Natural 1's, Missing by 5 or less, and other similar effects

[Darksol the Painbringer]

This is a FAQ thread in relation to how Natural 1's function in accordance with effects that trigger based on missing by a numerical value (or fewer/more); the original discussion started from this thread here, and because I noticed that there were several abilities and effects that had a similar clause to the original topic, I felt that a catch-all FAQ would be the smarter course of action.

To expand upon this originally specific topic, we have several examples where this mechanic is referenced in some shape or form:

-Fighting Defensively via Crane Wing Style (if an attack from an enemy misses by 4 or less)
-Upsetting Strike feat via Upsetting Shield Style (if an attack from an enemy misses by 5 or less)
-Mirror Image (the spell) (if an attack from an enemy misses by 5 or less)
-Failed Trip/Disarm attempts (if a combat maneuver check fails by 10 or more)

And multiple others.

The relevant Combat Rules from the PRD have this to say:

Attack Roll wrote:

An attack roll represents your attempt to strike your opponent on your turn in a round. When you make an attack roll, you roll a d20 and add your attack bonus. (Other modifiers may also apply to this roll.) If your result equals or beats the target's Armor Class, you hit and deal damage.

Automatic Misses and Hits: A natural 1 (the d20 comes up 1) on an attack roll is always a miss. A natural 20 (the d20 comes up 20) is always a hit. A natural 20 is also a threat—a possible critical hit (see the attack action).

Emphasis mine.

Text that mentions "missing by 5 or less" and other similar clauses can be found amongst the effects I've described, but the most common appearance is with Mirror Image from the PRD, and it has this description:

Mirror Image wrote:

This spell creates a number of illusory doubles of you that inhabit your square. These doubles make it difficult for enemies to precisely locate and attack you.

When mirror image is cast, 1d4 images plus one image per three caster levels (maximum eight images total) are created. These images remain in your space and move with you, mimicking your movements, sounds, and actions exactly. Whenever you are attacked or are the target of a spell that requires an attack roll, there is a possibility that the attack targets one of your images instead. If the attack is a hit, roll randomly to see whether the selected target is real or a figment. If it is a figment, the figment is destroyed. If the attack misses by 5 or less, one of your figments is destroyed by the near miss. Area spells affect you normally and do not destroy any of your figments. Spells and effects that do not require an attack roll affect you normally and do not destroy any of your figments. Spells that require a touch attack are harmlessly discharged if used to destroy a figment.

An attacker must be able to see the figments to be fooled. If you are invisible or the attacker is blind, the spell has no effect (although the normal miss chances still apply).

Emphasized relevant portions.

The FAQ question is: Does a Natural 1 interact with effects that trigger on a miss "by 5 or less"? In other words, do you still calculate the total result and compare it to the target's AC and determine if it triggers the effect, or does the fact that it's a Natural 1 mean that you don't follow the "miss by 5 or less" trigger?

Personally, I feel that this is something that can come up occasionally in mid-level games, between Crane Style and Mirror Images being the biggest culprits.

This thread will also serve as discussion thread until the Pathfinder Design Team decides to deliver us a FAQ (if at all). If you want an official answer for this, I suggest you hit the FAQ button on this post, and hopefully we'll get an answer within a reasonable time frame!

[Diego Rossi]

Let's keep this on the front page. :)

There are enough situations where this matter to make it a worthwhile question.

There is the opposite question too:
if you hit with a natural 20, by how much you have beaten the target AC?
The example used in that thread was Bull rush, but probably there are other situations where your margin of success matter.

[dragonhunterq]

Diego Rossi wrote:

if you hit with a natural 20, by how much you have beaten the target AC?
The example used in that thread was Bull rush, but probably there are other situations where your margin of success matter.

