Natural 1/Critical Failure rules

[Lord Il Palazzo]

The D&D 3.5 group I DM has traditionally played using the house-rule that natural 1s on skill checks and the like represent automatic failure when there's some plausible way the check could fail and that natural 20s produce success in a similar manner.

Specifics and details spoilered for length:

Spoiler

Show

Some of my players feel like it adds drama to the game to be able to fail even when using skills you're really good at and nobody besides me has expressed any strong opinion against this rule.

The natural 1 part has always bothered me as it just doesn't make sense that a person who has spent years studying magic can fail to identify even the most basic spell 1/20th of the time or that someone who is a "master of disguise" will totally fail to disguise himself 1/20th of the time.

In our most recent session when the party's swordsage tried to sneak past the party's wizard (in kind of a PvP sparring match). The swordsage rolled a nat 1 on Move Silently which came out to a modified 21 and the wizard rolled a modified 8 (I don't remember the natural roll, I would guess 6 or 7.) A discussion ensued over how much of a setback a nat 1 should represent, whether the wizard would have heard footsteps the whole way or possibly just a slight rattle of armor from the point where the swordsage started and on whether our nat 1 rule was reasonable at all. We also got to talking about what other systems or rules could be used to make a nat 1 a bad thing without necessarily being a complete failure.

I've got a discussion going by email with the players but I'm curious what other options are out there. Do you have any house rules or suggestions about how natural 1s are handled? While we play D&D 3.5, I'd be interested to hear your rules and suggestions regardless of the systems they come from.

[yougi]

I personally do it that way:

On skill checks, I use the die result as is, for the reasons you said.

On saves, I rule that a nat 20 gives you a 30, and a nat 1 gives you a -10 before your bonuses (from an optional rule in the books). Sometimes, that -10 is not enough to make you miss your save, or the 30 is not enough to make it.

On attack rolls, I say that if you have to confirm your crits, well you also have to confirm your crit misses: if you roll a 1 on an attack, you reroll the attack: if your second roll would hit, you simply miss, but if your second roll also misses, then something bad happens to you (like rerolling your attack against an ally, or getting an attack of opportunity from the guy you're meleeing with, or being flatfooted for a round).

[navar100]

Chance of failure is not necessary for fun. Characters can be just that good they can't fail. The fun is earning that expertise over time then enjoying the spoils.

Additionally - Rogue Hides. Rolls a 1. Wizard Spots. Rolls a 1. Both fail. Now what?

[Kurald Galain]

I enjoy fumble rules on skill checks, but I find 5% is too large a chance. Furthermore, I have always ruled that if you're well trained, a fumble doesn't represent you botching the job, but that something unexpected happens from outside factors. For instance, fumbling a roll to forage for food might mean that you find a bear. Then it's up to you whether you'll run or fight or do something else; it's meant to be interesting, not crippling.

[Lord Il Palazzo]

I enjoy fumble rules on skill checks, but I find 5% is too large a chance. Furthermore, I have always ruled that if you're well trained, a fumble doesn't represent you botching the job, but that something unexpected happens from outside factors. For instance, fumbling a roll to forage for food might mean that you find a bear. Then it's up to you whether you'll run or fight or do something else; it's meant to be interesting, not crippling.

This is the direction I really hope I can steer my group in. In my opinion, having something go wrong and need to be dealt with (now or later) adds challenge and is fun and failing what should be an easy task and having the whole game/encounter/plan go belly up isn't.

[TheOOB]

Personally, I loath critical failure and fumble rules. Remember in base D&D 3.5, the natural 1 only for sure fails on attack rolls and saving throws, and even then it's a normal failure(They also auto succeed on a natural 20). This is to always keep there some danger in combat, a weak foe can always hit, and a strong foe can always miss.

Now critical failures and fumbles on the other hand I find silly. Every D&D PC, even wizards and sorcerers are trained combatants. They are people who know how to fight and rely on their skills to survive the most dangerous job around: adventuring. There is no way that an adventurer has a 5% chance of dropping their sword or hitting their allies.

Now, if we where using a different dice system, say 3d6 instead of 1d20, I'd say maybe, as a "3" would only happen about 0.5% of the time, or less than 1 in every 200 attacks, but 5% is too high of a chance to make the rules something feasible to use unless you want your game to be completely silly.

