Fumbles

[Karl Aegis]

If there's a friend don't fire. Friendly fire is 100% avoidable if you're making logical decisions. If you do hit a friendly target you have made a bad decision. Or an Evil(tm) decision. Even demons think twice about hitting a peer.

[Ryton]

To give a simple example: in practice i hit the target most of the time and i frequently hit a bullseye.

When playing paintball i have both done and received accidental (as in not intentional) friendly fire.

Yes, but how many fumble charts account for fighting a stationary target, and how many simply require a 1 to be consulted?

And more specifically to your example, in all the presumably hundred of paintball rounds fired at friends (or strangers, depending on the range) how many times have you dropped your gun? Had all of the balls spontaneously explode in the hopper? Because those are just as likely to be found on a fumble table as friendly fire.

I'm not saying fumbles don't happen. I'm just saying if you want a realistic fumble, a master fencer might overextend and leave a hole in his defense that a similarly master fencer might exploit. He's not going to accidentally cut his own arm off.

[PoeticallyPsyco]

Not quite fumbles, but my current table has the current house rule for skill checks: a nat 1 isn't an auto-fail, but imposes a -10 penalty to the role, and a nat 20 isn't an auto-succeed, but grants a +10 bonus to the check.

[NNescio]

Not quite fumbles, but my current table has the current house rule for skill checks: a nat 1 isn't an auto-fail, but imposes a -10 penalty to the role, and a nat 20 isn't an auto-succeed, but grants a +10 bonus to the check.

By default, Nat 20s don't auto-succeed on skill checks anyway.

[Asmotherion]

The main issue that I have with this reasoning is that there are already mechanics to account for the difficulty of hitting your intended target, such as cover from other creatures and the penalty for shooting into melee. Yes, it would be more realistic if you could hit somebody other than your intended target rather than being limited to hitting your target or nobody at all, but I find the resulting gameplay of hitting an ally every other encounter less fun.

Depending on the tone and setting i don't disagree with this either. i however still enjoy the mechanic in a horror/gritty realism game a lot.

Yes, but how many fumble charts account for fighting a stationary target, and how many simply require a 1 to be consulted?

And more specifically to your example, in all the presumably hundred of paintball rounds fired at friends (or strangers, depending on the range) how many times have you dropped your gun? Had all of the balls spontaneously explode in the hopper? Because those are just as likely to be found on a fumble table as friendly fire.

I'm not saying fumbles don't happen. I'm just saying if you want a realistic fumble, a master fencer might overextend and leave a hole in his defense that a similarly master fencer might exploit. He's not going to accidentally cut his own arm off.

You're right. The examples given are far less likely to occure. This is why they'd be represented as a smaller percentage on the d100. For more realism you could add a seccond table referanced to by the initial fumble chart; however this would be overcomplicating things.

[Divine Susuryu]

Since this is the d20 forum, the majority fumble rules are easy to show as bad, especially for skills. After all, they come up 1 in 20 times in that most of them work off rolling a 1.

For attacks, there's already a critical failure state, the automatic miss - that works because all it does is waste a single attack that may not have even hit in the first place. If on top of that you add in additional consequences, you're making attacking more frequently a curse, not a blessing. If a Fighter 20 with his 4 attacks has a higher chance of humiliating or worse hurting himself than a Commoner 1 does, it's a bad rule. Simple as that.

For skills, this ties into the rules for taking 10. In activities like skills, it should be possible to perform consistently and even perform under pressure. That's why taking 10 is a thing, to prevent someone who is knowledgable in a field from randomly not knowing basic facts about that field. If you even use the most basic of critical failure states - failing on a 1 - what you're saying is that in a dangerous situation, even a 20th level 30 STR Barbarian fails at jumping a foot upwards 1 time in 20. I'm hardly a paragon of fitness nor would I be particularly high level, but in a dangerous situation, I absolutely do not have a 1 in 20 chance to fail to jump a foot into the air.

If you're using another probability system, such as d100s, then the issue is lessened but not absent.

[FaerieGodfather]

The use of supplemental, unofficial or-- worst of all-- homebrewed critical fumble rules is an instant dealbreaker for me. Not only are they a complete violation of everything I want from my roleplaying time, but long experience discussing fumble rules has taught me that their presence invariably means that the game will be filled with all kinds of other intolerable stupidity.

