Critical hit/fumble mechanic

[MKV]

How do you handle critical hits and fumbles in your game? I have seen several mechanics and am curious what others exist. Here are a few I have seen
1. Natural 20 auto-hits
2. Natural 20 auto-hits and deals max damage
3. Two natural 20's in a row allows you to roll on a table of positive attacks
4. Natural 20 and a failed save vs. paralysis by the enemy allows you to roll on a table of positive affects
critical fumbles are much the same replacing "natural 20" with "natural 1", "positive" with "negative" and so forth

[Jay R]

I'm currently running a 2E game, using the crit and fumble tables in The Dragon #39. If you roll a 20, then you roll again to confirm the crit. If you roll what you needed to hit the second time, then we go to the critical hit tables.

The same procedure in reverse is used for the fumble table.

The table is strong enough that crits and fumbles really matter, and any combat is dangerous, but it has not killed or maimed any PCs.

Gets kind of funny when you drop your weapon or lose your footing, though.

[Sir Chuckles]

The table is strong enough that crits and fumbles really matter, and any combat is dangerous, but it has not killed or maimed any PCs.

Gets kind of funny when you drop your weapon or lose your footing, though.

Cut down for emphasis.

I've really liked this kind of fumbling, as it really screws over specific characters far more than others. A Blaster Sorcerer laughs at fumbles, where as a two-weapon fighting anything will spend a surprising amount of time cutting their attack routine short in order to pick up their dropped weapon. Not to mention the fact that a trained and hardened warrior, regardless of experience will always have that chance of dropping their weapon. I just can't imagine a Warblade 20 going "Oops, it slipped!"

[DrMartin]

I like or dislike random tables for criticals and fumbles depending on the tone of the game. For games with epic and world-shattering scope I generally don't like them, the experienced warrior slipping in combat more often than the rookie just because he has more attacks per round does not sit well with me :D
I'm running a 3.P game at the moment where i allow players to raise the stakes of their actions in combat, essentially a bet where a better than normal outcome is levied against a potentially catastrophical (or worse than usual) failure. As a GM i can accept the bet or not, depending on the benefit or hindrance it would provide against the probability of success, and/or depending by how awesome is what the character is trying to achieve :)
It is working quite well so far, even if at the moment only some of the players are using this mechanic on a regular basis.
This moves the "out of the ordinary" contributions to the gameplay of fumbles from a random table to an area under player's control, and in a sense that makes the rule just as good as your players are.
Another side benefit is that id adds a bit of variety to martial classes, since they can improvise more freely their actions in combat and in a sense opens up a lot of versatility.

This is on top of the normal rules for criticals and natural 1s from 3.p (multiply dmg on a critical, always miss on a nat 1).

[Red Fel]

In my games, I prefer to go with what's written in the books, without adding more.

Generally, I find that the rules explicitly lay out what a natural 20 on a hit does. Assuming a d20-style system, they generally simply deal whatever extra damage the book says they do - double, or occasionally triple, or what-have-you.

I don't use critical fumbles. I don't find that they add much, I find that they penalize some classes more than others, that they penalize higher-level characters more than lower-level ones, that they penalize PCs more than NPCs, and most importantly, they're not a core mechanic. They're an optional variant, if that. The rules generally provide only that a natural 1 is a failure. Not a magnificent failure, not a cut-off-your-foot failure, just a failure, even if you have a +500 modifier to your roll. I'm fine with that. I acknowledge that any undertaking has some risk of failure. I think it's absurd that "risk of failure" can include a trained warrior accidentally disemboweling himself.

I've used some systems where there are more elaborate rules for critical failures and successes (e.g. "exploding dice"). Where the rules provide them, I go with them. But where the rules don't explicitly state them, I don't feel compelled to add rules for critical success, or particularly for critical failure. I just don't find that they add a substantial amount.

[Jay R]

Cut down for emphasis.

I've really liked this kind of fumbling, as it really screws over specific characters far more than others. A Blaster Sorcerer laughs at fumbles, where as a two-weapon fighting anything will spend a surprising amount of time cutting their attack routine short in order to pick up their dropped weapon. Not to mention the fact that a trained and hardened warrior, regardless of experience will always have that chance of dropping their weapon. I just can't imagine a Warblade 20 going "Oops, it slipped!"

