Mr Reynolds math is quite good.
The trouble with crit-stacking isn’t that it’s unbalanced, or even that it’s “too good”. It’s that it’s incompatible with the assumptions that D&D runs on.
From way in the beginning, D&D combat has been ablative. No, that’s not especially realistic – but it made managing your slowly-dwindling supply of hit points an important skill. It usually meant knowing that you were going to be overwhelmed several rounds before it actually happened, giving you time to look for an escape. It meant that the fighter could sacrifice some hit points to buy time for the wizard to work on a powerful spell with a long casting time. It meant that ambushes and first strikes weren’t all that important. It meant that letting your friends cover for you while you took a healing potion or stepped back to let a cleric heal your wounds was a sensible move. It gave you a rough idea of how long you could last, and how long your various friends could last, against a given opponent.
If and when you “encountered” something like a medusa, or a dragon, or any other creature with the capacity for one-round kills, the “encounter” often started long before you got to the creatures lair. It started with a rumor in a tavern, or a tale of adventure that served as a warning. After that, it was time to collect more information – often by visiting a sage or looking for people who’d encountered the creature and survived. Based on what they told you, your party selected appropriate spells, gathered supplies, and took special precautions before you went scouting. Depending on what you found, you might refine those preparations, choose to opt out, come up with a clever scheme to win without fighting, or (if you were unlucky) get caught. If it came to a fight at all, it would be a lot shorter than the investigation had been.
For game purposes, all of that was good. Player-characters would be in a lot of fights over their careers, so they’d be ambushed, and take critical hits, and make tactical decisions a lot. They might even have to run away a lot. Making sure that combat was ablative (even if it did play pretty quickly) was the combat equivalent of avoiding “save or die” spell and power mechanics. It let sensible players keep their characters alive.
The trouble with crit-stacking isn’t anywhere in the math of average damage. It’s in the actual play. Lets say you’ve got sixty hit points, and you’re taking some damage in a fight. What you’re inflicting doesn’t matter that much, so I’ll only consider your opponent…
You’re fighting a strong guy with a big weapon who can hit you.
- Round One: Miss and Hit for 15.
- Round Two: Hit for 10 and Miss.
- Round Three: Hit for 16 and Miss.
- Round Four: Miss and Hit for 13.
- Round Five: Time to fall back, let the rogue take a hit or two to cover you – his 38 hit points should be enough to handle this guy for a round, and probably for two – and get some healing. With only 6 hit points left, you don’t want to risk taking another hit right now; another sixteen-points (which, according to my dice, was coming up next) would kill you.
Now, against someone focusing on Crit-stacking, that might go…
- Round One: Miss and Miss
- Round Two: Hit for 6 and Miss
- Round Three: Critical Hit x3 for 18 and Miss
- Round Four: Miss and hit for 10
Now that leaves the character with 26 hit points, and no real reason to retreat. Even the critical hit wasn’t anywhere near that bad, so you should have time to fall back next round if you need too, right?
- Round Five: Critical Hit x3 for 36. Your character is instantly dead – yet the average damage he or she was taking was exactly the same.
Admittedly that’s assuming that the big hit comes only after four or five rounds – but player-characters do a lot of fighting. It will come up – and that’s why people don’t like crit-stacking. It’s also why you don’t see many monsters that rely on similar attacks and why save-or-die effects have gotten de-emphasized over the years. Both players and game masters usually want combat to be predictable enough so that they can make strategic decisions without too much complex math and without worrying about the idea that random fighter opponent #3 just might occasionally burst out with some enormous damage score. The same sort of reasoning underlies the unpopularity of wild magic and similar mechanics. It has nothing whatsoever to do with all those nicely-worked-out averages.
There are a couple of factors that tend to obscure the point in 3.0 and 3.5 D&D though.
The first major point is the emphasis on “balanced encounters” with monsters. Fundamentally, what game designers mean by a “balanced encounter” is one that the player characters will almost certainly win if they throw themselves into it and really make an effort. (For why that’s a bad idea, see HERE).
That, however, gets hidden in all the talk about “game balance” – wherein they mean that all the character types should be equally effective. They certainly don’t mean that a “balanced encounter” is one where “the player characters have only a 50-50 chance of winning”. You want a genuinely “balanced encounter”? Try an old-style Mirror of Opposition.
The second is simply that monsters don’t generally rely on crit-stacking, and neither do most NPC’s. Like other save-or-die effects, you see it being used a lot more often by PC’s than by NPC’s because randomly killing off characters is disruptive to the game – whereas the monsters are THERE for the PC’s to cleave their way through.
Now making combat less predictiable and ablative – and thus both far riskier and a thing to be avoided if at all possible – does make it far more realistic. Few D&D games, however, make combat a rare final resort – and it changes their tone a great deal if you play that way.