D&D 5th Edition: 8th Revision and Counting

[Djinn_in_Tonic]

You mean the unwinnable one? Or mob of certain death? Those demonstrably do exist and do emerge in actual game. The only thing that prevents them from occuring more often is how some GMs actively try to avoid them.

I meant the scenario of every attack made against you missing, and thus the damage being equivalent. Given that a roll of 20 hits, it would be EXTREMELY unlikely to actually happen.

[Kurald Galain]

"Misses doing ½X damage" works out mathematically to the same end result as "50% of attacks hit, doing X damage".

Only if you assume that 50% of attacks hit. That is not at all the case. If a low-level kobold attacks an armored experienced fighter, then a 10% chance to hit is more reasonable.

[1337 b4k4]

That's an Oberoni fallacy. Just a because a good DM can overrule the bad math doesn't magically fix the bad math. I expect WOTC to be better at math than the average DM, not worse.

No it isn't because by the rules as written you aren't supposed to be using the math in the first place. It's as if we had rules for holding your breath under water, and a paragraph that says "if you're not under water, you don't need to roll to hold your breath". You're arguing that it's an Oberoni fallacy for me to argue that the breath holding rules aren't broken for holding your breath while in the vacuum of space because according to the rules, you shouldn't roll (or at the very least shouldn't be using that system). The skill system is designed to be used to resolve checks where there is a "significant" and interesting chance of failure. What is an interesting and significant chance of failure? By the rules, at least a 10% chance or so.

The problem is, you're looking as the skill system as "this is how skills are resolved, so this is the minimum and maximum chances of failure for all things." On the. Other hand, I'm looking at the rules as saying that "if you have a scenario where you need to randomly resolve a chance of failure within this given range, you should roll the dice like this"

I mean lets be perfectly honest here, almost every scenario that people bring up for how ridiculous the skill system is are almost always ones where no one would roll anyway.

[noparlpf]

You're saying "there's no need to roll when it's pro vs. newb because there has to be at least a 10% chance of failure for it to matter". There is more than a 10% chance that the newb beats the pro by the current rules for skill modifiers. Therefore something is definitely wrong.

[Morty]

Perhaps the "something happens even on a miss" rule would be the default, but certain abilities, equipment and other factors could make misses have no effects. Heavy armor, for instance, could make a miss deal no damage at all, because the blow only scratches ineffectually against the metal. Who knows, maybe it'd even make armor relevant. I'm speaking theoretically, of course - the system would have to be balanced around this kind of thing from the ground up.

I do agree, of course, that making hordes of low level enemies a threat is a very good thing. The nigh-invulnerability of characters above certain level against low-level opponents in 3.x wreaks havoc on tension and any GM's attempts at introducing it.

[1337 b4k4]

You're saying "there's no need to roll when it's pro vs. newb because there has to be at least a 10% chance of failure for it to matter". There is more than a 10% chance that the newb beats the pro by the current rules for skill modifiers. Therefore something is definitely wrong.

I'm telling you that the rules aren't designed to model skill checks between complete newbies and masters where there isn't at least a 10% chance of failure. Put another way the rules don't say that a newbie has a 10% chance of beating a pro, but rather the rules give you a method for resolving situations where the newbie has at least a 10% chance of success. The DM (and the players) are supposed to decide BEFORE ANY DICE ARE EVEN ROLLED, whether or not there is at least a 10% chance of failure, and therefore the rolling mechanic is applicable to the situation at hand. You are not supposed to use the mechanic to determine the chance of success.

[Frozen_Feet]

Do you think a battle between 4 PCs and 65 mooks, or a similar ratio, is unbeatable? Reality says otherwise.

First, I said a PC. If there are four of them, there should be 80+ enemies. Second, I said some threat, not unbeatable. You could adjust the numbers until it models reality reasonably well.

However, the point is not to make PCs winning against numerically superior foe impossible - it is to make losing against numerically superior foe possible. Bad tactics should lead to defeat, but that doesn't happen if PCs are arbitrarily immune to weaker NPCs regardless of numbers and tactics.

[deuterio12]

Perhaps the "something happens even on a miss" rule would be the default, but certain abilities, equipment and other factors could make misses have no effects. Heavy armor, for instance, could make a miss deal no damage at all, because the blow only scratches ineffectually against the metal. Who knows, maybe it'd even make armor relevant. I'm speaking theoretically, of course - the system would have to be balanced around this kind of thing from the ground up.