[Doug Lampert]

Chance of failure is not necessary for fun. Characters can be just that good they can't fail. The fun is earning that expertise over time then enjoying the spoils.

Additionally - Rogue Hides. Rolls a 1. Wizard Spots. Rolls a 1. Both fail. Now what?

This is (part) of why opposed checks especially should NEVER use a 1 autofail or 20 autosucceed rule.

Solve that and you hit the next problem: The chance of 1 or a 20 on one of the two characters is a 19% chance, basically 1 time in 5 SOMEONE will have a roll which renders skill TOTALLY IRRELEVANT. The GOD of sneaking is caught by someone who's blind, deaf, and actually in another country 9% of the time because either he rolls a 1 (and the differently abled person doesn't) or the differently abled person rolls a 20 and the GOD doesn't.

Hey, blind and deaf is only -10 or -20 or something to the checks (blind people can tell the direction of the sun, deaf people can pick up very loud bass sounds). Distance is only -1 per 10', so it's just a set of big penalties, 1 and 20 don't care about skill or penalties or you don't need the autofail rule.

Both problems get far worse when one or both sides are groups. 4 adventurers and 4 orks all roll initiative (which is just a modified dex check and thus uses the same rules as skills which are also just modified ability checks). Congradulations, there is a whole 43% chance that you will NOT have a 1 or a 20 rolled, but don't worry, you'll get two or more 1s or 20s only about 19% of the time.

Lesson: Don't use autofail rules for opposed checks. You don't need them anyway, even if you think the variation on a d20 is too small (and I have to go "HUH?" to that in a game where a +2 is considered a big deal), then you still have a much larger variance on 2d20 being compared. Instead of + or - 10 being enough to put you off the range of the roll now it can take up to + or -19 to put you off range. Seriously, if the difference is 19 then the better guy wins. An exceptionally strong man is at a whole +9 vs. a 90 year old 90 lb weakling, +19 ought to be impressive.

[Codyage]

Attack Rolls
I rule that a 1 will cause you to fail despite bonuses, and a 20 will make the attack a success. I then make them roll to confirm a crit, and if that ends up as a 20 I let them get one more chance. If they roll three 20's in a row, I give them the kill on the creature. Unless it is immune to crits, then I just don't offer the third roll.

Saves
A 1 will always fail, and a 20 will always succeed. I give them the 20 in case they ever go against something that requires an astronomical save, they can still make it. The 1 failing is the trade off. But in most cases, a 1 or a 2 won't have enough with bonuses to match the DC anyways. But my players have agreed to that rule, because it gives them a slim chance of hope.

Skill Checks
The method I use is if they have the ranks for the skill maxed, when they roll the 20. It is a success. If they don't have ranks maxed, then it is a + 20 instead. If the person rolls a 1, it will always fail unless they have max ranks in the skill they rolled a 1 on. If they do, have max ranks it is just a +1 instead. If they can't get the DC with the +1 then they fail like everyone else. If the +1 does match the DC some how by favorable conditions whatever they were trying to do barely works. Such as a small gust of wind being enough to push you to the other side of the chasm as you barely grab the side. A natural 1 Jump check but with max ranks in jump.

If the situation occurs with two people rolling a 20 with max ranks in a skill, they need to just re-roll instead. Only if the they have the same amount of maxed ranks in a skill. If one of the characters has more ranks then the other person in their skill, they win. Example is the rogue with +11 to hide rolls a 20, but a wizard only has +10. They have maxed ranks, but the rogue still has one more then the wizard.

[Lord Il Palazzo]

One suggestion I made was coming up with a rule that makes players confirm nat 1s. If you nat 1 a skill check but would succeed with your modifier, you roll again. Failing the confirmation roll would mean you fail the skill check; maiking the confirmation roll means you succeed, possibly with some other consequence for the nat 1.

Has anyone ever played like this? How do you handle the confirmation? (Do you use the same modifier, same DC and all, like confirming a critical or do something else?)

To be clear, we don't have any consequence of a nat 1 on an attack roll besides missing with that attack; you don't accidentally throw your sword or hit an ally or anything. Our only modification in combat is that we don't roll to confirm nat 20 critical hits (a simplification our first DM used that kind of stuck.)

