there's a good chance that the Players will find a contingency that you didn't plan for, and do that. If you're figuring out what happens next in "real time" role playing, they can never catch you off guard.

Adding to what Vedex said, creating a bunch of branching plots for this would be some work.  It may be a lot easier to break things down as follows:

• Who is the person that is sending the guards?
• What is their relation/impression of the player characters and what they have done (for good or bad) in the general area?
• Are they sending the guards as an "escort" to then have a discussion, or are the guards coming to bring the party in and throw them into a dungeon right away?
• Create a few major NPCs and try to understand their motivations, then you can make decisions based on what the PCs do.

Knowing some of the above you can act based on the motivations of whomever is in charge.  I would add a couple of personalities for some guard types also. Additionally if the PCs are now "wanted" and refuse to come in for questioning etc,  the person in charge could issue bounties which could attract rival parties etc.

my $.02

Quote from Koradgee >>

 Create a few major NPCs and try to understand their motivations, then you can make decisions based on what the PCs do.

I was going to give this advice. Recently I struggled with the same thing - unpredictable party members. Having a strong idea of the motivations of your key NPCs and their goals is probably the most useful thing to have. The rest comes from improvising and responding to what they do. Have a few ideas in your head, but be prepared to pick and mix.

Also see if there are links into character backstories as well as the main plot. These things can help nudge them in a certain direction.

The problem is - in my opinion - not that Party members are unpredictable, but the fact that some DMs need them to be predictable.

To be clear - I'm not accusing anyone in this thread, I'm just opining about a general trend that does touch lightly on this issue ( next I'll be telling the kids to stay off my lawn ).

If you cannot change circumstance ( and you can't here - apart from re-programming your Players, which I think has more problems than it's worth :p ), adapt to your game design to the existing circumstances.

I think that too many DMs expect that they can replicate the predefined structure of video game design in their adventure planning, and that they will - and what's worse should - be covered, and that the failing of an impractical and badly fitting design approach is somehow a problem created by their Players.

TTRPGs are not video games. They have a flexibility and open ended aspect not available in computer games - or we'd all just play video games instead - the video/audio production values are often better :p

TTRPG adventure design and execution needs to take this into account.