I quite like your line of thinking there.

I do agree, of course, that making hordes of low level enemies a threat is a very good thing. The nigh-invulnerability of characters above certain level against low-level opponents in 3.x wreaks havoc on tension and any GM's attempts at introducing it.

Well, problem here is that a non-insignificant number of players likes nigh-invulnerability against much lower level enemies. And it should at least be possible for organized high-level PCs to butcher their way trough literal thousans of lesser opponents.

On the other hand, there should be some half-decent mass combat mechanics where an organized squad/unit of 1st level kobolds (perhaps a mid-high level kobold chieftain leading them) is a lot more threatening than an unorganized rabble with the same numbers.

[Morty]

Well, problem here is that a non-insignificant number of players likes nigh-invulnerability against much lower level enemies. And it should at least be possible for organized high-level PCs to butcher their way trough literal thousans of lesser opponents.

It should definietly be possible for high-level PCs to fight their way through innumerable opponents. It's well within D&D's narrative space. But the problem with it in 3.x is that it's not possible but trivial, and you don't even have to be very high level. Facing down an army should be a mighty deed for the greatest of heroes, not something mid-level characters can do without much effort.

On the other hand, there should be some half-decent mass combat mechanics where an organized squad/unit of 1st level kobolds (perhaps a mid-high level kobold chieftain leading them) is a lot more threatening than an unorganized rabble with the same numbers.

I agree, certainly. Coordination, morale and tactics should make a world of difference in how threatening a large group of combatants is, either against another group or powerful individuals. I'm interested in the idea of representing large groups as single entities for the purpose of the rules - it has the additional benefit in making them much easier to keep track of.

[deuterio12]

I wouldn't say it's that easy for mid level PCs in 3.5 right now to become immune to damage. Natural 20s still hit you automatically, and damage reduction is rare. Even spells like stoneskin and protection from arrows can be "overloaded" with enough damage. Ethereality would be the "easiest" way to become army-immune, but that's more of a problem with ethereality itself being kinda borked, and something like an Allip will take down not only mooks but high level brute monsters as well.

On the other hand, enemies 8 levels lower than you don't grant you any experience.

Not very sure about 4e, never played it much, but I remember mooks being a bit too easy to butcher compared to suposedly equivalent encounters.

And yeah, actually running thousands of single entities isn't very kind on the side of the DM, so some abstraction should be the way to go.

On the earlier D&D editions, literal armies of low level enemies were completely official possible ecounters, usually backed up by siege engines and some high-level guys. You can still see a leftover from that in 3.X's monster organizations:

tribe (40-400 plus 1 3rd-level sergeant per 20 adults, 1 or 2 lieutenants of 4th or 5th level, 1 leader of 6th-8th level, and 5-8 dire weasels)

[Stubbazubba]

Just say that if you miss by 10 or more, then nothing happens, but 9 or less and you get your back-up effect. I'd have to do the math to confirm, but my gut says that'll significantly reduce the number of times PCs have wasted a turn, but will still allow them to go against 70 Kobolds at level 10 and not feel terribly worried.

With that, I agree that a mob needs to become a single unit. Perhaps for each MM entry, you have rules for "mobs of _____," where each +X individuals adds +Y to attack bonus, damage, HP, etc., for the mob, but not AC. Every time the PCs deal Z damage, those values are reduced accordingly. Yes, that is somewhat a lot of bookkeeping to do, but it's at least a starting model to be refined upon.

For instance, a mob of Orcs is defined with a base of 40 HP, attack bonus of +5, damage of 1d8+4, and AC 14. This represents up to, say, a dozen Orcs, and every dozen Orcs above that gets +10 HP, +1 attack bonus, and +1 damage.

In that case, 130 Orcs would have 140 HP, +15 attack, and 1d8+14 damage. At full strength, I'd peg that at CR 9, give or take, but the fact that it still has a low AC and these stats theoretically decrease along with its HP, I'd knock it down a few CR steps. So while you take on 8 Orcs at level 1, you can take a whole camp of 130 by level 6-8. That seems relatively appropriate, given 3.5's power scale.

That still may not be the best way to model mobs, though. It's possible that individual units are individuals until they achieve a certain size, in which case they voltron into a mob unit, but doubling that number doesn't just increase the stats in the mob unit but creates a second mob unit, and eventually enough mob units voltron into an army unit.

