The DM Rolls

[Jason]

The Angry GM did a piece on the Insight skill in D&D and had quite a bit to say on the virtues of the Passive Perception mechanic. I think it's worth a read: theangrygm.com/insight-into-insight
The article on social interaction that spawned this article is a good one too.

[Democratus]

I like a lot of the advice on Angry GM.

In particular, his articles on using 4th edition monsters to create inventive 5th edition foes is fantastic.

Looking forward to reading the article you linked. Thanks!

[Quertus]

The Angry GM did a piece on the Insight skill in D&D and had quite a bit to sayn on the virtues of the Passive Perception mechanic. I think it's worth a read: theangrygm.com/insight-into-insight
The article on social interaction that spawned this article is a good one too.

Wow.

You know, usually, I look at the things Angry writes, and conclude two things: Angry is the absolute *best* at asking the right questions, and *horrible* at answering them. He has amazing, unparalleled insight into the questions to ask… and then always gives the wrong answer.

For once, reading through one of his articles, I am left perplexed, because my passive perception didn't flag anything as wrong. I may need to read more carefully, but there was nothing in his analysis that I strongly disagree with.

More on topic, I think that utilizing passive perception can be fine… but it doesn't model reality, or all of the stories that I want to tell. But I agree with Angry that it solves a lot of other problems, and that those are problems that ought to be solved.

[Darth Credence]

Never heard of, let alone read, anything by the Angry GM before. After reading that, I will do my best to forget I have now heard of him. That was nothing more than an example of covering up an inability to write coherently with "shocking" language and insult comedy. Why anyone would attempt to wade through that mess is beyond me.

[Jason]

More on topic, I think that utilizing passive perception can be fine… but it doesn't model reality, or all of the stories that I want to tell. But I agree with Angry that it solves a lot of other problems, and that those are problems that ought to be solved.

How does it not model reality?

He's not saying "don't let your players ever make active Perception checks if they suspect something is present". He's saying "give them the clues they notice through passive perception and then let them decide where to look or what to do next."

Edit:

Never heard of, let alone read, anything by the Angry GM before. After reading that, I will do my best to forget I have now heard of him. That was nothing more than an example of covering up an inability to write coherently with "shocking" language and insult comedy. Why anyone would attempt to wade through that mess is beyond me.

Because behind the shtick is usually some very good advice. This is one of his "angrier" pieces. In fact he's taking a break right now because people said he was laying it on a little too thick in this specific article. I find the shtick a little off-putting too.

[False God]

I think as long as there is trust between the DM and players, this is a fine approach. I also find it works better for more narrative play, and helps eliminate "rolling for everything" approaches in favor of "rolling when necessary", which makes the rolls themselves more valuable and success, failure or "other" more meaningful. It is also an effective play method in simpler systems (I'd argue 5E is even too complex).

Personally, I am still of the 4E mindset that aggressors should roll attacks, and defenders should have static defenses. "Random defense" for saves makes no sense to me in a system with fixed armor class that represents a combination of reflexes, physical and supernatural defenses.

Alternatively, I quite enjoy opposed rolls in dice-pool-based systems far more than rolled attacks vs. static defenses. But I find that pool-systems result in much more reasonable levels of success and failure, while a single d20 with modifiers and a simple pass/fail system often trends towards extremes.

And I find Angry's articles insufferable.

[Quertus]

Never heard of, let alone read, anything by the Angry GM before. After reading that, I will do my best to forget I have now heard of him. That was nothing more than an example of covering up an inability to write coherently with "shocking" language and insult comedy. Why anyone would attempt to wade through that mess is beyond me.

I'll not deny that his writing style is… not for everyone. But his *content* is good. No, not just good - I have never met his equal at *asking* questions, at understanding how to break down RPGs. And I've rarely meet anyone as bad as Angry at answering those same questions. So, win some, lose some.

How does it not model reality?

He's not saying "don't let your players ever make active Perception checks if they suspect something is present". He's saying "give them the clues they notice through passive perception and then let them decide where to look or what to do next."

Happily (despite reading comprehension not being my strong suit), I got what he was saying. And I still don't disagree with it, per se (although, now that I've read it, I do disagree with a few things from the spawning article).

No, my issue is that "passive perception" does not model (my experiences with) passive perception in this world.

