Static vs active defensive rolls.

[stoutstien]

Just wondering what the general feeling is for players.

Assuming the math is the same and we are working with the common dice+value roll over/under, do players rolling of given the option?

Pros:
Spreads out the cognitive load.
Players roll more dice.
Everyone sees the "randomness" so they can react to it without stopping to add color.
Makes combat feel more dangerous

Cons:
Players have more to do.
Hard to fudge
Not as common so could be confusing.

[kieza]

Depends on exactly how it's set up.

Sometimes the attacker rolls to attack and other times the defender rolls to defend: I find this inconsistent and annoying.

The attacker always rolls to attack and the defender always rolls to defend: involves more rolling, but doesn't do anything significant to the odds of success unless you're using a resolution mechanic more complex than d20.

The attacker always rolls to attack, and the defender sometimes rolls to defend if they have set up an active defense beforehand: this, I like. There is a consistent baseline in "attacker always rolls," and players can add to that if they so choose.

[GloatingSwine]

Depends how satisfying it is to win the rolls.

I'd say for standard defence with a null outcome (you didn't get hit) then it doesn't add much. If the outcome for the defender of winning is something better than that (can make some kind of responsive action or riposte) then it does.

[Xervous]

Rolling all the dice makes me feel like I’m playing an asymmetrical video game more than I’m interacting with a world through a character who resides in it.

[stoutstien]

Depends how satisfying it is to win the rolls.

I'd say for standard defence with a null outcome (you didn't get hit) then it doesn't add much. If the outcome for the defender of winning is something better than that (can make some kind of responsive action or riposte) then it does.

Hmm good point. If the roll itself was interactive on some level it could be seen in a similar vien as attacks and saves. Roll a 20 for defense then you can do X immediately and roll a 1 the inverse occurs.

[warty goblin]

Mathematically roll d20 + mode, compare to d20 + mod2 for defense is never equivalent to compare d20 + mod to a static number.

This is easy enough to see with some simple algebra. Writing X for the d20 roll, m for the modifier, and T for the target number, the second, standard D&D case is

P(success) = P(X + m >= T) = 1 - P(X < T - M)

But, writing X1, m2 for thr first roll and modifier and X2, m2 for the second, the active defense case is

P(Success) = P(X1 + m1 >= X2 + m2) = 1 - P(X1 - X2 < m1 - m2).

Now here's the key point, the difference of two d20 rolls has the same distributional form as the sum, which is to say the PMF is a pyramid; where the sum is centered at 21, the difference is centered at zero. The PMF of a single d20 is of course flat. There is no way to get these two distributions to behave the same way, in particular the effect of a +1 to your modifier on the probability of success is constant with 1d20, but in the active defense scenario it varies based on your opponent's defense. Getting +1 against an enemy with a defense one higher than your current attack is way, way more impactful than that same +1 against somebody with a defense that's 10 higher than your attack.

So rolling active defense is probabilistically the same as a system where the attacker rolls 2d20 + mod1, and compares to a static mod2 + 21. It enjoys an at the table advantage since active defense requires adding a lot fewer double digit numbers, but the success curve is the same. All of which is to say that this isn't an aesthetic change, it's a move to a more normalized distribution, similar to (though less extremely normalized) switching to 3d6.

[stoutstien]

I never intended to use 2 rolls just flipping the roll to be made by the player(s) rather than the NPCs. It will remain Dx+ mod compared to static value.

[GloatingSwine]

So in effect "save vs sword"?

Then yeah, I think if you do that for every attack roll made against players there definitely wants to be something more on the line than the null outcome. Whilst that's the case with other saves, they're rarer than attacks and generally attached to fancy things so it doesn't become obvious that you're rolling to have nothing happen.

Players rolling their attacks NPCs rolling theirs has the advantage that the roll is made by the active doer and the outcome of success is always "something happens". Which is nice and intuitive consistency.

[warty goblin]

I never intended to use 2 rolls just flipping the roll to be made by the player(s) rather than the NPCs. It will remain Dx+ mod compared to static value.

Then I guess the only question is does your game have monster critical hits, and if so, are the players going to the change from

"The orc warlord swings its axe is a brutal arc with all its strength, hitting you for 2d12 + 4" when the orc rolls a 20, to

"You seem to have two left feet as you stumble into the oc warlord's axe for 2d12 + 4" when they roll a 1.

[kyoryu]

I think "hard to fudge" is a pro, as is "players have more to do".