That just doesn't match what I've seen happening under this system. Suppose a high-level fighter hits his opponent on a 3 or better. Then to drop his sword, he has to roll a one, following by a one or two, followed by rolling for the possibility of dropping a sword on the table, after which he has to roll over his DEX. The chances are vanishingly small.

In the last eleven adventures, with adventurers from 1st to 4th level, I think we've had one dropped sword and one dropped arrow.

[Arbane]

The chances are vanishingly small.

Ha. Ha. Ha.

Since Seerow is right about everything (or at least about fumbles), I'll just quote him on this:

I have a few litmus tests for critical fumbles. I have yet to see any rule that passes all of them, but failing any is enough for me to say "Nope, not playing with this".


1) The "Are you a moron? Stop trying to gimp the weak classes" test. Does this fumble rule hurt a Fighter more than a Wizard? This includes can a Wizard's save based spells fumble, but also considers things like action economy and the penalties inflicted. If any of these things favor the Wizard being less effected, this rule should have been put out to pasture before you even started thinking it up.

2) The "Epic Hero" test. Does a higher level character have the same or worse chance of fumbling as a low level character? As an untrained character? Your higher level characters should be less likely to screw up. That's literally the function of level. If your fumble rules are arbitrarily punishing high level characters, this rule is not going to work.

3) The "target dummy" test. If you have a squad of soldiers (of any level) training by hitting something completely non threatening for an hour or so, how many will be injured, maimed, scarred, or dead as a result? If that number is greater than 0, there is a problem with your rules.

4) The "test of time". How long does it take for your fumbles to resolve? If it takes longer than it takes to confirm a critical hit (say by adding a step to reference a chart, draw a card, or anything else along those lines), this rule is a problem.

5) The "three stooges" test. Does the result of your fumble chart bring to mind the antics of the three stooges or any other slapstick comedy? This can also include things that trigger the target dummy test, but goes beyond that. If you have characters slipping and falling, dropping weapons, dazing themselves, etc, I do not accept this rule.



I haven't found a rule that satisfies all of these criteria, but I'm sure if someone cared enough they could design one. The real question is, why would you really want to?

[Kelb_Panthera]

I -loathe- fumble mechanics. A nat 20 is an automatic success on saves and attacks, a nat 1 is an automatic failure on the same and that's it. I even dismissed the actual written rule that a nat 1 forces a save from your equipment as unnecessary extra dice rolling.

Sometimes a blind squirrel finds an acorn and occasionally monkeys fall from trees. The above captures this sentiment adequately IMO and that's more than enough for me.

[Forrestfire]

At least in D&D, critical fumbles in their most commonly-used form are a sign to me that the D&D fundamentally misunderstands the system and are generally a dealbreaker for walking from a game for me. I don't use them, and I expect to not have them forced on me.

If someone wants them, then I would offer them the ability to critically fumble in exchange for some bonus (possibly counting as two flaws in 3.5, or giving some sort of other benefit in other editions).

The only critical fumble mechanic that I've seen and been decently ok with is "if you roll a natural 1, roll again, and if you still miss, a minor fumble happens."

[Jay R]

Ha. Ha. Ha.

Since Seerow is right about everything (or at least about fumbles), I'll just quote him on this:

OK, I'll answer his questions.

Are you a moron?

Your "expert" does not impress me by opening with an insult.

1) The "Are you a moron? Stop trying to gimp the weak classes" test. Does this fumble rule hurt a Fighter more than a Wizard?

No, the rule is much better for fighters. The tables I referred to include critical hits and fumbles, and the results of the critical hits (which wizards get hit with) is far greater than the results of the fumbles.

2) The "Epic Hero" test. Does a higher level character have the same or worse chance of fumbling as a low level character? As an untrained character? Your higher level characters should be less likely to screw up. That's literally the function of level. If your fumble rules are arbitrarily punishing high level characters, this rule is not going to work.

No, since crits and fumbles must be confirmed with a second roll, higher level fighters have more crits and fewer fumbles.

However, that also depends on the opponent's abilities, since the chance to hit does. A level 1 fighter fighting somebody he hits on a 14 or higher has exactly the same number of fumbles as a level 20 fighter against somebody he hits on a 14, which is quite reasonable. The greatest football players have about as many fumbles in games against the greatest opponents as schoolboy players have against other schoolboys.

3) The "target dummy" test. If you have a squad of soldiers (of any level) training by hitting something completely non threatening for an hour or so, how many will be injured, maimed, scarred, or dead as a result? If that number is greater than 0, there is a problem with your rules.

