After writing my post on double damage and "instant death," I started looking into the history of critical hits in roleplaying games. In doing so, I came across an installment of Gary Gygax's "From the Sorcerer's Scroll" column in issue #16 of Dragon (July 1978). Among many other topics, Gygax touches on the topic of critical hits. This is, I believe, the first time he specifically addresses the topic in print (though I'm prepared to be corrected, if I've overlooked an earlier text on the subject).
Purely from a historical point of view, Gygax's position strikes me as odd. Firstly, there's the matter of Empire of the Petal Throne, published by TSR in 1975. EPT includes critical hits, the first published roleplaying game to do so, and yet I can find no evidence that Gygax was particularly exercised about the inclusion of this mechanical innovation. I suppose it's possible that his opinion on the matter changed. After all, the section about appeared in 1978, which is more than enough time for him to have decided, on reflection, that critical hits were a problem.
Alternatively, it's possible that Gygax's condemnation of critical hits was a narrow one. He calls them "particularly offensive to the precepts of D&D," not "to the precepts of roleplaying games." He may simply have felt that Dungeons & Dragons was designed with a particular type of experience in mind and that critical hits ran counter to that design. Thus, the presence of critical hits in EPT was of no concern to him, since his opinion was solely concerned with D&D.
However, if that's the case, we have to reckon with the existence of the sword of sharpness and vorpal blade, two magical weapons introduced in Supplement I: Greyhawk in 1975, the same year as Empire of the Petal Throne was published. These weapons include the possibility of instant death by the severing of an opponent's neck and the possibility of this happening is much greater than that of an instant death critical hit in EPT or even the more traditional "double damage on the roll of a natural 20" favored by most versions of the rule in wider circulation. No doubt Gygax would (reasonably) say that these weapons are rare and their inclusion is entirely up to the individual referee. Still, I think there's more than a little mechanical similarity between the way these weapons work and critical hits and that somewhat undercuts Gygax's stated position.
From a purely personal perspective, I can't quite recall when I first encountered the concept of critical hits, but I suspect it was quite early in my introduction to the hobby. By the early '80s, critical hits were one of those rules that everyone knew about and many used, even without being able to point to a section of D&D's actual rules that supported them. For many years, I carried around a photocopy of the critical hit tables from issue #39 (July 1980) and occasionally made use of them. I've never had any really strong feelings for or against the concept, which is why I find Gygax's vehement denunciation of them so odd.
When I joined my group they were already using critical hits and fumbles and had been for some time. This was before the MM was introduced. Personally, I like them- they make for higher highs, and lower lows during combat, which spurs action and creativity.
ReplyDeleteGary seemed particularly miffed when it came to people tinkering with the combat system, in OD&D or AD&D. Of course by the time of AD&D, we were more often than not "listening" to Gary the Business Man, Not Gary the Gamer, even when he was talking about gameplay and game rules.
I like the sweet little backhand slap at "would-be game inventors".
ReplyDeleteGygax at his petty, bitter, childish worst, yeah.
DeleteYeah, it is sadly the Gygax that was a total asshole pontificating. It's not worth analyzing or trying to defend. The man was a total dick sometimes, like many of us.
DeleteWe used double damage on a natural 20 from day 1.
ReplyDeleteI learned about critical hits before I even knew who Gary Gygax was.
It not odd at all. By beginning of 1978 he and the rest of the TSR STaff were sick and tired of being bombarded by phone calls and letters explaining (often in great detail) of how D&D ought to be and how TSR was doing it wrong.
ReplyDeleteThe unfortunate side effect of which is that Gygax and the staff was negative of anything that wasn't invented or vetted by TSR.
A good published example of this attitude is Tim Kask's intro to Deities, Demi-gods, and & Heroes.
This is a good point, yet Gary took his attitude to the APAs which to my knowledge were MOSTLY keeping to themselves (though, and I'm not sure of the timing, but the Illusionist was contributed by Peter Aronson who also contributed to the Wild Hunt - but maybe later). I personally didn't really notice Gary's attitude until reading his diatribe in The Wild Hunt. And then later I read about some of his anti-social antics at conventions.
DeleteIt's sad that running a business so often runs afoul of hobbyist tendencies. And it's particularly sad considering the hobbyist environment where Gary got his start.
But the gaming hobby is better for the competition and alternate ways of doing things. And a good business person recognizes that.