The only games in which fumble rules are even theoretically acceptable are games like Paranoia and TOON and other games that are supposed to be ridiculous, slapstick farces... and I generally do not prefer those sorts of games as a rule.

You're right. The examples given are far less likely to occure. This is why they'd be represented as a smaller percentage on the d100. For more realism you could add a seccond table referanced to by the initial fumble chart; however this would be overcomplicating things.

Which of those examples do you think should even occur one time in a hundred after a character rolls a 1 on a d20? Once in every 2,000 attempted attacks?

There's simply no way to make the "realism argument" for critical fumbles even remotely valid.

in gritty realism the possibility of things going terribly wrong in battle should be there.

Other people have already pointed out how comedically unrealistic most fumble rules systems really are and how completely inappropriate they are for maintaining a gritty, serious tone, but I'm just going to share my Rolemaster story.

Now, I grew to love Rolemaster eventually, years later, even if I do still strip out the goddamned fumbles. But a couple of people in one of my old D&D groups, a long time ago, were Rolemaster people and they really went hammer and tongs into convincing me how much better Rolemaster was. (They were playing Rolemaster Fantasy Roleplaying, for reference, the 3.5 to Rolemaster Standard System's 3.0.)

I'm going to state this up front, that I had a real hard time with the system going in. I didn't have enough time to read and digest the rulebook, and Rolemaster Fantasy Roleplaying has an especially steep learning curve for people who aren't already well-versed in the ICE way of doing things.

So it took me almost two hours to finish up my character, holding up the rest of the game. I made some kind of lizardman fighter.

In our very first combat, finally, I rolled pretty poorly on my initiative and ended up going last... after a couple of orcs had disarmed me and wrestled me to the ground. This, admittedly, was fun and exciting. But pinned to the ground, I only had one real maneuver available to me... so I attempted use my reptilian nature and bite one of my enemies.

And I rolled a 03 on d%, a critical fumble. I had to roll again and subtract the result from 03.

And I rolled a 97, for a total of -94. Which means... I had to roll again and subtract the result from -94.

The guys from my D&D group are giving me knowing glances. The guy I just met chuckles. I get another pair of d10s out of my bag, and my girlfriend (at the time) blows on them.

00. -194.

The DM loans me his "lucky" dice. 96, total of -290, and just barely forcing another reroll. Everyone is now watching with rapt interest.

99 brings me to -389. On more reroll. 95. I utter the foulest curses I know and go to pick up the dice, but the DM stops me: "You only have to reroll on 96 or higher."

Total of -484. One in 3.2 million chance. We consult the fumble chart, and indeed, I have injured myself... to the tune of an E Critical and another E Critical on the Bite chart.

Got identical results on both critical rolls: Instant decapitation.

In the most realistic combat system in any roleplaying game ever, my character bit his own head off, twice, on his first attack roll.

I've never played a more realistic fantasy battle since.

[RatElemental]

The only fumble rules I've ever seen be mostly functional are the ones from shadowrun 4e, mostly because the more skilled, better equipped and more apt you are for what you're doing the less and exponentially less likely you are to glitch. You can also succeed while simultaneously glitching, and glitches are decided by the GM and supposed to be not disastrous: knocking over an object you were jumping over and falling on a nail you didn't see is an example given.

[NNescio]

=


You're right. The examples given are far less likely to occure. This is why they'd be represented as a smaller percentage on the d100. For more realism you could add a seccond table referanced to by the initial fumble chart; however this would be overcomplicating things.

For skills, this ties into the rules for taking 10. In activities like skills, it should be possible to perform consistently and even perform under pressure. That's why taking 10 is a thing, to prevent someone who is knowledgable in a field from randomly not knowing basic facts about that field. If you even use the most basic of critical failure states - failing on a 1 - what you're saying is that in a dangerous situation, even a 20th level 30 STR Barbarian fails at jumping a foot upwards 1 time in 20. I'm hardly a paragon of fitness nor would I be particularly high level, but in a dangerous situation, I absolutely do not have a 1 in 20 chance to fail to jump a foot into the air.

If you're using another probability system, such as d100s, then the issue is lessened but not absent.