Zero, of course. Nobody would apply fumble or critical hit tables for this. I can't imagine a DM willing to roll for hits at all in this situation.

4) The "test of time". How long does it take for your fumbles to resolve? If it takes longer than it takes to confirm a critical hit (say by adding a step to reference a chart, draw a card, or anything else along those lines), this rule is a problem.

It can take one step longer than resolving a critical hit, because many of the fumble results have saving throws. My players don't mind this, even if Seerow, who has never played in my game, makes up the guess that it is a problem.

5) The "three stooges" test. Does the result of your fumble chart bring to mind the antics of the three stooges or any other slapstick comedy? This can also include things that trigger the target dummy test, but goes beyond that. If you have characters slipping and falling, dropping weapons, dazing themselves, etc, I do not accept this rule.

No, it does not come even close to looking like the three stooges, in which every action causes a fumble.

It looks far more like the 1973 Michael York Three Musketeers, in which excellent fighters have occasional problems in the middle of very intense combat. In the first battle Porthos loses his sword. In the duel at the inn, Porthos is knocked on the head by a large weight he accidentally dislodged. In the battle by the mill, Athos's cloak gets caught in water wheel. In the final climactic battle, D'Artagnan throws the bundle into the wrong window. Yet Bill Hobbs's fight direction in this movie usually gets very high marks for accuracy and suspense.

Anything else?

[TheCountAlucard]

No, the rule is much better for fighters. The tables I referred to include critical hits and fumbles, and the results of the critical hits (which wizards get hit with) is far greater than the results of the fumbles.

That's assuming that the Wizard's even in a position to be hit (hint: he doesn't have to be, and typically, he isn't). And since he's capable of using spells that don't require attack rolls at all, he never fumbles. The fighter, who's generally-understood around here to be the guy who closes with people and fights them with some sort of weapon, will both eat more enemy crits and suffer more fumbles, even if he gets a secondary roll to avoid it.

And the point of the training dummy isn't that a DM and player would roll these out in a real play session, it's there for verisimilitude - it's generally assumed that characters always play by the rules of the game even when the DM fast-forwards over what they're doing: if a group of mundanes travels to a mountaintop castle, the DM skipping the travel sequence doesn't mean they suddenly sprouted wings and flew there. Hence, if your fumble rules won't allow for hitting a padded straw man for an hour without someone in the group decapitating himself, you have a problem.

[oxybe]

Jay R, is this the chart you're talking about?

[GraaEminense]

I don't DM many d20-games, mostly d100-below-target-number-which-is-Skill-with-modifiers (CoC and WFRP).

The idea is modified from BRP and should be transferable:
If you roll 10% or less of the target number (say, 1-3 of target 36%), you roll again. If you succeed again, you get a critical success.
If you roll over 90+10% of the target number (94+ on 36%), roll again: If you fail again, you get a critical failure.
Out of combat, critical successes and failures get really good or bad results depending on the situation.
In combat, critical success allows you to choose from a number of bonus effects (usually more damage) while critical failure forces a roll on an Epic Fail Table, which generally involves action-economy-losses.

It works pretty well: If you are skilled or circumstances are in your favour (high target number), you are more likely to get positive effects and less likely to get bad ones.

Translated to d20, it would be something on the lines of "natural 1 -> roll again, if a miss, roll on Bad Things Table" and "natural 20-> roll again, if a hit, roll on Good Things Table". A skilled fighter would be vastly more likely to get good effects than bad ones. No reason to limit the system to combat, either.

Of course, D&D is a special case because spellcasters don't use skill rolls in combat and can't get any critical effects. To avoid upsetting the balance, Good Things need to have much more impact than Bad Things.

[Jay R]

Jay R, is this the chart you're talking about?

Yes, it is.

That's assuming that the Wizard's even in a position to be hit (hint: he doesn't have to be, and typically, he isn't).

Enemy archers aren't looking for and aiming at the caster? And you're never ambushed? Why not? Seems to me that the unrealistic situation is one in which an adventurer, and the one that should be the highest priority target, can always avoid being attacked.

And since he's capable of using spells that don't require attack rolls at all, he never fumbles.

True. And he also never crits. A system in which crits add more pluses than fumbles do minuses is a net gain for the fighter.

