(Dialogue) skill checks: which is better any why?

[Shemar]

Also, Bloodline's system of a low "points" (0-5) system benefited avoided the "50 points needed at 49 points", as 4 points in presuade was vastly better at it than having 3 points.

Exactly. Skill ranges like that (1-50 or 1-100) are not only completely unrealistic (oh gee, I went from stutterng moron to the great philosophical debater in the space of 2 months! And I learned it all by bashing orc heads!) but they serve no purpose other than to present the usual leveling treadmill for those that are easily pleased or see upping their stats as a motivation to play. It is not so bad in PnP because 20-30 levels can represent 10 years of adventuring (and on occasion actual playing!) but in a cRPG where you get 20 levels in the space of a single storyline, it is quite ridiculous. So the problem is not really in how one implements skill checks but rather in how one implements skills and skill progression.

 

[Azrael the cat]

This is considerably better than the previous way, but I prefer the D&D way, which is kinda in-between. Say, the skill check has a difficulty of 50, so you throw a D20 to determine success. you need to have at least 30 speech to be successful, and then only if you throw a 20. With 50 speech, you will always succeed. But with 49 speed, you will also succeed almost always - except if you throw a 1.

49+1=50

I think a more important remark is why would 49 speed win you a speech check.

Obvious, isn't it? An average human in the lore has a speed somewhere between 0 and 1 (you increase it by 0.1 per skill point - speed is kind of difficult to get better at without magic, and small reaction time increases can still be beneficial). With speed 49, you can vibrate between time-streams, ala The Flash, and go back to try the other dialogue options.

 

[JarlFrank]

Giving it them a special color like Bloodlines did is okay, too, but more than that

I disagree. Hide the skill check completely. If a player is raising their speech skill, it should be organic and completely invisible to the player (as in special colors or [SPEECH])

This sounds almost as dumb as Bioware's "hide what your character is actually going to say" system. Hiding their own rolls from the player is newfag derp. If I'm playing an RPG, I want detailed output of my actions easily accessible (a concept that seems to be long dead). Your system would at best work for an adventure game with light RPG elements. But solving the problem of save scumming like this is like hiding damage rolls to solve Fallout's combat save scumming. It's the equivalent of removing a pimple with a chainsaw.

What.

Even in D&D the DM can make a hidden check. Like, when rolling for a sense traps check, the DM just says "You don't see any traps." after rolling, and not "You failed your roll therefore you don't see any traps, OH HEY THIS IS LIKE TOTALLY A HINT THAT THERE ARE TRAPS AND YOU JUST DIDN'T FIND THEM"

In dialogues, it makes sense, too. You don't always need to see what you rolled, or how high your chance to succeed is, or how many skill points you lack to actually succeed in the check.

Because failing and then having the game tell you "sorry, you need 5 more points in speech to succeed" just makes the player reload, go elsewhere, level up, and then return when he's raised his skill.

 

[Azrael the cat]

Also, Bloodline's system of a low "points" (0-5) system benefited avoided the "50 points needed at 49 points", as 4 points in presuade was vastly better at it than having 3 points.

Exactly. Skill ranges like that (1-50 or 1-100) are not only completely unrealistic (oh gee, I went from stutterng moron to the great philosophical debater in the space of 2 months! And I learned it all by bashing orc heads!) but they serve no purpose other than to present the usual leveling treadmill for those that are easily pleased or see upping their stats as a motivation to play. It is not so bad in PnP because 20-30 levels can represent 10 years of adventuring (and on occasion actual playing!) but in a cRPG where you get 20 levels in the space of a single storyline, it is quite ridiculous. So the problem is not really in how one implements skill checks but rather in how one implements skills and skill progression.

I went from L1 to L50 this campaign and got my speech to 50. I needed 40 in speech to convince the end boss to surrender.

v

I went from L1 to L5 this campaign and got my speech to 5. I needed 4 in speech to convince the end boss to surrender.