"Not as common" is... meh.

Most of the games I play use active defense and it works fine. It doesn't add a lot of time, either.

[stoutstien]

So in effect "save vs sword"?

Then yeah, I think if you do that for every attack roll made against players there definitely wants to be something more on the line than the null outcome. Whilst that's the case with other saves, they're rarer than attacks and generally attached to fancy things so it doesn't become obvious that you're rolling to have nothing happen.

Players rolling their attacks NPCs rolling theirs has the advantage that the roll is made by the active doer and the outcome of success is always "something happens". Which is nice and intuitive consistency.

Is it intuitive or is it just what is? If combat is occurring relatively simultaneously the attack <action> is merely a mechanical interface nested in combat that has its own mechanicals interface (turn/ rounds).

I think we frame attacking as more active is because it's an action in the meta.

Then I guess the only question is does your game have monster critical hits, and if so, are the players going to the change from

"The orc warlord swings its axe is a brutal arc with all its strength, hitting you for 2d12 + 4" when the orc rolls a 20, to

"You seem to have two left feet as you stumble into the oc warlord's axe for 2d12 + 4" when they roll a 1.

I'm thinking on keeping the 1 and 20 as auto miss hits and adding from there. Critical hits as a general rule is out as the individual player options will dictate how that acts.

[LibraryOgre]

Which variety do you mean?

Do you mean opposed rolls... i.e. Bob rolls a d20 to hit, the orc rolls a d20 to avoid, both add their bonuses, and the higher wins?

Or do you mean Bob rolls to attack the Orc against its AC, and rolls to defend against the orc against its Attack Value?

The first provides more variety and tactics (Hackmaster does this, and your defense die varies with your gear, your stance, and how many times you've been attacked since your last action), but it is also complicated and can really slow down the game.

The second can speed things up, engage the players more, and takes some burden off the GM... but it can also slow things down as players struggle to keep up with the math.

[stoutstien]

Which variety do you mean?

Do you mean opposed rolls... i.e. Bob rolls a d20 to hit, the orc rolls a d20 to avoid, both add their bonuses, and the higher wins?

Or do you mean Bob rolls to attack the Orc against its AC, and rolls to defend against the orc against its Attack Value?

The first provides more variety and tactics (Hackmaster does this, and your defense die varies with your gear, your stance, and how many times you've been attacked since your last action), but it is also complicated and can really slow down the game.

The second can speed things up, engage the players more, and takes some burden off the GM... but it can also slow things down as players struggle to keep up with the math.

I want to like hack master but I just can't get into it.

It's the ladder. I'm not to worried about the minute differences in resolutions times as much as the "feel" from allowing the players to "see" attacks.

*You can also run simultaneous resolutions this way with multi attack or more than one npc.*

[LibraryOgre]

I want to like hack master but I just can't get into it.

As much as I like it, it's very crunchy, and tends to require a big buy-in from all your players because of it.

It's the ladder. I'm not to worried about the minute differences in resolutions times as much as the "feel" from allowing the players to "see" attacks.

*You can also run simultaneous resolutions this way with multi attack or more than one npc.*

I like the "players roll defense against attack values" well enough; it keeps them more engaged, and helps address the problem of "Why can't I do anything to defend myself" question that I hear often... you're being attacked, you want to react, and you don't like that your defensive moves are abstracted into your other choices (i.e. weapon and armor, stats, etc.)

[CarpeGuitarrem]

It's alright but adds a lot of mechanical overhead in exchange for a small bell curve (but one that doesn't create significantly more stable outcomes, especially with evenly matched characters) and the potential for really high differences between attacker and defender, which can impact the game if you use thresholds of success. (E.g., rolling at +0 against a target of 10 with a d20 means you can at most exceed the target by 10, but if your opponent also rolls with +0, you can go up to beating them by 19.)

Personally I don't necessarily love having all that extra rolling but at the end of the day I guess it's fine. I like it in Cortex games because they're all about building big pools of dice, and you don't have to bother with difficulty modifiers.

[Easy e]

I prefer to actively do something as a player, rather than be a victim to something. To me, games are fun because they force you to make decisions.

So, to use D&D as an example; I do not enjoy being attacked, getting hit, and simply removing hit points. Boring.

I would prefer an opposed roll as an in-between option.

My preferences is to get a choice or option to avoid the attack. Is it better to parry, straight up dodge, tank it, or do something else. Of course, the option I chose will impact me later by either depleting future options, reducing resources, etc.