A lot of critical hit tables are terrible (IMO) because a) they short circuit the concept of hit points and b) they happen way too often. So crits like "abdomen turn open, your intestines pop out, roll to see if you trip on your tripes", or "genitals torn out" (both from Arduin crit tables) are terrible because they wreck characters, and crits (& fumbles) that happen 5% of the time can be bad because they just up the foes' damage potential too much.
ReplyDeleteThat said, AD&D is a good game for critical hits and fumbles because making a new 1st level character doesn't take much time. In my first (and only) experience with Runequest it took a long time to make a character and he died in his first battle when he fumbled and stabbed himself in the belly with a 2-handed spear - somehow.
Isn't short circuiting the hit point system the whole point of critical hits? To provide the possibility of swift & decisive combat resolution instead of slowly chipping away hits points? Low level characters may have few hit points, but at higher levels they often do and it can make for boring combat.
DeleteTrue. That is the single most obnoxious thing in 5e in my eyes ... the eternity it takes to whittle down a monster hit by hit, which I often liken to felling a huge tree. Not even double damage leaves much of dent in many cases.
DeleteOn the other hand, Stars Without Number has specific sniper rules that require the target to save or be mortally wounded. A breath of fresh air!
We've been using a "20" indicates MAX damage (no damage roll necessary) and a "1" is an absolute miss (not a fumble).
ReplyDeleteThe only problem is that I seem to roll more 20s than my players, resulting in a lot of dead PCs...
The excitement from a natural 20 and the agony from a natural 1 far outweigh considerations of 'realism' in my opinion. I don't like critical hit or fumble tables, I just do double damage, and I use a very simple percentile system for fumbles.
ReplyDeleteThe group I played with never did critical or fumbles for D&D and we played from summer 84 onwards. We did however play a lot of MERP/RM from 88 and we did enjoy the enjoyment of dishing out E criticals to enemies and scoffing at our colleagues fumbles. We ayed small amounts of WFRP and RQ and Stormbringer and combat was deadly from criticals in all of these.
ReplyDeleteI agree that the natural 1 and 20 are exciting and would use drop weapon/lose initiative for a "1" and do maximum damage.
Thinking a bit more modernly, for a 20 you could give maximum damage and advantage on your next round and for a 1 you go at a disadvantage next round.
But I do like the narrative and snark in the MERP and RM tables. The SpaceMaster ones were better!
The Iron Crown games were all designed from the ground up to be *about* the crticals, though. The interaction of different weapon and armor types and the way it influenced the crit charts you were rolling on was way more important than the number of concussion hits you were taking in a fight, and it was almost always crit effects that dropped a combatant. With D&D crits were at best a homebrewed afterthought, at least until WotC actually went through and looked at the math of combat so that they could include x2 and higher damage crits without breaking the game. Took till 3rd for the game to let you specialize in being good at landing crits, which is kind of a sad commentary on the TSR days.
DeleteSpacemaster crits were indeed the best for sheer excess and snark. Vader would have been very upset at the number of disintegrations our games had. :)
I always thought it was a missed opportunity that Dragon Magazine didn't have a monthly article on house rules with followups based on letter input on how they might have worked.
ReplyDeleteIf I remember correctly Gary was pretty free-wheeling in the OD&D days and very strict in the AD&D days so just make sure the house rules are OD&D only (wink, wink).
Interesting stuff! I do think that a critical on every 20 can lead to overkill. I've just come up with a house rule to get round this: essentially, a natural 20 allows for an exploding damage die - so that the door is opened, however narrowly, to infinite damage.
ReplyDeleteThat's… rather brilliant.
DeleteOf course, as Delta demonstrates with statistics and stuff, an exploding die is roughly equivalent to, well, less than 1 point of extra expected damage (closer to +¾ for d4s and d6s, closer to +½ for larger dice). So even though exploding dice are exciting and can occasionally lead to some memorable chains of rolling continual 6s, in the vast majority of the time, upping a d6 to 1d6+1 or 1d8 is actually the more impactful outcome. Compared to double damage crits, exploding dice crits barely register as doing anything at all.
… Which, don't get me wrong, is very much a reason to like the idea! They won't really disrupt the combat system as it's intended to work! (I do find exploding dice kind of personally annoying, though, because they leave some damage values within a given range impossible to roll: if your dN explodes, you can never roll any multiple of N for damage, unless you subtract 1 point from the total every time the die explodes. A petty pet peeve, but it bugs me anyway.)