The fighter, who's generally-understood around here to be the guy who closes with people and fights them with some sort of weapon, will both eat more enemy crits and suffer more fumbles, even if he gets a secondary roll to avoid it.

And deliver more crits to the enemy and be helped with more enemy fumbles.

And the point of the training dummy isn't that a DM and player would roll these out in a real play session, it's there for verisimilitude - it's generally assumed that characters always play by the rules of the game even when the DM fast-forwards over what they're doing: if a group of mundanes travels to a mountaintop castle, the DM skipping the travel sequence doesn't mean they suddenly sprouted wings and flew there. Hence, if your fumble rules won't allow for hitting a padded straw man for an hour without someone in the group decapitating himself, you have a problem.

People don't fumble as much when they are not under stress. But they still occasionally fumble. A DM who uses the same fumble table for practice as for real combat is being just as simplistically inaccurate as a DM with no fumbles at all.

If there was a reason to roll an hour's practice, I would include possible fumbles, because I have seen people drop arrows on an archery range, and I've twisted an ankle in fencing footwork drills. Professional athletes are often injured in practice, and some have had career-ending injuries in practice. One of the purposes of practice is to catch the fumbles and learn how to prevent them. But I would reduce the probability tremendously, because practice just doesn't equal the stress and confusion of battle.

My first approximation would be to require two twenties for a fumble, and then one or two regular rolls to confirm. I'd go down the list of potential fumbles to see which ones could reasonably happen in practice, and probably change the probabilities. But I'd discuss it with the players first, two of whom are competent DMs.

But the crucial fact remains this. Many people on the internet complain about my fumbles chart, but nobody actually playing the game does. And in fact, I got the chart from one of my players.

[MKV]

Boy! sure seems like a lot of people have very strong opinions on this. A lot of these complaints seem to revolve around two problems firstly that the probability of a critical hit is too high and secondly that the effects are too complicated and overpowered. So what are some thoughts on acceptable solutions to these problems, mine are below.

MKV's awesome critical hit/fumble mechanic.

1. get yourself some old-school d20s numbered 0-9 twice but instead of coloring them two different colors to distinguish 1-10 from 11-20 instead roll a d6 along with the d20 with 1-3 on the d6 meaning read the d20 as written and 4-6 meaning add ten
2. now a critical hit is scored whenever both a 0(10) and a 6 are rolled together thus no need for an extra roll to determine if you scored a critical
3. critical hits auto-hit except if the player is facing a god or some other uninhabitable entity in addition the player may improvise some feat for him to preform with this strike
4. if no feat is chosen the result is simply max damage
5. critical fumbles are the same except rolling a 1 on both dice which auto-misses with no additional mechanics

[Kelb_Panthera]

However, that also depends on the opponent's abilities, since the chance to hit does. A level 1 fighter fighting somebody he hits on a 14 or higher has exactly the same number of fumbles as a level 20 fighter against somebody he hits on a 14, which is quite reasonable. The greatest football players have about as many fumbles in games against the greatest opponents as schoolboy players have against other schoolboys.

That's not true and I'll prove it because I'm in a mathematical probability and statistics mood.


Fighter 1 gets one attack and, as you gave, hits on a 14. Since a 1 is a fumble that's 5% and a 65% chance to confirm. That's a 3.25% chance to fumble on any given turn.

Fighter 20 gets four attacks. The first matches our level 1 fighter's attack at 5%X65% for a 3.25% chance to fumble that attack. Then he gets his second attack. Fumbles on a 1 and misses on anything below a 19. That's 5% by 90% for a 4.5% chance to fumble. The last two are both only hitting on 20 so that's 5% by 95% for a 4.75% chance to fumble each.

Then we combine those odds for the full attack. The odds of -not- fumbling at all are 96.75% by 95.5% by 95.25% by 95.25% which comes to 83.83%.

This means that on any turn where both the fighter 1 and the fighter 20 can use their full compliment of attacks the fighter 20 is 13.92% more likely to fumble at least once while the fighter 1 is unable to fumble more than once at all unless the enemy provides an AoO window.

[LibraryOgre]

In Hackmaster, a 20 auto-hits, but you have to beat their defense roll to actually crit. Crit severity is based on how much damage your attack did (before the crit is applied), and how much you beat their defense by. Combine that severity with a d10,000 roll to determine location, and you'll find your crit effect.