This one is easy. 20+CMB -CMD. There are no rules or even implications within the rules that rolling a 20 increases the effective result of a CMB. The rules for CMB tells us clearly that:

CRB, combat manoeuvres wrote:

Rolling a natural 20 while attempting a combat maneuver is always a success (except when attempting to escape from bonds), while rolling a natural 1 is always a failure.

I am not aware of anything else that would even hint at a contrary answer.

[Talonhawke]

dragonhunterq wrote:

Diego Rossi wrote:

if you hit with a natural 20, by how much you have beaten the target AC?
The example used in that thread was Bull rush, but probably there are other situations where your margin of success matter.

This one is easy. 20+CMB -CMD. There are no rules or even implications within the rules that rolling a 20 increases the effective result of a CMB. The rules for CMB tells us clearly that:

CRB, combat manoeuvres wrote:

Rolling a natural 20 while attempting a combat maneuver is always a success (except when attempting to escape from bonds), while rolling a natural 1 is always a failure.

I am not aware of anything else that would even hint at a contrary answer.

The issues with the other thread is it's being stated that a nat.1 or 20 is never compared to the targets AC/CMB and there for doesn't trigger such effects.

[dragonhunterq]

Talonhawke wrote:

dragonhunterq wrote:

Diego Rossi wrote:

if you hit with a natural 20, by how much you have beaten the target AC?
The example used in that thread was Bull rush, but probably there are other situations where your margin of success matter.

This one is easy. 20+CMB -CMD. There are no rules or even implications within the rules that rolling a 20 increases the effective result of a CMB. The rules for CMB tells us clearly that:

CRB, combat manoeuvres wrote:

Rolling a natural 20 while attempting a combat maneuver is always a success (except when attempting to escape from bonds), while rolling a natural 1 is always a failure.

I am not aware of anything else that would even hint at a contrary answer.

The issues with the other thread is it's being stated that a nat.1 or 20 is never compared to the targets AC/CMB and there for doesn't trigger such effects.

Seriously?!

um, citation then I guess...

It will be interesting to see if 'they' can come up with anything that can be taken seriously...

[Talonhawke]

I know right, their whole arugment seems to hinge on the fact that since no amount of raising your total to hit could hit with a roll of a 1 then it can't trigger.

[Darksol the Painbringer]

I agree with your guys' sentiment, but I would appreciate someone from the other side of the argument to come in and give a precise explanation.

It gives the Devs some food for thought, and maybe it can help us at least understand where they're coming from.

[The Black Bard]

Haven't read the other thread, so I guess I'm coming in as the quasi-impartial observer. The rules for hits/manuevers states the 20 and 1 are always a hit/success or miss/failure. There is no other text implying a bonus or penalty, like some of the +20/-20 houserules out there.

So to me, if I have a +3 bonus when I bull rush someone, and I get a nat 20, I have both gotten a result of 23, and gotten an auto-success. So, if my target has a CMD of 16, I beat him by 5. If he has a CMD of 19, I beat him, but not by 5. If his CMD is 25, I still beat him, but obviously not by 5. For counter-manuvers like trip/disarm, the manuver was a success, so even if I got a nat 20 result of 21 against a CMD 32, there would be not be a counter.

In reverse, if I have a bonus of +20 and roll a 1, the auto-fail overrides whether his CMD is anything 21 or less. And since the check is automatically a failure, then anything that would meet the conditions for counter-maneuvers (10 or more on failure) would apply. Same for regular attack rolls.

To sum up, as posted in the starting post:

When you make an attack roll, you roll a d20 and add your attack bonus. (Other modifiers may also apply to this roll.)

Absolutely nothing in the following sentences precludes/alters/or invalidates that statement. In fact, the literal order of operations in the rule tells us the answer:
1. Roll a dice.
2. Add your bonus (with modifiers).
3. Check if it equals or beats, if so you hit.
3a. A Nat 20 always hits.
3b. A Nat 1 always misses.

I would be honestly surprised if this is officially ruled as any other interpretation. Won't be surprised if this gets the "working as intended" treatment.

