I am trying to adapt D&D 5e to a Harry Potter setting,* basically creating an adventuring group of wizards.

For the most part, I've matched up Harry Potter magic with D&D schools and have a basic set of spells available at each year/level. Eg: Lumos is obviously an Evocation cantrip/level 0 Light spell.

So, how would I go about classifying a more nebulous spell like Expecto Patronum?

Expecto Patronum (a summon spirit guardian) is a spell that most wizards find hard to cast because they have to get into the proper frame of mind (concentration) as well as have the appropriate level/ability. Also creating a fully material version of the patronus is even more difficult. So, there would have to be either two spells (like invisibility and greater invisibility) or there would be greater effect if cast at a higher level (like magic missile cast at 1st vs 5th level). I have gotten this far with deciphering this spell.

The real problem I have is with the varying success of casting this spell by different wizards. It was viewed as a super advanced spell, but Harry was able to cast it earlier than any other wizard and more successfully than most, but he was also able to train other wizards to use it at varying degrees of success.

My question: How can I use D&D 5.0 mechanics to simulate varying degrees of success with a harder spell like Expecto Patronum?

Note: All answers beyond “here is the D&D 5e rule for that” must be backed up with experience.

* I chose D&D 5e because 1) I have the books, 2) all of my players have the books, 3) I love the system, 4) it seemed at least possible that it could work.
† I meant to use a party of all wizards mostly because they learn new spells on their own merits without gaining a level (and no dealing with demons or communing with deities or whatnot required), so the wizard class seemed to fit best.

  • 3
    \$\begingroup\$ Note to answerers - any suggestions you have must be based on approaches that have actually been tested \$\endgroup\$
    – Wibbs
    Commented Jul 16, 2016 at 11:15

3 Answers 3


Look at other spells in the book

I'm under the impression that, as Harry's classmates attempt to cast the spell, they are not quite getting it right some times. Normally you cannot cast a spell you don't know or one that you don't meet the requirements for. But, as nitsua60 points out, if the spell is a scroll, you can cast it at any level as long as it's on your class spell list. There is a save DC for casting it successfully (DMG pg. 200) and a list of mishaps for failure (DMG pg. 140). For teaching and learning a new spell as you've described, you can use this system to simulate failure.

However, there are many spells in dnd 5e that have variability of effect beyond the damage dice. I think if you take a look at these spells, you should be able to come up with a system that works for you. This wouldn't be unprecedented and as far as creating a spell it would be fairly "vanilla", without teaching your players additional concepts from other games.


I am shying away from directly suggesting how your spell should work since I've no experience with creating such a spell. Rather, to answer your direct question of "How can I use D&D 5.0 mechanics to simulate varying degrees of success with a harder spell like Expecto Patronum?", I suggest you can do so by looking at other spells within the book. What you want to get out of the spell is a little difficult for us to answer, so I am attempting to lead you to the answer you seek. I hope this answer can be a stepping stone for you.

Caster's Choice: One of the more straight-forward ways is to determine the outcome based on what the caster is trying to accomplish. Look at Animate Objects, notice the chart gives different effects for how large the objects are. Larger objects are better, but you get less of them. This is pretty common among minion-controlling types of spells, where you can pick few large creatures or many small ones. This concept can be tweaked to work for your spell.

Target Specific: Similarly, Creation allows you to create something, but based on its composition it lasts from as long as 1 day for vegetable matter to 1 minute for adamantine. This variable length puts limits on the caster based on the difficulty of what they're trying to simulate. In the same way, your spell could be castable by all wizards, but it's effect diminished based on what they're doing with it.

Stages: You can affect it in stages, as is done with Control Weather and Prismatic Wall, where you must use the spell to change the outcome gradually over time. This makes the spells last/take a long time, but without being "on/off" in effect.

Variable DC: You can also change the DC, as is done with Scrying. Scrying specifically changes the DC based on how well you know the target and whether or not you've got something of theirs. You can modify this system to change the difficulty of the spell. This behavior is quite common for the DM as they create DCs for skill checks and the like. You can find better explanation of such effects in the DMG.