In Bloodlines each point of increase was a big jump. That means the opposite of what you were saying, as it makes going from 3 to 4 in a 5 point system the same as going from 30-40 in a 50 point system. The only way for the skill increase to be smaller, like you are saying, is to only be able to go to 5 in a 50 point system. And that only works if you set the skill checks much higher, otherwise you are just using a 5 point system with some extra space on the end.

The size of the numbers has absolutely nothing to do with how fast the skill is going up. All it changes is how big a number you use to represent each 'notch'. For the progression to be slower, the notches (whether that's 1, 2, 3, 4 or 10000, 20000, 30000, 40000 - it makes absolutely no difference) have to be much lower than the skillchecks. If the player can go up 5 notches for speech, and the skillchecks progress through to requiring 5 notches to pass, then you've got the same progression regardless of what numbers attach to each notch. Even if you say 'oh, but really the player can only get 5 notches this campaign, and there's a lore max of 500000 notches', if you're only USING the first 5 notches for your skill checks, then the other 499995 notches are just decoration on the end.

Trouble is, that if you DO use notches higher than that which the player can attain, you're making the content literally unattainable. No player can unlock that door or trigger that speech option. You've achieved the exact same affect that you would have by simply leaving that option out. Which means, that it's literally impossible to have more notches than the player can achieve, because any excess notches over the levels available just take the option away altogether (if a player who invests the maximum possible points in foozle-bending can reach a maximum skill of 5 foozle-bending-notches, and a skill-check requires 6 foozle-bending-notches with no way available for even a foozle-bending-specialist to buff his stats and get there, then what you've really done is make something that doesn't have the option of being foozle-bent).

The only way you can make levelling MEANINGFULLY slower is to make the campaign longer. That way the player still has maximum-achievable-notches=maximum-skillcheck-notches, but the player takes longer to get them and so his earlier skill is crap for longer. E.g. if you have a trilogy, then the player can spend a whole game with skills that are in the bottom third for the trilogy as a whole.

 

[Mastermind]

What.

Even in D&D the DM can make a hidden check. Like, when rolling for a sense traps check, the DM just says "You don't see any traps." after rolling, and not "You failed your roll therefore you don't see any traps, OH HEY THIS IS LIKE TOTALLY A HINT THAT THERE ARE TRAPS AND YOU JUST DIDN'T FIND THEM"

That's not hidden. You can see your own roll. So if you get a 2 you could always roll again until you get something close to 20. If you still don't find traps then there's only one other way to find them if they're there anyway.

A better analogy to what I was responding to would be to have the player check for traps automatically and never bother to tell him what he's doing. Either he find the trap automatically with his eyes or he finds the trap automatically with his foot. That's considerably different from how D&D handles it (assuming it's the same as in the video games)

In dialogues, it makes sense, too. You don't always need to see what you rolled, or how high your chance to succeed is, or how many skill points you lack to actually succeed in the check.

Actually due to how simplistic dialogues are handled in every cRPG, you do. There's no way to reasonably gauge what skill level you would need through experimentation or observation. Sure, a trial lawyer is likely to require a good speech check since they talk for a living, but something like a mechanic or a soldier would be up in the air. They could be completely socially awkward or conversational geniuses (speech skills can be useful to both professions but are hardly vital).

Because failing and then having the game tell you "sorry, you need 5 more points in speech to succeed" just makes the player reload, go elsewhere, level up, and then return when he's raised his skill.

There's nothing wrong with that. It's acceptable to go and fight weaker monsters to become stronger if you tackle something that's too strong for you early on, so why not do the same for speech?

 

[Undead Phoenix]

Neither. What I prefer is hard borders - but different levels of success or failure. So using your example of failing a skill check because of having 49 in Speech and need 50 to pass you'd still fail, but only a 'minor fail' as such, so the consequences wouldn't be as severe as if, say, you hadn't put any points into Speech.

 

[Darth Roxor]

What.