[1337 b4k4]

Additionally - Rogue Hides. Rolls a 1. Wizard Spots. Rolls a 1. Both fail. Now what?

Wouldn't this be resolved the same way you resolve any other opposed check where both opponents fail their skill check?

[Lord Il Palazzo]

Wouldn't this be resolved the same way you resolve any other opposed check where both opponents fail their skill check?

That's the problem; on an opposed check, nobody "fails" just for rolling low. They fail for rolling less than the other guy. In theory, a roll of 1 can win on an opposed check if the other person managed to get even lower (a low roll with no ranks and/or a penalty for a low ability score and/or some circumstance penalties, for example).

By the rules, a tie on an opposed check goes to the person with the higher modifier. If that's tied, you reroll. However, if you start house ruling certain rolls as automatically succeeding or failing, things aren't quite so clear.

[Kurald Galain]

Additionally - Rogue Hides. Rolls a 1. Wizard Spots. Rolls a 1. Both fail. Now what?

If it were my game, I would have something unexpected happen that interferes with both their plans. Depends on the situation they're in, of course. My point is, if you treat fumbles like "on every attack you have a 5% chance of stabbing yourself", they're going to suck and players tend to strongly dislike this. If you treat fumbles as a slim chance of something interesting happen... well...

One of the more memorable scenes I've run was a fight on horseback; at some point a PC knight and an NPC orc lancer charge each other, and both manage to fumble their attack rolls. I ruled that they both leaned close while their horses veered away, so they both fall off and end up grappling each other on the ground.

Or, there was the elf with very high constitution and a maxed out "resist toxin" skill (not D&D, but the skill lets you avoid negative effects from poison and alcohol) always bragged about how he could outdrink any dwarf. And statistically, he could. Now after a few sessions he finally meets a dwarf and gets a chance to show this in practice... and rolls a fumble, spends a fate point on a reroll, and fumbles again. Yeah, that was funny :)

[A Tad Insane]

I play with confirms. But I also narrirate the failures of those with maxed stats in a way to say it's not their fault, e.g. the sneak rogue rols a confirmed one when sneaking past a lazy guard, a crow screeches, causing the guard to look at see the rogue.

[Khedrac]

A reasonable Nat1/Nat20 option for skill checks is Nat 1 is an extra -10 and Nat 20 an extra +10.

[Erik Vale]

Chance of failure is not necessary for fun. Characters can be just that good they can't fail. The fun is earning that expertise over time then enjoying the spoils.

Additionally - Rogue Hides. Rolls a 1. Wizard Spots. Rolls a 1. Both fail. Now what?

"I think I heard something"
[Right near the end.]

[Deophaun]

Solve that and you hit the next problem: The chance of 1 or a 20 on one of the two characters is a 19% chance, basically 1 time in 5 SOMEONE will have a roll which renders skill TOTALLY IRRELEVANT.

In truth, unless the person making the spot check is alerted to the presence of something unusual, they should be taking 10.

Characters should only roll while not threatened if they are trying to do something out of the ordinary. This solves a lot of the problems with critical fumbles/successes on skill checks. Not all, but a lot.

[Ravens_cry]

Well, we just had a threads about this and similar ideas, like a -10, +10 rule.
I like neither to be frank. They don't add anything to the game in my opinion that needs to be added. Look at an acrobatic, the worlds greatest acrobat. Do they fail at walking across a balance beam 5% of the time? Could a paraplegic beat the world high jump record if they tried, on average, 20 times?
Another thing worth noting is that ,unlike attack rolls, some skill checks require multiple successes to be a full success. For example, balancing as one makes their way across a long ledge, or moving silently through a castle. The fact that every check has a 5% chance of failing, and, in the case of opposed checks like move silently, the guards a 5% chance of hearing them, no matter the skill level, turns these activities from being something risky but potentially rewarding, to downright suicidal.
This take away from the variance of the game as it makes players more reluctant to use their skills. it also takes away from the DM toolset as situations that call for these skills become much more potentially and, worse, arbitrarily, lethal.
In short, though the people I play with use the -10 +10 for skill checks, I hate it and I believe I have elucidated why.

[Killer Angel]

The D&D 3.5 group I DM has traditionally played using the house-rule that natural 1s on skill checks and the like represent automatic failure

(snip)

I've got a discussion going by email with the players but I'm curious what other options are out there.

