Spell Casting Options

There are many ways spell casting may be modelled in Echelon.

Goals of a Casting System

I have several desirable goals for the casting system I want to use.  Others may have other goals, though.

  • Power gained should be on par with that gained from combat, for the same cost in talents or other resources.
  • For simplicity, at least at first, it would be good if existing spell descriptions can be used.
  • Decision-making should be simple, with character design choices being in reasonably-sized chunks and with simple choices when casting.

“Class-Based Casting”

This model uses existing spell descriptions.

This model of spell casting is very simple, but probably broken.  For each D&D 3.x casting class there are two talents (“Wizard Casting” and “Improved Wizard Casting” that produce a Training Bonus equal to the tier of the slot (so Expert Wizard Casting and Expert Improved Wizard Casting would give +2 each, for +4 total).  The character has spell slots and spell knowledge as a character with a number of levels in the chosen class equal to his Level Bonus plus his Training Bonus.

This is broken in quite a few ways, I think.  For a single low-tier talent a mid- to high-level character can have significant casting ability.  For example, a Champion-tier character — 18th level, say, so LB +9 — with Expert Wizard Training would have the casting ability of a Wiz11 in D&D 3.x.  This is out of line.  Even if the benefits are reduced by one (no Basic casters at all, so Expert Wizard Training gives +1 instead of +2) the Champion would have the casting ability of a Wiz10.


This is wickedly overpowered and I can’t imagine a character above mid-level who doesn’t take at least a little bit of caster training associated with at least one class (and dedicated casters will likely take as many of them as possible — as written it would be possible to keep the casting ability of three classes at close to the character’s level).

Unified Casting

This model uses existing spell descriptions.

There’s a variation not too far from the first that might work better.  All characters (start with) the same spell slot table (wizard, for sake of simplicity).  A pair of general Caster Training and Improved Caster Training talents allow a character to develop his Caster Training Bonus.  The character’s caster level is equal to his Level Bonus plus Caster Training Bonus, and he has the spell slots of a Wizard of that level.  Additional spell slots might be gained through the use of other talents.  There are some limitations on how these may be used.

  • The character is limited to spells of a level equal to or less than his caster training bonus
    • Optionally, the character may be able to use higher-level slots for spells of the levels he has available.
  • The character is limited to ‘common spells’.  For simplicity, these are spells available to at least three core base classes (and to both arcane and divine casters).  Additional spell knowledge will be gained from other means (Spell Knowledge talents, for example).
    • Alternatively, use the ‘common spells’ defined in Arcana Unearthed and related books.  Complex and rare spells will be gained as above, from other talents.


This was my first cut at it, but I was never quite happy with it.  The core caster talent might be a little too much, but the Spell Knowledge talents I have described are kind of bland and uninteresting.  Also as above, the number of spell slots available seems much too high.

Eldritch Weaving

This model uses existing spell descriptions.

This variation is based on the Eldritch Weaver class from Green Ronin’s Advanced Player’s Manual.  There are a number of spell threads, each with about three spells per spell level (except cantrips, which are all in a single thread; I would fold them back in) and a lesser and greater power.  I imagine that each thread is implemented using two talents.  Together they provide a Training Bonus, and they have the following effects.

  • The Training Bonus controls the maximum spell level castable from the thread.  All spells accessible on the thread are known, all spells on the thread are considered ‘class spells’ (for the sake of spell completion and spell trigger items).
  • The character has one spell slot per spell level accessible, that may only be used for spells from the thread.
  • An ability appropriate to the thread and tier.  The abilities granted may be based on the Training Level (add them together before determining abilities), but I think prefer to have them considered independently.  This keeps each ability appropriate to the tier slot (adding two Heroic slots gives two ‘third tier’ abilities, not a ‘sixth tier’ ability).  This arrangement is subject to change.
    • Note that the ‘Improved Thread’ talent abilities may be a little better than the ‘Normal Thread’ talent for the tier (something I expect to do fairly regularly, actually).
    • The individual powers should be a little less powerful than normal for the tier, since the talent functionally also grants spell knowledge and spell slots.
  • The lowest tier of at least one of the talents might not give a Training Bonus at all, instead granting a different power.  This allows the use of cantrips (first-level spells) to Basic characters without allowing second-level spells to Basic characters.  This keeps things a little more in line with D&D 3.x availability.

Casters may also get an additional spell slot per spell level, up to their maximum castable level, that may be used for any accessible spell.  Caster level is as above, but on a per-thread basis.  Dabbling with a thread, spending only a low-tier slot on it, means you get a smaller number of spell slots and cannot cast the spells on that thread as effectively as those you have focused more on.

