[X] My Assistant

[deek]

Looks like someone found out the mathematical foundation for these skill challenges are based on Paschal's Triangle and WotC only used a third of the triangle instead of half, thus reducing all probablility. It seems all that needs to be done is get successes back to half the triangle...and the least obtrusive way is to allow players the ability to remove remove failures.

I believe this is going to be the official fix, allowing players to recoup failures in some form, whether that be side challenges, bribes, action points or healing surges...

Hehe...assuming anyone here cares:)

[FrankTrollman]

Looks like someone found out the mathematical foundation for these skill challenges are based on Paschal's Triangle and WotC only used a third of the triangle instead of half, thus reducing all probablility. It seems all that needs to be done is get successes back to half the triangle...and the least obtrusive way is to allow players the ability to remove remove failures.

I believe this is going to be the official fix, allowing players to recoup failures in some form, whether that be side challenges, bribes, action points or healing surges...

Hehe...assuming anyone here cares:)

Uh... no.

Pascal's Triangle is a useful and efficient method of memorizing and generating combinatorials. Since you're dealing with dice series, you are perforce using combinatorials in great abundance. If you choose to use the Triangle, that's fine. Personally I just do all my combinatorials in my head out to about 11 or so. For example, there a 11 ways to get 10 successes and 1 failure, 55 ways to get 9 successes and 2 failures, 165 ways to get 8 successes and 3 failures, 330 ways to get 7 successes and 4 failures, and 462 ways to get 6 successes and 5 failures. Now 11 attempts is the model for any set where the number of successes and number of failures allowed together equal 12, so in the basic rules it models a Complexity 3 challenge (and you disregard the possibility of 4 or 5 failures when calculating your chances of overall success). With the revision Moon Hawk was talking about earlier it models a Complexity 2 challenge (and 4 or 5 failures is OK).

But there's nothing magical that happens when you grab the 330 combinations of 7/4 and 462 combinations of 6/5. They are just combinations. It's... just faces of the dice and nothing more.

I mean yeah, the probability is busted, but going "Bam! Pascal's Triangle!" isn't going to solve anything. Pascal's Triangle is just a way for people with middle school mathematics to calculate probability in long form. If you can't do it any other way I guess that pulling out the triangle as a study aid isn't the worst thing you could do, but starting to look at the real probabilities generated and required by the system is the beginning of making a decent system, not a magic panacea for the broken system already in place.

-Frank

[raphabonelli]

Hey Frank. Did you read Keith Baker's answer to Skill Challenge problems?
http://gloomforge.livejournal.com/12135.html

A long story short... even he use house rules to this system. ^_^

[deek]

Maybe I am explaining it wrong, but here is the discussion: http://www.enworld.org/showthread.php?t=231832

"For me, the big breakthrough was realizing that the permutation distribution for the possible rolls fell on Pascal's Triangle.
(See: http://en.wikipedia.org/wiki/Pascal's_triangle )
Once you spot this, you see that you can get your players back to even odds (if their individual skill checks are even odds) by giving them the remaining half of the triangle, rather than just a 3rd off to one wing."

There's a quote. So, I'm not saying bam, its solved. I was trying to say the distribution was modeled after Pascal's Triangle and knowing that, you can bring the distribution back to a fair level (assuming 50/50 is fair given his other assumptions) rather easily.

[Moon-Hawk]

Maybe I am explaining it wrong, but here is the discussion: http://www.enworld.org/showthread.php?t=231832

"For me, the big breakthrough was realizing that the permutation distribution for the possible rolls fell on Pascal's Triangle.
(See: http://en.wikipedia.org/wiki/Pascal's_triangle )
Once you spot this, you see that you can get your players back to even odds (if their individual skill checks are even odds) by giving them the remaining half of the triangle, rather than just a 3rd off to one wing."

There's a quote. So, I'm not saying bam, its solved. I was trying to say the distribution was modeled after Pascal's Triangle and knowing that, you can bring the distribution back to a fair level (assuming 50/50 is fair given his other assumptions) rather easily.