[Talonhawke]

I hope for more than "Working as intended" since that just leaves both sides assuming they are right.

[Enchanter Tim]

I think there's an implicit thought that a natural roll of a 1 should never benefit the attacker. Likewise, a natural roll of 20 should never hinder the attacker. The idea that a natural 1 would help by destroying an image is antithetical to these ideals.

[dragonhunterq]

Enchanter Tim wrote:

I think there's an implicit thought that a natural roll of a 1 should never benefit the attacker. Likewise, a natural roll of 20 should never hinder the attacker. The idea that a natural 1 would help by destroying an image is antithetical to these ideals.

Oh! I was hoping for, you know, an actual rule or something rather than "it feels wrong".

[Enchanter Tim]

The closest I can think of is the rule for damage being negated. If your damage is reduced to 0, no other effects apply even if you "hit."

[Talonhawke]

Enchanter Tim wrote:

I think there's an implicit thought that a natural roll of a 1 should never benefit the attacker. Likewise, a natural roll of 20 should never hinder the attacker. The idea that a natural 1 would help by destroying an image is antithetical to these ideals.

But in the case of certain feats and weapon properties rolling a 1 in this case would be more beneficial than a 2 since you would avoid AoOs and such by not checking the AC.

[Forseti]

Enchanter Tim wrote:

I think there's an implicit thought that a natural roll of a 1 should never benefit the attacker. Likewise, a natural roll of 20 should never hinder the attacker. The idea that a natural 1 would help by destroying an image is antithetical to these ideals.

There's not always a straightforward way to determine what's a benefit to the attacker.

What about fighting a target under the effect of Shocking Image? What would you call a benefit here?

Do rolls of '1' vs. Shocking Image destroy an image when avoiding the damage would be more beneficial than bringing the number of images down? Does my barbarian destroy an image when he's at 1 hp, and not destroy an image when he's at his full 280 hp?

[Ravingdork]

F@Q'd.

[Diego Rossi]

dragonhunterq wrote:

Talonhawke wrote:

dragonhunterq wrote:

Diego Rossi wrote:

if you hit with a natural 20, by how much you have beaten the target AC?
The example used in that thread was Bull rush, but probably there are other situations where your margin of success matter.

This one is easy. 20+CMB -CMD. There are no rules or even implications within the rules that rolling a 20 increases the effective result of a CMB. The rules for CMB tells us clearly that:

CRB, combat manoeuvres wrote:

Rolling a natural 20 while attempting a combat maneuver is always a success (except when attempting to escape from bonds), while rolling a natural 1 is always a failure.

I am not aware of anything else that would even hint at a contrary answer.

The issues with the other thread is it's being stated that a nat.1 or 20 is never compared to the targets AC/CMB and there for doesn't trigger such effects.

Seriously?!

um, citation then I guess...

It will be interesting to see if 'they' can come up with anything that can be taken seriously...

The basic argument in favor of "missing with a natural 1 is similar to missing by infinitely large number" is that you can add as much as you want to the attack after rolling the 1 and you will sill miss.

Same thing for the natural 20. You can add as much as you want to your defense after the attacker has rolled a 20, and you will still be hit.

The he counter was "then with a nat 20 in a bull rush you have beaten the target CMD by an infinitely large number".

[Talonhawke]

What's funny I would be willing to put money down that if the first question asked had been about Jarring armor and not mirror image the question would have been an overwhelming yes you have to make the save vs sickened. But since the original question was a benefit to the attacker people started coming up with weird logic to make it work the way they wanted.

[Gallant Armor]

Talonhawke wrote:

What's funny I would be willing to put money down that if the first question asked had been about Jarring armor and not mirror image the question would have been an overwhelming yes you have to make the save vs sickened. But since the original question was a benefit to the attacker people started coming up with weird logic to make it work the way they wanted.

