That's not so. There are plenty of games -- not just RPGs, but games like Metroid or Zelda -- where you have some degree of nonlinearity without scaling. I don't recall appreciable level scaling in Fallout (though there may have been some?) or Arcanum or Vogel's games, not to mention a lack of level scaling in many older RPGs that were also nonlinear. Level scaling is only needed for a completely open world where the player can go anywhere in any order without a difficulty spike. It may also be necessary in a really large game where, absent level scaling, a large percentage of the world will be trivially easy by the time the player gets there.You can't have any degree of non linearity in exploration, without some sort of level adjustment for the enemies. If done right, nobody even notices it exists.
So while you were adventuring, the bully was lifting and got some mad gains.
Sounds like Oblivion. Robbers in full daedric armor worth a whole kingdom.
Finally, someone with brainz and vision.
I'm a huge proponent of difficulty scaling, despite not ever seeing it (properly) implemented:
1. level scaling (enemies increase their damage and/or health)
2. equipment scaling (enemies wear better armor and/or weapons)
3. numbers scaling (more enemies of the same/different type)
4. AI scaling (enemies becoming better organized and/or individually smarter)
...all of it should be, of course, narratively supported and limited to/consistent with in-game logic of the world in question.
Sounds like Oblivion. Robbers in full daedric armor worth a whole kingdom.
Only if executed poorly. I like how you deliberately ignored my comment above that explains how it should work. Haters gonna hate, I guess.
how it should work
That's not so. There are plenty of games -- not just RPGs, but games like Metroid or Zelda -- where you have some degree of nonlinearity without scaling. I don't recall appreciable level scaling in Fallout (though there may have been some?) or Arcanum or Vogel's games, not to mention a lack of level scaling in many older RPGs that were also nonlinear. Level scaling is only needed for a completely open world where the player can go anywhere in any order without a difficulty spike. It may also be necessary in a really large game where, absent level scaling, a large percentage of the world will be trivially easy by the time the player gets there.
A key aspect of great nonlinear games is the sense that some areas are nominally accessible but perilous. The player has to make cost-benefit decisions as to whether it's worth going someplace dangerous now, or whether it is better to come back later. It's particularly nice if you introduce some cost to abstaining -- thus, the player can't always be confident that he'll be able to come back later. But you definitely want at least some portions where you power up, then come back to challenging areas and crush your foes -- at least you do if you want a traditional hero-story arc to the game.
Level scaling is basically an easy way to insure that you don't need to balance your game taking account of nonlinearity -- it's like throwing up your hands and saying that everyone gets a passing grade because grading is too hard: it doesn't solve the problem so much as obliterates the very possibility of a solution.
It is tempting to fantasize about how these things could be done in a perfect world, but I suppose there's probably good reason that what one tends to see is either simplistic level scaling or nothing at all.How would you approach a situation where the player skips a part of the world that is supposed to be "easy", and then cones back to it? Fallout didn't do level scaling, except in some random encounters (if that at all). Your answer made me wonder if Tyranny will have Kotor-style hubs, where the player can choose the order of visit in a large area, and some kind of level scaling seems inevitable (to me).
Great post, very informative.
Sorry, missed this post until today.Regarding your four points about approaching the task at hand
Actually, oddly enough, I think of it as being even more centered on the player -- or maybe, better said, a "playing area" that is more centered on the "player character." Having things happen in the world independently of player action isn't something that I would necessarily endorse for its own sake (I might, in some games, but not in every instance in every game), but I definitely think that if you're doing a narrative game, you need to be thinking about what content is going to best serve the themes and protagonist's character growth. One real challenge in letting players explore freely is figuring out how to make sure that each area provides something appropriate for when the character arrives there not just in gameplay terms but in thematic terms. (At least, in a narrative/thematic RPG. The same might not be true in an open-world RPG or a dungeon crawler or whatever.)(1) So in short a world which is less centered on the player. This is excellent.