Actually, I think he's explaining it wrong.
The distribution isn't modeled after Pascal's Triangle. Pascal's Triangle can be a handy way of coming up with the coefficients of a binoimal distribution, I guess, if you don't know how to calculate n-choose-k, iterative addition is certainly simpler (although much slower) than factorials, but, but, it's not like anyone sat down and said, "Hey, i think we should make our binomial distribution coefficients use Pascal's Triangle, that'll be great." It's just how it is.
It's as if, um, as if you're trying to show how to integrate the sin function and then spend an hour ranting about how you had to realize that tangent = sin/cos. Yeah, it's true, but what does that have to do with anything? Maybe that's a bad example.
He seems to be making a big deal about Pascal's triangle, but the fact that rows of Pascal's Triangle correspond to the coefficients of a binomial distribution....that's not news.
So either he doesn't know what he's trying to say, you don't know what he's trying to say, or I don't know what you're trying to say. Or something like that. Know what I'm sayin'? (IMG:style_emoticons/default/wink.gif)

[FrankTrollman]

Yeah, watching baudot flail around over there is pretty sad. The fact that permutations can be generated with Pascal's Triangle is no basis for an epiphany. Basically he's so far lost into minutiae that he isn't abstracting this problem in any meaningful fashion. It's just a multiple coin flip problem where the individual coins are weighted by the chance of an individual skill test succeeding or failing. The fact that you need more successes to get the good result than you need failures to get the bad result isn't the source of the low number of positive outcomes. It's that you need more successes and you don't actually get more successes because characters actually have bonuses of +7 to +10 and they set the DCs so high that barely gets each roll into coin flip territory. So yeah, if you wanted to fix the math up to coin flip land, you could set the required successes and failures equal. Or you could raise skill bonuses or lower DCs such that the prospective number of successes required was the actual expected number generated. Or some combination.

Or hell, you could just reduce it to a single roll or even a coin flip.

Hey Frank. Did you read Keith Baker's answer to Skill Challenge problems?

Ha ha ha ha!

Holy shit. That is some funny stuff. He gives out up to +10 in extra bonuses (literally half the die that people aren't supposed to expect), and he gives people extra bonus free successes from time to time. And he insists that he doesn't consider the system to be broken. He isn't playing anything vaguely recognizable as the packaged system, how would he even know?

-Frank

[Fuchs]

It does offer a lot of advice though that should make running a game easier. Most of that many of us already do, but a newer GM might not think of that.

[deek]

I am probably saying it wrong then. The distrubtion DOES match up to that 1 5 10 10 5 1 row of Pacal's Triangle. Using that as the example, it can bee seen that all that is needed is to get the successes higher...which I think everyone who sees a problem with the current probabilities is saying, just in different ways.

I mean, we all have pointed out the probability for success is too low. So, this is just another way to compare a model to the system and use it as a basis to fix it.

I do agree with Fuchs. I mean, any DM that has run these challenges enough, found that everyone is failing at an alarming rate and then decides he wants to fix that, will do so by either lowering DCs, giving more bonus to rolls, giving automatic successes, allowing players to erase failures or adjusting the failure count.

I think by that, yes, the system is "broken". I don't think that means doing the above makes the system not even vaguely resemble the current system or is grounds to state its wretched and start crucifying people. But I think this is a fundamental philosophy difference and maybe why I am more lenient that some other posters. I don't want a system that doesn't allow the DM to alter the variance, give a bonus here or there. I want that built in so he can award some tremendous roleplaying. And maybe that should go without saying and the system should have been built to at least allow a 50/50 change out of the gate and at all levels.

I still go back to saying that the ultimate failure of a skill challenge is not an ultimate failure in the game. It should be designed as a setback. If you are using it to fail the party at each challenge, than yeah, at 1st level, the 11% success rate is way too low!

[raphabonelli]

@Frank: I didn't read the PHB entirely... reading some Shadowrun Missions right new... but aren't those bônus (teamwork bônus and so on) that he said on the rules? Or he house ruled everything?

[Nightwalker450]

I haven't messed with the calculations on skill challenges myself yet. But here's an interesting way to fix it.

If they beat the base DC (before +5 for skill check is added) it counts as neither a failure or a success. So it gives them that "stalling" window. Instead of straight up pass or fail.

[FrankTrollman]

I haven't messed with the calculations on skill challenges myself yet. But here's an interesting way to fix it.

If they beat the base DC (before +5 for skill check is added) it counts as neither a failure or a success. So it gives them that "stalling" window. Instead of straight up pass or fail.

Yeah, because that would make you succeed 46.1% of the time at Complexity 1 and 47.3% at complexity 3 with a +9 bonus. Also it would make every single +/-1 to your roll even larger because only 15 spots on the die actually matter.

I've got an idea: how about if you haven't looked at the math, you don't post your fixes for it? I know this may seem novel, maybe even rude, but if you seriously have no idea how the math works, your suggestions on how to fix it have at least an equal chance of making things worse as they do of making things better. It's like offering people medical advice or car mechanics when you have no basis in the field at all.

First figure out what the problem is, then come up with something to address that problem, then check your answer mathematically, and then post it. Writing down random mathematical gibberish and then selling it for real money is what got us into this situation in the first place.