Divine Intervention: Some spells' effects are determined entirely by the DM. Spells like Divination and Contact Other Plane are about asking the DM questions, which the DM answers in his/her own way. More famously, the Wish spell creates an effect by request of the caster that the DM carries out (or doesn't) in his/her own way. Your spell could also behave this way, but in my mind that might be less fun.

Random: Finally, you can go random with it: roll a dice to determine the outcome. Spells like Prismatic Spray, Reincarnate, or Teleport use this system. Teleport is by far the most variable: There are 4 different outcomes based on 7 degrees of familiarity and a percentile roll. These are reconciled in a chart, so if you do go the route of extreme variability, a chart might also help you.


It may seem like there is no such thing as a partially casted spell in 5e Dnd, but there is an established way you could model a partial success.


A major problem with the Expecto Patronum spell was not just casting it, but maintaining it in the presence of Dementors. A possible way of modeling this (and other "partial successes") in 5e is via Concentration.

Certain spells (like Witch Bolt cast with a higher level slot) have different effects on the round they've been cast, and on subsequent rounds. Perhaps Expecto Patronum could be a concentration spell with different effects on subsequent rounds (for example: first round, all undead within 100 feet become frightened of you, second round no undead within 100 feet may deliberately target you with attacks or spells, third round all undead must use their move and actions to retreat from you for the next 2d6 rounds, and remain frightened of you until the next sunrise).

What will complete this equation is if your target is somehow damaging your caster. For example, if the Dementors have the property "When a hostile creature is within 60 feet of a Dementor, the hostile creature takes 1 psychic damage at the end of the hostile creature's turn". This passive damage would force a concentration check every turn, since according to the PHB p203:

Whenever you take damage while you are concentrating on a spell, you must make a Constitution saving throw to maintain your concentration. The DC equals 10 or half the damage you take, whichever number is higher.

The practical upshot would be that it would seem that many wizards could not cast the spell to its full effect, while in reality many of them could cast it, but few could maintain it long enough for it to have its full effect.

  • 2
    \$\begingroup\$ Storm of Vengence I think behaves this way (different effects per turn). There is a precedence for that behavior, although it is a 9th level spell \$\endgroup\$ Commented Jul 16, 2016 at 22:01
  • 1
    \$\begingroup\$ I considered Concentration (it was included in my question), but in Harry's DADA class in the room of requirement there was no Dementor to cause damage and force the concentration check. I am definitely giving my Dementors a psychic attack and the AoE is mentioned in the books. Maybe the Patronus itself is fighting the spell and triggering the check? I like the idea of multiple rounds, though Harry was able to produce a Patronus instantly. \$\endgroup\$
    – LeHill
    Commented Jul 18, 2016 at 18:02
  • \$\begingroup\$ I'm glad you like the idea of multiple rounds. Sorry I didn't address the fact that you'd mentioned concentration earlier. To your point that Harry produced a Patronus instantly, the spell would produce a Patronus on the first round, but it wouldn't be able to fully repel them for about 12 seconds: but the Dementors might flee anyway, and certainly would be weaker. When it comes to the DADA class, keep in mind that Harry was faced with a Boggart. It had a weak version of Dementor magic, plus they show you your worst fear. That might warrant a concentration check (DM's discretion). \$\endgroup\$ Commented Jul 18, 2016 at 18:14

Varying Successes

Most D&D Spells have some form of varying success; e.g. even a simple fireball does 8d6 fire damage, Reflex Save for half. So the variable for success is encoded in the damage, modulated by the Reflex Save. Since you want something outside of damage to be the variable, two options come to mind.