Glad you like it! We tested it out last night, and it seemed to work pretty well. The thing is, as only a 20 allows the possibility of exploding dice, you still get plenty of hits that do 6 points of damage and no more. We had one crit with exploding damage yesterday, but plenty of genuine 6s on rolls that connected on lower rolls. But of course the impossibility of 6 could arise when a 20 is needed to hit the foe in the first place.
DeleteI do like the idea that a first-level character *might just* behead a dragon with a lucky sword stroke - even though the luck would have to be considerable!
http://hobgoblinry.blogspot.com/2022/02/house-rule-for-critical-hits-plus-str.html
ReplyDelete(Meant to post this link setting it out!)
I am pretty certain this is a reaction to the growing popularity of the Arduin Grimoire rules, which were first published about a year before the article and had taken the gaming world by storm (well, relatively speaking).
ReplyDeleteThough originating in the West Coast game node, I have found that the Grimoire quickly spread across the country, as did their influences. Gary would bash on Arduin and Dave Hargrave for many years; from my understanding his opposition to the material was mostly Business Gary, not Gamer Gary, as anything that leveraged sales away from TSR was considered anathema.
Remember, this is the same time that TSR realized that Judges Guild was making bank through their D&D license, and then jumped on the module bandwagon.
Also, new fantasy RPGs were really competing with D&D for the first time... the first ad for RuneQuest appeared in that same issue!
"Accept No Substitutes" is essentially what Gary was telling people with that article.
I can confirm that Arduin was happily mutating local D&D games here in upstate NY by the time this article ran - some of the earliest games I played at the big local club used AG rules, which confused the heck out of kiddie me after starting on pure TSR rules in my afterschool group.
DeleteThen again, Gygax might have been just generally pissy about any competition at that point. He'd been snarling about Tunnels & Trolls since it showed up too, and from what I heard back then Ken St. Andre was giving it back with interest.
This excerpt reveals some fascinating things about early roleplaying games. To which version of D&D does he feel critical hits would be "offensive to the precepts"? Is he talking about the original D&D, AD&D 1st ed. or AD&D 2nd? This era was before my time (I was born in the 70s), so most of what I know about early D&D comes from things I've read after the fact.
ReplyDeleteAs someone who became interested in the hobby in the 80s and 90s, I thought Gygax viewed early D&D as primarily a tactical wargame with a thin roleplaying veneer. However, Gygax's comments here show that he expected players to identify with their characters.
Instant death crits play merry hell with character advancement. It's already (when played by the RAW) torturously slow, and anything that makes it even more difficult to survive long enough to reach the next level is a thoroughly unwelcome addition to the game.
DeleteThen again, I can and do make the exact same argument about all the save-or-die BS that clots up TSR-era editions of the game, particularly the absurdly commonplace poison saves. Your character's existence should not depend on a single d20 roll in a game with attritional hit points instead of "injury level" mechanics.
I'm increasingly sympathetic to the idea that death shouldn't hinge on a single die roll, though I am not yet fully convinced.
DeleteMost of the poisonous monsters can be dealt with by ranged attacks. I tend to give a lot of foreshadowing for poisonous traps.
DeleteRather than just damage, stat drain for failed saves, with different types of poison draining different stats.
I’ve never been a fan of critical hits bolted onto a system because as DM I felt they were too harsh on the players and the few times I got to play my PC was repeatedly mauled. I went from a Monk where the Dragon Magazine tables gave him a knock on the head that reduced his Wisdom below class minimum to two successive 3rd level characters killed by triple damage 27 HP critical hits.
ReplyDeleteThe classic critical hit tables short circuit the players ability to gauge risk and perform logistics on their combat ability that to me is so key to the dungeon crawl as long term endeavor. If there’s a 5% chance per attacker per round of your Pc being removed from play it’s too great. The existence of random damage at all - that the 2d6 roll could come up 12 rather than 7, doubling the damage - is enough.
I think that this is a very good point.
DeleteBy my reading, Gygax seems to be more focused on monsters critting players (see his note #1), and that every monster being able to deal a potential death blow waters down those monsters with a death-dealing capacity (and leaves the poor players few means to prevent it). Am I reading that incorrectly?
ReplyDeleteStarting off all characters with a weapon that could instakill on a 20 would be equally offensive to D&D. Some very few characters eventually finding a weapon which can do that, and is subject to destruction from item saving throws, isn't offensive to D&D.