I'm a big believe in the maxim that it's more important that something feel "true" than that it be "real." I'm not sure villains always need to behave like a logical strategist because we're so accustomed to the idea of illogical villainy it can still ring true despite being unrealistic. That said, there is ample evidence in history of powerful empires continuously deploying forces that are just a bit too weak to overcome an insurgent threat.(2) I say this in relation to the rest of your post. Would it be wrong for the big baddie/antagonist to be meta enough to not to send its weaker minions to their deaths multiple times and instead opt to use a stronger underling for ie 'guarding the map?
Well, maybe yes, maybe no. Exile 3 is pretty high quantity. I think you can do non-stupid forms of scaling even in large games, though of course it would be very hard in an open-world game. (Although wasn't the whole damn point of Oblivion gates -- I never played the game -- that it would provide a scaling threat that fit narratively?)(3) To me your post in general speaks of a quality over quantity type of design though that might mostly be the part about Fallen Gods. That is why this point, despite being a good point, is somewhat irrelevant unless aimed at someone making a game where the amount of content is more important than what said content consist of.
My neurotic approach to writing FG's texts essentially forces me to aim for quality. Though the nature of the game (a procedural FTL-like set up) requires a certain quantity to work at all.(4) In TloZ:Twilight Princess and Metroid Prime 3: Corruption the fundamental aspect of what you described had not changed. Item A allows progress through obstruction A, item B allows progress through obstruction B etc. I disagree with you on the fun of blasting through trivial enemies though. In early game enemies in these two series of games are not push-overs only because of the scarcity of resources the player has at hand for example health. By late game these enemies are mostly in the way and make backtracking the prevalent feature of these games a slog. In the case of your own game I urge you to stay on the path of quality>quantity.
Thanks. Though that is a large if, one that you shouldn't set aside in the face of my hyping FG. :DRegarding only the first point you make about Fallen Gods since the other two are to put it lightly awesome if they come into fruition.
Generally speaking, they are incapable of detecting the god when he's too weak. In some instances (for example, with some kinds of wights), it's a little like saying, "Shouldn't moths go to flames before they get big and hot enough to burn the moths?" In other instances (such as wizards), it's a matter of their detection capabilities. And with godly enemies, it has to do with a mix of lassitude and fair play.(1) Just repeating what I said about meta knowledge. Could said magical creatures that you give as an example be smart enough to act before it is too late and the player can go toe to toe with them?
Thursday?By the way is there an ETA for Fallen Gods?
And with godly enemies, it has to do with a mix of lassitude and fair play.
We're having recurrent coder problems. I'd love to have something totally playable by the end of the year, though.
I'm from mobile now and I can't reply properly, but I always had a (possibly stupid) thought that leveling should be somehow "horizontal" instead of "vertical". I don't even know where to begin describing exactly what I mean, so I'll come back at it tomorrow, but the gist is that the whole idea of hp is going away, that a level up makes you do what you do better or more efficiently and it means nothing for your total survival of someone gets a fork in your eye. The same applies to your enemies. Bruce Lee at 19 is dangerous, at 25 he's deadly, but a fireball doesn't care either way. When he develops, he does so "horizontally", he expands what he can do, but always stays within the relative limits of his phenotype. An old dragon can barf in your face and kill you, even if you've been training as a warrior your whole life. That was the greatest disappointment I got with the Witcher 3. Since they went for action convert, they could have gone with an area-hit system with no HP bloat, where what mattered would be what actually happened and not when an HP sack was depleted. Leveling up should just be about unlocking ways for Geralt to fight more efficiently, instead of a gate for areas with the same enemies with larger numbers on top of their heads.It is tempting to fantasize about how these things could be done in a perfect world, but I suppose there's probably good reason that what one tends to see is either simplistic level scaling or nothing at all.How would you approach a situation where the player skips a part of the world that is supposed to be "easy", and then cones back to it? Fallout didn't do level scaling, except in some random encounters (if that at all). Your answer made me wonder if Tyranny will have Kotor-style hubs, where the player can choose the order of visit in a large area, and some kind of level scaling seems inevitable (to me).
