A discussion of Keith Baker's post regarding the Skill Challenge system

[Gladius Legis]

Jesus, these threads are crazy.

Gambler's fallacy, anyone?

No. Just, no.

 

[Nail]

...is it me, or does every one of these threads end up including the inevitable post about Gambler's Fallace from someone who doesn't get the statistics?

Nope, it's not you. There really is that kind of trend: Ignorance masquerading as piercing logic.

Guys, please: the Gambler's Fallacy has not occurred with respect to modeling Skill Challenges. Put another way: "yep, we can google too."

 

[saitir]

It's because statistics and probabilities are among the least grokkable things in mathematics.

Added: The maths behind this post (Stalker0's thread is somewhere here) was checked by several people. And its right, because he got it by working from base principles, and you get the same if you use the cumulative negative binomial probability (Pascal distribution/waiting process) for the chance of failure (and invert it to get the chance of success).

Cheers, LT.

I give up, but one last try, I suppose...

Yes, his maths are correct. But he's started from a false premise. Therefore, its invalid.

First falsehood: The DC table isn't absolute. Its short table of of examples derived from the 'rule of thumb' DC calculation. Definition of a rule of thumb is 'not intended to be strictly accurate or reliable for every situation'.

Second falsehood: Any DM, other than a psycho working against his players, is actually designing skill challenges geared up to the party that's actually going to be doing it. Consequently, a higher proportion of the time skill checks will be against 'good' skills.

Third falsehood: Situational modifiers and 'aid another' bonuses aren't 'kludges'. They're an inherent part of the D&D game that partially help promote player creativity and interaction.

Does this mean the skil challenge system is perfect out of the box? Nope. Its poorly defined so that people don't understand the flexibility it really has for aiding roleplaying and building fun scenes.

Are the numbers borked? Possibly, but only in a sense of they just based them to exactly match the skill progression and that probably means individual DCs are too easy or too hard. Possibly they even erred on the side of making the unconsidered default DCs too hard. Especially given that once people see numbers in a table, they become blind to that lovely sentence right above it that says its 'just a rule of thumb'.

Is the number of success to number of fails ratio out of whack? Possibly, and if so, this is the only absolute flaw in the system as presented.

Of course, the converts to the church of Stalker0 won't take any of this at all. But you guys go have fun, and as long as you are, thats great, because thats supposed to be the point.

 

[Gladius Legis]

This might seem really weird, and I understand if people don't agree with me. I don't think criticizing how skill challenges are described in the DMG has anything to do with criticizing 4th Edition. What the issue has been is not whether skill challenges are a good idea or not; just whether the rules in the book work the way they are supposed to, and if they have given good examples of how to make them work the best way possible. As far as I can see everybody who has been suggesting changes to skill challenges in these threads, has done it because they like the idea so much that they wish they had a system that is good enough for it.

If I didn't like the new edition, I wouldn't bother looking for ways to patch up little problems in it. The skill challenge concept has the potential of being a really major improvement on the game, which is why it seems to have become some of us "4E defenders" have been so bothered with that it is not really working so well as written.

I really hope people don't think that someone who likes something has to be blind to defects in it. Or that people would think that a great new mechanic that needs patching to live up to its potential is an excuse to bash a game. That would just be lame.

I'll cosign on this.

 

[Oldtimer]

I give up, but one last try, I suppose...

And again someone refuses to read the thread and understand the math.

*sigh*

 

[Tervin]

I give up, but one last try, I suppose...

Yes, his maths are correct. But he's started from a false premise. Therefore, its invalid.

First falsehood: The DC table isn't absolute. Its short table of of examples derived from the 'rule of thumb' DC calculation. Definition of a rule of thumb is 'not intended to be strictly accurate or reliable for every situation'.

Second falsehood: Any DM, other than a psycho working against his players, is actually designing skill challenges geared up to the party that's actually going to be doing it. Consequently, a higher proportion of the time skill checks will be against 'good' skills.

Third falsehood: Situational modifiers and 'aid another' bonuses aren't 'kludges'. They're an inherent part of the D&D game that partially help promote player creativity and interaction.

Does this mean the skil challenge system is perfect out of the box? Nope. Its poorly defined so that people don't understand the flexibility it really has for aiding roleplaying and building fun scenes.

Are the numbers borked? Possibly, but only in a sense of they just based them to exactly match the skill progression and that probably means individual DCs are too easy or too hard. Possibly they even erred on the side of making the unconsidered default DCs too hard. Especially given that once people see numbers in a table, they become blind to that lovely sentence right above it that says its 'just a rule of thumb'.

Is the number of success to number of fails ratio out of whack? Possibly, and if so, this is the only absolute flaw in the system as presented.

Of course, the converts to the church of Stalker0 won't take any of this at all. But you guys go have fun, and as long as you are, thats great, because thats supposed to be the point.

First of all - I am not running Stalker0's system, as I think it is too far away from the written rules for me to use this early on. I agree with his math and analysis, but have another solution. So, not a church member here. Just someone who has similar opinions about a problem.

Secondly, if the system works as written if you just don't use the examples and tables that they have provided - I just want to ask you a simple question...

Lets say I am making a 6/3 challenge for a level 6 party. They don't have any relevant items or utility powers and I want them to have about 80% chance of success if they are able to get +2 bonuses (through cleverness) to about half their rolls? The challenge is aimed at skills that they are decently good at, but not really where they are experts. Aid Another will not be applicable, unless they think of something real clever - which will then just lead to bigger chance of success.