-Frank

[Aaron]

I'm not convinced that improves the system. I mean, it takes away fail states, so it improves the overall chance of success, but at the cost of drawing the process out even longer. I think that'd only be worth the cost if you've seriously got your drama on at the table.

[Fuchs]

Well, while I didn't study maths, I am perfectly capable to calculate the odds of succeeding with a single die roll vs. a set DC.

So, pardon me while I forget all the stuff about sequences and triangles, and just do skill challenges one test a time, one test after another, and eyeballing the general situation and the resulting DC and modifiers for each test when it is time to roll it, not before.

That will work out exactly like I want it to, enable me to run an encounter very smoothly, and change it on the spot, avoid the "ok, just let the high mod guy roll all 11 checks" mentality, and generally make a social encounter run more like combat is run, where you don't set out each attack in advance, but decide on actions round after round.

It's not as if we particularly need skill challenges to be presented as a set number of rolls for an encounter, that looks too much like railroading anyway.

Problem solved, without any higher math to boot.

Oh, that's how I have been running skill use already, what a surprise.

[deek]

Well, while I didn't study maths, I am perfectly capable to calculate the odds of succeeding with a single die roll vs. a set DC.

So, pardon me while I forget all the stuff about sequences and triangles, and just do skill challenges one test a time, one test after another, and eyeballing the general situation and the resulting DC and modifiers for each test when it is time to roll it, not before.

That will work out exactly like I want it to, enable me to run an encounter very smoothly, and change it on the spot, avoid the "ok, just let the high mod guy roll all 11 checks" mentality, and generally make a social encounter run more like combat is run, where you don't set out each attack in advance, but decide on actions round after round.

It's not as if we particularly need skill challenges to be presented as a set number of rolls for an encounter, that looks too much like railroading anyway.

Problem solved, without any higher math to boot.

Oh, that's how I have been running skill use already, what a surprise.

Yeah, you can certainly break it back down to a series of individual checks and feel out the encounter round to round. I guess I just fell in love with the initial concept.

The DM picks 3 or 4 primary skills to start. He also chooses one or two secondary skills with a higher DC than the primary. Then plant a couple of those skills that either open up a new secondary skill, provide a bonus, or automatically succeed/fail. All ahead of time. That part of it I really, really like, because the DM can consistently design encounters using that method and the players have a few choices and a reason to try an outlandish use of a skill (again, assuming some level of predictability at your table). And obviously, they are modular, so you can reuse (wholesale or with a tweak) many times for the same scenario.

As a DM, I could create, say 20 or 30 of these ahead of time, and then use them whenever the situation calls for it. Now, you can replicate that with a series of individual checks, but from a planning standpoint, I think having this "web" of skills for the encounter ends up being more fun.

[Cheops]

@Deek:

and what you just said is the crux of why the Skill Challenge rule is so different and new. I can think of very few systems that actually present a rules method and guidelines for how to run a non-combat encounter. The only one I can think of that comes close is Exalted but that is still Social COMBAT and mechanically works the same as combat but with a different charm set.

nWoD did some interesting things with Extended Tests but never really fleshed it out to try and include the whole party or make it much more than one dice pool vs. another. I'm honestly very surprised that SR didn't rip it off (must have been after they'd already made the system).

Using the D&D Skill Challenge system, here's a SR4 chase sequence:

Complexity 12/6, Initiative is Driver and then by roll.

Pilot: Driver makes an Opposed Pilot Roll (Opposing pilot +1 per extra vehicle).
Navigation: Anyone makes a Intuition + Navigation roll. Success adds +Hits dice to the next character if they Pilot or Attack.
Attack: A successful attack (Physical, Drone, Matrix, or Magical) removes one pursuit from the chase. Failure is nothing but a glitch or critical glitch could mean collateral damage or penalty to next character to act.
Command: Someone with drones subscribed can use a drone to distract the pursuit. Roll Pilot + Con (whatever resists Con) opposed test. Each Hit removes one extra pursuit from the chase (min 0) for one action. Hits can instead be used to make pursuits stay out longer at 1 hit = 1 pursuit/round.
Spoof: Take a Spoof test against one pursuit. If you successfully spoof a command you may give the Pilot +1 to his next Pilot check. Multiple people can take Spoof test prior to the Pilot test to a maximum = Extra vehicle bonus for pursuit.

If the PCs win the challenge then they evade pursuit without any penalties. If they fail then they arrive at their destination but must fight a combat against their foes and gain 1 Public Awareness.

that's off the top of my head so you can probably refine it and make it work mechanically a bit better but as you see it gives everyone in the team a role to play in the chase. makes it much easier to play. and adds excitment.

