So I am interested in playing a character who uses thrown weapons with the returning infusion offered by the latest artificer, but I want to focus more on throwing than artificery. Towards that end, I am planning on using the battle smith specialty, and then on top of that, I was thinking I would multiclass. Eldritch knight or arcane trickster seem particularly synergistic.

Since the unusual artificer multiclassing rules say that I should round my level up, I’m thinking three levels of artificer to go with either of those classes, though the fourth for an ASI or the fifth for an extra attack are tempting, of course. And then I’m not sure which of eldritch knight or arcane trickster is a better fit.

  • Eldritch knight would be the more direct upgrade to martial skills, of course, but Extra Attack isn’t going to stack with Arcane Armament, and Weapon Bond is maybe redundant. War Magic seems really good though.

  • Arcane trickster gets Sneak Attack, which possibly allows for better damage if I already have an extra attack and fighter won’t give a third for a long time, and well, I like “tricks.”

The goal, overall, is combat capacity, but it doesn’t have to be straight DPS. As I said, the arcane trickster’s tricks sound pretty good.

My somewhat-naïve sense here is that eldritch knight is perhaps closer to what I want to do, but some of its anti-synergy means that arcane trickster is more competitive than it might be. What I’d like is a compare/contrast of these two options so that I can make a judgment between them.

We are starting play at 12th level, and while I only own Player’s Handbook, Xanathar’s Guide to Everything, and Wayfinder’s Guide to Eberron, I’m not necessarily opposed to getting more. The DM is fairly open to content, even some Unearthed Arcana content.

I consider 2 levels of artificer, for returning, to be a minimum as far as this question is concerned. Beyond that, while I am most interested in 3-5 levels of battle smith paired with one of either arcane trickster or eldritch knight, I am not opposed to someone challenging my assumptions that these are the best options.

  • \$\begingroup\$ Comments are not for extended discussion; this conversation has been moved to chat. \$\endgroup\$
    – V2Blast
    Commented Jun 15, 2019 at 10:16

1 Answer 1



EK is more flexible and has more races to play with than AT. AT does more consistent damage without spending a lot of resources. Both spell list (AT and EK) have nice perks and can be used anyway. EK are more tanky and AT are more mobile. AT have more uses outside of battle than EK, but if build for it, EK is a close one, particularly with variant Human. The synergies with Artificer are not that great for both, there are some redundancy in both clases.

Arcane Trickster vs Eldritch Knight

(AT = Arcane Trickster, Ek = Eldritch Knight)

Let's start with something simple: unless you are maxing DEX or using expertise, Mage Hand Legerdemain might not be as useful as it should, making the only feature of the AT (aside from magic schools) not that useful in battle (unless the campaign would reach level 17 for Versatile Trickster). Outside of battle is another story.

Weapon Bond: Although a bit redundant, it is not completely useless. Returning Weapon only returns the weapon after it is used. Weapon Bond can return the weapon even if is not used in battle (and it is not sent to another plane). So, if you are captured, you might still have your weapon back. Disarmed? No problem. Stolen? Same (maybe). A prank from a teammate?... no comment, bag of holding exists so...

War Magic: Ohh, War Magic, this could be very, very controversial. On one hand, this could make an EK a great tank by casting Blade Ward (but you are throwing stuffs, thus is less useful, unless your party needs an extra tank for an encounter). It might increase the damage of the EK by casting cantrips, but I feel that it might lose the "throwing weapons" theme that the build is going for. On the other hand, there are extremely situational and kind of controversial combos that can be done with it. For example, An Elf can use Elven accuracy and True strike to have a pseudo super advantage (3d20 dices), while useful against very high AC creatures (or when you have to hit that attack), it might be marginally worse than attacking twice in the long run. Nonetheless, War Magic is a big advantage over Mage Hand Legerdemain in battle.


As far as I read, I didn't found strong synergies between classes that would tip the balance too much. Both of them have some redundancies. For example, the rogue Cunning Action and the Boots of the Winding Path are somewhat redundant bonus action (dash/disengage vs 15ft restricted teleport). Same with Weapon bond and Repeating Shot. Other compliment both quite well because or are utility or are plain defence increase that both clases have their uses. Regarding spell, the artificer spell list is well balanced and can be tinkered as needed for any class without too much difference.

AT enchantment and illusion school vs EK abjuration and evocation school

This one is quite complex to address and I think can be very subjective. A well placed illusion can destroy an encounter so easily. But EK (and AT) can choose a spell from another school at level 8th, reducing a tiny bit the gap between them. Nonetheless, Evocation can be useful for the damage it can deal... and most evocation spells can be used with higher spell slots.

The same can be said enchantment vs abjuration (or any combination). There are going to be situations that can benefit one over the other inside and outside of battle. Just that some schools have more spells at first and second level than other schools, with Evocation and illusion being on top.

The only spell that might be of interest (thematically speaking) is Shadow Blade and it is from the illusion school. Giving a small edge to AT because AT can access it easily, while EK has to use its 8th level spell choice. But, again, it is situational and there are other spells that might be better than this.