So up to 11 Orcs are just Orcs, but a dozen becomes an Orc squad, and then you get 2 squads, 3 squads, 4, until 5 squads (60 Orcs) becomes an Orc warband, and 5 of those (300 Orcs) becomes an Orc camp, and 5 of those (1500 Orcs) becomes an Orc army. Each step has its own MM entry that can be better tuned to level appropriate-ness. If an Orc is CR 1 (or 1/2), a squad could be CR 4, a warband CR 8, a camp CR 12, and an army CR 16.

This way you can give larger groups access to siege engines or spells (since every huge army is bound to have some spellcasters) that wouldn't make sense to include in the write-ups for the individual monsters. This does, however, multiply the number of entries in said MM. But really, there's enough dross in the MMs that I think we can make room for some of these.

[Draz74]

Out of combat, you need to determine if a skill roll is needed. This is not a fallacy as this is the exact role of the DM. A DM runs the game, not just by applying the rules, but by applying human judgement.

This is bad game design IMO, and in any case it is certainly lazy game design. Even if all DMs could somehow agree on what contests are worthy of rolling for, and which ones have an obvious outcome that shouldn't require rolling, then there are by necessity going to be some situations that straddle the line between the two. Ones where the DM thinks the underdog should have a tiny chance of success, but greater than zero.

This is not how Next currently works. The variability of the d20 compared to the ability scores means that "a tiny chance of success" is a rarity -- and only occurs in situations that most of Next's supporters (at least on this Forum) say shouldn't require a die roll anyway. I seldom see any acknowledgement of borderline cases (that should require a die roll, but with minuscule chance of success), but I can only conclude that they would be situations where the Next mechanics would actually give the underdog a rather ridiculously large chance of success.

TL;DR saying that DMs can and should rule about whether die rolls are necessary at all really doesn't fix the problem of the d20's wide spread. This statement is entirely independent of my feeling that a well-designed game should lead to reasonable results even if a hypothetical (inexperienced) DM did ask for rolls on everything.

On the other hand ...

Then there's modeling. If you simplify all challenges to one skill roll, variability makes a mockery of the system. If you break a challenge down to two, three, or four rolls, then the absurdity goes away. Best two out of three is used in real life for just such a reason.

This. If WotC is determined to stick to the d20 mechanic (which I like), the six ability scores & checks system (which I don't like), and the bounded accuracy philosophy (which I'm on the fence about), then the only way they're going to get a game with reasonable outcomes is to demand that a lot of contests -- the sort that should generally be won by the more skilled contestant, with little luck involved -- are decided by not one, but multiple opposed checks.

This will, in turn, slow the game down significantly when it comes to something like arm wrestling. (It may bother players less when it represents a 45-minute chess game.) But it's the only way, if they want to cling to these other design fundamentals they've set up.

Like you say, Clawhound, this is precisely why the variability of the d20 has not historically been a big problem in the combat rules. The length of combat, the way it involves numerous d20 rolls, is what has mitigated the randomness enough to make the game work. WotC needs to realize this.

[Stubbazubba]

Like you say, Clawhound, this is precisely why the variability of the d20 has not historically been a big problem in the combat rules. The length of combat, the way it involves numerous d20 rolls, is what has mitigated the randomness enough to make the game work. WotC needs to realize this.

Indeed, and this is why the non-combat game has gone through many iterations, while the tried-and-true combat engine has remained fundamentally untouched. So the moral of the story is to make a universal d20 RNG work, there needs to be a robust non-combat system that works. That is obviously incompatible with Next's plan to go with single-roll, ability bonus-based checks.

The other side of this coin is that whatever activity in the game requires rolling dice, that is what the players' attention is drawn to and what the plot slows down to focus on. This is a problem when every arm wrestle requires three rolls to resolve. The plot slowing down to focus on everything simply creates a slow-paced crawl through plots mundane and exciting. We want to get to the good stuff, to the meaningful conflicts and cool battles where we get to feel like a powerful mage or warrior. In that case, the ability bonus roll is light enough to provide a quick resolution and move on. It doesn't do so very well, and there are a great stack of corner cases that defy expectations, but most players won't notice, much less get hung up on, these, and they blissfully move on to something more worth caring about.