This variation has characters gain spell casting ability and powers appropriate to the spells they learn.  Characters with broader knowledge (who spend more talents learning spell threads) have more casting capacity and more caster powers.


Of the three options presented so far, I like this one best.

Thieves World (Accumulated Mana)

This model uses existing spell descriptions.

Green Ronin’s Thieves World setting describes a mana-based system where casters must build up power until enough is gathered to fuel the spell.  I’ll describe it in general terms before describing how it might be implemented in Echelon.

There are three primary casting classes (Mage, Priest, and Witch), each with different casting ability scores (Intelligence, Wisdom, and Charisma respectively).  Each of these classes has a different class spell list.  There are also a couple of secondary casting classes (Godsworn and Initiate, using Wisdom and player’s choice of Intelligence and Charisma respectively).  The primary classes have spell casting and ritual casting bonuses, Godsworn can cast priest spells but only as rituals (no spell casting bonus), Initiates can learn and cast spells from both the Mage and Witch class lists but cannot use rituals — and for the purpose of spell completion and spell trigger items only has known or familiar spells on the ‘class list’).

When casting, any spell level may be attempted.  However, only spells up to a level equal to one-half the character’s caster level (minimum first level… though I’d be tempted to change that) for the relevant spell list (highest applicable in the case of multiclass casters) may be done ‘safely’.  Casting spells has a price.  On completion of casting a safe spell the caster takes nonlethal damage equal to the spell level (a number of free cantrips per day equal to the casting ability score, one point of nonlethal damage per cantrip after that).  On completion of an unsafe spell the caster takes lethal damage equal to twice the spell level.  You can overcast, but it hurts, literally.

When casting a spell, the character makes a casting check (d20 + spell casting bonus + key ability modifier + other modifiers) for each time unit of the spell (base casting time, usually a standard action) and adds the check total to his accumulated mana.  However, if the base die roll (the value on the d20) is too low you instead subtract this base value from your total.  The die range that causes this ranges from 1-19 in an area of ‘no mana’ through 1-10 for ‘low mana’, 1-5 for ‘normal mana’, 1 for ‘high mana’ and no chance in an area of ‘maximum mana’.  There is also the possibility of a spectacular success (again based on mana level) where you get to immediately roll again and add that amount to the total (including all modifiers and bonuses).

When the accumulated value crosses a threshold (equal to ten times (spell level + 1)) the spell goes off.  If the accumulated value becomes negative (failure on the first check or through repeated failure) the caster suffers a spectacular failure (which may lead to madness, a curse, strange weather, inflicted condition such as fatigued or nauseated, unexpectedly summoned monster, and so on).

Familiar spells take even longer, the caster must prepare (mentally review the spell, focus his divine power, etc.) for a number of standard actions equal to the spell level (or full-round actions equal to one-half the spell level) before starting to cast.  Unknown spells require two rounds of preparation per spell level.

Ritual casting works in a fairly similar manner, but checks are made once per minute (or the base casting time, if longer).  However, spell duration (if measured in rounds to hours, but not days) is increased to the next-higher unit (rounds -> minutes, minutes -> turns (10 minutes), turns -> hours, hours -> days), multiple casters may cooperate (a number of helpers equal to the primary caster’s key ability score modifier, who each make a check with the primary does and on success add their key ability score modifier +2 and on failure subtract 2 from their primary’s result; helpers need not be the same class as the primary but must have the spell being cast as a known or familiar spell), and the price (nonlethal or lethal damage) is split evenly between all participants, with excess going to the primary.

There are also rules for casting options such as hasty casting (increased risk of failure for each check in exchange for bigger check results on success) and cautious casting (decreased risk of failure per check in exchange for lower check results on success — if you can completely remove the failure range you can and will eventually cast the spells).  There are rules for varying mana levels (the duration of ‘permanent’ spells is reduced if they are too high-level for the local mana level), and so on… but honestly I’ve concluded this particular implementation is way more hassle than I want for Echelon.

Implementation in Echelon

That whole mess is pretty convoluted, rather more complex than I really want for Echelon.  However, my analysis suggests that low-level familiar spells can usually be cast in a round or two (even by low-level casters) while familiar high-level spells cast by high-level casters can usually be cast within three or four rounds (it doesn’t take long for d20+20+othermods to add up to 100, even with 25% chance of subtracting 1-5 from the chance).  So, a single check may be sufficient to show enough power has been collected.  Set the DC such that low-level spells are easier than high-level spells (at the appropriate caster levels) because it takes longer for higher-level spells and you should have it.

