I’m pretty confident that demi-human level limits, at least in AD&D 1st edition and onwards, never worked effectively to balance demi-human powers because the experience point system was rigged to ensure that multi-classing as a demi-human was massively unbalanced. I am writing this theory out by memory, and I’m aware that it has been done over a bit in other blogs recently, but I can’t find any discussion of the effect of the early incarnations of the xp system on multi-class PCs. I recall when I was playing AD&D back in the day that I became acutely aware of the imbalance in the system pretty quickly, but it hadn’t occurred to me for years since I played it, and I’m going on my memory of the advancement tables in writing this post so I may be completely wrong, but here goes…

Level advancement in AD&D was basically geometric, so for example you needed 2500xp to get to 2nd level, 5000 xp for 3rd, 10000 for 4th, etc. The amounts required doubled each level until about 12th, though there were a few levels in the middle where some classes went up in an arithmetic fashion, the most striking being (from memory) wizards, who between 6th and 9th level didn’t double in required xp. Multi-class characters required twice as many xps to gain a level. The big upshot of this is that a single-class character will, for the majority of the game, be only 1 level higher than the two classes in a multi-class character. For example, you could be a 4th level human fighter for the same amount of experience points as you could be a 3rd/3rd level Dwarven Fighter/cleric. In general, though, the benefits of these levels are essentially additive[1], and the natural assumption most people make is that a a 3rd/3rd level multi-class character is closer to a 6th level single-class than a 4th level single-class.

This means that until about 9th level (roughly when the geometric advancement stops) a multi-class character will be approximately twice as powerful as a single-class character. A triple-class character will be even better – a 5th level single-class fighter would be roughly equivalent to a 4th/4th/3rd level fighter/cleric/thief by dint of this geometric progression. Once the characters reach (about) 9th level this distinction stops, and the multi-class characters go up in levels at the same rate as the single-class characters, so a 15th level fighter would be (roughly) the same as a 12th/12th level fighter/cleric.

Demi-human level limits serve to “balance” the extra powers of demi-humans by stopping them from achieving epic levels. But consider even harsh limits like those on a halfling fighter/cleric (roughly 7th level and 6th level, I recall). The halfling fighter/cleric will reach this limit when a human roughly reaches 8th level. In order for the human to gain a numerically equivalent set of levels they will have to adventure for another 5 levels, i.e. the human remains weaker than the halfling until 13th level. For an elf magic-user/thief with good stats, the elf can probably reach 11th/10th level, so from 12th level until 22nd level a single-class human will be underpowered relative to this elf. Given most campaigns never reach 20th level, but the power imbalance starts at 3rd level and just gets worse until 9th level, this is a pretty blunt and ineffective tool.

I think D&D 3.5 fixed this by making the level progression arithmetic and making multi-classing possible for everyone. This is a much more effective balance on the power of demi-humans than giving them level limits which occur too late to practically affect the most significant problems, and probably never become practically applicable for the majority of parties.

I’d like to add the disclaimer that this post is based on my memory of a game I haven’t looked at in maybe 15 years, so any mistakes in the content should be seen as speculative revision. I recall being really vexed by the simultaneous problem of demi-human level limits on the one hand, and overpowered multi-class characters on the other, and I may be wrong in all the particulars. A lot of beer has flowed under the bridge since my last chaotic neutral magic-user thief freed the prisoners and killed the ogre…

fn1: in terms of THAC0 and hit points it wasn’t, but in exchange you got to start the campaign with spells and combat powers. So a 1st/1st level fighter cleric was basically a cleric with better THAC0 and hit points than a cleric; or a fighter with 1 less hp (on average) and cure light wounds. A 2nd/2nd level fighter cleric has the THAC0 of a 2nd level fighter and but has the spells of a 2nd level cleric. This fighter’s single-class comrade will be a 3rd level fighter, so his/her THAC0 will be one better and he/she will have 1d10+2 more hps on average. At 7th level, this fighter has 7d10 hps (mean 38.5 hps) and a THAC0 of 14, with no spells. The 6th/6th level Dwarf has 6d9 hps (mean 30hps), a THAC0 of 15, and the spells of a 6th level cleric. I’d rather play a character with 8 less hps (on average), 1 higher THAC0, and about 7 spells, as well as undead turning abilities,  personally. There is almost 0 mechanical advantage to any other choice. And I think it’s even worse if you play 3 classes, because the reduction in hps and THAC0 is negligible but you gain all the 3rd classes abilities. The classic would be a cleric/magic-user/thief, so you get double the spells of a mage, and better THAC0 and hps than a single-classed thief with the same xps.