However, I realize that this does slow things down a bit, but my engagement in a combat is so much better. I could have to make a decision at any moment.

[Pauly]

This comes down to “what is a hit?”.
The answer depends on the system, and from there the desirability of defensive rolls.
At a simplified level the attack sequence is
1) Did the attack hit the target?
2) Did the target’s defenses take effect?
3) What is the damage?

Generally speaking D&D combines steps 1 and 2 by making the target’s AC the difficulty of the attack roll. Other systems will separate steps 1 and 2 into different rolls.

Getting back to “what is a hit?”
A hit can either mean
1) the attack lands on the hit box inside of which the target resides, or
2) a blow on the target itself.
Option (1) is normally resolved by having a fixed difficulty level for an attack, followed by an active defense roll.
Option (2) is normally resolved by having a floating difficulty level with many situational modifiers.
Generally speaking in game design option (1) is preferred for games where combat is primarily ranged and (2) is more often used in melee heavy games.

Another consideration is the damage/HP relationship.
In games where combat is lethal and each hit can potentially kill the target active defenses to negate hits are highly desirable. In games where HP bloat turns targets into bullet sponges active defenses are less desirable.

[gbaji]

It's alright but adds a lot of mechanical overhead in exchange for a small bell curve (but one that doesn't create significantly more stable outcomes, especially with evenly matched characters) and the potential for really high differences between attacker and defender, which can impact the game if you use thresholds of success. (E.g., rolling at +0 against a target of 10 with a d20 means you can at most exceed the target by 10, but if your opponent also rolls with +0, you can go up to beating them by 19.)

Yeah. Was going to respond to an earlier post about the difference, but your statement is (IMO) a more clear explanation.

As long as the outcome is binary "You hit, or you missed" (which is the case for D20 systems), then there is functionally no value to having both attacker and defender roll dice. The only effect the defenders die has is to aplly a variable to-hit modifier to the attackers chance to hit. That does create a bell cuve probability outcome, and some odd effects based on relative starting points (base tohit vs base AC say), but that's a lot of extra work for the same outcome:

You either hit, or you didn't. So I'm also in the camp of "don't bother".

And yeah, someone mentioned the whole "I want to be able to do something to defend myself". Which is 100% valid. But you have to do a fair amount of modification to a D20 system to incorporate this. At the end of the day, defensive actions in those systems simply act as adjustments to effective AC which the attacker has to overcome with their roll. You *could* produce a die roll for that adjustment instead of a flat value. Or you could create some sort of skill roll attempt to get the adjustment in the first place (maybe an even better way to go). But you're still "stuck" with a game mechanic that says everything is either a hit or a miss, with nothing in between.

This comes down to “what is a hit?”.
The answer depends on the system, and from there the desirability of defensive rolls.
At a simplified level the attack sequence is
1) Did the attack hit the target?
2) Did the target’s defenses take effect?
3) What is the damage?

Generally speaking D&D combines steps 1 and 2 by making the target’s AC the difficulty of the attack roll. Other systems will separate steps 1 and 2 into different rolls.

Yup. In RQ, a hit means you hit the target. There is no concept of AC making this somehow "harder". You then roll damage and apply it to the armor points based on how much actual armor the target is wearing (and both weapon damage and armor may be enhanced via magic spells). Any damage beyond that blocked by armor, goes though as HP damage to the target. Simple.

A parry can be attempted by the target. If successful, the defender gets to add the armor points of the parrying weapon/shield to the worn armor and reduces the damage done as a result. Also very simple.

Of course, to have this level of comparison, you have to already have a system that represents "1", and "2" as discrete elements. D20 systems do not, so there's not really any value to having an extra roll.

In RQ, you can also dodge, which basically makes a successful hit into a miss. Um... There's some rules for comparing relative success levels for the hit vs the dodge though . The game system has rules for critical, special, and regular hits, plus misses, and fumbles, so a simple way to look at it is that a successful dodge subtracts its level of success from the success level of the hit. So a regular dodge reduces a regular hit into a miss. The same dodge would reduce a special hit into a regular hit, and a critical hit into a special hit. A special dodge will reduce the hit by two levels, a critical by three (a critcal dodge always dodges, since you can't get better than a critical hit). So yeah. Dodge acts a bit more like D20 systems. Sorta. But then again, when what you're trying to do is "not get hit", that's kinda part of the deal. The point is that you still get your actual worn AP to potentially block the damage just as well as if you hadn't dodged. There's just no concept of "Armor Class" as something that changes the opponents actual to hit roll.