Which of those examples do you think should even occur one time in a hundred after a character rolls a 1 on a d20? Once in every 2,000 attempted attacks?

Obviously the solution is to roll d1000(0)s, FATAL style. It's not too hard; just take 3 or 4 d10s!

(Don't.)

[Kesnit]

If there's a friend don't fire. Friendly fire is 100% avoidable if you're making logical decisions. If you do hit a friendly target you have made a bad decision. Or an Evil(tm) decision. Even demons think twice about hitting a peer.

So you would be OK with the ranged fighter sitting out most combats because there is almost always going to be an ally near an enemy? And the higher level the ranged combatant, the more they should sit out since they would be making more attacks, making it more likely to hit an ally?

[ekarney]

It's been said before, but I'll say it again.

It punishes martials for trying to do the one thing they're good at. I have the mentality of 'punishing' players for bad tactical decisions, the fighter swinging a sword at whatever needs to die, is a pretty sound tactical decision in my books. Arguably the fighter should get a bonus for swinging weapons, possibly in the form of some sort of good base attack bonus.

[HeraldOfExius]

If there's a friend don't fire. Friendly fire is 100% avoidable if you're making logical decisions. If you do hit a friendly target you have made a bad decision. Or an Evil(tm) decision. Even demons think twice about hitting a peer.

Since most fumble rules are applied to melee attacks as well as ranged, does this mean that everybody should just break off into 1v1 duels all the time (or 2v1 with flanking if your ally needs to be in reach to risk getting hit)?

Depending on the tone and setting i don't disagree with this either. i however still enjoy the mechanic in a horror/gritty realism game a lot.

That seems fair enough. I've just had bad experiences with people sticking fumbles into heroic fantasy, which left me rather soured towards fumbles in general.

[GloatingSwine]

the possibility of things going terribly wrong in battle should be there.

It is traditionally the function of the enemies to provide this though.

If you want "gritty realism" you shouldn't want the kind of slapstick that most fumble tables devolve into as people find ever more goonish nonsense to fill them up with.

(Also you don't need fumbles for friendly fire. The best friendly fire rule I ever saw was in the skirmish game Infinity. If you fire into a melee you take a -3 penalty on a D20. If you miss by 3 or less you hit the wrong target.)

[Biggus]

A level 20 Fighter with 4 attacks will mathematically fail 1/20 attacks, or fumble once every ~30 seconds. A level 1 Fighter will also fumble 1/20 attacks but that only happens once every 2 minutes on average. Essentially a highly trained warrior general will fumble more times/minute than a green behind the ears new recruit.

I've been thinking a bit more about this, which as far as I can see is the main mechanical problem with fumbles (most of the other objections seem to come down to "I find they make the game less fun for me" or "they're nearly always badly implemented"). I think it could be largely solved by using a slight variant of the fumble mechanic in the DMG. If instead of always making a DC10 Dex check, you make either a DC10 Str check (for most melee weapons) or a DC10 Dex check (for ranged or finesse weapons) to avoid fumbling. This would mean that martial characters who specialised in either Str or Dex would get less and less likely to fumble on their chosen attack style as they go up levels, with the chance becoming zero if the relevant ability score reaches 28.

1) In 3.5, you roll to confirm a critical hit. No one ever rolls to confirm a critical failure (at least not at any table I have played at). Instead the moment a natural 1 occurs, you immediately begin rolling on whatever chart is present to determine your impending doom. The simple act of rolling a d20 a second time-- and using a miss to "confirm" the critical fumble-- would be a mild improvement, however...

Not only do I always use some kind of confirmation roll as I already noted, in the current version I'm using the penalty for a standard failed confirmation is pretty mild: lose your next attack (if making a full attack) or half of your next move while you steady yourself, or take a -2AC penalty when next attacked if you don't have any actions left to lose this round (if you're not attacked before the start of your next turn, congratulations, you got away with it). I only get the "catastrophe table" (drop weapon, fall over, hit self, hit ally, break weapon) out if they get a natural 1 on the confirmation roll as well as the original attack roll.

[Karl Aegis]

So you would be OK with the ranged fighter sitting out most combats because there is almost always going to be an ally near an enemy? And the higher level the ranged combatant, the more they should sit out since they would be making more attacks, making it more likely to hit an ally?