Dude, you are trying to argue with Mastermind

 

[Mastermind]

Neither. What I prefer is hard borders - but different levels of success or failure. So using your example of failing a skill check because of having 49 in Speech and need 50 to pass you'd still fail, but only a 'minor fail' as such, so the consequences wouldn't be as severe as if, say, you hadn't put any points into Speech.

I like this. Much better than Darth Bioware's "hide everything from the player" derp. Sadly most RPG developers only put in the absolute minimum when it comes to speech so it's unlikely to be implemented.

 

[Tigranes]

At this stage of CRPGs and what most players expect in the gameworld, hiding [SPEECH] carries various problems with it. If we use New Vegas as an example, hiding % would have been infinitely better, going some way to preventing metagaming with magazines and reloading. Ironically, they actually did this for one convo (Legion spy), where it worked very well.

But if you hide [Success/fail] and even [Speech] as a whole, then there's no way for the player to feel that he is getting good (or bad) use out of his Speech skill throughout his playthroughs. There's no good indicator for whether he should invest more, and how much more, in the skill. This would only make sense if every dialogue line, or every significant decision made through dialogue, could be affected by Speech, so that you could tell 'organically', as you might a skill like Outdoorsman or Steal in FO1/2.

I'd think that a better way to do it is to keep [Speech] tags, but not success/fail (which should be bloody obvious anyway ) or percentages, and in important speech decisions, i.e. ones where you can get away with a lot or resolve an entire quest by speech, selecting the [Speech] dialogue option leads to multiple choices where you have to consider the situation and the personality of the NPC. For instance;

"There's no way I'm giving up the hostages."
" [SPEECH] Let's not be so hasty. There's no harm in hearing me out, is there?"
"Well, I suppose not, but this better be good."
(A) "You were once a soldier, and a good one. You might have left that life behind now, but you know what can happen when you drag innocents into a war."
(B) "The men out there are getting their orders from General George. The same one that once said negotiation is the first step towards defeat. Keeping the hostages will just push him towards drastic measures."

In ideal situations, (A) or (B) might then branch off to a second tier of Speech choices. This turns speech not into a single do or die check, but a conversation and negotiation where uncertainty and risk plays a huge role in your decision-making on whether to go for the biggest rewards or play it safe. It also makes you pay more attention to the actual content of the conversation and the NPC - you can't skip all the lines then hope to play a Diplomatic character.

The only downsides I see to these kind of approaches are that (a) you need good writing, (b) initially audience expectations will be broken, meaning some idiots will complain, (c) pure Speech Boys who think their high skill can get them past every firefight will find it harder. I don't consider (c) a big issue because the whining about can I beat the game without touching a fly is retarded - as for (b), you might be able to get around it with good presentation (i.e. hire Pete Hines, call it Radiant Speech = )

 

[Shemar]

I went from L1 to L50 this campaign and got my speech to 50. I needed 40 in speech to convince the end boss to surrender.

v

I went from L1 to L5 this campaign and got my speech to 5. I needed 4 in speech to convince the end boss to surrender.

Your premise is wrong because you are assuming a linear skill progression. In my favorite system it would take 1+3+5+8+12=29 skill points to get from Unskilled (level 0) to Master (Level 5). Given that your character starts with 30 skill points total and gains 5 skill points per level it takes almost 4 levels of doing nothing else than pumping that one skill to get from 0 to 4 and almost 6 levels to go from 0 to 5. Skill levels do a whole lot more than just give you +1 to your skill checks, they open up new actions that were previously impossible or make actions that required skill checks now be automatic successes. Skill checks are themselves based on the main attributes that are more or less static through the lifetime of the character (so you don't start with strength 12 and end up with strength 48, you don't get to be 4 times stronger/faster/smarter just because you are more exprienced).

 

[Mastermind]

The only downsides I see to these kind of approaches are that (a) you need good writing,

You need good writing either way. In this case though, you also need writing in quantity, and for options most players are probably not gonna see too. It would make a better game, but I'm not sure how many developers (especially indie ones) would have the resources to pull it off.