Were I to attempt to model reality into 3e, "passive perception" would be a die roll; active perception would be "take 20". Maybe.

That is, if you have 5 equally hidden things *that no-one is looking for*, hidden well enough that, say, in a class of 20-30, it would be really odd for anyone to spot more than one, they are not all going to spot *the same one*.

That probably didn't make enough sense. Let me try again.

We're I too misspell 5 works were autocorrect to fail me 5 times, and everyone noticed the same misspelling / error, that pretty solidly indicates that it has the lowest DC. However, IME, there have been *many* casual observations where the single thing noticed, or the set of things noticed, did not match up with the "lowest DC gets spotted" theory of casual perception. And it's much more telling for those who only spot a single thing, because presumably they might be using active perception after that.

Or, at least, that's my take on it.

But, if you ignore that, then, yeah, Angry's advice is uncharacteristically spot-on!

[Jason]

Were I to attempt to model reality into 3e, "passive perception" would be a die roll; active perception would be "take 20". Maybe.

That is, if you have 5 equally hidden things *that no-one is looking for*, hidden well enough that, say, in a class of 20-30, it would be really odd for anyone to spot more than one, they are not all going to spot *the same one*.

That probably didn't make enough sense. Let me try again.

We're I too misspell 5 works were autocorrect to fail me 5 times, and everyone noticed the same misspelling / error, that pretty solidly indicates that it has the lowest DC. However, IME, there have been *many* casual observations where the single thing noticed, or the set of things noticed, did not match up with the "lowest DC gets spotted" theory of casual perception. And it's much more telling for those who only spot a single thing, because presumably they might be using active perception after that.

Or, at least, that's my take on it.

Hmm. I'm not sure I understand you. Are you saying that in a room with 5 hidden objects, someone who has a high passive perception check shouldn't spot all of the hidden objects, despite having a passive perception score high enough to beat the DC of all 5 hidden objects in the room?

[Democratus]

Hmm. I'm not sure I understand you. Are you saying that in a room with 5 hidden objects, someone who has a high passive perception check shouldn't spot all of the hidden objects, despite having a passive perception score high enough to beat the DC of all 5 hidden objects in the room?

Seems a better approach is to give them clues that there are multiple hidden objects in the room. Then let them investigate further to get to all of them.

It's more engaging than "score high enough...all things found" and allows the party to interact with the environment. A win-win.

[Quertus]

Hmm. I'm not sure I understand you. Are you saying that in a room with 5 hidden objects, someone who has a high passive perception check shouldn't spot all of the hidden objects, despite having a passive perception score high enough to beat the DC of all 5 hidden objects in the room?

Something like that, yes.

Have your "passive" skills never had you miss really obvious typos / stains / jokes / whatever? It just feels really odd, the kids (and adults!) who will go to work/school and not notice obvious things, like that their clothes are inside out, or the obvious mistakes in their work, that it makes it difficult to believe in passive perception as "taking a 10".

It also doesn't model those rare times when particularly skilled people lose out on happening to notice something (when none of them were actively looking).

Surely you've seen things that at least *looked like* what I'm describing.

Now, one might well say that I'm wrong, and explain those outcomes via different underlying mechanics. OK, but would we model those mechanics in such an RPG?

Still, as I said, if, unlike me, you are happy with passive skills (passive skill checks? Passive skill use? Whatever), then Angry's article is spot-on. I happen to like to play "stories" where, in 3e parlance, the less skilled characters can *sometimes* win… until the more skilled character just becomes *that good*. That transition itself is important, IMO.

Seems a better approach is to give them clues that there are multiple hidden objects in the room. Then let them investigate further to get to all of them.

It's more engaging than "score high enough...all things found" and allows the party to interact with the environment. A win-win.

Given that I was taking simulation… unless you hear voices giving you clues as to how many things to look for…

[Jason]

Have your "passive" skills never had you miss really obvious typos / stains / jokes / whatever? It just feels really odd, the kids (and adults!) who will go to work/school and not notice obvious things, like that their clothes are inside out, or the obvious mistakes in their work, that it makes it difficult to believe in passive perception as "taking a 10".

It also doesn't model those rare times when particularly skilled people lose out on happening to notice something (when none of them were actively looking).

Surely you've seen things that at least *looked like* what I'm describing.

Okay, I, think I got it now. You're basically saying "I find it more realistic to always have a chance of doing really badly on a spot check and miss something that should have been obvious." And yeah, that happens in real life sometimes.