There is almost always going to be an enemy without an ally near them. There's only a limited number of spaces between the player characters and monsters, so if all those spaces are filled with allies all of you have made bad decisions.

[Voidstar01]

Not only do I always use some kind of confirmation roll as I already noted, in the current version I'm using the penalty for a standard failed confirmation is pretty mild: lose your next attack (if making a full attack) or half of your next move while you steady yourself, or take a -2AC penalty when next attacked if you don't have any actions left to lose this round (if you're not attacked before the start of your next turn, congratulations, you got away with it). I only get the "catastrophe table" (drop weapon, fall over, hit self, hit ally, break weapon) out if they get a natural 1 on the confirmation roll as well as the original attack roll.

That still punishes martials way harder than it punishes casters:
Without fumble:
level 20 wizard rolls a 1: spell misses
level 20 fighter rolls a 1: attack misses

With fumble rule, no confirm:
Level 20 wizard gets an AC penalty: the miss chance they have from any number of the spells that do that don't care, their AC was uselessly low anyway, and they shouldn't be anywhere near enough to melee combat for an enemy using a single move action to endanger them.
Level 20 fighter gets an AC penalty: is now significantly more likely to die because any number of enemies can move in to kill them since they're at the front lines.

Lost movement:
Wizard: less likely to need to move, since they have ranged spells, if they absolutely need to move 30ft right away, can use any number of teleportation spells they can get.
Fighter: unless you're an archer, sucks to be you I guess.

Lost attack:
Wizard: never takes full attacks
fighter: functions at half effectiveness this turn since lost attack + the auto miss equals 2/4 attacks doing no damage

Crit fail with a confirm:
Drops weapon:
Fighter: looses 6 attacks worth of damage and opens themselves to opportunity attacks.
Wizard: lol wut weapon

hits someone else:
Fighter: someone else takes some damage
Wizard: someone else now needs to save or die, or any number of other nasty things (exhaustion, sickened, panicked etc.)

This gets slightly better with the introduction of caster implements, but it's still 2 lost spells vs. 6 lost attacks. add in the fact the martial characters are still more likely to induce a fumble, and that wizards can just use AoEs to never risk it and you have a supremlyy unbalanced system.

[Voidstar01]

There is almost always going to be an enemy without an ally near them. There's only a limited number of spaces between the player characters and monsters, so if all those spaces are filled with allies all of you have made bad decisions.

having played an archer, I can confirm that's not true. Encounters rarely outnumber you more than 7:1, which is the amount it would need to be for each square next to your allies to be full.

of course you could get in an archer battle, but in that case it's better to let your melee rush the archer, since archery provokes AoO in melee combat.

[Telonius]

Fumbles are for bad guys, and mook bad guys at that. When I DM, Random Orc #12 might roll a fumble on a 1, and have some hilarious catastrophe happen to him. That doesn't happen to PCs or main bad guys.

[Adiart]

When I DM, Random Orc #12 might roll a fumble on a 1, and have some hilarious catastrophe happen to him. That doesn't happen to PCs or main bad guys.

I find this idea quite interesting, but I fear it happening too often (as is the case in a d20 system) might make the opposition look absurdly incompetent, which is not always a good thing in more serious campaigns.

[Quertus]

Fumbles are for bad guys, and mook bad guys at that. When I DM, Random Orc #12 might roll a fumble on a 1, and have some hilarious catastrophe happen to him. That doesn't happen to PCs or main bad guys.

Fumble Resistance (feat): allows you to make <insert check here> to avoid fumbles. Does not work against provoked fumbles

Fumble Immunity (feat, requires Fumble Resistance): you no longer roll for normal fumbles, and allows <roll> (as Fumble Resistance) against provoked fumbles.

Humiliating Critical (feat, requires Improved Critical): whenever you would score a critical hit - even against foes normally immune to critical hits - you may force the foe to roll on the critical fumble table instead. If your weapon has more than a x2 crit, they roll an additional fumble for every multiplier of your weapons crit above x2.

Obnoxious Target (feat, requires Dodge): wherever the target of your Dodge bonus attacks you, they must make <roll> or suffer a fumble. This counts as provoking a fumble.