Posted in ,

15 responses to “Were AD&D demi-human level limits irrelevant for game balance?”

  1. Paul Entwisle Avatar
    Paul Entwisle

    My memory of the multi-class and demi-human systems was that the level limits were there to compensate for the life span, the special abilities and bonus stat points of the demi-human characters as much as the ability to have multiple classes.

    This was because the D&D designers didn’t want D&D to suffer from the “Elder Problem” the way Tolkien’s work (or White-Wolf’s oWoD Vampre) does. Legolas is a thousand years [1] older than the rest of the fellowship (barring Gandalf) and he’s roughly that many levels higher too. Imposing a cap on demi-humans (which, from memory appled regardless of whether they were single or multi-class) means that the DM doesn’t need to explain why the world isn’t ruled by level 50 elvish wizards that humans can never catch in their lifetime.

    Your analysis also overlooks the dual classing rules for humans (AD&D 2nd Ed). Those rules made human’s vastly more powerful – demi-humans could multi-class, sacrificing some mechanical edge to have a grab bag of powers, but humans could take a class and grab enough of it to be useful then move on without having to sink more xp into it. So given the geometric xp progression a human player who thought his game would go to (say) level 12 would be best of starting as a fighter, going to level 9 (maximising his hp and getting access to the fighter exceptional strength bonus) then move to become a mage. He’d end up with enough xp to be about level 11 as a mage (and may even make 12 depending on how deep into that level the game got).

    The downsides of dual-classing were 1. You didn’t have access to your first class while leveling the second (but you just rode your parties coat tails to blow through the low levels) and 2. You had to have high stats to do it (another example of D&D rewarding you for aready having a kick-ass character).

    Also, taking a greater range of multi-classes actually damaged the character. A fighter/mage/cleric had an average of 4hp per level, which crippled them as a fighter and couldn’t wear armour when casting spells.

    [1] Arbitrarily chosen number.

  2. faustusnotes Avatar
    faustusnotes

    As I understand it the level limtis were simply there to achieve “balance,” due primarily to the stats and other advantages, though the Elder problem was mentioned, I think[1]. Level limits definitely prevent the Elder problem, though some kind of age-based xp penalty could work there, or introducing the geometric rate of advancement after level 20 rather than having it stop at level 10[2]. Also, the only reason for this is being an issue is the desire to balance human and non-human power, which in practice doesn’t work because no-one ever games to the level where it’s relevant[3].

    Dual-classing rules did give humans powerful options but they also came with the massive downside of being very weak for a long period of time – I don’t know if I ever met anyone who used these rules except to gain 1 level at first level, because they’re so crushingly disempowering at high levels. Can you imagine how boring it would be to play a 1st level mage when everyone else has 10 levels of spells and abilities?

    The hit point example is interesting. In your example it’s made up for by the mage having stoneskin (for an 11th level fighter/mage that’s like a 110 hp bonus). The better comparison is to think of the fighter/mage/cleric as being a mage with really good missile weapon skills, more hit points than a mage, healing powers, and a great deal more defensive powers. It’s true that multi-classing might weaken a fighter/mage relative to a fighter (in melee) until 7th level (and stoneskin) but it boosts the mage massively. Also, at early levels Mage Armour is better than the armour most fighters can afford. (Also, 1.5 hps average difference over 11 levels is only 18 hps, for the sacrifice of which your fighter/mage/cleric gets turning, and about 30 spells, many of them healing/protective).

    Level limits are a very blunt instrument, in short, to solve a problem much better solved by careful consideration of the level progression process.

    fn1: note that no-one makes this complaint about MERP or RM – it’s always glossed over in practice.
    fn2:Actually, now I think about it, it’s awesomely stupid to have a system where going from 2 9th level spells to 3 9th level spells is an arithmetic process, but going from 2 3rd level spells to 3 3rd level spells is geometric. Surely it would be better to have the earlier levels arithmetic and the later ones geometric? You’d kill both problems with one stone by having the xp requirement for every level over 20 being a doubling. Then realistically no elf would be more than a few levels higher than a human, and it would be practically irrelevant and solve the early problem of mage survivability/fun.
    fn3: And I happen to know that you prefer practical outcomes over symbolism…

  3. Paul Entwisle Avatar
    Paul Entwisle

    I played a number of games when I was younger where we regularly got to high levels (largely to to a total munchkin-ist approach to gaming, but still).