My first thought is "okay, that's reasonably realistic, but is this sort of thing something that should be modelled in a heroic fantasy game?" I'm not sure I want to have to one day tell a player "I'm sorry you died in the obvious DC5 pit trap the goblins set, but you rolled a nat 1 on your Spot check."

My second thought is "this sort of thing can probably still be modelled using Angry's approach." The approach is to give clues for passive checks and then let the players draw conclusions as to what the clues mean. If they draw the wrong conclusion then they will have sometimes missed the obvious, and the players will have no one to blame but themselves.

Plus there is still room for whiffing or rolling really well on follow-up search checks once you have the initial "passive" clues.

[PhoenixPhyre]

Okay, I, think I got it now. You're basically saying "I find it more realistic to always have a chance of doing really badly on a spot check and miss something that should have been obvious." And yeah, that happens in real life sometimes.

My first thought is "okay, that's reasonably realistic, but is this sort of thing something that should be modelled in a heroic fantasy game?" I'm not sure I want to have to one day tell a player "I'm sorry you died in the obvious DC5 pit trap the goblins set, but you rolled a nat 1 on your Spot check."

My second thought is "this sort of thing can probably still be modelled using Angry's approach." The approach is to give clues for passive checks and then let the players draw conclusions as to what the clues mean. If they draw the wrong conclusion then they will have sometimes missed the obvious, and the players will have no one to blame but themselves.

Plus there is still room for whiffing or rolling really well on follow-up search checks once you have the initial "passive" clues.

In 5e (specifically), a high passive Perception (PP) doesn't give you everything all by itself.

Instead, PP > DC means that you notice that something is out of place or unusual. So someone with high PP will have a lot of mental "that's odd" pings and get a lot of "hints" from the DM. But won't necessarily know all the significance of those pings/hints.

Significance is modeled by an Intelligence (Investigation) check (for generic cases). This can be passive as well if it fits the normal criteria for passivity.

So with a trap, the flow goes like
* PP < DC, no active check => trap triggers, you missed it entirely.
* PP > DC or active Wisdom (Perception) > DC => you notice something (e.g. "One of the flagstones has a slightly larger gap between it and the others than the rest.")

If you noticed something, you can investigate further or trust in your passive Intelligence (Investigation). If that beats the DC or you investigate in the right area (which usually provides automatic success), you may learn that the flagstone is slightly higher and seems to give very slightly under your weight and discover the concealed slots in the wall for blades to come out. It's a blade trap!"

Generally, high PP means a higher floor for not missing hints. But it doesn't give the whole game away, unless the "traps" (speaking generically) weren't really important to begin with. Things that are trivial to deal with as soon as they're noticed are pretty pointless IMO unless there are circumstances that mean that the party can't take time or attention to focus or where it's a one-and-done (no re-attempts, one failure = bad things) check.

The only case where high PP is enough alone to give a large advantage is with hidden creatures (ie ambushes), where having PP > min(Dexterity (Stealth)) is enough to prevent being surprised. But each individual has to make that check individually--the perceptive cleric can't help the bumbling wizard.

IMO, DMs should err on the side of giving the party more information than less information, but designing scenarios where even perfect information isn't enough to resolve the situation (because then there's no agency, the "right answer" is pre-determined).

[Quertus]

The whole "hints" thing is where I start actively disagreeing with Angry. Somewhat.

So, I can accept passive perception in a game. But, much like I need *many* GMs to actually properly run a character¹, so too would "passive perception" make the experience feel incomplete.

But just hints? I can get where Angry is coming from on this one - I honestly 100% can, as I've both seen it and done it - but, at the same time, it's a disastrous strategy when one or more of "the GM doesn't have the skills to pull it off", "the players have insufficient ranks in 'Knowledge: GM' to have a clue what's going on", "the players and GM are operating at (vastly) different levels of skill in the subject matter".

Now, some of that can be remedied somewhat by *training* the players, by making all such comments initially… not "TPK-inducing", and explaining when they fail how those hints work.

Where you run into real problems is when the logic of the hints is lost on the player(s) - or, worse, the (il)logic of the hints has no connection to anything but the GM's insanity-indicted misunderstandings of how the real world operates.