D&D has thought outside the box on this one and I don't think they should be penalized because they = no good math.

[Wolfx]

@Deek:

and what you just said is the crux of why the Skill Challenge rule is so different and new. I can think of very few systems that actually present a rules method and guidelines for how to run a non-combat encounter. The only one I can think of that comes close is Exalted but that is still Social COMBAT and mechanically works the same as combat but with a different charm set.

nWoD did some interesting things with Extended Tests but never really fleshed it out to try and include the whole party or make it much more than one dice pool vs. another. I'm honestly very surprised that SR didn't rip it off (must have been after they'd already made the system).

Using the D&D Skill Challenge system, here's a SR4 chase sequence:

Complexity 12/6, Initiative is Driver and then by roll.

Pilot: Driver makes an Opposed Pilot Roll (Opposing pilot +1 per extra vehicle).
Navigation: Anyone makes a Intuition + Navigation roll. Success adds +Hits dice to the next character if they Pilot or Attack.
Attack: A successful attack (Physical, Drone, Matrix, or Magical) removes one pursuit from the chase. Failure is nothing but a glitch or critical glitch could mean collateral damage or penalty to next character to act.
Command: Someone with drones subscribed can use a drone to distract the pursuit. Roll Pilot + Con (whatever resists Con) opposed test. Each Hit removes one extra pursuit from the chase (min 0) for one action. Hits can instead be used to make pursuits stay out longer at 1 hit = 1 pursuit/round.
Spoof: Take a Spoof test against one pursuit. If you successfully spoof a command you may give the Pilot +1 to his next Pilot check. Multiple people can take Spoof test prior to the Pilot test to a maximum = Extra vehicle bonus for pursuit.

If the PCs win the challenge then they evade pursuit without any penalties. If they fail then they arrive at their destination but must fight a combat against their foes and gain 1 Public Awareness.

that's off the top of my head so you can probably refine it and make it work mechanically a bit better but as you see it gives everyone in the team a role to play in the chase. makes it much easier to play. and adds excitment.

D&D has thought outside the box on this one and I don't think they should be penalized because they = no good math.

Just want to say that this is awesome. I really like the way it is laid out. I will have to remember this for my games. Now I just need to set it to paper in a format that lets me plug in what I need when I need it.

Aric

[Cheops]

Here's another one.

Quick and Dirty Shadowrun: for those times when you need to simulate a run against a facility but don't want to have to Mission: Impossible it

Climb: (Group Test) Get your fat arses over the fence.
Infiltration: Opposed versus Patrolman, Critter, Drone, or Sensor. First success counts as success for challenge. Only characters that have made a successful Infiltration test may take further actions against the facility.
Exploit: (Group Test) Can only be done once. In addition to a success you can now use other matrix actions in the challenge.
Edit: Does not add successes to challenge. However, may remove a failure that resulted from an Infiltration or Social test. You hacker erases the incriminating evidence from the camera feeds.
Spoof: The hacker provides overwatch for the group and magically doors open for you. Adds +Hits to any targeted PC's next action.
Intercept Wireless: The rigger listens in on the opposing radio freqs. Adds +2*Hits to any targeted PC's next Inflitration or Social action.
Crash Program/OS or Matrix Attack: Counts as automatic failure no matter what as you alert security.
Astral Combat or Astral Spellcasting: The mage reduces the Ward to Mana Motes. $Counts as automatic failure for challenge but the Mage can now use all his foci etc during run.
Spellcasting: The mage casts an appropriate spell to befuddle the opposition. A failure doesn't count against the total in the challenge. However, any Spell that deals damage (S or P) to a target adds a failure regardless of success or not as the target's biomonitor alerts security.
Etiquette: Does not count as a success but a failure still adds to total. Success allows you to use other social actions in the challenge. As if he actually works for the corp the Face blends in and uses its own social rules to take it down.
Con: The Face bluffs his way past the fierce guardians. A success adds to the total. Failure does not add to total but forces PC to make another Etiquette test before using Social skills again.
Attack: Your Sam takes down a guard to give better access to the facility. A successful attack adds to the total but regardless of success/failure make a Perception test for a Patrolman/Sensor/Drone/Critter. A success on their test also adds a failure to the total as Security is alerted to your presence.

Those are examples for the actual run. There are plenty more that you can come up with and as usual if they have a good idea give them some dice or lower the threshold. However, for the Run Challenge to succeed, their LAST SUCCESS must be from one of the following tests:

Assassination: With a final strike you fell the target. A successful attack knocks the fucker on the head.
Extraction: Make some sort of attack, social roll, or whatever to coerce the target into following you. A successful roll that enables you to leave the facility with the target ends the challenge.
Data Steal: Your last success must be a Hack + Edit test. Beware! Sneaky GMs may put Data Bombs or Encryption that will ruin the file if you don't remove them first as Free Actions.
Robbery: Make an Agility + Palming test against the Patrolman/Drone/Sensor/Critter as you sneak the targeted item past.