And yeah. I suppose this does fullfil the need players might have to be more active in terms of defending themselves. D&D has this odd sort of "you're just walking around, and you take damage based on what you are wearing, and not a lot else" concept to it. It does have the benefit of being much simpler. But it does mean that for the most part, you (the player) are a passive participant when being attacked. While in other systems, you are actually allowed to "do something" in that situation.

Some players like that. Some don't. Seen it go both ways. There's lots of people who try out RQ (or other similar games), and their response is "I'm having to rolll so many dice. It's too much work. Just tell me how much damage I take this round". Different strokes and all that, right?

[stoutstien]

This comes down to “what is a hit?”.
The answer depends on the system, and from there the desirability of defensive rolls.
At a simplified level the attack sequence is
1) Did the attack hit the target?
2) Did the target’s defenses take effect?
3) What is the damage?

Generally speaking D&D combines steps 1 and 2 by making the target’s AC the difficulty of the attack roll. Other systems will separate steps 1 and 2 into different rolls.

Getting back to “what is a hit?”
A hit can either mean
1) the attack lands on the hit box inside of which the target resides, or
2) a blow on the target itself.
Option (1) is normally resolved by having a fixed difficulty level for an attack, followed by an active defense roll.
Option (2) is normally resolved by having a floating difficulty level with many situational modifiers.
Generally speaking in game design option (1) is preferred for games where combat is primarily ranged and (2) is more often used in melee heavy games.

Another consideration is the damage/HP relationship.
In games where combat is lethal and each hit can potentially kill the target active defenses to negate hits are highly desirable. In games where HP bloat turns targets into bullet sponges active defenses are less desirable.

In regards to my framing:
-combat is dangerous. Few things can weather more than a single well placed blow let alone eatting dozens of them in a single scene.
-HP are low across the board and scale slowly.
-total attacks are low. Getting more than 1 is rare.
-a portion of melee attacks deal damage even on misses ( shields can prevent this once per round).
-AC are relatively lower.
-tougher armor and avoiding attacks will be modeled to be separate but overlapping for ease of use. *Early testing*
-npcs in general are designed with ease/speed of use first. If for any reason it needs to be converted to function like the PCs it should be seamless and thier creation should be just a fast. Will Include a tabled version to have X variants of a similar theme and thanks to not needing to roll as often the GM can run them "Character" forword.

The game world doesn't have a different "maths" for the PCs compared to the NPCs it's just presented in a different format.

[animorte]

Of course, along with everything else that's been said it does depend. Active vs passive defense is basically equivalent to rolling a save vs having a static number. 5e clearly has both weaved in separately. Utilizing an active defense would require a consistent combination of the two, and preferably valuable mechanical (and perhaps flavorful) implementation. If you view attacks as something that can be actively nullified, and explain it as such, then it can certainly feel more engaging to the players.

On could even include a < 5/10 > for additional effects. Maybe reducing damage, make a counter-attack, disarming (or various other combat maneuvers) for rolling that much better than your opponent. This should work as a built in part of the equally active offense, I would think.

[Quertus]

I think "hard to fudge" is a pro, as is "players have more to do".

Pretty much this. The one con I'll add... OK, fine, there's... gosh darned Spanish Inquisition, I keep coming up with more cons!

First off, not everyone comes to the table with 100 d20's or programs to automate all these rolls. IME, a good GM is prepared for the adventure, whereas the players don't know what the adventure is going to be. A good GM can be prepared to roll for 100 orcs throwing spears at the party or whatever; the players shouldn't be expected to be so prepared. So it's often (much) faster to just let a good GM roll for everything.

Then there's the consideration of cheaters. Gee, Bob, you got missed by everything again? How lucky for you. Shrug. I don't actually care about this one much, but some people might.

Then there's the really lucky and really unlucky players. One of my players, the one time I let them roll for enemy attacks, they nearly killed themselves in the opening salvo, whereas everyone else was barely bruised. Yeah, it really wasn't kind of me to let them touch the dice on the enemy attacks.

One last little dumb con is that everyone has to know the enemy stats in order to make the rolls. Which means that this isn't compatible with a game that involves hidden information (like, "the PCs shouldn't know that there's an Orc Shaman, who secretly cast Stilled Silent Prayer last round, boosting the attack bonus on the second round of attacks", for example).