Possibly worse worse, when the GM just whole cloth makes up fantasy physics that obviously inherently have no connection to how the real world operates, then operates solely on these hints ("you feel a tingle", "you taste purple", "their aura is mauve", etc).

IME, most GMs trying to describe the real world fail as badly as those who expect "your fingernails begin to vibrate" to work as a warning that you have been poisoned by a supernatural toxin. IME, most GMs who rely exclusively on hints need to be hit with a (verbal) clue-by-four repeatedly until they operate at a different level of abstraction in communicating what the PCs know, because their "hints" are so disconnected from reality as to be actively detrimental.

Or, as a shopkeep once asked, "what do you know about a customer when they ask, 'how much does this cost?'?"? (I was the only one at the table to answer correctly)

I disagree that you have to play as yourself. I think you can explicitly tell a player with sufficient knowledge skills, "the giant animated fire appears to a fire elemental; if it is, you know that such creatures live on the elemental plane of fire, and cannot be harmed by fire".

I think that that's better than, "you hear a crackle from the glowing red monster as it approaches you" in many ways.

¹ to run them under "a GM who does conversation well", and "a GM who does traps well", and "a GM who does feels well", and "a GM who does '5d chess' well", and… only much finer granularity than that.

[Pleh]

Okay, I, think I got it now. You're basically saying "I find it more realistic to always have a chance of doing really badly on a spot check and miss something that should have been obvious." And yeah, that happens in real life sometimes.

My first thought is "okay, that's reasonably realistic, but is this sort of thing something that should be modelled in a heroic fantasy game?" I'm not sure I want to have to one day tell a player "I'm sorry you died in the obvious DC5 pit trap the goblins set, but you rolled a nat 1 on your Spot check."

My second thought is "this sort of thing can probably still be modelled using Angry's approach." The approach is to give clues for passive checks and then let the players draw conclusions as to what the clues mean. If they draw the wrong conclusion then they will have sometimes missed the obvious, and the players will have no one to blame but themselves.

Plus there is still room for whiffing or rolling really well on follow-up search checks once you have the initial "passive" clues.

The whole "hints" thing is where I start actively disagreeing with Angry. Somewhat.

So, I can accept passive perception in a game. But, much like I need *many* GMs to actually properly run a character¹, so too would "passive perception" make the experience feel incomplete.

But just hints? I can get where Angry is coming from on this one - I honestly 100% can, as I've both seen it and done it - but, at the same time, it's a disastrous strategy when one or more of "the GM doesn't have the skills to pull it off", "the players have insufficient ranks in 'Knowledge: GM' to have a clue what's going on", "the players and GM are operating at (vastly) different levels of skill in the subject matter".

Now, some of that can be remedied somewhat by *training* the players, by making all such comments initially… not "TPK-inducing", and explaining when they fail how those hints work.

Where you run into real problems is when the logic of the hints is lost on the player(s) - or, worse, the (il)logic of the hints has no connection to anything but the GM's insanity-indicted misunderstandings of how the real world operates.

Possibly worse worse, when the GM just whole cloth makes up fantasy physics that obviously inherently have no connection to how the real world operates, then operates solely on these hints ("you feel a tingle", "you taste purple", "their aura is mauve", etc).

IME, most GMs trying to describe the real world fail as badly as those who expect "your fingernails begin to vibrate" to work as a warning that you have been poisoned by a supernatural toxin. IME, most GMs who rely exclusively on hints need to be hit with a (verbal) clue-by-four repeatedly until they operate at a different level of abstraction in communicating what the PCs know, because their "hints" are so disconnected from reality as to be actively detrimental.

Or, as a shopkeep once asked, "what do you know about a customer when they ask, 'how much does this cost?'?"? (I was the only one at the table to answer correctly)

I disagree that you have to play as yourself. I think you can explicitly tell a player with sufficient knowledge skills, "the giant animated fire appears to a fire elemental; if it is, you know that such creatures live on the elemental plane of fire, and cannot be harmed by fire".

I think that that's better than, "you hear a crackle from the glowing red monster as it approaches you" in many ways.

¹ to run them under "a GM who does conversation well", and "a GM who does traps well", and "a GM who does feels well", and "a GM who does '5d chess' well", and… only much finer granularity than that.

I agree with Jason here when questioning how much realism is good or worthwhile to pursue. It goes along with what Angry was saying about how realism is actually terrible game design.