Instead of causing nonlethal or lethal damage, have the spell draw from the appropriate ability pool (the spell level for safe spells, (level)d6 for unsafe spells.  A first-level caster could conceivably try to cast fireball (fourth-level spell in Echelon) but if he beats the spell DC he spends 4d6 Intelligence pool to do so.  Since he likely has only 7-9 (assuming a ‘smart’ wizard… though I question that) odds are good he’s toasted his brain a bit — if he has 7 Int Pool and takes 14 damage he’s down two conditions and might be whatever-is-down-there, if he’s unfortunate enough to roll 24 he’s taking hit point damage because he exceeded the double-condition threshold.  That would suck.  The same character, had he tried it at eighth level, can safely cast fireball and pay four points from his Int Pool (which is now probably 11-13 points anyway).

Varying mana levels can be modeled as modifiers to the check (if maximum mana areas give +10 to all casting checks, casting gets rather easier) or something similar to the Thieves World model where they may cause spectacular failure.  In the latter case we might still allow modification of casting chances — increasing the risk (chance of spectacular failure) you might get a bonus to the check itself… which suggests that you could now even have mixed results, where the die value says “spectacular failure” but the DC is met and the spell goes off.  This sounds interesting to me.  Conversely you could reduce risk at the cost of lowering your check result and thus the chance you get the spell off in any particular round.  So… masters can more reliably get spells off in similar time if even after minimizing risk they can make the check DC.


I don’t think I’ll go this way for a few reasons.  Primary among them, players expect spells will work (just as swords don’t fail — their wielders miss) and taking that away probably would not be popular… even though it just means you spend longer casting each spell since you don’t spend slots or anything casting them.  I do like how certain parts interact (such as the ‘catastrophic success’ that is now possible) and may explore it somewhat further.  I think this option would not be a good choice for the default rules and would be better kept in reserve for settings with “gritty magic” such as Thieves World.

Black Company (Tiered Freeform)

This model uses freeform spell construction, it does not use existing spell descriptions.

Green Ronin’s Black Company setting describes a free-form system with tiered access to spell elements, apparently somewhat similar to the epic spell seeds system.  On the face of it, this looks promising (already has tiers and talents, even!).

A caster has a casting bonus, and tiers ranging from Student Wizard or Dabbler (though they may be basically trained caster class vs. not-trained caster class) through First Tier and on to Fourth Tier.  Each tier provides bigger bonuses to the Magic Use check needed to cast a spell and access to higher-tier talents.

There are a number of talents, each of which is available only to characters of a certain tier of ability, and higher.  Each caster gains knowledge of some number of them.  Each also gives a continuous effect, such as a bonus to particular skills.

When casting a spell the caster chooses a talent and gets the base effect of the talent, then augments it in various ways (increases size of effect, such as doing more damage, increases range, increases area of effect, increases duration, and so on).  This leads to a calculated Magic Use check DC.  Casting time is derived by subtracting the caster’s ranks in Magic Use and then comparing to his tier.  On casting a certain amount of ‘drain’ is done based on…


Okay, this is way more complicated than I want to get in Echelon, and I didn’t even get into ‘Blended’ spells that use multiple talents.  It makes HERO look simple for creating powers on the fly.  Which it often is, really — pick the power, the advantages and limitations, a bit of math, and you’re done… often, even most of the time, and a lot of it can be done on the fly with a bit of practice without too much looking up at the table (common modifiers are easily remembered, HERO’s a very regular system that way).

In any case, this might be worth more research if nothing better turns up, but if that happens I suspect I might look at Hero or Elements of Magic first.  Even though this has tiers and talents built into it.

Iron Heroes (Simplified Freeform)

This model uses limited freeform spell construction, it does not use existing spell descriptions.

Malhavoc’s Iron Heroes has a freeform system that’s quite a bit simpler than the Black Company version.  An arcanist gains spell mastery in each school (at different levels and grades — one primary and one secondary at first level, one tertiary at third, and access to all other schools at fifth).  The grade (primary, secondary, tertiary, other) controls the mastery rating that may be applied for spells from the school.

Each spell school (same schools as D&D 3.x) has one or more methods.  Each method is only available to arcanists with relevant mastery rating equal to or higher than the mastery rating of the method.

To cast a spell the arcanist must gather mana, shape the spell, and channel the spell.

Gathering mana just seems to happen, as long as the amount needed is less than the amount of mana available to the arcanist.  If the available mana would be negative after casting the spell the arcanist much make a Fortitude save or take temporary ability score damage to all ability scores equal to the negative value of the available mana.  I would think this is best avoided.  However, you can continue to cast even when your mana is negative, it is just more likely to hurt, more, especially since the Fortitude save DC is 15 + abs(negative mana available).  I see nothing here about recovering mana available.