ReplyDelete"If something exists at all, why isn't it available all the time" isn't pointing out a structural flaw in the argument.
RE: Gygax, as Rob says, I'd love to see someone publish all the communications TSR received from gamers telling them what TSR should do, as if TSR's role in the hobby was to publish Gamer X's perfectly-fitted game system (which would, of course, be unacceptable to every other "Gamer X" writing the same sort of letter to TSR at that time.
I suspect it would be eye-opening.
Even if you concede Gygax's points about game balance and continuity, you still might rather play with critical hits and fumbles because they given combat something he never figured out how to introduce through official channels: excitement. The rules in D&D present a treatment of combat that is actuarial and gradual, and played as-is often feels like slowly rubbing an eraser across a page until worn it down to a nub. Players naturally want to have combat involve excitement, risk and dynamic action. Apologists for the standard rules always offer the explanation that the DM and players are supposed to fill in those details through ad libs and descriptions (this is usually offered up with a healthy side helping of sneers). But 48 years of house rules and alternate systems tell you the game always should have treated combat more dynamically.
ReplyDeleteI just wanted to highlight the massive difference between a combat mechanic, and an item mechanic. the combat mechanic will proc several times a night. the item mechanic, much less, but also, don't like it? make the vorpal sword a longsword +5. fixed. no proc.
ReplyDeleteWeird... it sounds like Gygax was against tinkering and making the game one's own while also advocating for preserving PC lives like some kind of 5e, story-first millennial.
ReplyDeleteI wonder what he'd say about Crimson Escalation?
I do not like critical hits in D&D--at all. When I am DM, nothing special happens when your attack roll is a 20--you just hit. Now roll your damage: If you roll maximum damage, THAT is something to celebrate, and the player can even call it a "critical hit" if he likes. Rolling maximum damage is its own reward.
ReplyDeleteI remember the time my fighter surprised an ogre and hit him with a mundane two-handed sword. Damage is 3-18 for large opponents (such as ogres). I rolled the three dice: 18! (Only a 1 in 216 chance of rolling that!) That plus my strength bonus was enough to kill the ogre with that single blow. It was awesome.
Back around 1978, our D&D group adopted a critical hit and fumble system inspired by the material in Arduin Grimoire volume 1 (1977). We required a secondary roll beyond the initial 1 or 20 in order to keep the frequency of crits and fumbles at a reasonable level. The secondary roll was tied to character class, so that fighters were least likely to fumble and most likely to crit whereas mages were most likely to fumble and least likely to crit on a melee attack. My impression at the time was that a lot of D&D groups were mining Arduin Grimoire vols 1-3 for modifications/new material. I don't recall anyone caring about Gary's opinions about anything.
ReplyDeleteI like a good critical hits table in D&D but I can respect those who reject one. I personally like adding a little more swinginess to D&D combat which can start to feel like slow attrition at times. I also can't imagine using one where monsters don't play by the same rules or where it's not balanced by a fumble table.
ReplyDeleteThe problem with fumble rules is that they turn the game into a clown show. In a real melee, your odds of hitting a friend or dropping your sword in any given minute are vanishingly small. Roll multiple 1s in a row small. Peg the odds at 5% per round for every attacker and you might as well accompany every combat with "Yakety Sax."
DeleteIf you require "balance," you need to either make both critical hits and critical fumbles exceedingly rare (at which point, why bother including them?), or you need to come up with a reasonably minor effect for what a fumble actually does (which, again, if it's suitably minor, why bother?) and balance crits against that.
Hitting oneself or an ally doesn't happen. Dropping your weapon basically doesn't happen. Opening yourself up to what amounts to an attack of opportunity from your opponent? Feasible, but then what happens if you roll a fumble in missile combat? Honestly, fumbles are just far more hassle than they're worth.
It seems highly likely in a crowded churning melee all sorts of mistakes could occur. Especially as combatants are jostled and everyone is moving separately. Missile weapons seem even more likely to cause damage due to friendly fire. Are you getting this from re-enactor experience?
DeleteI don't like the every time a 1 is rolled thing either so I do a second prove roll (a hit or miss) which adjusts for combantant skill and ac of the target.
A data point about fumbles is that the RuneQuest fumble table was based off real results of SCA combat. Now granted, SCA combat is not real sword play, and our RPG characters are living their profession as opposed to it being a weekend hobby, but it's still a reality.
