Friday, July 30, 2010

Level Scaling - Alternatives and Issues

In this final post, I'll discuss some of the means for implementing scaling apart from pure level scaling.  I must firstly recommend this post from a friend discussing level scaling and leveling in a very original fashion. It's definitely a novel approach that I'm very interested in.

But sticking within the conventional realms of the "standard" RPG, we're a little more hamstrung by the currently accepted mechanics. So what else can be done besides pure level scaling?

Scaling by Number
Instead of making enemies more powerful, let's just make more of them! This might sound like a good idea, but generally it's not that useful except to make the player feel as though they are really powerful. Weakling monsters, even in numbers, do not pose a threat to high level heroes. They simply won't do enough damage, won't hit, or be killed too quickly. Imagine your party of adventurers from Baldur's Gate 2 being swarmed by a thousand gibberlings. The warriors would kill several each round automatically because the creatures would be such a low level, and mages could obliterate hordes at a time with a few cloudkill spells. Of course, I've picked an exaggerated example here, but the principle is the same. Unless the weak monsters are made more powerful, it's possible that they won't be dangerous anyway.

Dynasty Warriors lives by the ethos "The more the merrier"

Also, as noted previously, there's the consideration of the comparative difficulty of a battle based on plot considerations, which is something that level scaling destroys. However, if scaling is done by number instead of level, then this problem isn't avoided anyway. If doesn't level up, but instead has a small army at his disposal, that's no less ridiculous than him being more powerful than a previous enemy. Arbitrarily increasing the number of enemies is no less plot/realism breaking than increasing the level of the enemies if the overall effect of retaining difficulty is the same.  Plot considerations and the "reality" of the game world both have to take a hit to prevent player boredom from having ridiculously easy encounters in a non-linear game. The only possible means to rectify that would be to have dialogue/descriptions change based on a players level as well. This would be an inordinate amount of a work and I'm afraid I'd have to strongly say it would be a complete waste of time.

The last point about increasing monster army sizes is a technical one. Each monster that is on screen requires both extra processing from the graphics card and the CPU to calculate and display their actions. If you have too many monsters, the game is going to slow down and become jerky and unresponsive. Bad performance is something that earns terrible ire from both reviewers and gamers, so much like boredom, it must be avoided at all costs.

Partial/Variant Scaling
In Mass Effect 1, higher difficulty levels introduced more scaling of enemies. However, this simply increased enemy hit points significantly, but this didn't increase the difficulty at all.  But, the thing that it did right was that scaling was performed relative to an enemy's own "power". That is to say, at lower levels, only bosses were scaled, then sub-bosses, then standard enemies.  By implementing a system that applies scaling using a multiplier against an enemy's "difficulty" or "ranking", you can provide a more even distribution of power.

Colour coded badness

Now, Dragon Age does attempt to do this, but I still don't think it succeeds entirely. In order to make the scaling a little more organic, I believe there should be some degree of randomness introduced into the system. All "grunt" enemies should not be created equal, not should all "sub-bosses" and so forth. Unless we're talking about unique creatures that should have a consistent(ly high) power level, the game engine should allow for some randomness in the difficulty of each creature, so that the difficulty of "standard" fights throughout an area varies. This is both more interesting and fun for the player, as each fight won't play out exactly the same.

Ability Scaling
Instead of simply making enemies more powerful by giving them hitpoints or making them do more damage, how about making them behave differently?  Note that I'm not talking about modifying AI here, as that is a horrendous task. Making decent AI is difficult enough, but then modifying it as the game continues is a nightmare, not to mention, what possible cause is there for the enemies to continually become smarter? (Yes, I'm sure you can think of a "logical" plot device to do so)

So giving enemies new powers or techniques for attack the enemy is one way of mixing up the difficulty of combat. This may even come down to things that force the player to adopt/use new tactics based on the items/skills they have acquired in their adventures. One option would be to implement resistances on enemies, forcing players to use items like poisons, traps or more powerful spells in order to deal (sufficient) damage to them.  Or perhaps environmental issues could be used, e.g. spawning more traps in an area for higher level parties.