 

[zeitgeist]

In general, I'd prefer conversation outcomes to be governed by a combination of character skills and in-character knowledge unlocking dialogue options that the player still has to choose between with. Just to (at least slightly) alleviate the player vs character intellect disparity which can be extremely frustrating.

Example: you could have a high [Persuasion] skill to unlock a couple of different persuasion dialogue paths, but you'd either have to find out some information in the game to point you to a specific conversation path you have to persuade the NPC with, or deduct this yourself (but have a higher chance of failure due to not knowing this information in-character - however, a combination of character skills could be used instead to deduct this information etc.)

I don't think a lot of effort should be expended on combating savescumming, metagaming, walkthrough reading and similar. That's always going to be there. But IMO one of the major reasons for savescumming to get the "optimal" outcome in dialogue situations is very simple: there are so few of them.

If there's only a handful of major NPCs in the game, only a handful of important conversations, and if all these conversations are presented in a win/lose context, getting the player a handful of important rewards... it naturally leads to savescumming or walkthrough reading. I'm pretty sure a lot of codex can recall the majority of important conversations from Fallout or PS:T.

In a gameworld that's built more organically where you never really know what the possible consequences of an action/dialogue choice will be, and where conversations aren't presented as the most important conversation yet which will surely decide your fate for the rest of the game, no matter what the conversation outcome is, players tend to accept their fate and move on.

This is somewhat comparable to letting XCOM rookies die during missions. The game fully supports this as a mechanic, and it's not as important to reload a lot (although some people will still do it, of course) to keep the entire squad alive in all missions, because of the steady supply of new recruits. Ideally, most players will only engage in savescumming when a really important squad member dies. On the other hand, if there was only, say, 12 characters in the entire game, each of them unique, savescumming would be the main way to play for practically everyone.

 

[Baron]

For those with the requisite Charisma I just unlock extra dialogue branches mixed in with the others. The chance of success of each line is kept hidden but should be reasonably apparent from the writing. Same with Wisdom, lines are added for those with low Wisdom, or high. Hopefully the player will click on the low Wisdom lines without thinking of the consequences (just like their PC); but stupidity is not enforced, just encouraged. I want players to choose a line of dialogue appropriate for their character, not pick a number with a higher probability of success. I think great damage was done to immersion by displaying the probability of every action, although obviously you need a subtle system to indicate low risk actions from nigh impossible.

One method I find useful to avoid the scum save is to not punish a player too heavily for choosing a line that might be in-character and fun for them. If failure is spectacular and beautiful, and hopefully even laugh out loud funny, the player might accept it and move on (obviously no instant death results.) I've also added situations where failure may be negative in the short term but beneficial in the long term. The player that is either freakishly lucky, or more likely a scum saver, will pay for it in the long term when it's too late to turn back. It's not necessarily about punishing these players just to retrain them that RPGs are not necessarily about 'winning.' It's about story, character and being kickass within your alignment until probability has you swallowed by an owlbear on dungeon level 4.

 

[JarlFrank]

What.

Even in D&D the DM can make a hidden check. Like, when rolling for a sense traps check, the DM just says "You don't see any traps." after rolling, and not "You failed your roll therefore you don't see any traps, OH HEY THIS IS LIKE TOTALLY A HINT THAT THERE ARE TRAPS AND YOU JUST DIDN'T FIND THEM"

That's not hidden. You can see your own roll.

No. No you can't. Please, go google "hidden dice roll" and check for the definition. It's when the DM rolls for you and doesn't tell you the result. This is often done with spot checks - the DM makes a roll without any of the players seeing, and depending on the result tells them something (you come into a room and see it's empty/you come into a room and see enemies hiding behind the furniture).

A better analogy to what I was responding to would be to have the player check for traps automatically and never bother to tell him what he's doing. Either he find the trap automatically with his eyes or he finds the trap automatically with his foot. That's considerably different from how D&D handles it (assuming it's the same as in the video games)

No, no it's not different from how D&D handles it. That's exactly what a hidden roll is. The DM rolls for the player and doesn't explicitly tell what he rolled or why he did so.

