So, the following issue has come up: the party has one rogue, Acavel, who's the best hider and silent mover (understandably). Barik, a ranger, could pour his skill points into hiding and moving silently, and together with a cloak of elvenkind or armor of shadows, could be as good at hiding, and a few ranks lower at moving silently. On the other hand, the one rogue could use the armor instead, and be _really_ good at hiding. So do we want one character to sneak around, recon, etc., or two?
Tactically, there's a lot to be said for having a team of two instead of one. If they find something, one can report back while the other keeps an eye on the situation; if they run into trouble, one can hold off the hordes while the other gets help, or at least the two can support each other in an organized retreat. Or, they could go off into different directions, surrounding an opponent or covering more area faster.
Here's the problem: suppose the two of them are headed down the corridor, and something might hear them. How's that get resolved? The way I interpret the rules (at least at first), they each roll their Move Silently checks, and the something rolls its Listen check, and if it gets a higher score than one of them, they're rumbled.
Probability-wise, this is quite a bit worse than one of them going alone, even if they have the same skill. In that case, it's the same probability as if the one going alone rolled his Move Silently, and Evil DM (tm), after seeing that roll was too good, said "naw, don't like that...roll again, see if you fail this time." To put it even more mathematically, if either scout had an M in N chance of being undetected (i.e., 1 in 2=50%), the two together have an M squared in N squared chance (25%, in the previous parenthetical).
That's a pretty vicious penalty. So PCs adopting the two-scout strategy should reconsider -- although a two-watchman strategy when you're camped out makes a whole lot of sense.
Unless you're a diehard fan of D&D rules trivia, stop reading now. The last paragraph was the most useful comment you're going to get out of this.
Group skill checks are one of the times when the situation for a group of NPCs need not be symmetrical to that of a group of PCs, although it's quite unclear. The DMG's section on Skill Checks (3.5e, p.30) indicates that for "influencing one person, creature, or group" or to "perceive one sound or sight", the "DM decides if NPCs are acting as individuals or as a group." How or if this applies in opposed checks is not at all clear; I go into detail about some possible schemes at the bottom of this post, below the line of asterisks. Don't read that unless even the other diehard D&D trivia fans think you're a freak. In the meantime, let's assume that this grouping thing doesn't apply; in opposed checks, every PC rolls and every NPC rolls, and each Spotter/Listener detects all the Hiders/Movers who scored less than they did. This makes it awfully hard for anyone to sneak past a group, and can generate boatloads of die rolls, but at least it's "fair".
How hard is "awfully hard"? Consider a "commando team" of rangers, sneaking up on a camp of orcs all gathered by the campfire singing "Kum-bay-ya". Let's say there's 5 rangers, around 9th level. Counting their +4 from dexterity and +2 for circumstances and the max +12 skill ranks, they're +18 on both Hide and Move Silently checks. These are just a score of grunt orcs, and let's give them a -2 for circumstances, so they're +0 on both the Spot and Listen checks. Unless a ranger rolls a 1 and an orc rolls a 20, the rangers won't get discovered.
The odds of one orc rolling a 20 and one ranger rolling a 1 are 1/400 (0.25%). But the chance of at least one of the 20 orcs rolling a 20 and one of the rangers rolling a 1 are 14.47%. And that's just the chance of the rangers being spotted. The chances of them being spotted _or_ heard are 26.9%, just over 1 in 4.
Here's a more typical example. Suppose a single rogue has a +10 to their Move Silently roll, and he or she rolls a 10 while trying to sneak pass a room. If it's got one guard in it, with a +6 Listen check, the rogue has a 70% chance the guard won't roll the 15 or more he needs. Change the number of guards to 4, and that chance drops to 24%. That seems low to me -- although if the "opposed check" were a tug-of-war, maybe that'd be about right. One way to try to fix this would be to give the guards all a -2 circumstance penalty (because they're distracting to each other) -- then the chance of the rogue's success improves some, up to 40%.
Or we could pretend three of the guards are really doing "aid another" attempts for the first one's listen check -- I don't even know how that would change the probabilities, but it could make things worse, not better. Or maybe after the second guard, the other two don't really help any, and can't make checks. And if that isn't complicated enough, how do we treat the inverse problem of a group trying not to be detected? Do we not have people after the first two make checks, because they don't contribute to the silence?
The simplest alternative system would be that in any opposed check, each side only makes one roll: the best Spotter/Listener rolls vs. the worst Hider/Mover, for example. But that ignores the contribution (or lack thereof) of the rest of the people entirely. Why is it always Hawkeye that catches the goblins? And does he see all the goblins, or just Klompy? What if some of the quieter goblins were coming from the other side? Do we treat them as a separate group? Oh geez, I thought this was the problem we were trying to simplify away...