    “I don’t know if I ever met anyone who used these rules except to gain 1 level at first level, because they’re so crushingly disempowering at high levels. Can you imagine how boring it would be to play a 1st level mage when everyone else has 10 levels of spells and abilities?”
    In practice, the geometric progression at low levels means that you can catch your 10th level friends again from 1st level in 1 of their levels – so say 4 to 8 sessions. This approach assumes you have an average number of sessions between leveling up, but I found it to be pretty true due to the increased xp awards at high levels.

    “The better comparison is to think of the fighter/mage/cleric as being a mage with really good missile weapon skills, more hit points than a mage, healing powers, and a great deal more defensive powers.”
    For me, this is an example of a synergy I can see the advantage to, but don’t really want. If you’re using missile weapons then your not casting spells. If you’re using missile weapons then wy do you care how many hit points you have – no one should be hitting you. In practice a mage/cleric has the same important advantages, which are a combination of damaging magic and healing magic, making them a great generalist, but the fighter element is just an xp sink.

    I also forgot to mention last time that I find the 3.5e approach to multiclassing to be OK for skill and combat based classes, but crippling for spell based classes. Taking one level in rogue is an OK idea for a fighter, you get more skill points for that one level, the hp difference is minimal and you might not even suffer a BAB difference [1]. By contrast a wizard making the same choice of a level of rogue is not sacrificing something that the rogue level can compensate for – they effectively trade the advantages of their highest potential level (i.e. being a 13th level wiard instead of a 12th) for the advantages of being a first level rogue. Given the tendancy for later levels to have cooler things (higher level spells, better feats that you meet the criteria for, other special abilities in classes like druid), you’re not really getting a good deal.

    [1] I can’t recall the BAB for the first level of rogue.

  4. faustusnotes Avatar
    faustusnotes

    I think I agree about the spell-based classes in D&D3.5. Whenever I think about multi-classing a spell-user, it’s either 1 level as fighter then a cleric (so the cleric can get cleave or such-like); or it’s a few levels of spell-user with a non-armour using skill class (either rogue or monk) to give them some additional powers. For example a 3rd level sorcerer followed by becoming a thief gives the thief access to mage armour, invisibility, maybe expeditious retreat, etc. These are really useful spells to have on the tip of your tongue. Similarly for the monk. I can see the value of getting to 7th level sorcerer/7th level monk for the stoneskin benefits; and obviously a few levels of cleric never hurt anyone. But in general one doesn’t want to hold back a spell-caster.

    I think this is partly the fault of D&D’s extremely weird spell system.

  5. Paul Entwisle Avatar
    Paul Entwisle

    “7th level sorcerer/7th level monk for the stoneskin benefits”
    So you can cast stoneskin what? Once a day? I know its a good spell, but your other option in 3.5 was to be a level 14 monk. I can’t recall the monk 3.5 abilities but I bet the ones between 8 and 14 inclusive aren’t too shabby. And if you’re playing with hard level caps (i.e. no epic levels) then in a long running campaign then you’ve actually denied yourself the monk 14 to 20 level abilities – this is more of a concern in a computer game, but there may be people who do table top gaming like that…

    I’d submit the problem arises due to the special abilities associated with the higher levels. These are balanced so that higher investment is inherently more valuable. Spell levels are one of the common examples where a 9th level spells are absolutely better than the 1st level spells, but the monk’s 20th level ability (Perfect body?) is better than any first level ability in the game – being monk 19/anything else 1 is basically a chumps game.

    Considering this I do have to note my feelings are highly biased by computer games. The high levels those tend to allow combined with level caps strongly encourage maximising your available special abilities (including spells) by specialising. Fighter-types by comparison can take anything that enables their core ability to advance (i.e. BAB must go up by +1).

  6. Paul Entwisle Avatar
    Paul Entwisle

    With everything I’ve said, I do have to admit that your central premise that demi-human level limits are stupid is right. The Elder problem is a valid concern, but it’d be better dealt with through other means (geometric progression above 20, story based reasons, insanity rules for those who gaze too long into the abyss, etc).

    I’d go further and say that the Effective Character Level system for powerful races in 3.5 also irritated me for the same reason I’ve explain about spell casters – sacrificing your highest level to get race abilities makes sense at level 5, but with level caps you find you’ve lost your best (specialised) abilities for stuff that’s a toy at higher levels.