I don't think of myself as much of a designer, but I guess this is the approach I'd take to the problem you describe:
(1) Is there any reason to permit the player to come back to a part of the world that was supposed to be easy? For example, say you have a standard Chosen One plot where the player starts as a farmboy in Peasanton and there is a gang of bandits lurking in the forest whom he's supposed to kill shortly after a stranger delivers to the farmboy his long-dead father's magic sword. If the player bypasses this "dungeon" (probably a small forest map with a handful of encounters and some trash gear that is only marginally better than a jerkin and homespun breeches) during the game's introductory phase and instead heads off to the nearby city where the quest begins in earnest, what reason is there to let him return to bandit issue later in the game? This is not a Scouring of the Shire scenario -- it doesn't really matter thematically. So I would probably simply close off the quest altogether. If you return to Peasanton you learn that a handful of farmers paid off the bandits and they left, or whatever.
(2) I would consider tying enemies' power advancement to something other than the player's level. For example, to the passage of time (in a DA:O or Exile 3 scenario) or, alternatively, to quest triggers. This would require something more than stat bloating. Ideally (as in Exile 3) you'd see actual manifestations of the enemies' advance, such as villages being destroyed, but at a minimum you'd want to change the kind of enemies and perhaps change some of the encounters. In a "gather the three maps" scenario, I might have the quest giver say something to the PC such as, "For now, you've got a lead on the Sith. But as soon as they know what you're up to, they'll be sending reinforcements to guard the map." Or whatever. Then I would scale the enemies' power to the number of map pieces you've recovered. I would use this kind of scaling sparingly, though, for main-quest segments where you know the player has to complete A, B, C segments but you don't know the order.
(3) I would consider having the way in which a player interacts with an area change depending on his power level. For example, if we have to let the farmboy go back to Bandit's Vale even after leveling up, what if upon his arrival, the bandits come out and greet him and promptly surrender? Or ask him to be their chief? The level scaling considerations are almost always treated as combat considerations, such that the logic runs something like: (i) this area is designed for a level 1 farmboy to fight in; (ii) at level 20, it will not be fun to fight in this area; (iii) therefore the enemies must be scaled up to be a challenge at level 20. But that's not an inescapable logical chain. Step (iii) could instead be "therefore, at level 20, the player should resolve this area with something other than combat." On a trivial level, Earthbound (the SNES jRPG) has enemies run at the player when the player's level is lower than a certain threshold, then run AWAY from the player when it is higher than a certain threshold, and I think (although perhaps I'm conflating it with another game) at a high enough threshold the combat just autoresolves with victory.
(4) I have long felt -- though I realize that this is not the market's preference -- that RPGs have too much character advancement. Rapid character advancements pushes stories almost ineluctably toward either a Chosen One plot (where the P.C. goes from zero to god) or an amensia plot (where the P.C. goes from [god before the story begins] to zero back to god). Moreover, the constant swapping out of gear deprives equipment of its important narrative role (consider, for example, the way light sabers are handled in KOTOR: they have an initial impact, and then they are trivialized). I think flattening the advancement curve for the characters and giving them new tools but not necessarily overwhelmingly more powerful tools is a better way to go. I think Zelda and Metroid (and I really only know these series from NES to SNES, not their more recent incarnations) are illustrative of this point. In both, you get more powerful but early enemies are not totally trivial. It's more that you develop new ways of interacting with them and the environment, such that it's fun to breeze through early areas. RPGs are a bit different -- I can't think of many RPGs where the actual interactive experience of exploration/combat is very fun, it's typically a more mental thing (either a reward loop or the pleasure of executing good tactics) -- but I think there's something to be done here.
For Fallen Gods, my plan is:
(1) There are some random combats (like in a jRPG) as you explore the world. These will "scale" to the player's level, but not by changing enemy HP or stats. Rather, the idea is that as you become more powerful, certain types of enemies who would otherwise try to kill you (e.g., a starving wolf or a couple poor outlaws) wouldn't dare try now, while other enemies that wouldn't notice you when you were weak now do so (e.g., magical creatures drawn to your godly power).
(2) The overwhelming majority of encounters in the game are not random combats but events, and these won't scale. For the most part, they are "random" as well (some occur as you wander (like random jRPG combat), some you see temporarily on the map as a place you can visit but which will go away if you don't, and some stay on the map forever. These all have a large variety of approaches. Generally speaking you can choose to fight or not to fight (though many of them don't have fighting an option at all), so you can make an intelligent decision as to whether you can handle the combat. But you also get new ways to address the encounters as you get new items and followers. These are not necessarily "better" options so much as different ones -- you get more control over what resources you spend/receive which gives you more tactical flexibility, and you can get more (or at least different) lore from the encounters.
For example, in one event you encounter a gang of dwergs (dwarfs) carrying off a sleeping maiden. Dwergs were spawned from the quicksilver that spilled from the broken corpse of Karringar where Orm split her open and scraped out the Karringold to build the Skyhold (a bit of a long story). They yearn to return to the threefold nature of Karringar (the Iron Hag, the Gold Maiden, and the Silver Man), and in fact their kidnapping of the maiden (like their love of gold) is part of this need (she is blonde). Dwerg don't like to fight, so when you encounter them, they offer to pay you to leave them alone. You can take the money (which requires no skill), try to negotiate for more, try to free the girl, etc. You might also have a coin made of Karringold (a cursed but magical coin that spawns offspring from time to time) -- you can trade it for the girl. You might have a witch who -- being, in their eyes, an incarnation of the Iron Hag -- can compel them free the girl. If you try to force them to free the girl, they might slit her throat -- with healing magic you can save her. Etc. So the encounter always has a "level one" solution (take the money) but as you become more powerful (new abilities, new items, new followers) you get a variety of different ways to approach it. You are also somewhat more likely to succeed at options that were always available (like haggling for more money).
(3) As the above perhaps implies, my approach to itemization is much more in line with Zelda/Metroid than the typical cRPG. I believe there are about two dozen items in the game. About half of those are swords or armor -- you can only have one sword and one armor at a given time. While some swords/armors are better than others, each is unique. For example, the sword Firebrand has inner heat that makes it more useful against trolls (as it dries up the Trundspittle that animates them) and woodwights, and it's also a sword of historic importance (having belonged to Eirik the Fair), which means that it can be used for symbolic effect in certain encounters (for example, with Songspeakers, who are said to have betrayed and doomed Eirik with a misleading prophecy). Or there's the Wightweave armor (a silken shirt, essentially, made from wights' tresses), which provides healing and has some role in events involving woodwights. The expectation is that the player may find three swords and a couple armors in the course of a game, not that he'll churn through a new one every five minutes. Beyond the weapons/armor, items are typically not expendables or stat boosters. Instead, they have unique effects: the Wurmskin Cloak lets you talk to birds, for example, and also protects you (to some degree) from wurms' breath; Skirfir's skull lets you move easily through hills and can be used to intimidate some foes. Etc.
I think if RPGs focused more on powering up the player in this way -- a smaller number of more meaningful items/skills/etc. -- then you woud have less of a problem with level scaling. But we'll see if it works in FG.
(1) Is there any reason to permit the player to come back to a part of the world that was supposed to be easy? For example, say you have a standard Chosen One plot where the player starts as a farmboy in Peasanton and there is a gang of bandits lurking in the forest whom he's supposed to kill shortly after a stranger delivers to the farmboy his long-dead father's magic sword. If the player bypasses this "dungeon" (probably a small forest map with a handful of encounters and some trash gear that is only marginally better than a jerkin and homespun breeches) during the game's introductory phase and instead heads off to the nearby city where the quest begins in earnest, what reason is there to let him return to bandit issue later in the game? This is not a Scouring of the Shire scenario -- it doesn't really matter thematically. So I would probably simply close off the quest altogether. If you return to Peasanton you learn that a handful of farmers paid off the bandits and they left, or whatever.
(4) I have long felt -- though I realize that this is not the market's preference -- that RPGs have too much character advancement. Rapid character advancements pushes stories almost ineluctably toward either a Chosen One plot (where the P.C. goes from zero to god) or an amensia plot (where the P.C. goes from [god before the story begins] to zero back to god). Moreover, the constant swapping out of gear deprives equipment of its important narrative role (consider, for example, the way light sabers are handled in KOTOR: they have an initial impact, and then they are trivialized). I think flattening the advancement curve for the characters and giving them new tools but not necessarily overwhelmingly more powerful tools is a better way to go. I think Zelda and Metroid (and I really only know these series from NES to SNES, not their more recent incarnations) are illustrative of this point. In both, you get more powerful but early enemies are not totally trivial. It's more that you develop new ways of interacting with them and the environment, such that it's fun to breeze through early areas. RPGs are a bit different -- I can't think of many RPGs where the actual interactive experience of exploration/combat is very fun, it's typically a more mental thing (either a reward loop or the pleasure of executing good tactics) -- but I think there's something to be done here.
I think this is a well-founded concern, and one that has an analogy in what I would term the "degenerate maze" problem, especially present in jRPGs but in western RPGs with non-tiny dungeons, too. There, the issue is that while supposedly the goal of a maze is to find the shortest route to the end, in an RPG dungeon (which is often a maze with trappings around it), the goal is actually to avoid the shortest route to the end because every dead end has loot or XP at the end of it, and there's seldom a resource-management risk to gathering all of it.In RPGs with open-ish, explorable worlds, arbitrarily closing off areas is almost invariably a design gaffe that should be entertained only if it can't be avoided. In fact, when it comes to games revolving around exploration, one of the unwritten rules (in my opinion) is that backtracking is always possible. This, I think, is something that a lot of designers strain against, because a world that allows backtracking is also by and large a very static world that harshly limits narrative options; Zelda games, for instance, often rely on gimmicks like time travel and alternative worlds to adhere to this principle while still having some plot progression. That said, there's a reason they stick to it; the fact of the matter is that much of the satisfaction of those games is going off the beaten path, and if you reduce the amount of content (even if said content has become, effectively, obsolete) available to the player due to skipping ahead in the sequence, there is a particular subset of players (that is to say, me) who will, deep in their bones, feel that they have been punished for not sticking to the designer's "sequence".
Of course, this gets worse the more narrative there is in the game, because while a dungeon with some random loot might be give-or-take, some RPG players (again, I'm mostly talking about me) are often very motivated to maximise the amount of story they get out of a game and will even meta-game to find an optimal path. I remember finding this very distracting in Bloodborne, where there's a distasteful dissonance between progressing through the game looking for fun stuff and figuring out the "correct" path to trigger various NPC quests and whatever, some of which were quite obtuse. The problem isn't even that the game had missable content, but that said content is particularly easy to miss while playing the game in what is meant to be the most fun way to play it.
I think FG will probably prove a disappointment (what doesn't!), but your description is more or less right.The reason why I say that this wouldn't be a problem in Fallen Gods is that, if I understand correctly, it's a game composed out of randomized (and unique) set-pieces, in which case there wouldn't really be a way to metagame in the first place; the player has no ultimate control over the order in which he encounters scenes, which, in my opinion, an excellent state of affairs. This would mean that the player's concerns would be more generally strategic, trying to increase the amount of options available to him overall and, in general, win the game, rather than to create an optimal path through a predetermined scenario. This is why I think that more RPGs, if their conceit has to do with "choices", should try mixing in significant strategy and sim elements with clear victory and fail conditions. Which is not to say that exploration-based RPGs with a fixed pre-determined scenario are bad (they're great, actually), but they should stick to what works in them, and arbitrarily or even purposely closing off significant content is at cross purposes with what makes them fun to play.
That's not so. There are plenty of games -- not just RPGs, but games like Metroid or Zelda -- where you have some degree of nonlinearity without scaling. I don't recall appreciable level scaling in Fallout (though there may have been some?) or Arcanum or Vogel's games, not to mention a lack of level scaling in many older RPGs that were also nonlinear. Level scaling is only needed for a completely open world where the player can go anywhere in any order without a difficulty spike. It may also be necessary in a really large game where, absent level scaling, a large percentage of the world will be trivially easy by the time the player gets there.You can't have any degree of non linearity in exploration, without some sort of level adjustment for the enemies. If done right, nobody even notices it exists.
A key aspect of great nonlinear games is the sense that some areas are nominally accessible but perilous. The player has to make cost-benefit decisions as to whether it's worth going someplace dangerous now, or whether it is better to come back later. It's particularly nice if you introduce some cost to abstaining -- thus, the player can't always be confident that he'll be able to come back later. But you definitely want at least some portions where you power up, then come back to challenging areas and crush your foes -- at least you do if you want a traditional hero-story arc to the game.
Level scaling is basically an easy way to insure that you don't need to balance your game taking account of nonlinearity -- it's like throwing up your hands and saying that everyone gets a passing grade because grading is too hard: it doesn't solve the problem so much as obliterates the very possibility of a solution.
I think this is a well-founded concern, and one that has an analogy in what I would term the "degenerate maze" problem, especially present in jRPGs but in western RPGs with non-tiny dungeons, too. There, the issue is that while supposedly the goal of a maze is to find the shortest route to the end, in an RPG dungeon (which is often a maze with trappings around it), the goal is actually to avoid the shortest route to the end because every dead end has loot or XP at the end of it, and there's seldom a resource-management risk to gathering all of it.In RPGs with open-ish, explorable worlds, arbitrarily closing off areas is almost invariably a design gaffe that should be entertained only if it can't be avoided. ... Of course, this gets worse the more narrative there is in the game, because while a dungeon with some random loot might be give-or-take, some RPG players (again, I'm mostly talking about me) are often very motivated to maximise the amount of story they get out of a game and will even meta-game to find an optimal path. I remember finding this very distracting in Bloodborne, where there's a distasteful dissonance between progressing through the game looking for fun stuff and figuring out the "correct" path to trigger various NPC quests and whatever, some of which were quite obtuse. The problem isn't even that the game had missable content, but that said content is particularly easy to miss while playing the game in what is meant to be the most fun way to play it.
As long as we're still in fantasyland, what I'd say is that the "degenerate maze" problem shows that the real problem is a lack of resource-management risk. If it is possible to fail to complete the maze because you hit too many dead ends, then players will actually be incentivized to solve the maze properly. Similarly, if dilly-dallying around just to vacuum up every bit of early content has some downside (for example, losing out on content or preferences down the line), then you can force players to make reasonable choices with how they behave, rather than encouraging them to play degenerately.
Of course all of this depends on what kind of game you're making. If you're primarily making a kind of pastoral exploration game without a sense of urgency, then obviously such a design step would be a bad idea. You'd also have to make sure that you avoid walking-dead scenarios, which I view as anathema in anything other than a procedural RPG.
Still, I guess I'd stand by my point, which is in any game, you have to ask, "Why should you allow the player to come here?" I don't think, "Because we designed a map and recorded voice over" is a great answer. There has to be some kind of engaging gameplay, or narrative, to justify it. Since level-scaled opponents tend not (IMHO) to produce engaging gameplay, and since level-scaled opponents tend to break the narrative, I don't think a solution to my bandit camp scenario is to scale the enemies. Changing the nature of the content there is a better approach, though, again, this is really fantasy land. Still, my view is that if revisting an area is just going to yield either a ghost town or filler crap, it's probably best to close it off.