Provoke Fumble (spell, Sorcerer/Wizard 2, Chaos 1, Luck 1, Hexblade 1): target must make a Reflex save, or their next attack within 1 round/level provokes a fumble.

[Crake]

Not quite fumbles, but my current table has the current house rule for skill checks: a nat 1 isn't an auto-fail, but imposes a -10 penalty to the role, and a nat 20 isn't an auto-succeed, but grants a +10 bonus to the check.

This is actually an official variant..... For checks that auto succeed/fail on a 20/1 (which does not apply to skill checks). I had a DM that ran this for skill checks, and it was universally loathed by everyone at the table. It gave rise to the joke about olympic swimmers drowning in swimming pools, because even with +14 to swim, rolling a 1 results in you sinking under the water.

More often than not, this poorly concieved houserule is actually detrimental as opposed to helpful. Since most people will stick to attempting skills they're at least moderately trained in, it's rare that people will attempt something that would require a natural 20 to succeed, with or without the +10 modifier, thus, in most circumstances, the +10 modifier will never actually be useful, since you will succeed on something less than 20 anyway, and scoring above the DC doesn't do much. On the flipside however, it is quite possible to have a modifier so high that it will succeed even on a 1, however with this rule, you need a whopping +10 more to achieve that. Thus, more often than not, the +10 goes unused, while the -10 acts as an annoying hindrance.

Don't be this guy's DM. If you're going to use that variant, use it where it's supposed to be used: attack rolls and saves.

I find this idea quite interesting, but I fear it happening too often (as is the case in a d20 system) might make the opposition look absurdly incompetent, which is not always a good thing in more serious campaigns.

To be fair, mooks are different from minions. Orc #12 may not necessarily apply, but kobold or imp #12 could make more sense.

[Drackstin]

For my games i run both fumble and crit charts, it adds a sense of mystery to the game and lot of comedy. the charts is only rolled on combat because skills can't fumble or crit, you just did really good or bad.

as for combat, my PC love it when they crit a enemy and kill it in one shot, slicing off legs or arms, crushing skulls or just plain decapitating someone. but they know this comes at a cost, because fumbles can happen, and you can end up killing yourself or allies. but you need to look at it this way, you have a 5% chance to roll a fumble, some classes can even reroll, and if you do fumble, the first 40% (my charts are both percentile out of 100 and only the top 70% are really bad) don't really do anything, trip and fall, reflex save to not, or drop your weapon reflex save to not. but if they kill a fellow player they know they messed up, but they like that about the game, stuff happens in real life that can be very bad, now lets say your always doing dangerous stuff, that amplifies it. this makes the game exciting.

[Voidstar01]

For my games i run both fumble and crit charts, it adds a sense of mystery to the game and lot of comedy. the charts is only rolled on combat because skills can't fumble or crit, you just did really good or bad.

as for combat, my PC love it when they crit a enemy and kill it in one shot, slicing off legs or arms, crushing skulls or just plain decapitating someone. but they know this comes at a cost, because fumbles can happen, and you can end up killing yourself or allies. but you need to look at it this way, you have a 5% chance to roll a fumble, some classes can even reroll, and if you do fumble, the first 40% (my charts are both percentile out of 100 and only the top 70% are really bad) don't really do anything, trip and fall, reflex save to not, or drop your weapon reflex save to not. but if they kill a fellow player they know they messed up, but they like that about the game, stuff happens in real life that can be very bad, now lets say your always doing dangerous stuff, that amplifies it. this makes the game exciting.

Please tell me you at least have players roll to hit against their ally's AC when they're going to hit them? Nothing sounds more suck-tacular than dying because your friends great axe somehow phased through your greatshield, +5 fullplate, and +5 defending shortspear to decapitate you

[Jay R]

The crucial insight is that special critical hits should give more than critical fumbles take away. When that doesn't happen, then all the bad things people are describing can happen.

1) In 3.5, you roll to confirm a critical hit. No one ever rolls to confirm a critical failure (at least not at any table I have played at).

Every time I've used special crits and fumbles, at many tables from 1976 to the present, the crits and fumbles have been confirmed the same way

2) Critical fumbles are mathematically awful for the players. On the one possibility of the presence of a table that offers permanent injuries or other penalties for a critical fumble, it is statistically definite that by the end of a campaign a combat character will be so stacked up with penalties it will be rendered all but unplayable. Meanwhile, the vast majority of injuries on NPC's are irrelevant because at the end of combat they will likely be dead.