Rogue vs Fighter

Given that the starting level is 12 we can build a character that might not be as fun to play at lower levels but can be effective at 12. Furthermore, I reduced the complexity by considering three choices: Fighter 8 vs Rogue 8 vs Rogue 7.

A note on Sharpshooter: Two of the three effects are very useful for the builds: Not having disadvantage on long range attack with ranged attacks (perfect for thrown weapons and their meagre distance) and reduce cover penalty. The other effect (+10 damage, -5 attack), sadly, cannot be used with thrown weapons.

Rogue 7

Let's start with the simplest: Rogue 7 / Artificer 5. This build has the least flexibility of races to play with. Given that Rogue 7 only gives two ASI and, if your DM takes very seriously cover (even from allies), you really want the Feat Sharpshooter and max INT. Thus, this lead to variant Human as the only race to choose that I know of, simple because of the extra feat.


  • 4d6 Sneak Attack
  • Arcane Armament
  • 2nd level spells (Artificer)
  • 2nd level spells (AT)
  • More utility outside battle than Fighter (Expertise, more skills, etc)
  • Better mobility than Fighter
  • 5th class level prepared spells (Artificer)
  • 3rd level spell slots


  • 5 spells known (AT)
  • 2 ASI (less flexibility on races)

Rogue 8

This one is a bit of a contrast from Rogue 7. Rogue 8 is less effective in combat of the bunch, given that it does not have Arcane Armament or Extra Attack. The advantage over Rogue 7 is the flexibility it has when considering races. It has 3 ASI, thus, it can easily achieve max INT and Sharpshooter. Races like High Elf for the +1 in INT, +2 DEX and an extra cantrip might be valuable. Or it can take variant Human and use the extra Feat for things like Proficiency on DEX saves by the Resilient feat (it has evasion, thus it might be a good addition) or Lucky to (try to) land that Attack when it counts.


  • 4d6 Sneak Attack
  • 3 ASI
  • 2nd level spells (AT)
  • 6 spells known (AT)
  • 1 Spell from any school (AT)
  • More utility outside battle than Fighter (Expertise, more skills, etc)
  • Better mobility than Fighter


  • 1 Attack per turn
  • 1st level spells (Artificer)

Fighter 8

Fighter 8 is the most flexible of all this builds. It has 4 ASI, which gives the most flexibility at the moment of choosing a race. Like a race that does not have stat bonus in INT? No problem, 14 INT + ASI (or 15 INT + feat with INT increase like Resilient, Elven Accuracy or Observant) and there are enough ASI to max INT and get Sharpshooter. Moreover, Extra Attack seems to work with Returning Weapon without downsides like Arcane Armament. The downside is that the +2 on damage from Dueling style is not near the 4d6 sneak attack from the Rogue (14 damage on average if my maths are correct). However, since INT is (might) be your primary stat, you are quite capable of dealing damage with the few spells and cantrips that the class has.


  • 4 ASI
  • 2nd level spells (EK)
  • 6 spells known (EK)
  • 1 Spell from any school (Ek)
  • Best race flexibility


  • Lower DPT (Damage per turn) than Rogue
  • 1st level spells (Artificer)
  • Less utility outside battle than Rogue (Not expertise, less skills, etc)
  • Worse mobility than Rogue

Rogue or Fighter 9

The discussion in regards to Rogue or Fighter 9 would be more a 3rd level spell slots and class feature VS 1 ASI/feats. Given how the multiclassing spell progression for the Artificer work, it is possible to get 3rd level spell slot (but not spell for that level) if going for Artificer 3.

For that, I'm going to start with the fighter Indomitable class feature because it is the easiest. The Lucky feat > Indomitable. 3 rerolls on attack rolls, ability checks or saving throws are way better than just one for saving throw, and both are on long rest. Thus, the question would be, how about 3rd level spell slots? Well, if we consider 1 lucky point = Indomitable, therefore, how 2 lucky points are vs 3rd level spell slots? I think, 3rd level spell slots might be better than 2 lucky points. The reasoning is that you may have more spells to cast and more flexibility in battle than just two rerolls. But, keep in mind that rerolling an attack or a saving throw may be enough to save a TPK (or the PC life) or kill the BBEG before he escapes.

For Rogue 9 it is a bit trickier. 1d6 + Magical Ambush can be really good. I may even say that Rogue 9 is better than Rogue 8 if you stomach variant Human for the max INT and Sharpshooter. I'd argue that Sharpshooter is even better than max INT in the long run considering cover and minimum distance for thrown weapons, but that is if there is no problem with a sub-optimal build. Rogue 9 might not have two attacks per attack action (see side notes), but a Phantasmal Force with advantage might it be better. There could be some feats that can be argue might be as good as 3rd level spells slots plus 1d6 and Magical Ambush. Magic Initiate might be one of them, particular with Bless, and being a back-liner, it is easier to maintain concentration. This would help up to 3 party members and frees the cleric (or similar) of its concentration slot (the same can be said with Fighter).