Bigger and badder spiders should have bigger and badder webs

Dedicated/Multi-part Scaling
As was suggested in a comment in my first post on this subject, another means is to specifically cater for the various points at which the player can encounter a fight.  Now, this is a great suggestion, as it offers the potential for offering customised encounters without the bland approach of standard scaling.  Unfortunately, it can be very time consuming.  Dedicated scaling is a manual process, going through a modifying all the basic elements of an encounter based on the level of the character. Balancing multiple factors via an algorithm while taking into consideration plot issues is typically not something that can be done algorithmically (at least not easily) so it comes down to a designer to edit them all by hand.

One of the boss encounters I created for my NWN2 module Fate of a City was customised based upon the level of the player and their party members (if any). This was probably one of the most difficult (if not the most difficult) and time consuming fights to balance in the entire 10-15 hour module. This was because I was changing the power, level, abilities and number of enemies based on those factors. Dealing with all those different variables meant working out what was reasonable and what was not took a lot of trial and error.

A boss from Fate of a City, possibly alone, possibly with friends.

Going through that process for multiple boss fights, not to mention minor encounters, across an even wider range of levels is not something I would particularly want to attempt. The tradeoff is that it potentially offers the best opportunity to provide a consistently difficult and interesting challenge, but the amount of time spent performing iterations of design and playtesting mean that it is largely infeasible except for very special encounters.

For better or worse, I haven't strayed too far from the "conventional" means of increasing difficulty in battles. Ultimately, combat in RPGs comes down to some sort of equation, and we have a limited means to play with the numbers in that equation. If we can have mechanics that remove the simple number aspect and require a little more from the player like the unpredictability of players in a FPS, or the varied strategies used in an RTS, then perhaps we could begin to experiment a little more liberally with our scaling options.


  1. "If we can have mechanics that remove the simple number aspect and require a little more from the player like the unpredictability of players in a FPS, or the varied strategies used in an RTS, then perhaps we could begin to experiment a little more liberally with our scaling options."

    We can! You see a lot of this in MMOs, where party size scaling becomes an issue as well as difficulty, gear etc. Debates about why 10 man raiding is inherently easier and less interesting than 25 are a current favourite, but it's been going on for some time. Encounter design has a lot of levers to pull.

    It's quite possible to implement those sort of mechanics in DA, too. Classic Week has six hours worth of MMO-style bosses, many of them remixes of specific WoW mechanics.

    Of course the problem you face then is you're really scaling for player skill rather than character power, particularly in DA where awareness of a handful of critical mechanics fiercely stratifies the player base. I'm planning on implementing Heroic and Casual modes in my next module to cater for that, but I'm not sure how sophisticated that will end up being. It's hard to see a way around hand-tuning.

    On the plus side I don't think it's important to tune trash - if it's not a boss, it's a pacing mechanic, not a challenge.

    - Mengtzu

  2. An interesting thought, though I'd potentially argue as soon as we bring MMOs into the discussion, we're talking a whole new ball game. Raids are designed with specific level ranges in mind, and you've got multiple players with a very broad range of talents to deal with situations.

    If we're talking encounter design in general, then I suppose there is a lot more we can include in terms of general scenarios - but that tends to be some fancy scripting and behavior to make individual boss fights more interesting. As far as I'm aware, that type of thing isn't scaled. That said, the only MMO I've played is WoW and that was before Burning Crusade, so I'll confess to being somewhat ignorant of encounter content. MMOs and my obsessive gaming nature do not mesh well together.

    I'd definitely like to see more unique fights introduced to single player RPGs. I had a planned mod for NWN2 that I eventually ditched due to the shrinking community and the imminent release of Dragon Age which had numerous "special" fights - I'd even done the scripting for quite a number of them. I think the only difficulty is that in an MMO, you expect to fail, and the games are set such that you typically will. In a standard single player RPG, the player can't die or their adventure is over. That expectation alone makes the encounter dynamics a lot different between MMOs and single player RPGs.

    Heh, but look at me, I'm now writing an essay on encounter design... I could have used this as part of a whole new blog post!