In fantasy games, we want the calculated imbalance of versimilitude, not actually balanced simulationism of realism.

I think I mildly disagree with Angry about Passive scores being equivalent to Taking 10. Mechanically, the structure is similar, but I think it's worth noting the Taking 10 rules implied it was a one and done Action that you took. Part of the genius of Passive scores is it expands Taking 10 into a more or less continuous state (again, limited by when you are not distracted or threatened, which preoccupies your senses).

Quertus, sure people aren't always good at giving or taking hints, but I would argue that's not a problem with playing this way.

"You could train your players." Well, yes, but actually no. Players are training themselves all the time. Everything we do, we learn more about that activity. Players will fail to take hints and GMs will fail to give them and by and large, that's just an effect of choosing to play a game with people. Bad pitches will be thrown and bad swings will miss even good pitches.

It seems to me like you are insisting the game be played at a certain skill level to be fun, but that's not true. The learning curve is only a problem when participants get stuck and stop making forward progress, developing cognitive defenses for their counter productive gaming habits.

A good GM knows how to read their players' proficiency and use their hints to teach the game as they go, gradually increasing complexity as the players get familiar with making investigative questions.

A good player can "lead" a rookie GM by demomstrating good investigation and prompting them with creative questions.

[Jason]

So, I can accept passive perception in a game. But, much like I need *many* GMs to actually properly run a character¹, so too would "passive perception" make the experience feel incomplete.

Are you talking about all the hats you as a GM have to wear to run a game, or actually running with multiple GMs?

Where you run into real problems is when the logic of the hints is lost on the player(s) - or, worse, the (il)logic of the hints has no connection to anything but the GM's insanity-indicted misunderstandings of how the real world operates.

I would call that "bad GM skills". The GM is the source of all of the players' perceptions of the world through the descriptions he or she creates. If he or she can't describe it in a way that makes sense to the players then either learn to do that or it's time to pick a new GM. If all of the players are constantly baffled by "hints" like "vibrating fingernails" or "tasting purple" then the problem isn't on their side of the table.

I disagree that you have to play as yourself. I think you can explicitly tell a player with sufficient knowledge skills, "the giant animated fire appears to a fire elemental; if it is, you know that such creatures live on the elemental plane of fire, and cannot be harmed by fire".

I don't think anyone said "you are playing as yourself." In fact, in this article Angry mentioned recognizing deadly nightshade instantly as a passive use of knowledge skills. So he might very well tell someone with the appropriate knowledge "that looks like a fire elemental. You know those are made of fire, and so won't be hurt by it."

There's another article where Angry points out that it's ridiculous to have long time players playing new characters pretend that they don't know what prevents a troll from regenerating until they make a die roll. If all the players already know that you need silver to fight lycanthrops then just let that be common adventurer knowledge, rather than letting a die roll dictate how they act.

[gijoemike]

This isn't a D&D 5e sub-forum. So much of that only applies to that particular edition of that particular game.

In OSR games, DM rolls for the characters aren't unusual at all. Many of the original DMs did a great deal of this, and the players still managed to have enough fun that it launched an entire industry.

I was referring to the reroll line of feats in 3.X & pathfinder, spycraft 1&2, powers/spells in D20 Modern, Destiny powers in 4th ed, and as you have pointed out the reroll powers in 5th, remember you can force rerolls in hackmaster but that is after the result of the roll in most cases (mostly used to avoid crits). In more recent eds of call of cathulu dice rerolls are part of the game when hunting for successes. Whitewolf <Storytelling> has several powers and cases were rerolls on failures can happen. An additional note, the action dice in Spycraft would allow for adding an exploding die to any d20 roll and the pointman could add his dice to your roll. So all d20 player rolls had to be made in the open.

So, that is a selection of games both d20 and not over the last 20 years. Stating that so much only applies to that particular edition of that particular game is completely not understanding the point.

[Democratus]

So, that is a selection of games both d20 and not over the last 20 years. Stating that so much only applies to that particular edition of that particular game is completely not understanding the point.

Indeed. And in each case, the mechanic and situation for rerolling applies to that specific edition of that specific game.

As was my point, games with re-rolls are a small subset of all Roleplaying Games.

Even in games with re-rolls after the fact, it's not necessary for the player to see the roll. I've played every game with reroll mechanics you mentioned below (plus others like Periphery, Ragnarok, Nephilim...) and never had issues with hidden rolls.