I suppose these climbers should fall once every minute... (no matter what's the effective difficulty of the cliff, mind you).

[Jay R]

Chance of failure is not necessary for fun. Characters can be just that good they can't fail. The fun is earning that expertise over time then enjoying the spoils.

Makes sense. I could see losing the chance to fail one a roll of 1, only if the characters started off at first level. Otherwise, by your logic, they aren't trying to earn it; they want to be given a perfect ability to start.

Additionally - Rogue Hides. Rolls a 1. Wizard Spots. Rolls a 1. Both fail. Now what?

Easy. The rogue failed to conceal himself; his body is not completely behind the rock. The wizard failed to notice the thief's head sticking up over the rock.

(It makes more sense when you realize that the rogue's roll can affect more than one seeker. Assume there is also a fighter. The rogue fails to Hide; the wizard fails to Spot; the fighter successfully Spots. So the fighter sees the rogue and the wizard doesn't. In your scenario, the thief is protected from his failed attempt to Hide by the fact that there was no competent seeker.)

[Krazzman]

Nat1/Nat20 for Saves and Attacks are Failures/Successes. Thus is mainly to balance the critical hit system dnd uses. 20 automatic hit(save) == 1 automatic miss(failed save). Critical Hits can actually be missed. If your threat range for crits is 15+, but you would need a 19 to hit you would still miss with your 15.
If you want to introduce a fumble system you should work it like crit's. If you actually confirm this fumble it should give you a minor disadvantage. Like roll on a table and in this table the worst thing is a "Your weapon's hilt loosens it's grippyness /you have a minor ache in your bones (for natural weapons) and as such can't give all your power into it. -2 to hit, -2 dmg and -2 against disarm checks." or "Your motions are such off that you provoke an attack of opportunity." or for ranged, your bowstring snaps/you misfire etc and now you need a fullround action to replace it.

You now, minor stuff like losing an action, having a situational negative modifier. Not something like you throw away your blade, no save.

If you want something similar for skillchecks let on a 1 confirm mishaps. For example while hiding a rogue sneezes. Happens quite often in media and while he still has a good hiding place he busted it by sneezing, IF the guards can pin-point where the sneezing came from. For Nat 20 a streak of luck could be while hiding successfully you find something or while hiding in a barrel the lid won't move from outside the barrel.

[Deophaun]

I suppose these climbers should fall once every minute... (no matter what's the effective difficulty of the cliff, mind you).

Why aren't they taking 10?

Nat1/Nat20 for Saves and Attacks are Failures/Successes. Thus is mainly to balance the critical hit system dnd uses.

Critical success is not balanced by critical failure. Critical success is balanced by both teams being subjected to its effects. Eliminate critical failure, and you will have no effect on the impact of critical success to the game. Make one side, and one side only, exempt from critical failure, and you will affect game balance.

[Lord Il Palazzo]

Why aren't they taking 10?

Critical success is not balanced by critical failure. Critical success is balanced by both teams being subjected to its effects. Eliminate critical failure, and you will have no effect on the impact of critical success to the game. Make one side, and one side only, exempt from critical failure, and you will affect game balance.

I assmume he mean "balances" as in "mirrors" or "is equivalent to"; he didn't seem to be saying that critical success make the critical system more fair.

[EarFall]

The D&D 3.5 group I DM has traditionally played using the house-rule that natural 1s on skill checks and the like represent automatic failure when there's some plausible way the check could fail and that natural 20s produce success in a similar manner.

Specifics and details spoilered for length:

Spoiler

Show

Some of my players feel like it adds drama to the game to be able to fail even when using skills you're really good at and nobody besides me has expressed any strong opinion against this rule.

The natural 1 part has always bothered me as it just doesn't make sense that a person who has spent years studying magic can fail to identify even the most basic spell 1/20th of the time or that someone who is a "master of disguise" will totally fail to disguise himself 1/20th of the time.

In our most recent session when the party's swordsage tried to sneak past the party's wizard (in kind of a PvP sparring match). The swordsage rolled a nat 1 on Move Silently which came out to a modified 21 and the wizard rolled a modified 8 (I don't remember the natural roll, I would guess 6 or 7.) A discussion ensued over how much of a setback a nat 1 should represent, whether the wizard would have heard footsteps the whole way or possibly just a slight rattle of armor from the point where the swordsage started and on whether our nat 1 rule was reasonable at all. We also got to talking about what other systems or rules could be used to make a nat 1 a bad thing without necessarily being a complete failure.