Most of the critical fumbles from the chart in The Dragon #39 are saving throws to avoid a minor annoyance -- "lose grip on weapon; roll dexterity or less on d20 or no attack next round". That means that often, nothing happens - just an additional moment of tension. If not, then the minor annoyance occurs. Many of the rest are a small amount of damage. None of them are permanent damage.

I agree with you that critical hits should give more than critical fumbles take away. Any rule that fails to do that is a bad rule. But that doesn't mean fumble rules are bad. It means that THAT fumble rule is bad.

3) Even without the presence of a permanent injury system, it still places a massive statistical burden on the players. In order for the players to finish the campaign, they have to win every combat. In order for the enemy NPC's to stop the players, they have to win once. Critical fumbles dramatically stack these odds in favor of the enemies the more often combat that occurs.

I agree with you that special critical hits should give more than critical fumbles take away. That doesn't say that fumbles are bad. It just tells you how to design fumbles that aren't wrong.

4) As with many similar combat changes, it greatly hoses martial characters and rarely affects spellcasters. Even if you include supplemental effects for spellcasters that aren't actually an additional punishment for non-spellcasters (for example one noteworthy table I was at for a very short time where a critical fumble on a ranged touch spell caused it to automatically hit a nearby ally instead), spellcasters always have the option to simply stop using spells that require attack rolls. Martial characters don't.

This is instantly fixed by ensuring that special critical hits give more that critical fumbles take away.

As with all other home-brewed rules, home-brewed fumble charts should be carefully designed to make the game fun for the players. If your DM isn't doing that with his fumble rules, then the problem isn't fumble rules; the problem is a DM who isn't making sure that his home-brewed rules are making the game more fun for the players.

5) Such systems are often grossly imbalanced compared against the critical hit rules already in place in the PHB. A successful critical hit results in multiplying the weapon damage roll by the particular weapon you are using. And that's it. The mathematical advantage gained by this varies greatly depending on the weapon, but at best it generally results in an extra round of successful attack rolls. On the other hand, critical fumble charts often produce a staggering array of effects, the most common of which in my experience is dropping your weapon (costing you a round of attacks in addition to opening you up to free AoO's from the enemy), to damaging or breaking your weapon which potentially removes you from the fight, to causing harm to an ally, or some other permanent effect which could not only remove you from the current fight but future fights as well.

I agree with you that stupid fumble charts are stupid. The answer is fumble charts that aren't stupid, not getting rid of fumble charts.

Using the special critical hit and fumble charts from The Dragon #39 has never cost my PC the fight. Mostly it just complicates the fight. And once it unambiguously won us the fight. [The DM later told us that he wasn't going to have the BBEG kill us, but my lucky strike prevented him from delivering his big speech.]

In the many, many years I have played tabletop RPG's I have never once played in a game that used a homebrewed critical fumble chart that was not ultimately detrimental to the game for one reason or another.

Then those are bad tables. That doesn't prove that all fumble tables are bad.

I have found the concept to be largely contrary to the point of playing an escapism game. And my advice on critical fumble systems begins and ends with "Don't."

If I had your experiences, I would probably agree with you. But I have used reasonable special critical hit and fumble tables, and have had great times.

[Netbrian]

Oh, is the presence of orcs a requirement for fumbles?

Yep. The PCs blow when orcs are near.

[Doctor Awkward]

If I had your experiences, I would probably agree with you. But I have used reasonable special critical hit and fumble tables, and have had great times.

Just curious, but why is your anecdotal evidence any more valid than mine?

And regardless of how much fun you might have (accidentally) had the math is not on the side of the players when critical fumbles are in play. Over the course of the campaign they will statistically harm the PC's far more than they harm the enemies.

[NNescio]

This is actually an official variant..... For checks that auto succeed/fail on a 20/1 (which does not apply to skill checks). I had a DM that ran this for skill checks, and it was universally loathed by everyone at the table. It gave rise to the joke about olympic swimmers drowning in swimming pools, because even with +14 to swim, rolling a 1 results in you sinking under the water.