DeleteNow it does help that RQ crits and fumbles are 1/20 the probability to hit or miss so more experience characters will crit more often and fumble less often.
p. 61 of the AD&D DM Guide further explains EGG's thinking on the issue.
ReplyDelete"In like manner, consider all of the nasty things which face adventurers as the rules stand. Are crippling disabilities and yet more ways to meet instant death desirable in an open-ended, episodic game where participants seek to identify with lovingly detailed and developed player-character personae? Not likely! [...] Such rules as double damage and critical hits must cut both ways ~ in which case the life expectancy of player characters will be shortened considerably - or the monsters are being grossly misrepresented and unfairly treated by the system. I am certain you can think of many other such rules."
The implication is not that instant death effects are out of the question, but that there are already enough in the system, and they do not need to be made a routine part of combat.
Mike! Long time, no see. Drop me an email at jmaliszeATgmail.com when you get the chance. I'd love to get back in touch.
ReplyDelete"instant death" no longer allows participants to use judgement when playing.
ReplyDeleteHilarious, coming from the guy who wrote Tomb of Horrors.
I like how 3e handles critical hits. A natural 1 on an attack roll is always a miss, with no further penalties. A natural 20 is always a hit and prompts another attack roll. If the second roll beats the target's AC, then it's a critical hit instead. This keeps the excitement that comes from rolling a natural 20, without 5% of all attack rolls becoming critical hits.
However, I dislike how critical hit damage is handled. By default, you simply roll for damage twice. This means you have the possibility of rolling snake eyes, which turns the whole thing into a letdown, since your supposed "critical" hit did less damage than an average one. I've seen some DMs use a hit location table to alleviate this issue; it might have only been two damage, but it was two damage to the eyes. Unfortunately, this creates its own problems because D&D's abstract hit points were never meant to model specific wounds.
My solution is to treat the first damage die as maximum, with the second die being rolled normally. That way, a critical hit always hits harder than a normal one, and we don't have to track injuries like "wounded sword arm: -2 attack penalty" and such.
Many 3/3.5 weapons have 10% base crit (or threat, more accurately) chance, and that's disregarding the many ways to improve your odds of crit-fishing. IIRC even 3.0 let some classes get down to a chance to crit on 12+ or 14+, although that took some serious dedication. The weapons that did threaten a crit only on a 20 were either "bad" non-martial weapons or they had x3 or even (in rare cases) x4 damage mods.
DeleteThat fact that you had to confirm a crit by rolling a successful (normal) hit against AC also helped fighters and other high-AC types take far fewer crits than they would otherwise. It's a nice touch that very few other crit systems bother to think about.
Never much cared about a crit still whiffing on damage, but one GM I played with house ruled it so that a second crit on the confirmation roll maximized the damage. That worked pretty well, although it rewarded crit-fishing quite a lot.
Another one had a crit on the confirm roll lead to another roll to confirm, which could also crit and trigger a yet another confirm, ad infinitum. Each successful confirm after the first added X1 to the final crit multiplier, and you could hypothetically score infinite damage. That one probably didn't actually reward crit-fishing as much as the first, and if you ever managed anything really impressive you probably should have played the lottery instead.
This is an area I love to point out Cold Iron which uses a clever randomizer technique to place results on the standard normal distribution with all it's infinite expansion to either end. Fumbles are moderated by having an absolute result value to occur that places them further and further down the "miss" side of the curve as the character is more and more experienced. On the other side, the open ended nature of the bell curve is used for open ended crits, however, they are pushed out by the defensive capability of the opponent. The biggest results:
Delete1. Fumbles become almost non-existent after the characters have gained some levels.
2. Crits most often occur when one side has used strategy and tactics to create a significant advantage, often overwhelming an opponents defenses. This allows for a quick win.
Also, since crits are some multiple of damage (and yes, there IS the possibility of a damage whiff), and characters hit points do scale up (though damage DOES also scale up), as levels go up, the likelihood of an instant kill goes down. And then add in the way negative hit points are handled, with the possibility of saving someone so long as the field is held, results in relatively few PC deaths so long as the PCs are being smart. Now if they let themselves be routed by a poor choice of engagement, yea, they will be losing characters.
Another driver here is the death spiral, when a Cold Iron character has lost half their hit points, their attack and defense ratings drop.
This creates an environment that truly supports strategy and tactics and meaningful choices.