To shape the spell, the arcanist chooses the method for the spell as the base for the spell (cost in mana, range, casting time, and so on).  He then augments it in various ways (based on school and method) which gives him a final mana cost (which is also used to determine the check DC needed to cast the spell).

To channel the spell, the arcanist makes a mastery check (DC = 5 + mana spent + Int modifier… I have to think this is an error — being smarter makes it harder to cast spells?).  If successful the spell works properly.  If the check fails by 10 or less the arcanist suffers a ‘moderate disaster’, which usually looks like a partial backfire — hits the wrong target, reverses the effects, and so on.  If the check fails by more than 10, the arcanist suffers a ‘major disaster’, which is much the same but worse.  Finally, a natural 1 that would otherwise succeed is a moderate disaster and a natural 1 that also fails is a major disaster.

Spell casting in Iron Heroes is deliberately risky.

Implementation in Echelon

This would be pretty easy to do in Echelon, though I’d shift things a bit.

First, each school has two training talents (Abjuration Training and Improved Abjuration Training).  The Training Bonus is used as the ‘mastery rating’, and the spell gets constructed and configured accordingly.  If you want an energy shield (Abjuration 2 spell) you can put up a force shield that fills five adjacent squares for four points, +2 per additional energy type, +2 per additional square.  Mana comes straight from an ability pool (possibly one for all spells, possibly different schools can draw from different pools).  No Fortitude save to avoid the ability score damage because I don’t have it, and because going negative automatically causes an effect.  Make the appropriate ability check and the spell goes off, if the check fails you may suffer the described effects (such as the force shield appearing in the wrong place, or inverting and causing damage).

The DCs would need to be reworked ; between the higher ability score ‘modifiers’ on checks and the Level Bonus the DCs from Iron Heroes would be inappropriate.

Alternate Implementation in Echelon

Much as above, but instead of using the Training Bonus to determine which abilities are available, use the specific entries (which is how I may do Iron Heroes in general).  If you just take Abjuration Training you could get access to the ‘odd-mastery’ powers (Arcane Shield, Absorbing Shield, Mana Vacuum, Elemental Fortification, and Arcane Membrane) and have a limited Training Bonus (which means you run greater risks when you use the higher-tier powers… but you can still try).  You have some raw ability but your training is incomplete.  If you also take Improved Abjuration Training you get the ‘even-mastery’ powers (Energy Shield, Elemental Shield, Instant Fortification, Impenetrable Wall, and Temporal Shell) and a potentially much better Training Bonus (making it safer to cast any and all abjuration spells.

It might be worth rearranging the powers slightly to better accommodate this change, but it could be a workable approach.

For what it’s worth, I am likely to take a similar approach converting the various Iron Heroes mastery feats and class abilities, so this would be consistent with that approach.


It may be just how the material was presented, but this method is much simpler than that described for Black Company (which has been described to me as “more detail than fun”).  It ties in quite well (possibly including casting costs compared to the ability pools they draw from, especially if different schools draw from different pools) with other structures already described.  It is evident that higher-tier access to a school provides better abilities (especially if the alternate implementation described above is used) but expanded mastery (improved training) is valuable as well.

I think this is my current second-place option.


I think I’d never consciously realized how many ways spell casting has been modeled in d20 rules.  I know there are some other variations I have overlooked.

Overall I see two immediately viable approaches, a couple more that are more complex than I want to get into right now but may be possible to simplify, and a couple of simple but broken approaches.

Because it allows me to reuse existing material, I am likely to keep my initial spell casting rules to the ‘Eldritch Weaving’ variation described above.  It has the benefits of using existing spell descriptions (which saves me a lot of time right now) and gives me a place to hang additional thread-specific abilities, so those who practice different types of magic actually have some different abilities outright.  I like that.

As a backup I am likely to look toward the Iron Heroes-flavored approach (in one of the two variations described above).  I lose the ability to use existing spells, but the dynamic spell generation is relatively simple to apply and the system described already fits reasonably well with existing structures.  I would likely want to adjust some details for balance reasons, and may wish to apply specific powers as with eldritch weaving, but overall this looks pretty good to me.

Third option… I”m not sure.  Possibly, just possibly, something based on the Thieves’ World model… but I think it more likely I’d look into one of the options I’ve so far ignored.

