D&D 4E Showing the Math: Proving that 4e’s Skill Challenge system is broken (math heavy)

[Stalker0]

I have written up a much more detailed analysis of the skill challenge system across all levels. You can find that here:

http://www.enworld.org/showthread.php?t=230357

Warning: It is a very long post.

 

[abeattie]

This hits the nail on the head for me. I dropped a c-note on a nice, pretty slipcase edition. I damned well better not need to be sticking multiple pages of errata in there with it. I'll take that from an indie company with lower production values and a lower price-point. But considering what they've held themselves out as...this is gettin kinda deep.

Here's my problem with this thread.

People said -- well, if you drop out the +5 it works better .. NO! We can't do that.

People said -- well, if you aid another to cover the +5 it works better... NO! That's boring!

People said -- we need to run skill challenges with easy and hard skill checks... NO! That's not the basic situation.

=-=-=

Now -- having read the section a few more times, I want to point out the following.

The default bonus for a skill in the pc's racial focus is around +11 (assuming 3:16 build, +2 to the attribute for the race, +2 to the skill for the race). Now -- that means that a moderate skill check for the one or two players with that skill is at a 60% pass rate.

In all of the examples given there is at least one easy skill which grants a +2 -- its usually perception or insight. So -- every group ought to have a perception or insight (trapspotter) adventurer. That person may not have a racial bonus linkage, dropping a high bonus down to around +9 -- at target 15, 70% passage rate, counts as a success, and nudges the skill monkey up to a 70% passage rate.

Now -- what about aid another. If aid another is "I roll to help out" that's boring, if aid another is players actively bolting on other points to the lead PCs act "My lord, it is only just that we do this!" says PC1, PC2 nods, interjecting "and its what your grandfather would have wanted you to do boss" -- man -- how boring. That +2 drives success rate up to 80% against a moderate challenge -- at level 1.

=-=-=-=

PCs should lose skill challenges -- not all of them -- but look at the emphasis in the text on limiting the harm caused by failure. In most of the examples failure results in an extra combat encounter (hey.. that would get me my xp I lost by failing the challenge too!) slows down progress, but doesn't toss the players off the trail.

I don't know if I would run it that way, or if that is a great system, but I do not agree that we can ignore the balances inherent to the system in declaring it broken. That seems, to me, inherrently unfair. At level 1 you have always needed to play the game tight to the wire to get by, why should skill challenges be any different. Except now, instead of rolling through packs of 1st level PCs, we just suffer the agony of losing face a few times before we get our big-boy pants and show the world what a paragon hero can do to a skill challenge.

---

2c

 

[Imban]

Here's my problem with this thread.

People said -- well, if you drop out the +5 it works better .. NO! We can't do that.

If you do this you get something functional but still wonky. That's a lot better than non-functional, but you get the issue of higher-complexity challenges now being both easier and more rewarding. It's a pretty good idea if you just have to fix the problem now, but some people think that the wonky result is bad for their game.

People said -- well, if you aid another to cover the +5 it works better... NO! That's boring!

Well... yes, fundamentally the idea of a skill challenge is that everyone participates, and everyone rolling pretty close to an autosuccess to boost the one guy who has a chance is boring. You might as well just not use skill challenges at all in this case. Plus, in some cases, aid another is obviously limited or not appropriate - the chase scene in Escape from Sembia, the challenge/response scene in Keep on the Shadowfell, et cetera. Likewise, Perception acts as "aid" in the Urban Chase skill challenge, which wouldn't make sense if they could just traditionally aid - that would both be easier and has no penalties for failure.

People said -- we need to run skill challenges with easy and hard skill checks... NO! That's not the basic situation.

From all published skill challenges, easy checks are always good choices when they're available, but they're rarely available consistently enough that you can complete the entire challenge on easy checks. Regardless, it boosts your chances of success to take easy options that contribute to success whenever they're available. Hard checks, especially if they don't give some outstanding bonus for success, just screw you for choosing that skill, and autofails are (obviously) even worse.

 

[Spatula]

People said -- well, if you drop out the +5 it works better .. NO! We can't do that.

No one said you can't do that. There's still some inherent weirdness but I imagine it would work well enough. It's not the RAW, however.

People said -- well, if you aid another to cover the +5 it works better... NO! That's boring!

If everyone is aiding one person who's making the checks, what is the point of the skill challenge? Why not just have that one person make the checks and reduce the DC by 2*(size of group-1)?

People said -- we need to run skill challenges with easy and hard skill checks... NO! That's not the basic situation.

Hard skill checks make winning the challenge... harder. Easy checks are there but if you look at the skill challenge templates in the DMG, you can't use easy checks to get very far into the challenge. Usually they're only good for 1 success.