The argument worded consistently is along the lines of: when an attacker rolls a natural one it is treated as a miss or roll + total attack bonus, whichever is worse for every effect. When an attacker rolls a natural 20 it is considered a hit or roll + total attack bonus, whichever is better for every effect.

[Talonhawke]

So like asked before what about Shocking Image? I mean seriously is it well 2d6 might drop you so yep it pops an image but nah you have 180 hp nothing happens to the image?

Your making it way more than it actually is nothing about a Nat. 1 suggest its some sort of cosmic botch. It's there so that there is a chance of failure for attacks and skills regardless of actual bonuses not to be Schrodinger's dice roll where we have to check what the outcome of both sides of the roll would be vs all situations to figure out what the worse thing would be. It's real simply it missed it can still pop an image it can still trigger upsetting strike it can force a fort save vs jarring armor it can end your crane wing bonus all depending on modifiers what can it not do "hit".

[Glorf Fei-Hung]

The fact that you miss on a nat 1, or hit on a nat 20 do not change the values of the attacks that you rolled. If you are attacking a mirror image and roll at nat 1 and your attack bonus is high enough that a nat 1 still means missing by 5 or less than you have successfully met the conditions to destroy an image, thus an image is destroyed. By the same token if you rolled a nat 20 and get the benefit of an automatic hit you still have to roll to see if that nat 20, best possible hit you could do, still hit an image or the real creature. From a balance perspective you can't say the 1 has no chance to even destroy an image while the 20 still has to find out if it hit an image rather than the target.

The same applies for exceeding a value, if your CMB is not high enough to exceed a targets CMD by 10 or more a nat 20 does not make 20+10(CMB) exceed 35 by 10+, it just means you automatically succeed at the Combat Maneuver even if you wouldn't have normally.

[Gallant Armor]

Talonhawke wrote:

So like asked before what about Shocking Image? I mean seriously is it well 2d6 might drop you so yep it pops an image but nah you have 180 hp nothing happens to the image?

Your making it way more than it actually is nothing about a Nat. 1 suggest its some sort of cosmic botch. It's there so that there is a chance of failure for attacks and skills regardless of actual bonuses not to be Schrodinger's dice roll where we have to check what the outcome of both sides of the roll would be vs all situations to figure out what the worse thing would be. It's real simply it missed it can still pop an image it can still trigger upsetting strike it can force a fort save vs jarring armor it can end your crane wing bonus all depending on modifiers what can it not do "hit".

Shocking image is the only case I've seen where there is a mix of positive and negative that could cause issues. There are issues with your method as well:

Disarm wrote:

If your attack is successful, your target drops one item it is carrying of your choice (even if the item is wielded with two hands). If your attack exceeds the CMD of the target by 10 or more, the target drops the items it is carrying in both hands (maximum two items if the target has more than two hands).

For disarming, if you roll a nat 1 you fail, but if you exceed their CMD by 10 they drop things from both hands?

[shaventalz]

Gallant Armor wrote:

Disarm wrote:

If your attack is successful, your target drops one item it is carrying of your choice (even if the item is wielded with two hands). If your attack exceeds the CMD of the target by 10 or more, the target drops the items it is carrying in both hands (maximum two items if the target has more than two hands).

For disarming, if you roll a nat 1 you fail, but if you exceed their CMD by 10 they drop things from both hands?

Personally, I'd tend to read the succeed-by-10 clause as being part of the "success" section. The fail-by-10 clause following it specifies failure, so there's less of an issue on that end of the roll.

[ryric]

I'm in the camp that you still generate a total on a 1 or 20 and compare it to the target number for effects that trigger off the target number. You just append "miss" or "hit" onto the effects as well.

Natural 1 is an automatic miss...it does not negate any other effects of the attack. Don't add extra stuff that isn't there. If you are so good at fighting that your nat 1 misses by less than 5, you pop an image, for example.