********************************************************************************
[This is the consideration of "asymmetrical" opposed checks, where a group of NPCs is treated differently from a group of PCs. If you haven't been riveted by the post so far, you're really going to despise the time you'd be wasting if you read this. You've been warned.]
The examples in the DMG indicate that a PC would get one check to spot a group of NPCs at a distance [Note: it's not clear if this is an opposed check, but it seems not], but that a PC Moving Silently would face the Listen checks of every NPC in a group. But what about a group of PCs Moving Silently? Do the NPCs get one number to beat, or the worst of two? There isn't a clear example I can find in the DMG of opposed checks where the NPCs are treated as a group -- would that mean they make one roll, or that they only have to oppose one roll, or that NPCs each make their own roll in opposed checks, and are never treated as a group? The third case makes the most sense, that I talked about above, although the first one isn't completely crazy. The second one is in fact completely crazy, and I'm worried about my own mental health just as a result of thinking about it.
The first case isn't symmetric, but it's not entirely unbalanced: if a group of NPCs rolls once for the whole group, that means they'd be harder for an individual to spot that a group of PCs (since they'd only have to succeed once, while each PC in a group would have to succeed), but they'd be worse at spotting an individual (since they've only got one chance to succeed, while the PCs would have a chance each.) If a group of NPCs and a group of PCs try to sneak past one another, there's still a good chance they'll run into each other, since the PCs have a lot of chances for high Spot/Listens, but just as many chances for low Hide/Moves. The chance would be considerably reduced from the "everyone rolls against everyone" scheme, however. The PCs would be wise to have multiple people on watch, but only send one out for scouting reports, so they're not fighting the probabilities.
The other asymmetric case is that the group of NPCs only opposes one roll (that is, all the NPCs make a check against one roll for the PCs.) This seems pretty antithetical to the idea of D&D, since the NPCs are being treated more individually than the actual characters with individual free wills, but let's put that aside for a moment just to consider this silly case for academic purposes. Then the roles are reversed, and the only interesting twist is how to decide what the bonus on the PC roll is: is it the bonus of the most skilled PC, the least skilled, an average, or is it different for different skills? (A group of generic NPCs are pretty close to the same skill level, so I didn't mention this for the previous case.) No matter the bonus selection mechanism, the PCs don't gain anything by having more than one night watchman; in fact, if the PC bonus is anything but "same as most skilled PC", it actually hurts them to have more than one! On the other hand, it doesn't hurt them to send out several scouts -- if several PCs have the same skill bonus, there's no penalty for numbers. Again, opposing groups (PCs vs. NPCs) are likely to spot each other, this time because the NPCs have lots of chances to get high and low rolls, but not as likely as in the "everyone rolls" scenario.
Click here to return to Fantasy - and not the X-rated kind home page
4 comments:
I think you're right about two scouts, and venture no real opinion on the group stuff. Seems to me that the only thing that makes sense is everyone vs. everyone on both listen and spot, which makes for many, many die rolls and, as you note, makes group sneakiness wildly improbable.
And it's probably worse than you note. A natural 1 will probably result in being heard regardless of the opposing roll.
Part of what's at stake here is the same as with the fumble rules. 1/20 simply isn't that low a frequency, not when actions get repeated as often as they do in a game like this. Even 1/400, which seems like a wildly improbable thing, simply isn't. These probabilities are too high for an expert to catastrophically fail, either at skill checks or at holding otno a weapon.
Is some of this a result of the shift to d20s for everything? Seems to me I remember more percentile rolls and more 3d6 rolls in older versions. A 3 on 3d6 is a lot less likely than a 1 on a d20; the d20 has no bell curve.
Another thought. Here there's an asymmetry between skills where one *success* is what matters and those where one *failure* is what matters. While there's reason to have one specialist in sneakiness, there's no reason to have just one specialist in spotting and listening. The specialist may be asleep, distracted, or looking the other way. Having a couple people with even slightly adequate spot/listen checks multiplies the chances of success. I'm going to beef up my spot ranks a bit. (Spotting the hourglass a bit earlier in that last big fight could have spared me a death and therefore a thousand years of torment!)
Anyone interested in Feat kibbitzing, as long as we're talking about dndblog instead of Bob's new campaign? I'm finally getting quick draw-- and that finally finishes my list of feats that seem like damn-I-can-hardly-play-without-this, the minimum-necessary feats. Later ones are gravy...
Most of the detection skill check issue is from Jacob's second comment, not the first: the big issue is that for virtually all situations, a group that doesn't want to be detected can't afford even one failure, and a group doing detection only needs one success. From there the wonders of exponential decay kick in.