KotS web enhancement: 8/4, 2 easy successes max
Negotiation: 8/4, 1 easy success max
Dead witness: 6/3, 1 easy max
Urban chase: 12/6, easy checks are only used for aid another, which kinda disputes the idea you can use aid another will-nilly, as the easy check DC is going to be at least 10 and likely higher. The easy check here also seems to penalize the party if you flub it.
Interrogation: 4/2, no easy checks
Lost: 6/3, see urban chase
Secret Lore: 8/4, no easy checks
Combat trap thing: 6/3, see urban chase

 

[Wulf Ratbane]

Everyone still questioning which bonuses apply where can save themselves a lot of trouble by going back and reading this overlooked but important post.

http://www.enworld.org/showpost.php?p=4279427&postcount=32

Down the left hand column is the skill check success rate. The actual skill bonuses don't matter, the DC doesn't matter, the success rate is all that matters. We don't care how the PC accumulates that success rate (whether by attributes, feats, powers, etc.). It doesn't matter.

Pay special attention to the .60 and .70 lines.

A .60 success rate means the PC is looking for a natural 9 or better. (It's a +11 bonus against a DC 20, for example; it can also represent a +6 bonus against DC15.)

A .70 success rate means the PC is looking for a natural 7 or better. (It's a +13 bonus against a DC20.)

A simple +2 bonus differentiates these two cases.

The .60 group succeeds (at best) 36% of the time and their success rate drops off dramatically as complexity (the number of successes required) increases.

The .70 group approaches 50% success at the simplest complexity and their success rate actually increases as complexity increases.

The .65 line represents the tipping point: Pretty much a flat 42%-43% success rate across the board.

If you can keep your game in that .65 range, success and failure will be predictable. But deviate to one side or the other of that tipping point and the results, and the effect it will have on your game, are not at all intuitive.

I can certainly accept/anticipate that WotC intended for skill checks to reliably fall into that .65 range, but the fact that the system becomes so swingy is well outside the expected performance of 4th edition. It is like a very precariously balanced spinning top, and the slightest push sends it wobbling out of control.

That kind of design is not very 4e-like when you compare it to the stability and robustness of the rest of the system.

 

[Spatula]

Something this basic and fundamental and, by your characterization, at best they've only become aware of it within the past day after months (at least) of design and playtesting?

When there's lots of different people all working on the same text, there's lots of opportunities for obsolete or unintended changes making it into the shipped product. Ideally that sort of thing is caught before shipping, but the designers and editors are human. That's why there's errata.

It's also possible it wasn't playtested. It's also possible that it was playtested, but someone made a late change without considering the ramifications. The KotS skill challenge (not the web enhancement), for example, allows for 4 successes before 4 failures, not 4/2 or 8/4 as in the DMG. An earlier iteration of the skill challenge rules? Perhaps. There's lots of different possible explinations. We don't know and have no way of knowing.

If they have thought this through, designed it, and playtested it, then they shouldn't have to confab over it. They should be able to trot out simple and numerous examples of how it's all supoposed to work.

They've already released two examples, one in KotS and one in its web enhancement. They don't agree, exactly, but they're examples.

I hope they do. After all, I just plunked down a considerble chunk of hard earned cash on the assumption that I would be getting a quality playable product--and by playable, I mean one not requiring house rules right out of the box in order to play.

Planning on an all-skill challenge game, were you? 4e isn't unplayable, this one facet doesn't work as advertised but has an easy kludge (use the non-skill DCs for your challenges) for the time being.

There was a 0% chance that the books would have no errors. Publishing game books seems to be a lot like software development, in that regard.

 

[Xereq]

<snip>

 

[Nathal]

I thought WOTC hired mathematicians for this very reason. Where did I read that? Was I dreaming it? It's odd that this problem happened.

 

[Ipissimus]

Really? You clearly understand it. Are you under the impression that anyone here doesn't? What makes you think it's not "adequately exlpained"?

It seems to me everyone grasped that just fine.

All due respect, M, I don't think I deserved that. And if everyone has grasped that what is the discussion? The system's broken because it isn't?

I was under the impression that the main thrust of the discussion was that the skill challenge system doesn't work properly because the party success rate doesn't sit on the magical 55%. And I was only trying to suggest that the way it is written works as it is intended to and fooling around with the numbers may not be in our best interests.

What makes me think it's not adequately explained? Mainly history. You can't really claim that the previous editions were the best at presenting information in a clear and concise manner.

If I was out of line, I'm sorry. If my opinion isn't welcome, I'll happily take it elsewhere.

 

[Ridley's Cohort]

I suspect there is a hidden assumption that the DM will throw in some number of Easy tests or aids, based on clever use of skills.

If the test is too hard by default, then the lifespan is extended -- the DM gives fewer helping hands.

Surely everyone failed most of the time during play testing otherwise.