Wild Magic

D&D 3.5e's sourcebook Complete Arcana contains Wild Mage Prestige Class. This class used Wild Magic rules to cast spells:

Wild Magic: A wild mage casts spells differently from any other arcane spellcaster. She reduces her caster level by 3 for all spells she casts from now on. However, every time she casts a spell, her use of wild magic adds 1d6 to her adjusted caster level. For example, an 8th-level sorcerer/1st-level wild mage has a base caster level of 6th, not 9th, but her actual caster level varies from 7th to 12th for every spell she casts. Caster level affects all level-based variables of a spell, including spell penetration checks.

Since 5e doesn't have a concept of caster level, you could replace this with a spell level modifier dice. You already suggested to have the different successes represented as different spell levels in your answer, so you could add this to the spell text:

This spell is very tricky and hard to get right. An additional d4-2 modifies the level of the spell. If the spell level is below 1, it fails automatically.

I've played the original 3.5e Wild Mage for over a year in a campaign, and really love the mechanic. When D&D Next came out, it was the first class I ported to 5e. In the original rules, a modifier of d6-3 is used, which is added to the caster level. This means in the average case (d6-3 ~= 0.5) it slightly improves the caster level. In 5e, there are no caster levels, only spell levels, with less than half the range, 9 instead of 20. Therefore I reduced the d6-3 to a d4-2. I added some further restrictions, but it worked really well. Lots of fun.

Arcana Check

Alternatively, you could also simply use an Arcana check:

Make an Arcana check when you cast this spell:

  • Below 10: The spell succeeds with minimal success.
  • DC 10: The spell succeeds with a minor success.
  • DC 15: The spell succeeds with a major success.
  • DC 20: The spell succeeds with extraordinary success.

This pattern is often seen in 5e adventures. It works well because it uses an existing, widely used mechanic that already solves the problem of varying successes in other contexts.


In D&D 3.5, there are some spells where casting them takes a round, longer than an action. The same is used in 5e e.g. in the spell Slow, which has a chance to make the affected creature take 1 round to cast a spell that takes 1 action usually.

This can be used to simulate the increased concentration demands during casting. If the caster needs to concentrate for a round while casting the spell, enemies can try to interrupt him, and only great wizards (with a great constitution save for concentration) can successfully cast this spell.

I like Slow and think it is one of the most fun spells in 5e. I've used it many times as a DM or player, since it puts the focus on spell casters in an interesting way.

  • 1
    \$\begingroup\$ I will have to use Wild Magic somewhere, it seems a good fit for HP. I'm not sure that it or an Arcana check will work for the Expecto Patronum spell though, since Hermione (the one with the highest Arcana check, I'm sure) had difficulty learning this spell and Harry couldn't fail to cast the spell once he mastered it (the d4-2 would have to go away at some point). \$\endgroup\$
    – LeHill
    Commented Jul 18, 2016 at 18:23
  • \$\begingroup\$ @LeHill If HP is mid- to high-level, even a 1 on the d4-2 would still let him cast the spell successfully, just not at the highest level... \$\endgroup\$
    – Mala
    Commented Jul 18, 2016 at 18:26
  • 1
    \$\begingroup\$ True, but this assumes the wizard would have to level up to master the spell and can't just practice it to get better at it. Also, assumes you have to be a certain level before you can even attempt this spell - could be why people were surprised that Harry could cast it. It is a good idea! \$\endgroup\$
    – LeHill
    Commented Jul 18, 2016 at 18:35
  • \$\begingroup\$ @LeHill If leveling up is a problem, does making each successful Arcana check provide an improved chance of success on the next attempt? (Running tally of success that eventually leads to "no failure" on a given attempt). That would seem to fit the HP story, but it isn't a 5e mechanic as such, it would be a homebrew built around a 5e arcana check. \$\endgroup\$ Commented Jul 19, 2016 at 2:40
  • \$\begingroup\$ Leveling isn't exactly a problem, just a consideration. There is a 5e rule about group checks that keeps a running tally of successes made by the group to overcome some obstacle. That could be useful. \$\endgroup\$
    – LeHill
    Commented Jul 19, 2016 at 17:22

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .