RPG theory: in-game balancing

[Celebrim]

As a guy that written an RPG but is not an RPG designer, I'll try to give you my take.

Games are most fun when you feel like you are making meaningful choices that will effect the outcome. This is true in tabletop RPGs pretty much irrespective of your aesthetics of play. If you are engaging with the fiction you are hoping to change it. But this goes double for people with challenge based aesthetics of play, who want to not only make meaningful choices but feel that their choices were clever, not obvious, and required study and analysis and skill to come up with. One of the many reasons combat is an enduring focus of table top RPGs is precisely that it allows for complex challenges to overcome.

But your two examples don't actually fulfill any of those needs in a strong manner. In the case of encountering something well above your ability to overcome either this negates your choices since it doesn't matter what you do you will lose, or else it turns into a color/stage setting/cut scene type encounter where the narrator describes what you see but you can't really interact with it in a meaningful way. Because there aren't meaningful choices here, a very little of that goes a very long way. Think about that famous scene in the original Jurassic Park movie, where the music swells and the protagonists look up and they see the immense herds of dinosaurs. It's not that there is no point in having that sort of stage setting scene in a tabletop RPG, it's that movies are much better suited to delivering sensation based aesthetic with music and visuals than a narrator in a table top game is. It's often a bad idea to encourage a GM to be overly wordy and prosaic in narrating because it's so hard to deliver a beautiful enough word picture to make that time well spent, and that's especially true for something where the PC has no agency over what he sees.

And the same sort of thing is true of pushover encounters. There just aren't any meaningful choices when you are so much more powerful than the threat that there is no threat. It's not that it's entirely wrong to occasionally have an establishing scene where you let the players show and feel just how dangerous and powerful they've become, but a little of that goes a long ways. After a while, killing kobolds is just rote rolling with no payoffs and no meaningful choices.

By comparison balanced challenge make for meaningful choices where the players have more emotional investment because they have to 'sweat it out' and work for their victory, and where the players have more satisfaction because they realize that they needed to make those smart plays in order to win whether against a particularly tough foe or a run of bad luck. There is just more "fun" to be had because you are embracing more aesthetics of play simultaneously in more compelling ways.

And again, it can't be overemphasized how poorly tabletop RPGs do sensation. They are hard to wring enjoyment out of compared to video games or movies played passively and casually. So whatever game experience you create needs to have as it's core experience balanced challenge, with only "outliers" rarely happening and those "outliers" carefully managed in order to be fun.

Lastly, this is a well covered ground in the game design community. You should be able to find a lot of discussion around why balanced game play is core to your experience and when and how you can break those guidelines profitably to achieve particular effects. You aren't covering ground no one has thought through before.

 

[aia_2]

As a guy that written an RPG but how is not an RPG designer, I'll try to give you my take.

Games are most fun when you feel like you are making meaningful choices that will effect the outcome. This is true in tabletop RPGs pretty much irrespective of your aesthetics of play. If you are engaging with the fiction you are hoping to change it. But this goes double for people with challenge based aesthetics of play, who want to not only make meaningful choices but feel that their choices were clever, not obvious, and required study and analysis and skill to come up with. One of the many reasons combat is an enduring focus of table top RPGs is precisely that it allows for complex challenges to overcome.

But your two examples don't actually fulfill any of those needs in a strong manner. In the case of encountering something well above your ability to overcome either this negates your choices since it doesn't matter what you do you will lose, or else it turns into a color/stage setting/cut scene type encounter where the narrator describes what you see but you can't really interact with it in a meaningful way. Because there aren't meaningful choices here, a very little of that goes a very long way. Think about that famous scene in the original Jurassic Park movie, where the music swells and the protagonists look up and they see the immense herds of dinosaurs. It's not that there is no point in having that sort of stage setting scene in a tabletop RPG, it's that movies are much better suited to delivering sensation based aesthetic with music and visuals than a narrator in a table top game is. It's often a bad idea to encourage a GM to be overly wordy and prosaic in narrating because it's so hard to deliver a beautiful enough word picture to make that time well spent, and that's especially true for something where the PC has no agency over what he sees.

And the same sort of thing is true of pushover encounters. There just aren't any meaningful choices when you are so much more powerful than the threat that there is no threat. It's not that it's entirely wrong to occasionally have an establishing scene where you let the players show and feel just how dangerous and powerful they've become, but a little of that goes a long ways. After a while, killing kobolds is just rote rolling with no payoffs and no meaningful choices.

By comparison balanced challenge make for meaningful choices where the players have more emotional investment because they have to 'sweat it out' and work for their victory, and where the players have more satisfaction because they realize that they needed to make those smart plays in order to win whether against a particularly tough foe or a run of bad luck. There is just more "fun" to be had because you are embracing more aesthetics of play simultaneously in more compelling ways.

And again, it can't be overemphasized how poorly tabletop RPGs do sensation. They are hard to wring enjoyment out of compared to video games or movies played passively and casually. So whatever game experience you create needs to have as it's core experience balanced challenge, with only "outliers" rarely happening and those "outliers" carefully managed in order to be fun.

Lastly, this is a well covered ground in the game design community. You should be able to find a lot of discussion around why balanced game play is core to your experience and when and how you can break those guidelines profitably to achieve particular effects. You aren't covering ground no one has thought through before.

Well, your comments are definitely right and i have nothing against your assumptions and conclusions. My point is slightly different: i believe that in-game world should be similar to reality. This means that, as an adventurer, a PC should find himself (along with the whole party) against a wide range of events... From the "boring" one (i.e. a fight against a considerable lower lever NPCs than the party) up to the "impossible" ones... According to the "good" mastering "unwritten" rules, a GM should always try to combine encounters where there is always a chance to have a success... In my mind this is not correct: a "good" GM should let the players have any kind of encounters... and a "good" player should be able to understand when it is the case to withdraw or to go into the fight...
According to my way of mastering, the GM should reward i the way a party which flees from a sure defeat exactly like a victorious fight with a "balanced" opponent.
That's the point where i wanted to get with my questions in the first post.

 

[Umbran]

Well, your comments are definitely right and i have nothing against your assumptions and conclusions. My point is slightly different: i believe that in-game world should be similar to reality.

Of course, you can play as you wish. It may be important to note, though, that the genres of fiction we tend to play in typically do not match reality in more ways than you can shake a stick at. Protagonists really don't live lives similar to reality, and we are largely okay with that.

So, I am not sure that "similar to reality" brings as much value or satisfaction to play as one might guess.

According to the "good" mastering "unwritten" rules, a GM should always try to combine encounters where there is always a chance to have a success...

Since you are citing "unwritten" rules, we can't exactly ask you to show an example.

Maybe we should look at the advice given in the 3e DMG on encounter design.... I quote the 3e DMG page 102:

Difficulty
Sometimes, the PCs encounter something that's a pushover for them. At other times an encounter is too difficult and they have to run away. A well-constructed adventure has a variety of encounters at several different levels of difficulty.
What follows is a table that suggests:
10% of encounters should be Easy
20% should be Easy (if handled properly)
50% should be Challenging
15% should be Very difficult
5% should be Overwhelming

So, we cannot really speak to your "unwritten" rules. But the actual written rules don't seem to agree with your claim.

According to my way of mastering, the GM should reward i the way a party which flees from a sure defeat exactly like a victorious fight with a "balanced" opponent.
That's the point where i wanted to get with my questions in the first post.

So, you asked a question to give yourself the opportunity to pontificate on your own answer?

 

[Celebrim]

My point is slightly different: i believe that in-game world should be similar to reality.

So, the first thing you should be asking is, "Why do I believe this? What is simulating of reality for? What good is it?"

There are several possible answers to that.

1) A game that is similar to reality except where its conceits are specifically called out (things like "There are large flying firebreathing lizards.") is one in which players can make natural propositions with some expectations that they can predict the results of their propositions. The less like reality your game is on a casual level, the harder it is to play because the player's expectations about cause and effect will be wrong and they'll not understand how to accomplish their intended goals. This is "casual realism". The results of interacting with your game world are close enough to realistic that they are familiar.
2) Another reason to simulate reality is reality has depth and feels real and believable and immersive. Simulating reality is the easiest way to bring the game to life such that the players are interacting with it emotionally and logically as if it was real, and for a lot of players that's a compelling and rewarding aesthetic of play. If you have hard gamist construct like what you see in something like World of Warcraft, where difficulty is strongly silo'd into regions of balanced encounters then there are limits imposed on how much you can believe in that world. Either you have to accept at all times the unreality of the world or you have to except at all times the unreality of game constructs (that is, you have to accept that level 20 bears are exactly as powerful as level 72 bears, and that the game numbers have no relationship to the world they ostensibly describe and can't be used to make predictions about how the world will behave).
3) A third reason might be that you are actually simulating a historical reality and not a fantasy.

However, all that said, there are a lot of reasons for eschewing realism because ultimately realism doesn't prove to be fun. Realistically speaking, interesting things shouldn't be happening to the PCs at all. The PC's live lives that are unrealistically exciting and filled with interesting events. Things happen to them all the time, or else you are going to be in a situation where you have a slice of life RPG about the trials of being an ordinary miller's son that never has anything more adventurous come his way than an ale and a smart mouthed farmer boy looking for trouble.

What typically you actually want is just enough realism to suspend the player's disbelief. As long as the player isn't taken out of the moment by the realization of how unreal everything is, you are good.

I'm old enough to remember the fetishization of realism in design discussions from the mid 80's through the early 90's. In the early days of RPG design, realism was often treated as the solution to whatever ailed or seemed to ail an RPG or gameplay. If the rules didn't seem to be working, there was an impulse to say that this was because they weren't realistic enough. If the table wasn't having fun, there was an impulse to say that it was because the game wasn't realistic enough. Lots of effort went into creating game constructs that met people's expectations of realism, right up to and including processes of play that looked like solving word problems in a physics or engineering class. And those assumptions that realism would fix the issues really weren't questioned much until we had systems taking those ideas to their logical conclusion and we started to realism that by and large not only did they not fix the problems that they were intended to fix, but they weren't necessarily any more realistic and often we were gaining nothing by being "realistic". The effort was in fact a waste.

And quite frankly, I just disagree with the principle you've laid out here. A good GM always, always, always has his finger on the scale to tilt the situation in the PC's favor ever so slightly in some regard. A good GM never hits the players with no win scenarios intentionally. If you are doing that, you have lost your way and your motives are questionable. Now, the players may find way to dig themselves holes they can't get out of, and if they do you aren't responsible for always or necessarily ever rescuing them. But there's no skill involved in hitting players with no win scenarios. It's zero skill GMing because as a GM you have infinite resources, so challenging the players is never a difficult to do that requires a bunch of skill, and it never shows off your skill as a GM to challenge the players. The skill is demonstrated by making something fair and challenging, where the player knows you aren't using a lot of force but still presenting difficult problems to solve.

In the case of impossibly hard encounters, the skillful GM is always got a flashing neon arrow pointing to the way out and providing fun for tucking your tail between your legs and retreating. In the case of death traps, the skillful GM always is telegraphing, "My god man, look at the bones!" so that if players want to push the big red button despite all the warnings not to, that's on them and they'll know it. One of the surest signs you've lost your way as GM is you start fantasizing about how much in awe the players are going to be of the terrible thing you are going to present to them. In fact, probably avoid spending too much time fantasizing about how a future encounter is going to go EXCEPT for thinking it through to make it balanced and interesting.

Remember, if you are talking about game design this even goes beyond good GMing. You have to as a game designer explain to your audience how to play your game in a way that is going to be functional and fun. And that means essentially treating unbalanced encounters as an exception case and providing guidelines for how you run them in a way that they are fun. For example, one of my campaign ideas that I've always wanted to run involves the town the PC's are in getting attacked by an impossibly high CR ancient red dragon, that they have no chance against at all. The PC's are essentially ordinary inhabitants of Laketown when Smaug attacks. And the start of the adventure is not fighting the dragon, but dealing with the consequences of a town on fire, buildings collapsing, livestock and people panicking and all that other mundane but exciting hazards that is in the scale of what starting PC's can handle, while the dragon is doing it's thing in the background and defeating and eating the town's BDH's. But if I ran that encounter as, "A dragon flies up to you and breathes fire, take 15d10 damage save for half", that's not the start of a campaign. That's not fun. And that's not high skill GMing (unless the PC's are meant in the game to play as ghosts or petitioners in the afterlife, neither of which is something I'd expect a novice to pull off without considerable help from a designer).

 

[Blue]

As a game designer, i don't completely understand the need in many games to preserve balance in encounters: why a 1st level party should never meet a dragon? This would result lethal for most of the characters but in any case it could happen (and let me add, the GM should reward any player who is able to let his PC survive the encounter, even if this is a total defeat for the party!).
This of course works also in the other sense: why a veteran party of level 20 should not be attacked by a group of bandits of level 1?
The balancing concept has been stressed out so much that in the (nightmarish) XP system of 3E/3.5 this was a cornerstone in terms of rewards...
I really miss the importance of looking after a balance encounter or to properly scale a module.
One simple question: why?

This is a product of games with a couple of characteristics. (And I use "commonly", as in more than half the time, but need not be all the time.)

For why you don't see nigh-unwinnable combats
1. Combat is used as a common way to overcome challenges.
2. The penalty for failure in combat commonly has severe repercussions on a player's fun, like character death, as opposed to other systems like Fate or Masks where it can be taken out of a scene in many different ways, many of which can be used to enhance the story the table is coming up with.

(There are plenty of other points that could come up as well. For example people enjoying OSR might want to add in lengthy character creation as a prerequisite for the focus on balance in challenges.)

Basically, if losing combat commonly leads to big deal issues, players and GMs want to avoid it. If if can lead to interesting issues then players and GMs are much more likely to embrace the down-beat and go with it.

For why you don't see very easy combats
1. When the system only provides one level of zoom to a high granularity of combat, which takes long enough that the system even enforces that all character types can contribute to combat in some way so the player is not bored, then "meaningless" combats in terms of the mechanical implications (resource attrition, whatever) become a waste of time. They take up a chunk of the session and with little mechanical change or consequence.

This is pretty straightforward. Personally, I love occasionally throwing a weak combat at my players to curbstomp and rememeber they are heroes. But I also use tools like the Montage from 13th Age for other things. "Okay, you high level adventurers are set on by a phalanx of hobgoblins. Why don't we go around the table and tell everyone the cool move you did as the party defeated them."

 

[aia_2]

Thanks for the last couple of comments, some side notes though:
1. Combat/fight is not the only context where the concept of balancing is used... I mention them as they are the most immediate and easy application of balancing
2. Prob being a non native english speaker, i misused some words and have been misunderstood, apologies: i used realism with a different meaning from what you understood. I just wanted to point out that in a RPG session, players' life should be open to any kind of event like in our everydays lifes. This means that you can have a lucky day (you made an heoric action with success) but you can find that s**t happens also to your PC.

Back to my previous post: imho the "good" GM should be able to present such a world to his players (here i misused the word realism) and the "good" players should be able to understand that when it pours it'd be the case not to go ahead in order to avoid bad consequences.
This of course has to be rewarded even if the players decision didn't bring any improvement but the fact their PCs survived.

 

[MGibster]

Thinking of other media, why don't super-hero comics spend a lot of time on Thor or Superman running into muggers or robbers? (Sure, it happens, but a lot of times its a few throw away panels to give the character time for some thought balloons or conversation with another hero. Is it worth having the party roll attacks when they'll clearly curb-stomp the other side? Or is it better to just give them the successes and move on after the brief bit of flavor is accomplished?)

There was an episode of Superman the Animated Series where he stops a crime in progress by the League of Assassins (Batman's enemies). These ninjas are ineffectively attacking Superman when he asks incredulously, "Do you know who I am?"

I'm with @Umbran, balanced encounters produce a more satisfactory game experience. But RPGs are a bit different from chess or Warhammer 40k. As a GM, I'm not going to purposely set up an encounter knowing the PCs will be curb stomped. On the flip side, sometimes players put themselves in situations where they're likely to be curb stomped.

 

[Celebrim]

I just wanted to point out that in an RPG session, players' lives should be open to any kind of event like in our everyday lives.

I think we know what you are trying to say. This is a very old subject that has been discussed many times and so the discussion is mostly moving over familiar points and arguments. And I think that a lot of people in the thread are sympathetic to the idea that not every encounter needs to be perfectly balanced.

I think there are principally two points of disagreement.

First, most rule sets don't actually enforce balance and do in fact provide for a wide range of encounters from trivially easy to overwhelmingly dangerous. Even your example system 3.X D&D suggests, encourages, and provides for this sort of wide range of difficulties in encounters. And even that is only a guideline and not a rule, like the common advice on how to write well in English, such as avoiding "-ly" adverbs or avoiding sentence fragments. However, that is just a guideline, and strong and experienced writers can break the guidelines if they know what they are doing. In the same way, you can break the guideline on balanced encounters, but you better know what you are doing if you do so.

And secondly, while I think most of us agree that there can be value in a wide range of challenge difficulties, I don't think you'll get a lot of agreement with the assertion that "players' lives should be open to any kind of event like in our everyday lives". Real life is not fair or just. It's quite possible to get up in the morning and die to some accident you couldn't have seen coming. For a soldier, the reality of war is not fair and balanced challenges. Sometimes it's a 105mm shell bursting 5m from you and your body comes apart before you have time to realize you've died. We could in real life go to sleep, be hit by an asteroid, and never wake up. But these sorts of things shouldn't happen in a game because the point of the game is to be fun. There may be a point in a 'Kobashi Maru No Win Scenario' when training Star Ship captains for the harsh realities of command, but there isn't much point in running the 'Kobashi Maru' as anything but a one shot. In a game intended to go more than one session, no win scenarios are to be avoided.

In reality most GMs very carefully but unconsciously steer way from realistic events because they are given examples of play that follow certain patterns. And ultimately, most successful RPGs are not in any way trying to be realistic and conform the game to the realities of everyday life. Instead, most RPGs actually are trying to have verisimilitude to fictional narratives. They want actions to play out the way they play out in some favored story.

My advice to people who want to design RPGs is for them to play as many different styles of game with as many different rule sets and processes of play as possible. Sit at a lot of tables. Run a lot of different games. Learn what works. Learn what doesn't. Try to figure out why things work, and how different styles and rules achieves different effects. So if you have this theory that a good game would involve the party of 1st level characters being attacked by a hungry ancient red dragon, start a campaign that way and see how it goes. Test your ideas. How much fun do your players report it is to make characters and then unexpectedly have them die in a brutal unfair encounter where they stood no chance of doing anything, not even running away? How many times can you run that encounter and people still be interested in it? How much less fun do you think it will be if instead of the first encounter of the campaign, a dragon that is not even part of the adventure comes out of nowhere and eats everyone unexpectedly after two years of developing characters. It could happen. Things like that do happen to NPCs, so in a sense it's realistic. Why do you think game designers discourage GMs from doing that in their games? Are they just dumb? Have they not thought this out enough?

 

[aramis erak]

(disclaimer: it will likely have already been discussed but i do not find this topic...)

As a game designer, i don't completely understand the need in many games to preserve balance in encounters: why a 1st level party should never meet a dragon? This would result lethal for most of the characters but in any case it could happen (and let me add, the GM should reward any player who is able to let his PC survive the encounter, even if this is a total defeat for the party!).
This of course works also in the other sense: why a veteran party of level 20 should not be attacked by a group of bandits of level 1?
The balancing concept has been stressed out so much that in the (nightmarish) XP system of 3E/3.5 this was a cornerstone in terms of rewards...
I really miss the importance of looking after a balance encounter or to properly scale a module.
One simple question: why?

For me, it's that it gives me a firm grim on what the designer expected to be suitable for a given range of play.

I'll note that it's largely the D&D tree that does it... BX/BECMI, it was (roughly) Monster HD = dungeon level...
here, quote time...

No. Appearing (Number Appearing) gives the suggested number of that monster type which will appear when encountered on the same dungeon level as that monster's hit dice (or monster level).

Bold & italic original. Generally the dungeon level is a guideline for the average party level where it will be challenging but survivable for a typical 4 key roles party (F, MU, Cl, Th. Ha, Dw, Elf can replace F, and Elf can replace wizard.)

A monster's level is only a guide, and a monster could be found anywhere in a dungeon, whatever the level. However, as a general rule, it is useful to limit monsters to 2 dungeon levels higher or lower than their hit dice.

That being the starting point for many of the OSR games, as well... they keep to it.

As a GM, when running Cyclopedia (the 3rd subedition of "Basic D&D"), I knew that, given a roughly balanced encounter, it was going to be a rough but winnable fight. Keeping to the ±2, I have fairly predicable challenges for a stnadard 4 party, and can limit the dungeon well to the party's level (adjusted for party size)...

I've found the 3E version didn't work well for me, but the 5E one does. And the 5E one worked brilliantly for me... but the math is often done wrong by others, so... YMMV.

 

[aia_2]

I think we know what you are trying to say. This is a very old subject that has been discussed many times and so the discussion is mostly moving over familiar points and arguments. And I think that a lot of people in the thread are sympathetic to the idea that not every encounter needs to be perfectly balanced.

That is interesting, if you have some links or hints, pls let me know for i'd be really interested in reading about such a topic!

I think there are principally two points of disagreement.

First, most rule sets don't actually enforce balance and do in fact provide for a wide range of encounters from trivially easy to overwhelmingly dangerous. Even your example system 3.X D&D suggests, encourages, and provides for this sort of wide range of difficulties in encounters. And even that is only a guideline and not a rule, like the common advice on how to write well in English, such as avoiding "-ly" adverbs or avoiding sentence fragments. However, that is just a guideline, and strong and experienced writers can break the guidelines if they know what they are doing. In the same way, you can break the guideline on balanced encounters, but you better know what you are doing if you do so.

Correct, but that is not the point i want to discuss: i am not complaining about rules... i am complaning about the fact that whenever you have a 7th level party, your adventure is always "railroaded" in terms of encounter: never a monster where the wisest thing the party can do is ti flee, never a useless group of 1st level npcs (who by the way have not a tag on their shoulders "1st level, hit here to go faster")... far lightyears an encounter with a mixed group of npcs (say 3 of 1st level and 2 of 11th level)... My point is also try to train the players to "read the situation" and take decisions that ordinary rules would consider negative in terms of experience (like withdrawing) whereas they should be calculated in the exact opposite sense: a combat would turn in a likely death and a withdraw would imply the pcs survives.

And secondly, while I think most of us agree that there can be value in a wide range of challenge difficulties, I don't think you'll get a lot of agreement with the assertion that "players' lives should be open to any kind of event like in our everyday lives". Real life is not fair or just. It's quite possible to get up in the morning and die to some accident you couldn't have seen coming. For a soldier, the reality of war is not fair and balanced challenges. Sometimes it's a 105mm shell bursting 5m from you and your body comes apart before you have time to realize you've died. We could in real life go to sleep, be hit by an asteroid, and never wake up. But these sorts of things shouldn't happen in a game because the point of the game is to be fun. There may be a point in a 'Kobashi Maru No Win Scenario' when training Star Ship captains for the harsh realities of command, but there isn't much point in running the 'Kobashi Maru' as anything but a one shot. In a game intended to go more than one session, no win scenarios are to be avoided.

Why do you tie fun with success? Even an episode where a PC dies could be perceived with fun. The assumption here is similar to the one i do not understand about the balance in encounters. If the assumption is this one, we can go ahead for eons in this conversation.
Fun could also come from unexpected situations, regardless or not they are "positive" for a character. Fun could be an encounter between a 4th level party with a 12th level group of bandits npcs (and sorry, they could be interested in robbing in the street... i do not see any illogical situation here, they're not living in a modern environment like ours). If a player plays under the assumption that "s**t happens" like real life, he will think two times before entering in a fight with the first group of npcs he meets. The fun is also "read" the situation and being a winner even in case you decide to flee.

In reality most GMs very carefully but unconsciously steer way from realistic events because they are given examples of play that follow certain patterns. And ultimately, most successful RPGs are not in any way trying to be realistic and conform the game to the realities of everyday life. Instead, most RPGs actually are trying to have verisimilitude to fictional narratives. They want actions to play out the way they play out in some favored story.

...ok and this brings me back to the original question: it not only a matter of mastering, it is also a matter of balancing in game design... and the question is again: why?

My advice to people who want to design RPGs is for them to play as many different styles of game with as many different rule sets and processes of play as possible. Sit at a lot of tables. Run a lot of different games. Learn what works. Learn what doesn't. Try to figure out why things work, and how different styles and rules achieves different effects. So if you have this theory that a good game would involve the party of 1st level characters being attacked by a hungry ancient red dragon, start a campaign that way and see how it goes. Test your ideas. How much fun do your players report it is to make characters and then unexpectedly have them die in a brutal unfair encounter where they stood no chance of doing anything, not even running away? How many times can you run that encounter and people still be interested in it? How much less fun do you think it will be if instead of the first encounter of the campaign, a dragon that is not even part of the adventure comes out of nowhere and eats everyone unexpectedly after two years of developing characters. It could happen. Things like that do happen to NPCs, so in a sense it's realistic.

Train and test, ok. Re to the dragon: less fun for what reason? If the way you present that event leaves room to survive with a huge defeat of the PC, why the fun should be wasted? Why the fun should be only meet and defeat a pack of rats when the party is at 1st level?

Why do you think game designers discourage GMs from doing that in their games? Are they just dumb? Have they not thought this out enough?

No, i do not think anyone is dumb. However my "why" is still unanswered.