[Xervous]

Indeed. And in each case, the mechanic and situation for rerolling applies to that specific edition of that specific game.

As was my point, games with re-rolls are a small subset of all Roleplaying Games.

Even in games with re-rolls after the fact, it's not necessary for the player to see the roll. I've played every game with reroll mechanics you mentioned below (plus others like Periphery, Ragnarok, Nephilim...) and never had issues with hidden rolls.

All rerolls my players have chosen to take were either heat of the moment do or die, or really important flavor scenes they wanted to drive home to align with their character concept. If I was to be rolling life or death behind the scenes and only narrating noteworthy results it wouldn’t appear much different from rocks falling. The other category is exclusively player opted actions.

[D+1]

Another old school DMing technique - the DM rolling for the player.

Old school is not having to roll any dice in the first place because verbal interaction between the player and DM probably resolved everything needing to be resolved.

The purpose of it is to avoid metagaming. PC searches for traps. The DM rolls for it so that when he says the PC doesn't find any traps the player doesn't know if it's because there really isn't a trap or there is a trap but the character missed it because of a low roll.

This, for example, would instead be something like: ALL the PC's search for traps (because looking for traps is something all characters can do and doesn't REQUIRE a thief character to be a party member to normally deal with traps). They explain what kind of traps they are looking for and where they are looking. The kind of trap that is actually present is in this case unusual; it doesn't fit their description of what they're looking for and the DM responds that no trap is found. A Thief player is still suspicious or wants to be thorough. The thief player gets to ROLL their characters chance to detect traps (which especially at lower levels is generally quite poor). This does not replace all other efforts to find traps, it is in addition to it. Players are still expected to verbally describe their searching process especially at lower levels because the description from even modestly experienced players is going to be more successful than the die roll. If the description succeeds no die roll or thief character is even needed. If the roll were successful then obviously the trap is found and there would be no reason for the roll to have been made by the DM or kept secret. If the description failed AND the roll failed then there is still no reason the roll would need to be made by the DM or done secretly because there is no way for the PLAYER to know if there really ISN'T a trap, or if there IS a trap and the players efforts to find it simply all came up short.

At higher levels, traps as a major element of gameplay has mostly been left behind. The process is technically still the same but the skill of thieves to find traps just by die roll has advanced to the point where IF there is a thief in the party it is far more expedient to simply have them roll. Even if there is no thief in the party, the players simply recite a now well-known litany of what their characters need to look for in order to locate and deal with traps. Rolling the dice is mostly just alleviating them of that chore of recitation.

THAT is the old school way.

[Pleh]

And I find Angry's articles insufferable.

Actually, on a side note, I just found a trick that helps make his shtick a lot more fun to read (for me at least).

I started imagining it being read aloud by Gordon Ramsey.

[OldTrees1]

Actually, on a side note, I just found a trick that helps make his shtick a lot more fun to read (for me at least).

I started imagining it being read aloud by Gordon Ramsey.

Interesting, but not a guaranteed technique. Gordon Ramsey is unbearable for me.

[Jay R]

Part of the problem is rooted in the 3.0 (and later) decision to make all rolls be good on the high end. This gives the player meta-information which the PC shouldn't know.

When I needed to make perception rolls to notice something in AD&D, I had them roll a d20, and if they rolled under their INT (sometimes modified by circumstances), they succeeded.

So in one case, I had somebody roll, he rolled a 4, and I said, "You see a force of archers on the hill." In this case, it is not obvious to the player whether he made his perception roll and saw what was there, or failed his saving throw and believed the illusion.

One player didn't like the idea of rolling and not knowing if he should be hoping for high or low, and asked me to roll any die when that was the situation. No problem. I'm happy to do that for the player who wants it, and to leave the roll ambiguous for others who want to make their own roll.

[Duff]

Actually, on a side note, I just found a trick that helps make his shtick a lot more fun to read (for me at least).

I started imagining it being read aloud by Gordon Ramsey.

Oh that's hilarious!!

[Democratus]

All rerolls my players have chosen to take were either heat of the moment do or die, or really important flavor scenes they wanted to drive home to align with their character concept. If I was to be rolling life or death behind the scenes and only narrating noteworthy results it wouldn’t appear much different from rocks falling. The other category is exclusively player opted actions.

