D&D 5E A more dynamic skill system?

[Beleriphon]

The dice control the flow of information, and some rolls should (statistically) fail, but to solve the matter requires the players, not the dice. In 5e, the dice are unlikely to fail, and the players always have the skills they need.

Did I explain it better?

Maybe, I'll lay out a scenario and tell me what you thinkl.

Lets take a look at a good solid murder mystery. I'll grab the basic plot from Agatha Christie's Murder on the Orient Express. We'll set it in Eberron and the adventure is now Murder on the Orien Express (I'm brilliant!). The lightning rail starts in Sharn, and doesn't stop until it reaches Flamehold in Aundair, roughly 1000 miles away.

Here's the scenario, largely cribbed from the Wikipedia entry for the novel:

Early in the morning, party is wakened by a cry from the vicitim's compartment next to theirs. The wagon conductor responds by knocking on the door and a voice from inside responds in Elvish, "It is nothing. I am mistaken".

In the morning the victim is discovered, the an Orien representative on board asks the party to investigate (who care why, its D&D lets just roll with it for now). The party determines that the victim has twelve stab wounds. The window is left open in the victim's compartment, presumably to make the investigators think the murderer escaped out the window, but there are no footprints outside the window in the snow. A handkerchief with the initial "H" is found in the compartment, a pipe cleaner, a round match different from the matches the victim used and a charred piece of paper with the name "Armstrong" on it.

I'll pause here, at what point do we use Skills to find any of these things? A search should make them obvious. The next section is where I'd use a skill though.

The piece of paper with the word Armstrong on it helps the party figure out who the victim really is and why someone would want to murder him. A few years back, a man named Cassetti kidnapped a three-year old girl, Illidra Armstrong. Cassetti collected a ransom from the wealthy Armstrong family, but killed the child within two hours of kidnaping her. The party concludes that the victim is actually Cassetti. The voice in victim's compartment could not have been the his, since the victim spoke no Elvish.

I'd use some skills to put together than 1) the burnt paper refers to the kidnapping and murder, 2) Cassetti is the victim and 3) Cassetti doesn't speak Elvish so somebody else was there.

We go on from there. I wont spoil the actual killer from the novel but skill uses should reveal facts about clues rather than obvious clues themselves. So, talking to a suspect should get some facts, it would be skill checks that help determine truth from falsehood. Going into a murder scene and searching should produce all available an obvious clues with skills providing less obvious links between clue or what the clue actually means.

 

[Jd Smith1]

Maybe, I'll lay out a scenario and tell me what you thinkl.

Lets take a look at a good solid murder mystery. I'll grab the basic plot from Agatha Christie's Murder on the Orient Express. We'll set it in Eberron and the adventure is now Murder on the Orien Express (I'm brilliant!). The lightning rail starts in Sharn, and doesn't stop until it reaches Flamehold in Aundair, roughly 1000 miles away.

Here's the scenario, largely cribbed from the Wikipedia entry for the novel:

I'll pause here, at what point do we use Skills to find any of these things? A search should make them obvious. The next section is where I'd use a skill though.

I'd use some skills to put together than 1) the burnt paper refers to the kidnapping and murder, 2) Cassetti is the victim and 3) Cassetti doesn't speak Elvish so somebody else was there.

We go on from there. I wont spoil the actual killer from the novel but skill uses should reveal facts about clues rather than obvious clues themselves. So, talking to a suspect should get some facts, it would be skill checks that help determine truth from falsehood. Going into a murder scene and searching should produce all available an obvious clues with skills providing less obvious links between clue or what the clue actually means.

I would never write or use a murder scenario where the killer left behind a monogrammed hankie. Or a killer in a captive audience.

 

[Beleriphon]

I would never write or use a murder scenario where the killer left behind a monogrammed hankie. Or a killer in a captive audience.

Okay, but specific details of the novel not withstanding how would you handle a murder scene? What do skills do for a character? Do they find clue, or do they explain clues? Do they do something different?

 

[Garthanos]

So basically no one can fail ever.

That is not understanding fail forward.
When people talk about Fail Forward in RPGs, they mean that failure should not stop the action, and failure should always have interesting consequences.

Fail Forward

A blog for practical advice, tools, resources, tips, and examples of how to run a roleplaying game (RPG).

www.runagame.net

They suggest calling it success at a cost but I am not sure that carries the meaning more clearly ;p or is actually implies just one variation

 

[Fanaelialae]

Basic competency, yes. But automatically establish a cause of death? No. That requires specialized training.