[Stellar_Magic]

In Hackmaster, a 20 auto-hits, but you have to beat their defense roll to actually crit. Crit severity is based on how much damage your attack did (before the crit is applied), and how much you beat their defense by. Combine that severity with a d10,000 roll to determine location, and you'll find your crit effect.

D10,000? Damn... the only time I've rolled more then a d100 is for one of the more ridiculous wild magic charts I found.

[LibraryOgre]

D10,000? Damn... the only time I've rolled more then a d100 is for one of the more ridiculous wild magic charts I found.

d10,000 assumes you're the same size; you might roll something different against a larger or smaller opponent.

[Jay R]

That's not true and I'll prove it because I'm in a mathematical probability and statistics mood.


Fighter 1 gets one attack and, as you gave, hits on a 14. Since a 1 is a fumble that's 5% and a 65% chance to confirm. That's a 3.25% chance to fumble on any given turn.

Fighter 20 gets four attacks. The first matches our level 1 fighter's attack at 5%X65% for a 3.25% chance to fumble that attack. Then he gets his second attack. Fumbles on a 1 and misses on anything below a 19. That's 5% by 90% for a 4.5% chance to fumble. The last two are both only hitting on 20 so that's 5% by 95% for a 4.75% chance to fumble each.

Then we combine those odds for the full attack. The odds of -not- fumbling at all are 96.75% by 95.5% by 95.25% by 95.25% which comes to 83.83%.

This means that on any turn where both the fighter 1 and the fighter 20 can use their full compliment of attacks the fighter 20 is 13.92% more likely to fumble at least once while the fighter 1 is unable to fumble more than once at all unless the enemy provides an AoO window.

Well, of course. When I say he gets no more fumbles than a lesser fighter, of course that means no more per unit attack. The best running back in the NFL gets more fumbles than I do per year - because he plays in lots more games.

If you fight four times as much, you will get four times as much of all the effects from fighting.

That's true if you fight for forty rounds and the other guy fights for ten. It's true if you fight four times as many enemies. And yes, of course it's true if you attack four times as often each round.

And you also get four times as many hits, and four times as many critical hits. And since the crits on that table have more effect on average that the fumbles, that's a net gain for the top fighters.

[Kelb_Panthera]

Well, of course. When I say he gets no more fumbles than a lesser fighter, of course that means no more per unit attack. The best running back in the NFL gets more fumbles than I do per year - because he plays in lots more games.

If you fight four times as much, you will get four times as much of all the effects from fighting.

That's true if you fight for forty rounds and the other guy fights for ten. It's true if you fight four times as many enemies. And yes, of course it's true if you attack four times as often each round.

And you also get four times as many hits, and four times as many critical hits. And since the crits on that table have more effect on average that the fumbles, that's a net gain for the top fighters.

That's just it, though. Except for his first attack the fighter 20 has a dramatically -better- chance of fumbling with each attack. HIs second attack is 38.46% -more likely- to fumble than the fighter 1 and his other two are 46.15% more likely to fumble.

The high level fighter isn't just more likely to fumble overall but he's more likely to fumble on 3/4s of his individual attacks, or thereabouts depending on how often he gets AoO's and whether either of them has combat reflexes.

[Jay R]

That's just it, though. Except for his first attack the fighter 20 has a dramatically -better- chance of fumbling with each attack. HIs second attack is 38.46% -more likely- to fumble than the fighter 1 and his other two are 46.15% more likely to fumble.

The high level fighter isn't just more likely to fumble overall but he's more likely to fumble on 3/4s of his individual attacks, or thereabouts depending on how often he gets AoO's and whether either of them has combat reflexes.

Oh - you're applying an AD&D rule to E3 or later. Yes, I agree that that doesn't usually work. When you mix editions indiscriminately, you get unintended results. I use this in AD&D 2E, in which higher level fighters don't get lower and lower chances to hit.

[Kelb_Panthera]

Oh - you're applying an AD&D rule to E3 or later. Yes, I agree that that doesn't usually work. When you mix editions indiscriminately, you get unintended results. I use this in AD&D 2E, in which higher level fighters don't get lower and lower chances to hit.

Well yeah, that'd do it. Nevermind then.

[YossarianLives]

I also despise critical fumbles. It just hurts melee characters.

[LibraryOgre]

I also despise critical fumbles. It just hurts melee characters.

Not as much as spell mishaps can hurt magic characters. When's the last time a critical failure turned you into an amphibian for a month?