Of course you can modify these based on the job (ie. Data Steal + Prototype Robbery).

[last_of_the_grea...]

Goddam, I hate math!

I get the point, Frank. You are clearly upset about the skill system and skill challenge rules. This is assuredly understood by all. You have made strong, clear points about how and why it is fundamentally flawed. You can stop pointing it out now.

Everybody else, please stop riling up Frank. It is obvious this is a passionate issue for him. There is no need to challenge his viewpoint or logic anymore on this issue. He has clearly and repeatedly stated and defended his position. Let him find other parts of the game to comment on, good or bad. I'm getting the (hopefully false) impression that a couple of people (not even most) are just trying to get under his skin which is sucky.

[Wounded Ronin]

The last of the great mikeys has caught t3h corr3ct. People keep busting out mathematically flawed "solutions". Just stop with the mathematically flawed "solutions" and that aspect of the conversation need not continue.

[FrankTrollman]

D&D has thought outside the box on this one and I don't think they should be penalized because they = no good math.

That is where we disagree. They have been working on this system for 3 years and are charging people $105 for their "final" product. If it doesn't work out of the box they have serious explaining to do.

The design concept is awesome. I can clearly see why they would have gone with the concept work of whoever came up with this stuff. Not just on skill challenges, but at every level of design in 4e. But it's not done. The math doesn't work. It doesn't look like a playtested and edited product, and the fact that they went to print with it anyway is something that they should be penalized for.

-Frank

[Kyoto Kid]

...105$ just for the core rules? Crikey, I can get a .5 TB external HDD for less than that.

[Cheops]

That is where we disagree. They have been working on this system for 3 years and are charging people $105 for their "final" product. If it doesn't work out of the box they have serious explaining to do.

The design concept is awesome. I can clearly see why they would have gone with the concept work of whoever came up with this stuff. Not just on skill challenges, but at every level of design in 4e. But it's not done. The math doesn't work. It doesn't look like a playtested and edited product, and the fact that they went to print with it anyway is something that they should be penalized for.

-Frank

And yet you still play SR?

[apollo124]

...105$ just for the core rules? Crikey, I can get a .5 TB external HDD for less than that.

Yup, just in case all the D+D bashing around here has you wanting to plunk down $105 American for this stuff, here's the link.
http://www.wizards.com/default.asp?x=produ...ndacc/222127600
That's $35 each for the Player's Handbook, DMG, and Monster Manual. Truly the bare minimum needed to run a game, although of course players can get away with just getting the PHB, or like my friends did, just bumming mine off of me. Not even a group discount for buying all 3 at once. (IMG:style_emoticons/default/sleepy.gif)

Note that I haven't even seen any of these books, much less read or played them. I'll leave the higher math to those who care about such things. Honestly, I wasn't ever even thinking of buying the new D+D, although I was hoping to maybe give a demo a shot if I make it to GenCon Indy.

[Bull]

I should note that D&D4 isn't any expensive tahn any other edition of AD&D or D&D3.0. The game has always had 3 "Core Books", each priced the same as what an equivelant "core rulebook" for any other game would run.

3.0 Came out initially at $20 each, whichw as a little cheaper than the average book ran then (Most were $25-30), but that only lasted through the first print run, and bumped to $30 with th second and subsequant print runs. 3.5 was likewise $30 per book.

$35 is actually a pretty good price per book for 4e, considering the size of the books, the quality, the full color and high quality paper, etc.

<shrug> Whether it's worth it or not is up to individual tastes, of course. But with everyone having some sticker shock here, I felt it necessary to note that this is not significantly more expensive than 3.5 was. At least this time it's all new material, and not just 95% recycled material that they just sold you a few years prior to that.

Bull

[Eugene]

I think that Savage Worlds (Chase System & Mass Battles) and Spycraft (Dramatic Conflicts) had decent ways of doing a "skill challenge". Instead of x successes vs. y failures, there was a track (called Lead in Spycraft) that represented how close/far you were from success. In a chase, for example, the lead represented how far apart the participants were. In a manhunt challenge in Spycraft, it was how close you were to finding someone who'd "gone to ground." The one negative about it is that it assumes that the PCs are a single entity and the opponents are a second, single entity. My players didn't like the chase systems for that reason (they all wanted to roll separately; the fast didn't want to be hindered by the slow, etc.)