Of course an investigation should not come down to a single roll; that is a scenario design flaw, not a system flaw. A decent murder investigation should incorporate 20-30 pieces of data or more, any one of which being equal to the others. The players could possible solve the matter with ten or more (or at least frame someone), but ultimately they will have to figure things out from the data; the dice just determine how many and which items they acquire.

A character trained in Medicine presumably has that training. Or at least a close enough approximation thereof.

In the real world I write software and have no medical background unless you want to count high school bio. But if you show me a body that's been stabbed in the chest, I can hazard a reasonable guess as to why the person died, even if I can't provide a textbook medical examiner's answer. Unless something weird happened, like the person was stabbed, but that didn't kill him so his assailant drowned him, I'm probably going to be correct.

Having 20 to 30 clues locked behind checks, many of which are redundant, isn't a terrible way to go. Odds are that the players will be able to succeed at 30-50% of the checks just based on the die roll alone (unless the DCs are really high).

However, you're still locking all the passages in the dungeon and requiring the rogue to pick them so that the party can proceed. There are lots of potential routes but if the party runs a streak of bad luck and flubs all of the doors that they need to proceed, then the dungeon is done.

I run a sandbox, as I've said, but even so I don't like it when my players are frustrated because they can't get any traction nor am I a fan of when my work gets wasted (although that's not as big an issue, since most anything that isn't used can always be recycled down the line). Especially when it's simply a bad streak of luck because no one is able to roll above a dang 5 (which has happened on multiple occasions) rather than any fault of the players.

In my current campaign, which is presently nearing its end, I started using handouts for the first time. On these handouts, I added rumors that the characters would hear around town in between sessions. Early on in the campaign, I decided that a barghest would be preying upon the goblinoids living in the town's ghetto. So I seeded a rumor about it into the handout, with the idea that the players would ignore it for now and be super impressed with my use of foreshadowing, once it built to an issue and NPCs asked them to get involved. Well, of course, they latched on to that. However, they hadn't made the NPC connections they really needed to figure out what was going on (akin to going to a dungeon they weren't appropriately leveled for).

It wasn't impossible for them per se, but in this case it was improbable as it would have required a few blind leaps of intuition. As I said, I hadn't anticipated them going after this job at such a low level, without even any impetus of reward.

To their credit they did their best, talking to NPCs and such, but were largely spinning their wheels, despite picking up a few decent clues, whose veracity they were unable to verify. Through sheer random chance, they decided to stake out the local tavern on a day when the payment to the local crime Lord was due. And ended up haring off on a totally different adventure that I also hadn't prepped, wherein they busted a pair of guards who were on the take and had a "friendly" chat with the local crime lord (who genuinely had nothing to do with the murders, though they only pretended to believe him). Later, when the NPCs asked the party to investigate the murders, they were convinced it was their frenemy the crime lord, so they decided to do other stuff instead, and eventually the barghest returned to hell and the situation resolved itself without their intervention, though there were a lot of unhappy goblins in the ghetto. (The characters eventually did sort of make it up to the goblins by donating a lot of gold to fix up the ghetto and make it a genuinely nice place to live.)

The point being was that I really lucked out. It was a fun session in which my players ended up thinking me a genius with deeply entangled plot threads, when in fact it was two loosely outlined NPCs and they managed to jump from one thread to the other without even realizing it, assuming connections that didn't exist.

That's a beauty in emergent gameplay, but it could have very easily been a terrible session where the players spun their wheels the entire time and couldn't make any headway. That's how I see it anyway.

 

[5ekyu]

Ok so while it still feels to me like a very amorphous search for an imprecise better, here are some things I do.

First, at any tier, remember that a failure on an ability check is not pass/fail - its options are complete success "overcome the challenge", fail "no progress" and "some progress with setback". That third option gives you a massive amount of flexibility.

As such, for a complex scene, I set the DC for "complete success" pretty high and then I judge yo what degree they get somewhere with setback or nowhere at all based on the events, the roll, proficiency or not etc.

That is all "by the book".

Second, if you look at the DMG on social encounters you find it's a lot more than just one die roll. Unless you start at friendly with no risk, you are likely looking at needuping to gather info to get leverage to shift attitude to get to yourcapptoaches getting you results. Indifferent requires a 20 DC to get even slight risk actions accepted.

Still "bumpy the book" here.

So, if you as GM present them with a variety of options to "work" an NPC, you pen up a lot of different paths for them to choose.

Third, off-book, I tend to treat any task that is not representing a less-than-minute activity as a race to three set. To pass you need to get to three passes before three fails. Each fail gives you a little and makes things harder if you keep going the same way (disad). Butbif you take the "little" and go a different tact, back to a regular toll.