  7. faustusnotes Avatar
    faustusnotes

    I agree with you about the spell sacrifice issue but I do think that you’re being too biassed by computer games with their level limits – the 19/1 exchange is not an issue if you can go (arithmetically) to level 21.

    The huge difference between lvl 9 and lvl 1 spells – or even between the power you gain in going from lvl 8 to lvl 9 vs. lvl 1 to lvl 2 – is a strong argument in favour of reversing the original D&D xp progression system, so it’s geometric at later levels.

    It’s not so much that I think the level limits are stupid, I just don’t think that the elder problem was the central concern of the designers and they could have done a better job by fixing the xp system. Then a 7th level fighter would match a 4/4 lvl fighter/magic user, which is more sensible, and the elder problem would go away.

    I agree about the effect of the system in 3.5 on multi-class wizards. Though I do think the main reason people multi-class is to have fun.

    I also think that in 3.5, multi-classing for fun became a waste of time because there were so many class choices in the supplement that a straight combination of those in the main book became a waste of time.

  8. Paul Entwisle Avatar
    Paul Entwisle

    Do you really think a level 4/4 fighter/magic user is the match of a level 7 fighter? I’d expect the fighter/mage to get steamrollered in a contest. But this is also a result of irregular jumps in power for some classes (especially mages) combined with some difficult evaluating the actual effective level of a multi-class character – especially given my note on irregular differences in power gained from levelling.

    For the xp system, I don’t think you can look at it in isolation. The actual question you need to confront is: How often do you want characters to level up? Then you need to design your xp requirements and monster/gold xp rewards to try to match that. The xp structure itself is then an additional factor for multi/dual classes so you can balance it against single classes, but it should be the final factor.

    For example if a level 8 character fights gorgons and can expect to level every 5 sessions and a level 20 character fights dragons and can expect to level every 5 sessions, then the Elder problem is a big issue as level 40 elves will fight gods/Sauron and level every 5 or so sessions…

  9. faustusnotes Avatar
    faustusnotes

    The 4/4 vs. 7 challenge depends on the conflict environment, I think. If they are warned of each others’ presence then the fighter/magic user has a fairly good chance at winning – having the ability to level some kind of combat levelling effect (like fear) at the other guy, and then to deal out hit point reductions and/or buffs. If they start in too close, though, it’s a big gamble.

    I wonder if this is part of the reason that supplements have always been a big part of D&D (besides the obvious)? Because it’s so poorly balanced that they are constantly coming up with fixes, then fixes for the fixes.

    My view on the speed of leveling is that the earlier levels should probably be easier to advance through, if the later levels have geometrically more powerful spells (like time stop). But the spells in D&D are quite haphazard, so maybe there’s no easy answer to this in that system.

  10. Paul Entwisle Avatar
    Paul Entwisle

    My leveling preference would be for slowing increasing amounts of time between levels. Say 4 sessions at 1st level and 8 sessions at 20th. But that should vary depending on how often you game and what story you’re telling. I red box D&D game where dungeon clearing is a major point probably wants more frequent advancement than a game (using the same system) where kingdom’s hang in the balance. But for me that’s a function of what I find interesting and the rewards that would be needed to keep my attention. I like dungeon crawls, but I like them for levelling up, by contrast giving me control over a kingdom’s trade policy would probably interest me more than getting the next level of spells.

    On that note, does anyone want to play Birthright? 😉

  11. Robert Fisher Avatar

    I read several of Gygax’s responses to the question of level limits and, to my recollection, he never claimed they were for balance. He said they were to create a human-centric world. i.e. The elder problem.

  12. faustusnotes Avatar
    faustusnotes

    As we have discussed, though, they fail at the elder problem too, or are at least a blunt stick, with the method better handled by reversing the geometric nature of the leveling system. I also think that the notion of human-centric in his responses was very much about balance – not wanting the elves to overpower the humans is about balance.

  13. Big Eye…

    […]Were AD&D demi-human level limits irrelevant for game balance? « Compromise and Conceit[…]…

  14. faustusnotes Avatar
    faustusnotes

    I’m preserving this trackback here as an example of the pointless madness that is the internet. Someone needs to update their machine learning techniques.

  15. Little One Avatar
    Little One

    3.5 sucks–the only reason the system is reasonable is because it makes all classes only slightly different from each other in certain terms. Sure, be a mage, but a cleric still gets magic missile.

Leave a comment