Elements of Magic, BESM d20, Expanded Psionics Handbook (which actually makes for a decent ‘magic’ system), Arcana Unearthed (which I considered as a resource for the Unified Casting option above and otherwise just a variation of the normal d20 rules… but I should re-examine it), possibly even Mutants and Masterminds or something derived wholly from HERO System.

However, for now I have a couple of immediate options to pursue.  I’d best get to it.

No related content found.


  1. hadsil

    I’m against any magic system that punishes you for the audacity of casting a spell, such as hit point damage, non-lethal or not. Just admit you hate magic and player characters doing nifty powerful stuff already and get rid of magic completely so that characters can only swing a weapon in combat in nice safe low numbers of mediocrity.
    (Colloquially speaking, not against Keith.) That is what 4E D&D did (got rid of magic completely). Pre-Saga Star Wars is what damaged you in hit points for the audacity of using the Force.

    Ars Magic, my favorite magic system, does have fatigue, but it’s not punishing. For one thing, it is possible to cast spells without causing fatigue at all, including the real powerful ones. Second, getting the intial status of “fatigued” is just a label. You suffer no mathematical game mechanics penalty of any kind. It’s repeated castings that each causes fatigue without rest in between where you start having penalties. This is necessary to prevent characters from casting spell after spell after spell after spell ad infinitum. The game needs to function despite “unlimited” spellcasting.

    I’m also against any magic system that forces you to do nothing for several rounds except casting a spell. I’m ok with particularly powerful spells taking more than one round to cast, if you must have that, but not as standard operating procedure of casting a spell. A round of a player character doing absolutely nothing is devastating. This is why in D&D stunning someone for a round is considered a powerful effect. Doing something, actions in 3E D&D terms, mean everything. Worse, by the time you finish casting the spell, the combat may be over, the reason you wanted to cast the spell may no longer apply, and/or the target you wanted to cast the spell against is dead or otherwise not in the combat any more. This is part of why GURPS magic sucks.

  2. Doug Lampert

    I tend to agree that taking HP is bad. But my pet peeve is backfires. Take a spell system with autofail and backfire mechanisms, and consider that you will probably have practiced a basic move you’re expecting to use in combat a hundred or so times, and ask how any mage survives training! These systems IME ALWAYS get the fumble chance too high for surviving apprenticeship to be plausible.

    A one in a thousand chance of an raw apprentice screwing up a spell in a disasterous way probably means there are no wizards in the world. (Ars magica has the nice feature that I can be in a magical aura area and divide by 5 rather than 2 on a spontaneous spell and there is no way I can actually screw it up, YEAH! The appretice can light a candle as soon as he learns creo ignem with no game mechanical chance that he blows up the covenant!)

    You could probably do Ars Magica style casting (but not lab research) in Echelon, a basic talent gives a score of 5 in one art, an expert talent a score of 10, a heroic talent a score of 15, ext…. This has the nice feature that with 15 arts you have a LOT of possible talents for the specialist mage to take, but that since you can max one type of magic with only two arts a gish is possible.

    You’d need to tie in with the fatigue rules and you’d need a way to learn non-spontaneous spells (that could be done using the Ars Magica rules). You’d also probably need to convert vis requirements to some more D&D flavored component requirements.

  3. On first look, Ars Magica-style magic access control (if not actual resolution) looks like it has some fair possibility. Five Techniques and ten Forms probably fit rather well with the number of talents available.

    One of the simplest models would be to define the Technique(s) and Form(s) required. Add the character’s scores in the applicable ones to determine whether the character is capable of casting the spell.

    A fifth-level Creo Terram might be castable by anyone having Creo + Terram >= 5. Whether this is Creo >= 5, Terram >= 5, or some combination. Or it might be a harder limit, where you must have 5+ in both Techique and Mode.

    This is all about access, though, and doesn’t address capacity (how many spells may be case)… but there are several possibilities there as well.

  4. Regarding hit points and backfires, a backfire or other botch doesn’t necessarily have to do hit point damage.

    Perhaps a normal failed spell just doesn’t work, and a botched spell means the spell doesn’t work but you end up paying the mana anyway (not directly harmful but may still be dangerous under the right circumstances).

    If spells can be directly harmful on a botch (fireball at ground zero on a botch rather than poor judgement and misplacement) then yeah, it gets pretty silly. A lot like combat botches resulting in self-injury (to the point of decapitating yourself with an axe!). Yes, many such injuries are possible in real life, but it doesn’t really work well to model it in-game in a simple manner, and probably not worth the complication to ‘get it right’.

  5. Pingback: KJD-IMC » Links of the Week, September 12, 2011

Leave a Reply

Your email address will not be published. Required fields are marked *