I've got a discussion going by email with the players but I'm curious what other options are out there. Do you have any house rules or suggestions about how natural 1s are handled? While we play D&D 3.5, I'd be interested to hear your rules and suggestions regardless of the systems they come from.

Auto failures on skills aren't much fun - but it shouldn't be too trivial either. I find a nice compromise is to treat a 1 as a -5. YOu might still succeed - and it helps with stuff like hide vs spot and bluff vs sense motive where one side "failing" doesn't seem right.

[Deophaun]

I assmume he mean "balances" as in "mirrors" or "is equivalent to"; he didn't seem to be saying that critical success make the critical system more fair.

"This is mainly to mirror..."
"This is mainly to make equivalent to..."

I'll let Krazzman speak for himself, but the above constructions are odd, and not what I'd expect "this is mainly to balance" to be referring to. If someone told me they were doing something "mainly to mirror x," I'd tell them to be more original. If something existed "mainly to be equivalent to x," I'd conclude it was pointless. ("This is mainly to offset" or "this mainly counteracts," however...)

[Bogardan_Mage]

In truth, unless the person making the spot check is alerted to the presence of something unusual, they should be taking 10.

Characters should only roll while not threatened if they are trying to do something out of the ordinary. This solves a lot of the problems with critical fumbles/successes on skill checks. Not all, but a lot.

Not using critical fumbles and successes on skill checks solves all of the problems with critical fumbles and successes on skill checks. I see a lot of suggested mechanics in this thread but no explanation of why the rule needs to exist in the first place.

[Kurald Galain]

Not using critical fumbles and successes on skill checks solves all of the problems with critical fumbles and successes on skill checks. I see a lot of suggested mechanics in this thread but no explanation of why the rule needs to exist in the first place.

Because they lead to interesting situations (assuming a competent DM, of course).

[Krazzman]

"This is mainly to mirror..."
"This is mainly to make equivalent to..."

I'll let Krazzman speak for himself, but the above constructions are odd, and not what I'd expect "this is mainly to balance" to be referring to. If someone told me they were doing something "mainly to mirror x," I'd tell them to be more original. If something existed "mainly to be equivalent to x," I'd conclude it was pointless. ("This is mainly to offset" or "this mainly counteracts," however...)

I actually meant something like it resembles a mirror that both sides of the game are subject to an at least 5% chance of failure (to hit/to save) or at least a 5% to succeed (to hit/to save). Both having a chance to critically hit CAN be anticlimatic or awesome but is so in a "reasonable" way.

And yes I tend to use odd constructions quite often. Dunno why, just happens from time to time, when I'm not thinking my posts through enough.

[PersonMan]

Why aren't they taking 10?

Because they can't. Unless they're on some micro-cliff where there is 0 danger, they can't. You can't take 10 in a stressful situation, and in my experience 'if you fall, you die' is stressful. Quite so, even.

[Deophaun]

Because they can't. Unless they're on some micro-cliff where there is 0 danger, they can't. You can't take 10 in a stressful situation, and in my experience 'if you fall, you die' is stressful. Quite so, even.

I'm sorry, I shouldn't assume 3.5 or 4e rules in the generic Roleplaying section. Naturally, if you're playing in a system where stress precludes taking 10 such as... well... you tell me... that would be an issue. Of course, such a system probably wouldn't have a take 10 rule to begin with.

[PersonMan]

I'm sorry, I shouldn't assume 3.5 or 4e rules in the generic Roleplaying section. Naturally, if you're playing in a system where stress precludes taking 10 such as... well... you tell me... that would be an issue. Of course, such a system probably wouldn't have a take 10 rule to begin with.

Taking 10

When your character is not being threatened or distracted, you may choose to take 10. Instead of rolling 1d20 for the skill check, calculate your result as if you had rolled a 10. For many routine tasks, taking 10 makes them automatically successful. Distractions or threats (such as combat) make it impossible for a character to take 10.

Emphasis mine. I did mean 3.5. Now, perhaps I wasn't clear, but when I said "stressful" I meant "failing means you fall off a cliff and die". Which is, in my view, a threat.