[Animidest]

I've been opposed to fumbles ever since my first campaign, when a party member's eye exploded from a "fumbled" Search check.

[oxybe]

For my actual feelings on Crit/Fumbles, in a game like D&D I use the rules presented (20 hits, 1 misses. the end.) unless I know you and your gameplay style very well, the inclusion of them is a good way of tipping me towards not playing in that campaign.

I've never played with crit/fumbles that were entertaining and i've never seen them presented in a way that would enhance my game play.

In a non-D&D game it's on a case by case basis, specifically depending on the overall theme and feel of the system.

[Pex]

Not as much as spell mishaps can hurt magic characters. When's the last time a critical failure turned you into an amphibian for a month?

Given the rule exists magic users can easily avoid the spell mishap by never casting a spell that requires an attack roll. It's disappointing to lose Disintegrate, but the wizard isn't crying. The druid can still wildshape into a bird and cast with impunity forgoing wildshaping into combat forms to avoid the fumbles. His Animal Companion and Summons deal with that. The cleric does have the harder time since there's no point to meleeing if he's going to fumble and many attack spells are touch attacks. Still, buffing party members is a valid tactic while Command, Hold Person and Blindness are good for low level attack spells. He can also Summon creatures if need be, and at higher levels Greater Command is there as well as Flame Strike if he just needs something to do.

Warriors can't avoid making an attack roll. Two-weapon rogues and rangers can kiss themselves goodbye, and the monk thought he had enough problems already with his flurry of blows. Flurry of "This Blows" now.

[azoetia]

At my table, 20 is an automatic success plus it's got some awesome added on top. Currently I've got new players in my group so I'm just doing old fashioned double damage while they get acclimated, but in some games I like coming up with more situational results like instant dismemberment, decapitation, knocking out, knocking prone, flipping over, throwing off a building, etc instead of double damage, though instant kill crits don't happen against PCs. I've never only done full damage; that's just not anywhere near gonzo enough for my tastes. Critical successes on things like skill rolls or saving throws are usually a bit uncanny. Not necessarily supernatural, but amazing to behold.

Rolling a 1 is a critical fumble; these events are treasured at my table. It's always situational and I make an attempt to make it seem reasonable (so no eyeball exploding from a fumbled search roll) but it's usually at least inconvenient if not downright dire. Someone who fails a sneak roll is noisy. Someone who fumbles a sneak roll is noisy and doesn't realize it, or perhaps trips on something and causes a domino effect of disaster that brings everything around him or her down in a symphony of crashes. Fumble an attack roll? Maybe you break a bowstring. Maybe you over-swing and fall over. Maybe you hit yourself or your ally with a spell. A trained fighter isn't going to cut off her own foot, but she might screw up in a way that a trained person reasonably might when under duress, or perhaps I might make an excuse for it like a bird flying in her face at the wrong moment. I sometimes let my players decide their own fumbles when they have hilarious ideas, but if I left it totally up to them it would become The Three Stooges within an hour.

I like the chaotic feel that everything could spin out of control at any time due to a couple of crazy rolls. It's one of my favourite things about playing D&D and sometimes leads to some very Tarantino-esque moments.

[NichG]

Generally I've done it by the book, but honestly I like Numenera's approach to them and would consider trying to port that over with some adaptation for D&D's particular idiosyncracies (e.g. different character types need to roll more or less and the usual sources of problems with fumble rules).

In Numenera, when a character rolls a 1, the DM can choose to consider it a fumble. In this case the player has an option - spend a resource (which happens to be XP, but XP is more fluid in Numenera, so some kind of 'action point' would be a good proxy) and prevent the crit fail, or they gain 1XP and can award 1XP to another player and something bad/unexpected happens. That bad thing may not be 'they screw up', but could be only tangentially related, such as 'reinforcements show up' or 'the monster uses a special ability next round'. Monsters in Numenera have powers which specifically and only activate in that sort of situation, so it's expected to occur once or twice per fight and is pretty much built into the fabric of the game, and there are a number of guidelines about how to design these complications.

The point is basically, okay, something worse than failing happens to you - but you get rewarded for it. And if you let it happen to you once when it's not such a big deal (such as dropping your weapon on an attack when the fight is almost over anyhow), you and another player can prevent it later when it would be really bad for you (such as spending an action point to boost a saving throw on a save-or-die). And if you try to milk it during downtime by doing hundreds of irrelevant rolls, the DM can simply choose not to invoke a fumble.