[DM Livgin]

Using the result of the 1/20 plus bonuses to determine secondary results defiantly keeps things clean and easy. This reminds me of a less extreme version of the crit deck problem, critical hits/misses are common enough that having ultimatum effects make it less fun for the players.

[skizzerz]

You are told it is an auto hit/miss but are not told to further adjust what you rolled. As such, I believe that by RAW you would still do the math. For example, if you had +20 to hit against AC 17 and rolled a 1, you miss by 17-21=-4. As -4 is less than 5, any condition that triggers on a miss of "5 or less" will apply. On the flip side, if you only had +10 against AC 17 you miss by 17-11=6, which is not "5 or less".

Same thing for auto hits, do the math, see if the actual number meets the threshold.

[Ascalaphus]

I'm with "math as normal". It works for each of the situations named. It doesn't pulling arbitrary numbers out of the sky to see how far you bullrush someone. It doesn't care whose advantage it is to pop or not pop a shocking image.

[KingOfAnything]

I could see treating 1s and 20s differently. Do the math for a 20, but treat a 1 on the die as a 1 total and calculate from there.

[Ascalaphus]

KingOfAnything wrote:

I could see treating 1s and 20s differently. Do the math for a 20, but treat a 1 on the die as a 1 total and calculate from there.

Why? Why complicate things further?

[Talonhawke]

Ascalaphus wrote:

KingOfAnything wrote:

I could see treating 1s and 20s differently. Do the math for a 20, but treat a 1 on the die as a 1 total and calculate from there.

Why? Why complicate things further?

Clearly ones need to be a fumble of some sort can't have people still being competent when they hit their forced 5% miss.

[Orfamay Quest]

KingOfAnything wrote:

I could see treating 1s and 20s differently. Do the math for a 20, but treat a 1 on the die as a 1 total and calculate from there.

Shrug. I don't see how this is anything other than an arbitrary screw-the-martials move.

"Math works as normal, except when I don't want it to because I think it will be hilarious to watch the fighter flail around."

[KingOfAnything]

I'm looking at Crane Wing's 4 or less clause. In the corner case that an enemy hits exactly on a 2, rolling a 1 makes it easier to hit the monk with their iterative attacks. That doesn't make much sense to me.

I agree that an internally consistent rule is more desirable. We should treat 1s the same no matter if a near miss provides benefit or detriment.

I'm convinced that "never check" isn't the correct method, but I do think that a 1 should always be a total miss, not a near miss.

I don't think there is a particularly strong reason to treat 1s and 20s the same. They are both automatic, but in different ways.

That's my reasoning for a system where a nat 1 is treated as a 1 that misses.

Quote:

"Math works as normal, except when I don't want it to because I think it will be hilarious to watch the fighter flail around."

Wizards make the occasional attack roll, too. Fighters may roll more attacks, but that means they get more 20s as well as more 1s.

Ultimately, I think automatic hits and misses increase the variability of the system. Even a mouse can bite the hand of gods with a little luck, and even the gods can completely miss the mouse.

[Forseti]

KingOfAnything wrote:

I'm convinced that "never check" isn't the correct method, but I do think that a 1 should always be a total miss, not a near miss.

Why shouldn't a '1' be a near miss? If the numbers are such that doing the calculations is relevant at all, you're looking at an attacker that can easily hit the defender.

If I attack your AC of 10 with my AB of 40 and I roll a '1', you better believe it's a near miss. I score a hit with 32 in excess of your AC if I roll a '2', but a roll of '1' is so terrible an attack that it can't even be considered a near miss? If that's the reasoning, what do I have to roll to have a near miss?

Am I doomed to either score resoundingly spectacular hits or laughably clumsy misses, just because I have a very high AB? I don't think so.

[KingOfAnything]

Forseti wrote:

Am I doomed to either score resoundingly spectacular hits or laughably clumsy misses, just because I have a very high AB? I don't think so.

You say laughably clumsy, the wizard says astoundingly lucky.

[Forseti]

But what would I need to do to get near misses? Surely I should be able to get those?