All results are noteworthy in a life or death situation. So that means narrating every blow-by-blow. Very engaging even if the random number generator is behind the screen.

[Nightcanon]

I'd rather have my DM roll for me (including making old-school Gygaxian 'keep' em on their toes' nonsense rolls for atmosphere) than have him working through a list of pre-generated rolls. There's a reason why clinical trials don't operate on the basis that you know the allocation for the next few patients, based on the order that you enroll them.
'Gee thanks, Dave, glad to know that my natural 20 enabled me to detect the elusive scent of snow berries from the field 600 yards away, but my 2 resulted in me not spotting the family of drop bears I just ride beneath...'

[farothel]

In our group the GM never rolls for the players, even if some rolls should be secret. Of course, we roleplayed almost 20 years together now, rotating the GM position so each of us has been a player and a GM and we're quite used to each other.
As to chain rolling things like knowledge checks, when we spot something, all players who have the appropriate knowledge skill and see the thing roll at the same time and all get info depending on their roll.

[Tanarii]

The Angry GM did a piece on the Insight skill in D&D and had quite a bit to say on the virtues of the Passive Perception mechanic. I think it's worth a read: theangrygm.com/insight-into-insight
The article on social interaction that spawned this article is a good one too.

Yeah, he got what passive checks are for all wrong. He's making the classic mistake that someone who has merely skimmed the rules instead of reading them makes. He assumes the "passive" in passive scores applies to the character being passive.

[Quertus]

Now, some of that can be remedied somewhat by *training* the players, by making all such comments initially… not "TPK-inducing", and explaining when they fail how those hints work.

Quertus, sure people aren't always good at giving or taking hints, but I would argue that's not a problem with playing this way.

"You could train your players." Well, yes, but actually no. Players are training themselves all the time. Everything we do, we learn more about that activity. Players will fail to take hints and GMs will fail to give them and by and large, that's just an effect of choosing to play a game with people. Bad pitches will be thrown and bad swings will miss even good pitches.

GM: "You watch as the ogre flips the lever, and the princess begins slowly descending into the vat of boiling oil. 50' of uneven floor stand between you and this scene."

Player: "I rush over and attack the ogre, hoping to push it back from the lever."

GM: "Rocks fall, everyone dies."

(later)

GM, to fellow GM: "They made their passive 'Find Traps' roll, and I told them that the floor was uneven, but they charged in and died anyway. Stupid players."

[Pleh]

GM: "You watch as the ogre flips the lever, and the princess begins slowly descending into the vat of boiling oil. 50' of uneven floor stand between you and this scene."

Player: "I rush over and attack the ogre, hoping to push it back from the lever."

GM: "Rocks fall, everyone dies."

(later)

GM, to fellow GM: "They made their passive 'Find Traps' roll, and I told them that the floor was uneven, but they charged in and died anyway. Stupid players."

Feels like a strawman to argue against disingenuous DMing, which I would hope was rather clearly not my point.

[Pex]

Part of the problem is rooted in the 3.0 (and later) decision to make all rolls be good on the high end. This gives the player meta-information which the PC shouldn't know.

When I needed to make perception rolls to notice something in AD&D, I had them roll a d20, and if they rolled under their INT (sometimes modified by circumstances), they succeeded.

So in one case, I had somebody roll, he rolled a 4, and I said, "You see a force of archers on the hill." In this case, it is not obvious to the player whether he made his perception roll and saw what was there, or failed his saving throw and believed the illusion.

One player didn't like the idea of rolling and not knowing if he should be hoping for high or low, and asked me to roll any die when that was the situation. No problem. I'm happy to do that for the player who wants it, and to leave the roll ambiguous for others who want to make their own roll.

That's the point. This was done even before 3E as you yourself just said you did. Why? To avoid metagaming. You didn't want the player to know if what he saw was real or illusion just because of a number on a die. This is game edition agnostic.

In my usual rhetoric I won't say DMs who do this are inherently tyrannical . Only as much as a DM is one independent of this, doing this is a one way to be one by faking rolls and dictate what they want to happen that hurts the player. But anyway, accepting the DM is Honest True trying to enforce a more realistic atmosphere for fun I still prefer the player rolls his own fate at the moment it happens. Having to teach players as necessary not to chain roll or refuse to roleplay the result is a price I'm willing to pay.