So really, we need different setups for non-combat actions and in-combat actions. I frankly think the old idea of skill tiers was, in this regard, a step in the right direction. It was woefully incomplete, but it delivered the exact kind of results that people here saying "The DM shouldn't have you roll for that" want while standardizing it in an actual mechanic, which is what the rest of us want. In such a system, the unimportant challenges take little to no table time, they're just there to move the plot along and make the heroes feel successful, which saves time for the important challenges that we want to slow down and roll out. And with more tweaks to the way those things are rolled (Advantage for trained skills instead of/in addition to a bonus, for one), plus a revamped skill challenge as an optional rule, (i.e. a module if those ever actually happen), I think the game would work great going back and forth between non-combat non-challenge, non-combat challenge, and combat.

[Morty]

I wouldn't say it's that easy for mid level PCs in 3.5 right now to become immune to damage. Natural 20s still hit you automatically, and damage reduction is rare. Even spells like stoneskin and protection from arrows can be "overloaded" with enough damage. Ethereality would be the "easiest" way to become army-immune, but that's more of a problem with ethereality itself being kinda borked, and something like an Allip will take down not only mooks but high level brute monsters as well.

Total invulnerability isn't necessary, though - you just have to be resistant enough. After all, you'll be killing several of the low level enemies per round at the very least.

Just say that if you miss by 10 or more, then nothing happens, but 9 or less and you get your back-up effect. I'd have to do the math to confirm, but my gut says that'll significantly reduce the number of times PCs have wasted a turn, but will still allow them to go against 70 Kobolds at level 10 and not feel terribly worried.

I think D&D could use some mechanics for degrees of success in general.

With that, I agree that a mob needs to become a single unit. Perhaps for each MM entry, you have rules for "mobs of _____," where each +X individuals adds +Y to attack bonus, damage, HP, etc., for the mob, but not AC. Every time the PCs deal Z damage, those values are reduced accordingly. Yes, that is somewhat a lot of bookkeeping to do, but it's at least a starting model to be refined upon.

For instance, a mob of Orcs is defined with a base of 40 HP, attack bonus of +5, damage of 1d8+4, and AC 14. This represents up to, say, a dozen Orcs, and every dozen Orcs above that gets +10 HP, +1 attack bonus, and +1 damage.

In that case, 130 Orcs would have 140 HP, +15 attack, and 1d8+14 damage. At full strength, I'd peg that at CR 9, give or take, but the fact that it still has a low AC and these stats theoretically decrease along with its HP, I'd knock it down a few CR steps. So while you take on 8 Orcs at level 1, you can take a whole camp of 130 by level 6-8. That seems relatively appropriate, given 3.5's power scale.

That still may not be the best way to model mobs, though. It's possible that individual units are individuals until they achieve a certain size, in which case they voltron into a mob unit, but doubling that number doesn't just increase the stats in the mob unit but creates a second mob unit, and eventually enough mob units voltron into an army unit.

So up to 11 Orcs are just Orcs, but a dozen becomes an Orc squad, and then you get 2 squads, 3 squads, 4, until 5 squads (60 Orcs) becomes an Orc warband, and 5 of those (300 Orcs) becomes an Orc camp, and 5 of those (1500 Orcs) becomes an Orc army. Each step has its own MM entry that can be better tuned to level appropriate-ness. If an Orc is CR 1 (or 1/2), a squad could be CR 4, a warband CR 8, a camp CR 12, and an army CR 16.

This way you can give larger groups access to siege engines or spells (since every huge army is bound to have some spellcasters) that wouldn't make sense to include in the write-ups for the individual monsters. This does, however, multiply the number of entries in said MM. But really, there's enough dross in the MMs that I think we can make room for some of these.

There was a pretty robust homebrew project for 'mobs' on these boards. Unfortunately, the author is now banned and with the search engine down, looking up the thread is problematic. But it had a template that could be used to make large groups of enemies into single entities and allowed to account for their level of organization, leadership et cetera. Which I think is extremely important - twelve convenienty evil humanoids could be a reasonably surmountable challenge as an unorganized mob and a deadly challenge as a disciplined, organized group despite their overall level of combat training (which is to say, stat blocks) remaining the same.

Of course, this kind of mechanics would be, by necessity, heavily abstracted. It would be up to the GM when and how to merge large groups into 'mobs'. But I think it's better than the current situation in 3.x, where big groups of enemies are both pain in the butt to run and largely unthreatening.

[Saph]

There was a pretty robust homebrew project for 'mobs' on these boards. Unfortunately, the author is now banned and with the search engine down, looking up the thread is problematic. But it had a template that could be used to make large groups of enemies into single entities and allowed to account for their level of organization, leadership et cetera.