[KingOfAnything]

Forseti wrote:

But what would I need to do to get near misses? Surely I should be able to get those?

If you'd rather have a near miss than a hit, be my guest.

[Forseti]

If I have to miss at all, which will happen to everyone, I sure would like some of those misses to count as near misses. So I can pop mirror images, for example. Why should I be denied the chance to have near misses?

[Samish Lakefinder]

It is easy to tell if you missed by five or less. Add five to your result and see if you hit. If you still missed then you did not miss by five or less, simple math. If you roll a natural 1 then add your attack bonus with an additional 5, you still get a miss, so you did not miss by 5 or less. If you ask what you need to change to get a hit you find that you missed by the rule saying natural 1s miss and that you did not miss by any number because no number added to your result causes you to hit.

[Fuzzy-Wuzzy]

Samish Lakefinder wrote:

It is easy to tell if you missed by five or less. Add five to your result and see if you hit. If you still missed then you did not miss by five or less, simple math. If you roll a natural 1 then add your attack bonus with an additional 5, you still get a miss, so you did not miss by 5 or less. If you ask what you need to change to get a hit you find that you missed by the rule saying natural 1s miss and that you did not miss by any number because no number added to your result causes you to hit.

When I hit, how do I tell whether I hit by five or more? Does this apply to natural 20s?

[Talonhawke]

Samish Lakefinder wrote:

It is easy to tell if you missed by five or less. Add five to your result and see if you hit. If you still missed then you did not miss by five or less, simple math. If you roll a natural 1 then add your attack bonus with an additional 5, you still get a miss, so you did not miss by 5 or less. If you ask what you need to change to get a hit you find that you missed by the rule saying natural 1s miss and that you did not miss by any number because no number added to your result causes you to hit.

So does a Nat. 20 on bullrush mean no extra 5ft or infinite 5ft?

[skizzerz]

Talonhawke wrote:

Samish Lakefinder wrote:

It is easy to tell if you missed by five or less. Add five to your result and see if you hit. If you still missed then you did not miss by five or less, simple math. If you roll a natural 1 then add your attack bonus with an additional 5, you still get a miss, so you did not miss by 5 or less. If you ask what you need to change to get a hit you find that you missed by the rule saying natural 1s miss and that you did not miss by any number because no number added to your result causes you to hit.

So does a Nat. 20 on bullrush mean no extra 5ft or infinite 5ft?

By that logic, infinite. I rolled a 20. I subtract 5 and still hit due to nat 20, so +5ft. Subtract another 5, still hit due to nat 20, +5ft. Repeat ad infinitum.

I'm curious as to where people are seeing the rule that normal math no longer applies to secondary effects when a nat 1/20 is involved. I certainly don't see one.

Nat 1 causes me to miss, nat 20 causes me to hit. This may mean that I miss or hit by a negative number when I do the math. Negative numbers are less than 5 (less than all positive numbers, really), so any effect that happens when I miss or hit by 5 or less will apply. Any effect where I need to miss or hit by 5 or more will not apply should I miss or hit by a negative number.

If my bonus was such that a 1 or 20 was a normal miss or hit (I missed by 1 or more, or hit by 0 or more), then I still do the math normally and compare the result, just as if I rolled a 2 or 19 with an extra -1/+1 bonus tacked on respectively.

[DmRrostarr]

I can see this giving more credibility to using the 1 counts as a -10 and a 20 counts as a 30 optional rule in situations like this.

[dragonhunterq]

DmRrostarr wrote:

I can see this giving more credibility to using the 1 counts as a -10 and a 20 counts as a 30 optional rule in situations like this.

In what way does some flaky half baked 'I don't like how this works' reading that has no actual rules support lend any credibility to anything?

[Talonhawke]

dragonhunterq wrote:

DmRrostarr wrote:

I can see this giving more credibility to using the 1 counts as a -10 and a 20 counts as a 30 optional rule in situations like this.