EDIT: What's with "Makes combat feel more dangerous"? IME, the unknown usually feels more dangerous than the known, and this method is incompatible with hidden information.

So, the new lists look like this:

Pros:
Spreads out the cognitive load.
Players roll more dice.
Everyone sees the "randomness" so they can react to it without stopping to add color.
Makes combat feel more dangerous (???)
Players have more to do.
Hard to fudge.

Cons:
Not as common so could be confusing.
Players have less cause and ability to be prepared than GM.
Cheaters exist.
Some players are just unlucky.
Incompatible with hidden information.
Maybe makes combat feel less dangerous?

[gbaji]

One last little dumb con is that everyone has to know the enemy stats in order to make the rolls. Which means that this isn't compatible with a game that involves hidden information (like, "the PCs shouldn't know that there's an Orc Shaman, who secretly cast Stilled Silent Prayer last round, boosting the attack bonus on the second round of attacks", for example).?

This isn't necessary at all. The defender has a skill at say, parry, or block, or dodge, or whatever. They succeed in their skill and the result applies some adjustment to the opponents to-hit roll, or reduces damage, or whatever. No need to know anything about the other guy at all.

If we're assuming some sort of "appying adjustments back and forth, but ultimately resolving a D20 style hit/miss calculation", then it doesn't really matter whether the attacker applies their own effects after adjusting to the defenders adjustments, or before.

And if you want to add such things, the GM simply stating "make your <whatever> roll at -X, or against DC=Y" (or whatever the game system uses), is not terribly burdensome, while not informing the player of any more information than "my skill was harder to make against this opponent", whch is already something that exists with every other "roll against difficulty" calculation you might use anyway. And yes. I'd argue that if this opponent is harder to parry/block/whatever than the last one, the character (and thus the player) would/should actually notice this. In the same way you'd notice that this lock is harder to pick than the last one, or this door is harder to knock down than the last one, or well, any other sutation in a game where you're resolving a skill against variable difficulty.

[animorte]

No need to know anything about the other guy at all.

or well, any other sutation in a game where you're resolving a skill against variable difficulty.

Well said. Basically what I was going to follow up with, but much better.

AC constantly changes between enemies, just like their own saves and attack modifiers. There's really not much difference. It just has the illusion of creating more agency in the results as opposed rolling against or just watching the rolls against your own static DCs, which is not necessarily a bad thing.

[GloatingSwine]

Is it intuitive or is it just what is? If combat is occurring relatively simultaneously the attack <action> is merely a mechanical interface nested in combat that has its own mechanicals interface (turn/ rounds).

I think we frame attacking as more active is because it's an action in the meta.

Yes. "I roll when I do things, you roll when you do things" is absolutely the intuitive way to do it. It's an "action" in that it's the thing that produces a tangible outcome in the game.

I think "hard to fudge" is a pro, as is "players have more to do".

Players having more to do is fine if their characters are also doing more because of it. Just flipping the maths and having the player roll when their character doesn't do anything new as a result gives them busywork without connecting at-table actions to in-game actions.

Active defence needs to be more than "save vs. sword", there needs to be a tangible character outcome to success not the null outcome.

[MoiMagnus]

Looking at D&D, I don't like NPCs making saving throws against spell.

So what I prefer is that either you go with "attacking party is the one rolling dice" (like in D&D4e, or regular attacks in any D&D), or you go with the asymmetric "players are always the ones rolling dice" (so they roll for attack and defence while the monsters have fixed values for everything)

[Note that in the latter, I assume that you're also using average values for monster damage rolls and any other similar randomness]

[stoutstien]

Yes. "I roll when I do things, you roll when you do things" is absolutely the intuitive way to do it. It's an "action" in that it's the thing that produces a tangible outcome in the game.


Players having more to do is fine if their characters are also doing more because of it. Just flipping the maths and having the player roll when their character doesn't do anything new as a result gives them busywork without connecting at-table actions to in-game actions.

Active defence needs to be more than "save vs. sword", there needs to be a tangible character outcome to success not the null outcome.

It's only intuitive if you're framing that when you take the attack action it is resolved immediately. If turns are mostly occurring simultaneously the attack action is committing to attacking for that entire turn.
It's just as easily framed as making some sort of save or test against the enemy's attack and your attacks are making a save/test against the enemy's AC/DR.
No one ever complains when they have to make a save against an effect because most been conditioned to realize that they're different but practically are the same. You just add some different number.