The 1/20 not being small enough issue is much more the "gotcha" for fumbles (and nonopposed skill checks). Certainly in the old days the percentile dice were the fallback choice for deciding a situation. Using 3d6 (to test against an ability) was after my time, and I haven't heard of a 3 being a guaranteed fail or 18 being a guaranteed success. (Of course, 1/216 is a lot less likely than 1/20.)
I think the simplification of relying on the d20 outweighs the drawbacks, as long as these instant fail/succeed cases are kept from disrupting the narrative style of the game. I want to run a game where PCs are, generally, cool -- you should be excited and happy with your character, if not yet satisfied. People that drop the tools of their trade every few minutes are not cool.
Scott's campaign, on the other hand, is one in which the PCs are thrown about in an unfriendly world -- the very ground reaches up to catch their feet, the weapons leap from their grasp. The narrative is about soldiering on in the face of hopelessness, and I'm working hard to come to terms with that and not pitch a fit when Barik drops his axe. Again. :)
Overall, skills are a tricky business, in terms of tactics and in terms of "realism". (Obviously someone can get better in a craft. After some basic training, can you really get better at not making noise? Or listening?)
Barring exceptional circumstances (which a DM might gleefully engineer), some skills are voluntary, and for some of those it's impractical for multiple PCs to be good at them. (Diplomacy, for example.) Listen and Spot are "involuntary" skills, and so having several people high in these wouldn't be a waste. Still other voluntary skills are based on tasks that often the whole party will have to do (Ride, maybe Climb) -- whether to specialize or generalize in these is a little trickier.
Moving Silently (and to a lesser extent Hide) could fall into this last category. If the party _could_ all move along like mice, that'd be awfully handy, and we'd be able to go through a dungeon without assuming that everyone already knows we're there. (The whole "send a scout ahead alone" strategy could be dispensed with.)
With the mithril armor (-2), and before spending his 12 available pts, Barik's "high" skill bonuses are: Climb and Survival (+8), Listen and Use Rope (+7), Concentration, Knowledges(dungeoneering, geography, nature), Search, Sense motive, and Spot (+5). His Move Silently is +4 and his Hide is +3. I have scrupulously been trying to avoid Scott's -5 penalty for not having a skill rank in something, so I have a skill rank in every ranger class skill except Heal (and the optional crafts and professions).
I'm inclined towards increasing Survival, Spot, and Listen up to +10 each, and putting the remaining 2 points into Move Silently. If we _all_ have crazy Spot and Listen, maybe that's overkill. The other appealing place to put points is Search.
Feats: I'm going to get Quick Draw too (which Barik needs desperately, for throwing multiple axes as well as switching to bow). The remaining ones I'm keen on are Two-Weapon Defense, Point Blank Shot, Rapid Shot, and I'll throw Far Shot in there too. (I'm planning to get 12th & 13th levels in Fighter, so I can get 3 of those 4.)
I find the skill system pretty realistic, given the adventure genre. High-level PCs are extraordinary figures in the world. I can't get much better at most of the skills with any training I'm likely to engage in; but Olympic atheletes can get Tumble and Jump and Balance and Climb, and I'll bet the CIA provides some pretty serious training in Search, Spot, Listen, Decipher, and a bunch of others. There are certainly some world-class Forgers, lock-openers, and sleight-of-hand practitioners. Shifting into genre fiction, any credible ninja is going to have Move Silently skills that are off the charts as far as normal people are concerned, which suggests that it can be trained into.
At extremely high levels, yeah, Spot and Listen risk turning into Superman's senses. But at those levels mostly what matters is opposed skill checks, and so no one ever becomes either de facto invisible or de facto all-seeing.
I've been gradually re-gearing my skill allocations toward game needs and away from character-conception stuff. I've got 28 ranks divided among Bluff, Diplomacy, Intimidate, Innuendo, Gather Information, and Sense Motive. Those, plus Disguise, were the core of the initial character conception-- more Nemesis than The Shadow, or more 1970s Batman than Urban Legend Batman. All of that, of course, did very little good in the initial stage of the campaign, and none at all after the Massacre (after which all the conspirators were exposed).
Since then I've been building up the other skills-- and, like Bob, been trying to avoid that -5 penalty. But there are still occasional role-playing choices that have to be made-- various knowledge and craft skills (I've now got Knowledge: Hell). As an overall result, I'm really not a specialist in anything; I'm not maxed out in any skill, and break 10 (before characteristic bonus) in only Tumble, Spot, and Move Silently. Tumble and Spot are largely the result of the two new levels; I dumped lots of skill points into them for sheer game reasons. Then, besides Knowledge: Hell, just made a few marginal improvements in the arms-race skills like hide and move silently.
The Spot and Listen improvements certainly seem right. I wonder whether, in game terms and for the next, primarily urban/ sewer/ fortress stage of the adventure, Search might not be more important than Survival-- but then, I think Survival is one of those skills I tend to underrate.
Post a Comment