In what way does some flaky half baked 'I don't like how this works' reading that has no actual rules support lend any credibility to anything?

And even then you would still have people flipping out that you popped and image even with a roll of 1 and -10 on your total if it was still high enough.

[Samish Lakefinder]

The bullrushing people for infinite distance is the same misunderstanding of math as dividing by 0. As you approach dividing by zero you get bigger and bigger numbers from the positive side and larger negative numbers from the negative side, but dividing by zero is actually undefined not infinite. People are looking at the numbers approaching a nat 20 and seeing that you are generally pushing things further and further, but when you get to 20 you switch to a different rule of "A natural 20 (the d20 comes up 20) is always a hit".

[Fuzzy-Wuzzy]

Samish, you still haven't answered this. Please do so without chatting about what is and isn't infinite, because I haven't asked about infinity. Just "how do I tell whether I hit by five or more? and how do I tell that when I roll a natural 20?"

Fuzzy-Wuzzy wrote:

Samish Lakefinder wrote:

It is easy to tell if you missed by five or less. Add five to your result and see if you hit. If you still missed then you did not miss by five or less, simple math. If you roll a natural 1 then add your attack bonus with an additional 5, you still get a miss, so you did not miss by 5 or less. If you ask what you need to change to get a hit you find that you missed by the rule saying natural 1s miss and that you did not miss by any number because no number added to your result causes you to hit.

When I hit, how do I tell whether I hit by five or more? Does this apply to natural 20s?

[Quantum Steve]

This is really something the devs need to weigh in on because the answer isn't in the RAW, so the only thing left is GM discretion.

It isn't even necessarily useful to look at nat 20's because there doesn't have to be parity between the two. 20's can work one way and 1's work another because they are different things.

[dragonhunterq]

Quantum Steve wrote:

This is really something the devs need to weigh in on because the answer isn't in the RAW, so the only thing left is GM discretion.

Of course it is. Add your attack bonus to the 1 and if you meet the triggering condition, the ability triggers.

Any ruling to the contrary has absolutely no basis in the rules at all. All efforts to elicit any support for any other position has been met with what amounts to "I don't like it because 1's should be bad" or logic that makes Fuzzy-Wuzzy look silky smooth, it's so woolly.

[With apologies to Fuzzy-Wuzzy - I couldn't resist...]

[Samish Lakefinder]

I have said several time that you do not hit by any number you are hitting by a rule.

Bull Rush-
If your attack is successful, your target is pushed back 5 feet. For every 5 by which your attack exceeds your opponent's CMD you can push the target back an additional 5 feet.

Lets check the first sentence: Are we successful.
"A natural 20 (the d20 comes up 20) is always a hit" = yes
Effect pushed back 5 feet.

Second Sentence: For every 5 by which your attack exceeds your
opponent's CMD you can push the target back an additional 5
feet.

The general rule for attacks says "When you make an attack
roll, you roll a d20 and add your attack bonus. (Other
modifiers may also apply to this roll.) If your result equals
or beats the target's Armor Class, you hit and deal damage."
The specific rule for Natural 20s says "A natural 20 (the d20
comes up 20) is always a hit.". Notice you do not add attack
bonuses to a natural 20 or compare it to the opponents AC.

Since we do not compare a natural 20 to the opponents CMD we can not exceed it.

Result: Pushed back 5 feet.

[Samish Lakefinder]

dragonhunterq said wrote:

Any ruling to the contrary has absolutely no basis in the rules at all

General rule for attacks "When you make an attack

roll, you roll a d20 and add your attack bonus. (Other
modifiers may also apply to this roll.) If your result equals or beats the target's Armor Class, you hit and deal damage."

Specific rule for Natural 1s "A natural 1 (the d20 comes up 1) on an attack roll is always a miss"

There is no mention of adding any bonuses to the natural 1 or comparing it to the AC.