Should have taken 10.

(Though yes, I agree. It's poorly conceived. Even with judicious use of Take 10, in combat this still leads to [master acrobats] faceplanting on any floors that are not smoothed out to professional standards [flagstone, rough hewn floor, slight slopes, etc.], and failing really obvious Listen checks like someone speaking right next to you. )

[Drackstin]

Please tell me you at least have players roll to hit against their ally's AC when they're going to hit them? Nothing sounds more suck-tacular than dying because your friends great axe somehow phased through your greatshield, +5 fullplate, and +5 defending shortspear to decapitate you

Yes they do have to see if they hit their adjacent ally, but most of my players crit fish so they end up criting and dealing a lot of damage. by the time a player would have all those items, some of my players have a 15-20 crit and about +30 or more to hit.

[Sereg]

I like fumbles because of my philosophy of "conflict is chaos" and the idea that it's fun to unexpectedly have things go badly sometimes.

There are two problems:

1. People often like to play as powerful people and don't like being forced to have their characters make stupid mistakes. This is a taste issue. Personally, I consider that part of the fun.

2. It widens the gap between melee and casters. Though this is fixable. Firstly, you can only fumble on the first roll of a round (high BAB doesn't screw you over) and secondly, you always have to roll to cast a spell (a fumble causes you to lose the spell AND zap yourself for damage scaling by spell level). That means casters can't avoid the consequences of fumbles.

[Jay R]

Just curious, but why is your anecdotal evidence any more valid than mine?

All the anecdotal evidence is equally good. My logic is more valid than yours because it is supported by all the anecdotal evidence, not just some of it.

You wrote, "No one ever rolls to confirm a critical failure (at least not at any table I have played at)."

This anecdotal evidence is perfectly good to describe your tables. But my anecdotal evidence shows that it cannot be generalized to all other tables. Hence my conclusion, "If I had your experiences, I would probably agree with you. But I have used reasonable special critical hit and fumble tables, and have had great times."

On the subject of whether fumble tables can ever work in a game, your anecdotal evidence shows that what you did worked badly. My anecdotal evidence shows that what I do works well.

My conclusion that it can work well or badly depending on implementation is supported by your anecdotal evidence as much as by mine. Your conclusion that it is always bad is disproven by mine.

And regardless of how much fun you might have (accidentally) had the math is not on the side of the players when critical fumbles are in play. Over the course of the campaign they will statistically harm the PC's far more than they harm the enemies.

First of all, so what? Living dangerously is the point of an adventure game. The way for a PC to avoid harm is to stay home. But nobody wants to play Hovels and Housework.

I have said it several times in my post. I will now say it so directly that you cannot deny that I have said it. I play with special critical hits and critical fumbles. I agree with you that adding critical fumbles without special critical hits is a net tactical loss for the PCs. [I also don't care, but that's a different issue.]

But that's not what I'm defending. I'm defending a system in which special critical hits add more than critical fumbles take away. By the exact same logic that you are using, this is a net gain for the PCs, and especially a net gain for the martial characters.

[I agree that fumble tables with instant kills are horrible. I don't do that, and I'm not defending that. A good fumble table should add complications and suspense, not random death.]

The tables I use are from The Dragon #39. More than half of the fumble results are a DEX roll to prevent a fall, dropped weapon, lost action, or the like. This is not a huge penalty even when you fail the roll. Many of the others are an automatic penalty of the same basic level. Actually doing damage to oneself or an ally is pretty rare (16% of fumbles, which are only 5% of all misses), and never overwhelming. Meanwhile, the additional crit damage from the special crit table is quite helpful.

In over 40 years, I've only had two PCs die, and neither of them came from a critical fumble.

I've certainly lost fights -- fleeing, being captured, etc. But I would have a lot less fun in a game I could never lose, so I reject your notion that making it possible for PCs to lose encounters makes all the fun "accidental".

A good fumble table should add the occasional extra moment of suspense, and occasional setbacks to overcome.

"Accidentally"? Accidents don't last over forty years.

Have fun your way, and that's great. But please recognize that there is not only one way, and that disagreeing with you does not make my fun accidental.

Have your fun, purposely using what you have made work, and I will have mine, purposely using what my DMs and I have made work.