Part of the reason why I'm so interested in this is it's a more tangible system for the players to interact with and in my opinion it's a lot smoother as well because I don't have to either double check against my cheat sheet for the players AC or ask them every time I make an attack with their AC is because they vary. So if an individual hero has a instant action where they can change that value there's no longer needs to be any exchange for it. They roll the dice see the dice and immediately decide what they want to do about it if they have the opposite available rather than 2-3 quips back and forth. A lot of games have grown accustomed to this back and forth as just a part of the game I just don't think it's necessary.

This entire idea came up to me because my wife asked me to make her a game that uses a single d6 for everything and the players and GM never roll that die on the same turn. It was a fun little puzzle and it ended up to be a nice little system but it got me thinking on expanding it. Could I make combat deeper by making it half as complex?

[Asmotherion]

I'm into active defensive rolls to the point of making them one of the selling points of my own system.

I like the idea that all conflict is resolved by opposed rolls.

[stoutstien]

I'm into active defensive rolls to the point of making them one of the selling points of my own system.

I like the idea that all conflict is resolved by opposed rolls.

I'm not opposed to the concept of opposing rolls but I've never seen it done well. Either it's horribly and balanced one way or the other or it is a net gain of nothing with a lot more maths. I do have an exception for solo RPGs. I think it's one of the best mechanics you can use for that design space.

I've experimented with dramatic DCs when checks are made. It kind of turns everything into an opposed roll but it never got past theoretical testing. I think it's doable for games that use dice pools because the curve is going to be more predictable.

I just don't like dice pools for anything that isn't cumulative and universal. Like I like pools that represent things like time, urgency, or threat because everyone can see and interpret a single pool rather have a bunch of individual ones.

[LibraryOgre]

I'm not opposed to the concept of opposing rolls but I've never seen it done well. Either it's horribly and balanced one way or the other or it is a net gain of nothing with a lot more maths. I do have an exception for solo RPGs. I think it's one of the best mechanics you can use for that design space.

One thing I find Hackmaster does well for its opposed combat rolls is to vary the defense die rolled.

For example, if you're "two weapons, attacking with both", your defense die is 1d10p; because of penetration (exploding), you can get a high defense, but if you're swinging two swords, you're going to be worse at defense than someone who is using the second weapon for defense. For a lot of weapons, your defense die after the first attack is d20p-4. If you're using a shield and the opponent misses by less than 10, they hit your shield, and you take less damage from the attack (because your shield's damage reduction is included, and the weapon damage is lowered on a shield hit).

Additionally, success isn't binary; there are shades of success. A 20 that beats the defender's defense (when bonuses and everything are considered) is a critical... but a 20 on defense is a "Perfect Defense", allowing you a free attack, and a 19 on defense is a "Near-Perfect Defense", which allows an unarmed attack on the opponent (and that ignores armor). If you're using two weapons, a near-perfect defense can be an off-hand weapon strike. Thieves are the masters of the NPD, getting it on an 18 or 19, and doing more damage with one.

It's a complex system, though it flows well once you've learned it.

[stoutstien]

One thing I find Hackmaster does well for its opposed combat rolls is to vary the defense die rolled.

For example, if you're "two weapons, attacking with both", your defense die is 1d10p; because of penetration (exploding), you can get a high defense, but if you're swinging two swords, you're going to be worse at defense than someone who is using the second weapon for defense. For a lot of weapons, your defense die after the first attack is d20p-4. If you're using a shield and the opponent misses by less than 10, they hit your shield, and you take less damage from the attack (because your shield's damage reduction is included, and the weapon damage is lowered on a shield hit).

Additionally, success isn't binary; there are shades of success. A 20 that beats the defender's defense (when bonuses and everything are considered) is a critical... but a 20 on defense is a "Perfect Defense", allowing you a free attack, and a 19 on defense is a "Near-Perfect Defense", which allows an unarmed attack on the opponent (and that ignores armor). If you're using two weapons, a near-perfect defense can be an off-hand weapon strike. Thieves are the masters of the NPD, getting it on an 18 or 19, and doing more damage with one.

It's a complex system, though it flows well once you've learned it.

I have a copy of 4 and both 5E hack masters on my shelf and I respect them but not for me at all. I still recommend it to people because it's actually a great system for people who want that crunch without going the full homogenized route.

Also have a soft spot for well done parodies.