There are some D&D modules that do the same. When I played through the World's Largest Dungeon a while ago, there was a 'horde' template for dozens or hundreds of lesser enemies. Got quite nasty when we ran into a horde of mohrgs, but luckily we were quite high level by then.

[navar100]

As an aside, this would be a nice time to mention another aspect of game design we could move away from: the blank miss.

Spoiler

Show

Probably as a holdover from D&D's wargaming roots, failing most tests means nothing has happened. The problem here is that "nothing happening" means the Players have spent their time without moving closer to an interesting result -- nothing is any deader, objectives are no closer to be completed, and disaster is no more imminent. This is one reason why combat can feel "boring" if it goes on for more than a few rounds -- the weight of those misses begins bearing on the mind of the combatants.

4e provided a glimpse of a better way -- the "effect on a miss" that most Daily Powers gained to compensate for the expenditure of a renewable resource. But why stop there? Why not have every "miss" more one step closer to some sort of conclusion? The easiest way is for all attacks to deal half damage on a Miss -- combatants get closer to death every action and even the "tank" cares about what happens in a given round. This will cut down on "dead rounds" where bad rolls mean little damage is actually done and time is effectively wasted and (importantly for some) it will give a "realistic" way for Mundanes to remain "fun" without abilities that can end a combat in a few rounds.

Tip of the hat to Deeper in the Game for raising the issue in my mind. If I were designing 5e I would at least consider removing The Blank Miss from the game.

For limited use abilities ok, but I think it's too iconic for regular combat. However, a Natural 1 no longer automissing on an attack roll can work if 3E's iterative attacks are used. A warrior's first attack will eventually can't miss because he's just that good. Spellcasters have been doing that since level 1. The iterative attacks provide the chance of missing necessary for a game to be fun. As an extra benefit, the concept of critical fumbles is disintegrated.

[Clawhound]

Single die rolls for skills work in many situations. The reason that they work is that you get one roll, but the character rolling will use the skill many time over his career. That works just fine for those frequently used but no downside skills. Such skills include knowledge, gathering food, and gathering information. If fact, you can fail those rolls until you succeed.

I'm happy with single die rolls in combat. The challenges are varied enough the underdog can win sometimes, which will often be the player characters. In this way, I'm happy with the underdogability of the current skill system. It can hurt or help players in equal measure.

Here's what I see as problems. Sometimes the die rolls gets absurd because skills get used. A few skills trigger total failure if only one check is missed, such as sneak or climb. This often makes the rogue's sneak and climb look like a failure. The more often the DM makes you roll, the more reliably that you eventually miss. So there are huge pressures behind these skills to simply max them out to no-failure.

Some skills shouldn't have die rolls. A blacksmith should not need to roll a die to make a sword. Die rolls are inappropriate for craft skills.

[Alejandro]

Do you think a battle between 4 PCs and 65 mooks, or a similar ratio, is unbeatable? Reality says otherwise.

While I can agree in principle, this is a bad example to use. It's a technological, cultural, and military ambush with a large amount of luck.

[Synovia]

I think you got your facts mixed up; until the very high levels (15+), or with the aid of spells, the bonuses to skills was pretty much the same as the dice spread.

No, I don't. You very regularly have low level characters with significant plusses in 3.5.

For example, a Level 1 can have +4 (Attribute) +4 (level) +2 (synergy) +2 (item) = +12 bonus. It can go higher if said character preps certain spells.

IE, a level 1 character is 60% more likely than the standard level 1. A typical DC for that level should be about 15. (or, 1-2 is a failure, 3-20 is success)

[deuterio12]

Of course, this kind of mechanics would be, by necessity, heavily abstracted. It would be up to the GM when and how to merge large groups into 'mobs'. But I think it's better than the current situation in 3.x, where big groups of enemies are both pain in the butt to run and largely unthreatening.

I did some google search about D&D mobs and found this on another forum's homebrew section. What do you think?

While I can agree in principle, this is a bad example to use. It's a technological, cultural, and military ambush with a large amount of luck.

Then check out the Battle of Watling Street. Boudica had aready faced (and defeated) roman forces and had plenty of opportunity to loot equipment from the previous battles. She also had around 23 to 1 numerical advantage, but still her army got routed, as she got overconfident and engaged the romans on the choke point they wanted.