Example - looking thru old newspapers for info on the family. One success gets you info on doctor, reference to another clue, court case dismissed (one success). Next fail results in realizing the papers in file are incomplete and likely not finding more (now at 1 success 1 fail.) If you keep at the library you roll at disad, taking more time, and maybe someone gets suspicious. But if you decide to follow the court records or go find the doctor or find out who could remove files from library... you get back to normal checks in another direction. Obviously, this is in addition to normal advsntsge and disadvsntage.

So, the race to three is kinda like skill challenge in that it keeps spawning new options likely involving different skills and specialties.

Finally, the biggest key is The Three Rs - robust, reactive, resilient. Each encounter needs to be robust (Offering a variety of approaches and risks), reactive (changing as events unfold - choices opening and closing as thing change) andcredilient (interesting or fun outcomes and results regardless of how things go.) Especially for "mysteries" and socials- that keys on complex NPCs eith thrirbiwn ideals, flaws, bonds, resources etc.

For setting DCs I tend to use the DMG guidelines, plus the 5 up down for advantage. That tends to give DCs between 5 and 25 based on the skill " opposing" you. So, if they were proficient in relevant skills, had abilities above norm and had a lot of planning or resources or time you are looking at DC 25. For each of those missing, drop by 5. So, not proficient, no aptitude to speak of, unplanned or rushed - likely 5. Circumstances for scene can adjust those.


One last thing, if multiple PCs try stuff "for more rolls" I resolve the failures first with those setbacks usually involving lost or destroyed clues or at least making things harder. "Found the murder weapon but smudged prints or contaminated blood."

Obviously, these all focus on the mechanics side and how they provide opportunities for the PCs to choosec between. These help feed the roleplay but the roleplay drives these. Mechanics add and remove trays from the buffet but it's the roleplay choices that still make up the meal you have chosen on your plate.

 

[dave2008]

Did I explain it better?

Yes, thank you!

Well, my view is this: in an investigation, the (successful) skill rolls give the players some raw data; the players use that raw data to formulate an expanded investigative plan, which requires role play, more skill rolls, and ultimately more raw data. The key is never to let the players get all the raw data, or clues, so that in the end, the players will have to work out the who, how, and why themselves.

However, you can still do this with the broad skills in 5e. There is nothing saying a success requires you give them all of the information required. There are some easy ways to handle this:

1. Require multiple checks. You can set this up similar to a 4e skill challenge if you want where ultimately you need multiple success before a set number of failures to fully complete the task. Of course with some successes along the players get some of what they need, but not everything. This can require multiple checks of the same or different types.
2. Set the DCs high, so you need more than a roll to get the answer. You need a roll + a situational bonus (good tactics, roleplay, clue, etc,) to provide a good chance of success.
3. Combine 1 & 2: the first part of the challenge is a DC 10, 2nd DC 15, third DC 20, etc. You can then provide situational bonuses as the checks get more difficult.
4. Degress of success / failure. If you succeed by 1-4 you get X info, if you succeed by 5-9 you get y info and if you succeed by 10+ you get everything!

 

[Tonguez]

Well, my view is this: in an investigation, the (successful) skill rolls give the players some raw data; the players use that raw data to formulate an expanded investigative plan, which requires role play, more skill rolls, and ultimately more raw data. The key is never to let the players get all the raw data, or clues, so that in the end, the players will have to work out the who, how, and why themselves. As it is, with the broad-application of the 5e skills and the unlikely chance of failure, the players will get too much information for zero role play.

In our last campaign, I kept a murder investigation loosely based on the Ripper killings going for three sessions (five players), 14+ hours of time 'at the table' plus the player discussions during the week.

The dice control the flow of information, and some rolls should (statistically) fail, but to solve the matter requires the players, not the dice. In 5e, the dice are unlikely to fail, and the players always have the skils they need.

Did I explain it better?

Are you not just describing a multi-layered skill challenge? That can be done with the current skills Investigate, Medicine, Survival - just dont give all the answers at once, scatter clues and red herrings around your setting and then require the PCs to go an look for them, sort which are relevant or not and put the pieces together.

 

[doctorbadwolf]

So basically no one can fail ever. How is that more dynamic?
Whats needed is that failure have consequences and that not all parties can automatically do everything by the virtue of existing. That would put an actual gameplay element back into skills.

Well, no, but if you’re only interested in making assumptions and arguing against them, have fun.

 

[Jd Smith1]

Okay, but specific details of the novel not withstanding how would you handle a murder scene? What do skills do for a character? Do they find clue, or do they explain clues? Do they do something different?

Perception, a forensic skill, an autopsy skill, an interview skill, an interrogation skill (for when the PCs get frustrated and start breaking fingers), a drawing skill (if an era without cameras), a camera skill (for an era with cameras), should cover the basics.