This is a hypothetical example, trying to be close enough to a real situation that it is worth looking at. Anyone who can give me a good answer to that problem basing it only on the rulebooks and common sense, will make me stop saying that the system needs work.

 

[The_Fan]

Thing is, the designers themselves have admitted thewe's something scwewy going on hewe. Seems that several different revisions of the system got combined together for the final draft, and it wound up not being what was expected. They had a meeting on Friday to discuss possible revision.

It's bad that errors like this are seeing print, but it's good that they're on top of it.

 

[grymckr]

You wouldn't know that the system needed you to be generous with the DCs, allowance of aid another and profligate with the circumstance bonuses without the mathematical analysis presented here. The problem is the system presented is so obtuse that even DMs that genuinely want to help the party out will have difficulty realize just how much "help" the PCs need in deviation from the presented rules. (Honestly, do you expect a DM, even one who wants to help the party, to know that he NEEDS to lower the DCs, give circumstance bonuses, and allow constant aid another cheese just to give the party a chance? Only if they've read this thread and learned that the system is broken.)

Read the sample skill challenge presented on pg 77 of the DMG, because it is the only guideline these non-ENworld reading DMs are going to get. In the example provided, the DM provides ONE circumstance bonus and there are ZERO aid another checks. On top of that, there is a check (intimidate in a social encounter) that is actually an AUTO FAIL. A DM reading this example is going to have an unrealistic expectation of just how much he needs to help out the party.

A DM who has read this thread, at least has an idea (though the system will remain obtuse, and it's hard to figure out exactly how much you need to help the players to give them, say, x% chance of succeeding). So math ftw. Personally I hate math, but I'm glad there are mathematicians around to crunch the numbers for me.

I give up, but one last try, I suppose...

Yes, his maths are correct. But he's started from a false premise. Therefore, its invalid.

First falsehood: The DC table isn't absolute. Its short table of of examples derived from the 'rule of thumb' DC calculation. Definition of a rule of thumb is 'not intended to be strictly accurate or reliable for every situation'.

Second falsehood: Any DM, other than a psycho working against his players, is actually designing skill challenges geared up to the party that's actually going to be doing it. Consequently, a higher proportion of the time skill checks will be against 'good' skills.

Third falsehood: Situational modifiers and 'aid another' bonuses aren't 'kludges'. They're an inherent part of the D&D game that partially help promote player creativity and interaction.

Does this mean the skil challenge system is perfect out of the box? Nope. Its poorly defined so that people don't understand the flexibility it really has for aiding roleplaying and building fun scenes.

Are the numbers borked? Possibly, but only in a sense of they just based them to exactly match the skill progression and that probably means individual DCs are too easy or too hard. Possibly they even erred on the side of making the unconsidered default DCs too hard. Especially given that once people see numbers in a table, they become blind to that lovely sentence right above it that says its 'just a rule of thumb'.

Is the number of success to number of fails ratio out of whack? Possibly, and if so, this is the only absolute flaw in the system as presented.

Of course, the converts to the church of Stalker0 won't take any of this at all. But you guys go have fun, and as long as you are, thats great, because thats supposed to be the point.

 

[xechnao]

Thing is, the designers themselves have admitted thewe's something scwewy going on hewe. Seems that several different revisions of the system got combined together for the final draft, and it wound up not being what was expected. They had a meeting on Friday to discuss possible revision.

It's bad that errors like this are seeing print, but it's good that they're on top of it.

Nope. They just botched this part of design due to incompetence and not bad organization. They just can't present a functioning model for what they want to do with the 20 sided die with simple common mathematics. It is not feasible. I hinted at this months ago -it sparkled a big thread here and an enormous one at rpg.net- and the community's conclusion was (at least from what I know) that the focus of skill challenges should not be the mathematic model but rather the meta-gaming model. They should have really only put advise on making a free-form roleplaying model that has the skill challenge effect they wanted, used in conjunction with a single effort skill roll and PC skill ability.

For example: provide DCs and complexities as normal. Only one dice roll should be made during the challenge though at a point chosen by the players that will have to do with how much they want to succeed (in this case roll say ending of challenge) or how few resources they want to sacrifice (say beginning). During the challenge players will have to combine efforts and based on their skill ability regards to the DC would provide hindrances or benefits for the challenge. Also their cumulative abilities by how they would be used and combined would provide the gravity of results of success and failure. The randomizer should only be of one effort: just choose when to make a roll and that specific roll's timing would influence in a specific way all the rest of the above.

 

[Dave Turner]

Jesus, these threads are crazy.

Gambler's fallacy, anyone? Someone might be able to reduce a skill challenge to a single x% chance, but that means next-to-nothing when one takes into account that what's being discussed in such a context is "the false belief that the probability of an event in a random sequence is dependent on preceding events, its probability increasing [or decreasing]* with each successive occasion on which it fails to occur."

Google saves, and ends the ongoing effect at the end of its round.

*added: mine

Pssthpok, rest assured that the Gambler's Fallacy has already been raised once as a possible challenge to Stalker0's math and has been soundly discredited by several mathematicians and laypeople. Unless you're trolling this thread, you should let this particular criticism go.