Anyway like the poster above me pointed out, facing enemies with diferent technologies/cultural/militaries is a pretty big selling point of D&D. Only very rarely will you face enemies of your exact same mentality and tactics.

[Morty]

There are some D&D modules that do the same. When I played through the World's Largest Dungeon a while ago, there was a 'horde' template for dozens or hundreds of lesser enemies. Got quite nasty when we ran into a horde of mohrgs, but luckily we were quite high level by then.

I wasn't aware of that, but it's good to know.

I did some google search about D&D mobs and found this on another forum's homebrew section. What do you think?

That is the very same system I talked about earlier, just posted on a different forum. I quite liked reading it, but I don't know how it works in practice and there are some things it doesn't cover. Still, it's a good example of how such rules might look like.
Mind you, mob rules wouldn't cover everything. They work well if the group is packed together, but what if they spread out on a large area and try to flank a lone opponent? What if, say, the PCs run down a street and assassins fire upon them from the rooftops? Such situations would also benefit from some systems to streamline them, but they can't really work the same way as a mob would. It would have to be addressed.

[Tehnar]

No, I don't. You very regularly have low level characters with significant plusses in 3.5.

For example, a Level 1 can have +4 (Attribute) +4 (level) +2 (synergy) +2 (item) = +12 bonus. It can go higher if said character preps certain spells.

IE, a level 1 character is 60% more likely than the standard level 1. A typical DC for that level should be about 15. (or, 1-2 is a failure, 3-20 is success)

When you have a dice spread of 19 points, then +12 is not significant when compared to it. Neither would +25 or even +30 be significantly higher then the dice spread to dismiss them out of hand.

[Rhynn]

When you have a dice spread of 19 points, then +12 is not significant when compared to it. Neither would +25 or even +30 be significantly higher then the dice spread to dismiss them out of hand.

I don't think I understand what you're talking about here at all. Can someone explain this? Why is +25 not significant on a 1-20 spread? Why is +12 not? (How is that a spread of 19 points, not 20, anyway? Is it "it can go up or down 19 points from the bottom/top" ?)

[Cavelcade]

I'm telling you that the rules aren't designed to model skill checks between complete newbies and masters where there isn't at least a 10% chance of failure. Put another way the rules don't say that a newbie has a 10% chance of beating a pro, but rather the rules give you a method for resolving situations where the newbie has at least a 10% chance of success. The DM (and the players) are supposed to decide BEFORE ANY DICE ARE EVEN ROLLED, whether or not there is at least a 10% chance of failure, and therefore the rolling mechanic is applicable to the situation at hand. You are not supposed to use the mechanic to determine the chance of success.

So if there's a 10% chance of the newb beating the pro, the newb should have a 20% chance of beating the pro?

[Doug Lampert]

I don't think I understand what you're talking about here at all. Can someone explain this? Why is +25 not significant on a 1-20 spread? Why is +12 not? (How is that a spread of 19 points, not 20, anyway? Is it "it can go up or down 19 points from the bottom/top" ?)

Spread is the difference from top to bottom, I consider that fairly obvious as what spread means, so that's why 19 points.

Edited: 12 strikes me as significant, but the person at 12 points lower still outright WINS a contest 7% of the time, so him winning or losing are both still significant chances (7% strikes me as a significant chance).

So if there's a 10% chance of the newb beating the pro, the newb should have a 20% chance of beating the pro?

Nah, we've established that something that comes out to over 20% if rolled isn't worth rolling because it's CLEARLY and INDESPUTABLY less than a 10% chance.

So something that IS worth rolling because it's a 10.1% chance must have a much higher chance than that.

Maybe you only roll when it's exactly 50-50, I'm not sure although "flip a coin" as the only action resolution mechanic would simplify things. But I'm not sure how I'm supposed to know what has a 10% chance with made up numbers with no clear real world references OTHER than looking at the rules. But I'm somehow supposed to decide PRIOR to looking at the rules and in contradiction to the rules.

Doesn't bother me, if I'm supposed to make **** up for basic action resolution without using the rules then I'll get a better rule set. I can make up circumstance modifiers, I can deal with "we're using a d20 so 'really unlikely but possible' means a 5% chance", but not "decide what the odds are without using anything you know about the odds".

But them I don't know if this 10% was from the playtest packet or the "rule" that you decide prior to looking at the rules, but then I didn't read the most recent one all that carefully.

[noparlpf]

Spread is the difference from top to bottom, I consider that fairly obvious as what spread means, so that's why 19 points.