 

[Mastermind]

No. No you can't. Please, go google "hidden dice roll" and check for the definition. It's when the DM rolls for you and doesn't tell you the result. This is often done with spot checks - the DM makes a roll without any of the players seeing, and depending on the result tells them something (you come into a room and see it's empty/you come into a room and see enemies hiding behind the furniture).

A better analogy to what I was responding to would be to have the player check for traps automatically and never bother to tell him what he's doing. Either he find the trap automatically with his eyes or he finds the trap automatically with his foot. That's considerably different from how D&D handles it (assuming it's the same as in the video games)

No, no it's not different from how D&D handles it. That's exactly what a hidden roll is. The DM rolls for the player and doesn't explicitly tell what he rolled or why he did so.

That's not how I've seen it handled in D&D games but I'll concede that it may be this way in some of them. Either way, "D&D does it" does not make it any less retarded.

 

[Shemar]

About half the time my players do a skill check they do not know the result. It depends on whether their characters would know the result or not. For example bashing a door is an obvious open roll as the result is immediately evident. Checking for traps is not, as they should not know if they do not see traps just because they failed their check or because there aren't any. Stealth checks and dialogue checks are handled without my players knowing whether they succeeded or failed (although it is usually obvious by the results).

It varies from DM to DM. but my philosophy is if the character does not know something the player does not know it either. I like the same philosophy in computer RPGs as well. I don't think it is just a DnD thing either. I would handle skill checks the same way no matter what RPG system I was running.

 

[Mastermind]

About half the time my players do a skill check they do not know the result. It depends on whether their characters would know the result or not. For example bashing a door is an obvious open roll as the result is immediately evident. Checking for traps is not, as they should not know if they do not see traps just because they failed their check or because there aren't any. Stealth checks and dialogue checks are handled without my players knowing whether they succeeded or failed (although it is usually obvious by the results).

It makes sense for someone not to know the result for certain, but they should know how well they performed a task. IOW, reveal the result of their roll but give a neutral result ("you didn't find anything") if they fail. This does not automatically mean they have to reveal something's there. The DM can just say "you don't find anything" if you don't meet the DC. the player can (and should) then be able to redo the check on their next turn if they did a shoddy job the first time.

 

[Shemar]

About half the time my players do a skill check they do not know the result. It depends on whether their characters would know the result or not. For example bashing a door is an obvious open roll as the result is immediately evident. Checking for traps is not, as they should not know if they do not see traps just because they failed their check or because there aren't any. Stealth checks and dialogue checks are handled without my players knowing whether they succeeded or failed (although it is usually obvious by the results).

It makes sense for someone not to know the result for certain, but they should know how well they performed a task. IOW, reveal the result of their roll but give a neutral result ("you didn't find anything") if they fail. This does not automatically mean they have to reveal something's there. The DM can just say "you don't find anything" if you don't meet the DC. the player can (and should) then be able to redo the check on their next turn if they did a shoddy job the first time.

That sounds pretty horrible. If the game mechanic is to keep rolling until you get a good result, why bother rolling anyway?

 

[JarlFrank]

Checking for traps for an hour isn't the point of the game. In my current game, the elven rogue (the one char in the party with the best search skill) just looks around and I throw a d20, then either say "You see trap X, X and X" or "You don't see any traps", without telling the exact result of the throw.

In some cases, I tell the throws, like when someone does something stupid and fails/succeeds. Or I let them throw their own dice. Like when our elven ranger tried to show off how awesome he is by jumping over our cleric's head while having the sorceress on his shoulders. I said "You threw a 7.", which already made everyone go "haw haw this is going to be fun now" before telling the results.