I like the critical hit system from Tom Moldvay's Challenges System. I’ve been tinkering with that in my own AD&D game.
ReplyDeleteA major problem with the D&D critical, which triggers on a natural 20, is that the worse you are at hitting and opponent, the greater share of attacks that hit will be critical hits This makes little sense
ReplyDeleteBack in the 80s, my group introduced criticals that had to be confirmed: if you rolled s natural 20, you had to rolle again, and only if the second roll would hit did you do double damage.
Yea, confirming crits (which is mathematically close to what RQ does) helps a lot in making things feel more sensible. Fumbles if used should be confirmed also.
DeleteAnother way to scale crits and fumbles is margin of success or failure, which is helped by some kind of open ended resolution system.
"...many would-be game inventors feel they have sufficient expertise to design a better system."
ReplyDeleteThis odious quote is actually right in line with Gary's advice in the Preface and Introduction of the DMG, which also talks about what is "contrary to the major precepts of the game."
Rather than the usual "These rules are merely guidelines, modify them as you see fit!", Gary counsels the opposite. He advises against tinkering with the rules, explaining that a lot of know-how went into them, and that if you tinker too much you're no longer playing AD&D, like that's some big deal.
There, he also writes to would-be DMs:
"Welcome to the exalted ranks of the overworked and harassed, whose cleverness and imagination are all too often unappreciated by cloddish characters whose only thought in life is to loot, pillage, slay, and who fail to appreciate the hours of preparation which went into the creation of what they aim to destroy as cheaply and quickly as possible. As a DM you must live by the immortal words of the sage who said: “Never give a sucker an even break.” "
Gary's always been the guy who wrote those letters to Dragon magazine.
He was in a chess tournament phase for a long time. I'm not sure if he truly wanted the same level of definition and even handed judging. Perhaps seeking more respect for the hobby in the same manner as chess. Perhaps he was seeking a one single set of rules that would eliminate the competition. Mastering AD&D as a wargame would take resources that might have been used on a competing product. Maybe a little of both.
DeleteNeither was ultimately successful. The only reason D&D survives is it is "Oreo" or "Kool-Aide". It's the standard in all editions. All editions have certain problems in common hit points are ridiculous at high levels, finding a way to keep them within reason would help but it's the one to beat in high fantasy.
I'd like to think AD&D would have turned into something like Castles and Crusades given time.
I was first introduced to them by a friend who bought ARMS LAW from ICE in 1984 (?)
ReplyDeleteSomething struck me in that excerpt - how are the players supposed to know which monsters cause instant death except by trial and error? I thought it was considered poor form for the players to read and memorize the Monster Manual, and then use that knowledge in game?
ReplyDeleteI feel like the expectation that players would remain ignorant of DM mechanics is part of a different playstyle. Back in the day almost everyone in my group was also a DM, so the expectation was that everyone knew the MM. I don't think our group was unique in this. And Gary, being Gary, would have wanted to sell more DMGs and MMs, so...
DeleteIn Gary's imperious Preface to the DMG, he tells DMs that if their non-DM players are caught reading the DMG, the DM should punish them by taking money or two magic items away from their characters.
DeleteBut Gary expected all players to memorize the Monster Manual?
Oddly, I object to the fumbles more than I do critical hits, and it's largely down to the name history has dealt the former. It implies that a bunch of trained warriors, possible even supposed heroes, are *fumblers*, constantly at non-negligible risk of hitting their allies or themselves, tripping over their own shoes, etc.
ReplyDeleteNow, if you want to treat a 1 rolled as something unforeseen or unfortunate befalling a character that isn't framed as stemming from his or her own incompetence, that's a bit easier for me to swallow. Maybe the opponent comes down with a might swing that splinters your shield and drives you to your knees, costing you the initiative next around. Same end result, but the PC isn't stuck playing the fool on top of being penalized. So unless you're running a straight up comedic game like Paranoia, I reckon it's best to include few, if any, true fumbles.
As to critical hits, I do think Gygax has a fair point. Resource management is a component of old school play many love, and anything that makes it tougher to manage one specific key resource (hit points) at random intervals may not be desirable under that mindset. Especially with pre-WotC hit point totals and death rules in effect. It can take the randomness already baked into attack and damage rolls and nudge it one step toward full roulette. Personally, I've been using the more measured method recommended in the Swords & Wizardry rules (auto-hit and +1 damage on a natural 20) and it works fine.