Edited: 12 strikes me as significant, but the person at 12 points lower still outright WINS a contest 7% of the time, so him winning or losing are both still significant chances (7% strikes me as a significant chance).

Yeah, considering every increment on a d20 is 5%, and 5% is usually the cutoff point for statistical analyses of things.

[Rhynn]

Spread is the difference from top to bottom, I consider that fairly obvious as what spread means, so that's why 19 points.

Edited: 12 strikes me as significant, but the person at 12 points lower still outright WINS a contest 7% of the time, so him winning or losing are both still significant chances (7% strikes me as a significant chance).

I don't think I quite understand the meaning of "significant" here. If your chance to win is 7%, you are 54% likely to lose 20 tests in a row. I think that's a pretty significant chance of losing a pretty huge number of tests, in a RPG context.

Heck, if 5% is the limit for significance, you're 5.5% likely to lose 40 tests in a row at those odds...

Is the argument here really about the specific lowest chance of success that can be had in opposed tests (apparently not accounting for level), and what's a good chance?

What do each of you consider acceptable numbers for that?

[1337 b4k4]

So if there's a 10% chance of the newb beating the pro, the newb should have a 20% chance of beating the pro?

I really don't know how I can make this any clearer. Let me try one more time though.

The rules given allow you to adjudicate a random outcome within some probability range (which from what other people have posted appears to be at a minimum 10% chance of the lowest ranked individual outdoing the highest ranked individual). The rules then also state that you should only use the mechanic to resolve situations that are statistically significant. Since the minimum thing the mechanic can resolve is a newbie with a 10% chance to beat a master, then by definition, that is the line for statistical significance within the mechanic. Therefore, if you have some situation you want to resolve where you feel that the newbie having a 10% chance of beating the master is no appropriate, then you have a situation where:

A) You have already determined at least generally what you believe the statistical result should be
B) You have determined that such a result is outside the range which is statistically significant to the model

Therefore, by the rules as written, you should not be using the mechanic in question to resolve the situation. Either you shouldn't roll at all, or you should use some other mechanic. Every single mechanic has an upper and lower limit which the mechanic will be able to resolve adequately. This does not make the mechanic broken, nor is it an Oberoni fallacy to suggest you should use another mechanic for this situation.

[Synovia]

When you have a dice spread of 19 points, then +12 is not significant when compared to it. Neither would +25 or even +30 be significantly higher then the dice spread to dismiss them out of hand.

The spread of the dice is only meaningful when in the context of the DC being rolled against, and the modifiers present.

If players have +14, and the DC is 15, spread is irrelevant.

In the example, the spread is larger than the modifier, but the modifier is the determining factor in the success.

[Oracle_Hunter]

The rules given allow you to adjudicate a random outcome within some probability range (which from what other people have posted appears to be at a minimum 10% chance of the lowest ranked individual outdoing the highest ranked individual). The rules then also state that you should only use the mechanic to resolve situations that are statistically significant. Since the minimum thing the mechanic can resolve is a newbie with a 10% chance to beat a master, then by definition, that is the line for statistical significance within the mechanic. Therefore, if you have some situation you want to resolve where you feel that the newbie having a 10% chance of beating the master is no appropriate, then you have a situation where:

A) You have already determined at least generally what you believe the statistical result should be
B) You have determined that such a result is outside the range which is statistically significant to the model

Therefore, by the rules as written, you should not be using the mechanic in question to resolve the situation. Either you shouldn't roll at all, or you should use some other mechanic. Every single mechanic has an upper and lower limit which the mechanic will be able to resolve adequately. This does not make the mechanic broken, nor is it an Oberoni fallacy to suggest you should use another mechanic for this situation.

This is vastly amusing

So, we're arguing that the average GM should be able to calculate statistical significance on the fly (answer B) or should abuse the terms of "statistical result" to involve no actual math (answer A)? And this seems like a sensible way to run a railroad?

It is simpler to say that the RAW (as you've interpreted it) is that GMs should only roll when they want the rolling side to have a chance of success. There is literally no math involved on the side of the GM.

So yes, if you're OK with the rules giving them GM only two options (i.e. to declare a 0% chance of success [no roll allowed] or a 10%-or-so chance of success [a roll allowed]) to resolve "edge cases" then you must be happy with the RAW as you've stated it. Some people are happy with that sort of choice, others would like some greater granularity.