But in a dialogue... nah. It would be stupid. It encourages better roleplaying if you hide the results. Also, it allows for better storytelling. Imagine one player has a conversation with a guard, trying to convince this guard to get them out of the city because they're wanted by the local lord, and makes a speech check. He rolls a 1, failing. Yet the guard pretends that he'll take the party to the place they wanna go, only to cross them and sell them to his lord.

Now, if you tell the player "you rolled a 1", he knows "I failed". Which means the player will notice that the guard wants to cross him instead of really trying to help. It just makes NPC reactions to player choices too fucking obvious and predictable.

So no, mastermind, it's not a good thing to have every single roll be transparent.

 

[shihonage]

There's no re-rolling in my game. All checks are either hard or pre-seeded (and hard).

No matter how much you reload, you can't open that safe, because the roll was decided at the beginning of the game. If you increase the relevant skill, it may open despite that. But savescumming is not an option.

I dither this inflexible scheme by providing more opportunities for various checks, such as picking the lock or shattering it if you're very strong (as the NPC does in the end).

 

[Mastermind]

That sounds pretty horrible. If the game mechanic is to keep rolling until you get a good result, why bother rolling anyway?

Depending on the task there could be time constraints. If you have to hurry through an area you might not be able to afford more than one or two checks. If there are no story related time constraints, why shouldn't the character be allowed to roll until they get a decent number? They have all the time in the world to do a thorough check, no?

 

[Mastermind]

But in a dialogue... nah. It would be stupid. It encourages better roleplaying if you hide the results. Also, it allows for better storytelling. Imagine one player has a conversation with a guard, trying to convince this guard to get them out of the city because they're wanted by the local lord, and makes a speech check. He rolls a 1, failing. Yet the guard pretends that he'll take the party to the place they wanna go, only to cross them and sell them to his lord.

Now, if you tell the player "you rolled a 1", he knows "I failed". Which means the player will notice that the guard wants to cross him instead of really trying to help. It just makes NPC reactions to player choices too fucking obvious and predictable.

So no, mastermind, it's not a good thing to have every single roll be transparent.

Your example's not particularly smart either. If the NPC is good enough to presumably pass the speech check if he rolls better he should be able to read the guard well enough to know if it's a trap or not as the two characters' skill in the matter should be similar. It would take an unrealistically stupid PC to fumble the persuasion attempt that badly and not know something went wrong.

 

[JarlFrank]

Yes, but would actually telling the roll add anything to the situation except for metagaming? No.

 

[Black Cat]

Even in D&D the DM can make a hidden check. Like, when rolling for a sense traps check, the DM just says "You don't see any traps." after rolling, and not "You failed your roll therefore you don't see any traps, OH HEY THIS IS LIKE TOTALLY A HINT THAT THERE ARE TRAPS AND YOU JUST DIDN'T FIND THEM"

Given we are playing a roll role playing game knowing how things are working is kind of essential, otherwise you can't really plan your character and build in any detail, and it is even more so in computer games, given the pretty vast amount of bugs, design idiosyncrasies, etc, that you can't actually discuss with the GM.

And let's be honest about this: Only GMs with kind and weak souls use hidden rolls. I prefer to tell them I'm going to roll for trap detection, that they failed, and that therefore not only they don't know there are any traps but that they believe there to be none if the roll was bad enough, and therefore any sucesive action that I believe is trying to detect a trap will make them lose experience permanently.

It adds a level of fatalism, and their terror striken faces every time they have to take a step and are waiting mechanical death to fall upon them is priceless. :3

 

[Shemar]

That sounds pretty horrible. If the game mechanic is to keep rolling until you get a good result, why bother rolling anyway?

Depending on the task there could be time constraints. If you have to hurry through an area you might not be able to afford more than one or two checks. If there are no story related time constraints, why shouldn't the character be allowed to roll until they get a decent number? They have all the time in the world to do a thorough check, no?

No. The one check they roll IS the best check they can do. They fail thay simply don't find it. The End. In a situation where they were meant to find it anyway I just tell them they found it without going through a pointless repetitive skill check.