In conclusion, on the fighter side, there is not much to gain but the fighter is more flexible, thus can be done without sacrificing anything too crazy. Those 2 3rd level spell slot are better than the lucky feat (but other feats might be debatable). For the Rogue, if not going for an optimal build (max INT or not having Sharpshooter) or don't mind variant human, Rogue 9 is better than Rogue 8. Rogue 7 have not only 3rd level spell slots but also 2nd level spells from the Artificer and two attacks per attack action.

Possible race combos

Rogue is the less flexible of the two, with Rogue 7 the least flexible in general (only Humans). Rogue 8 can max INT and having Sharpshooter with any race that improve INT like High Elfs, Half Elfs and Humans. Obviously Rogue 8 and variant Human are quite flexible and can have 1 feat like Lucky or other more thematically interesting one.

Fighter 8 is the most flexible, almost every race can be used and have max INT. Races that have a natural high INT can give more flexibility as an extra Feat, but a flying crazy Aarakocra throwing axes from the sky can be quite amusing. Keep in mind that variant Human can have 4 ASI and a feat (2 ASI and 3 feats, for example), thus, combinations like Magic initiate (bless or Silent image, for example) and Spell Sniper can net you 3 cantrips, 1 1st level spell and other goodies (improving the utility inside and outside of battle depending on cantrips and 1st level spell chosen). Or even being a pseudo skill monkey with Prodigy and Skilled.


From my point of view:

AT vs EK in terms of archetype features. The winner undoubtedly is the EK for battle purposes. AT for outside of battles is more useful in general if using expertise or having nice DEX. But, EK if built for it, can be really good.

AT vs EK in terms of Races selection flexibility. The winner is EK in general, but Rogue 8 is a close one.

AT vs EK in terms of spells. I give it a tie here simply because battles are not just pure raw damage. Creativity can change the course of a battle. Both clases have a good spell selection. However, only Rogue 7 have two sets of 2nd level spells (one from AT another form Artificer) and 3rd level spells slots. That gives Rogue 7 an advantage.

AT vs EK in terms of damage. AT have an edge in this category. 4d6 sneak attack is not joke, and Rogue 7 have the most consistent damage output with Arcane Armament. Moreover, AT sneak attack does not use resources, while EK spells does (cantrips aside). However, EK with War Caster might have something to say in this regard. An 11th level cantrip can have between 3d6 to 3d12 damage output. For example, a firebolt can have a 3d10 with an average of 16.5 damage, thus a Fighter attack using a magic Spear +1 (1d6) and War magic can do: 3d10 + 1d6 + 1 + 5 (INT) + 2 (Dueling style) = 27.5 damage. A Rogue 7 can do with a dagger +1: 2d4 + 4d6 + 10 (INTx2) + 2 (+1x2) = 31. Quite close, BUT, that is if we use firebolt (with fire being a common resistant is not that great). Ray of frost does 3d8, thus, the total would be 24.5 damage. One thing to consider with War caster, aside from the damage, it breaks a bit the throwing feeling of the build. A fighter without War Caster would do: 2d6 + 2 + 10 (2xINT) + 4 (2xDueling) = 23 on average. Thus, Rogue 7 having the advantage here.

AT vs EK in terms of build flexibility: without a doubt, EK. Rogue 8 as a second good option.

AT vs EK in terms of utility outside of battle: Hard to say. Rogue is the most consistent without investing in feats for it. EK can be quite effective if built for it (not as much as rogue, though)

AT vs EK in terms of general class features. I'm not going to touch it because the analysis is quite extensive as is. Rogues are more mobile and skilled, Fighters are more tanky and with Action Surge can deliver a nova attack for one turn. Fighters have more ASI at level 8 than Rogues but at 10 they have the same numbers. Rogues have Evasion, etc.

AT vs EK in sheer battle prowess and spell flexibility. Rogue 7 is the best in this regard: 3rd level spell slots, two sets of 2nd level spells (AT and Artificer), more spells known, cantrips, 4d6 sneak attack, and Arcane Armament. From the pure view of battle optimisation, Rogue 7 is the best. But it is a very rigid build.

Thus, for simple battle and utility outside of battle optimisation, Rogue 7 is the best option, but variant Human is almost a must. Fighter 8 is quite flexible and can open a lot of opportunities to RP with a lot of races, or be a powerhouse of feats with variant Human. Rogue 8/9 is a nice middle ground between them. In terms of magic, both have their uses. AT can be very nasty with its tricks, while EK can have a lot of damage potential with the right spells at the right moment. EK's War Magic can also add dept in combat depending on the build. Most cantrips have secondary effects that might be more beneficial than just plain old damage.

Side notes

Although it seems that EK have the edge here, there are things that I didn't considere. The reason is because the Artificer's spells are quite versatile and can perform really well with any build. Who does not like feather fall, blur, etc? They are utility spells that are useful for both builds. Only Rogue 7 can use 2nd level Artificer spells, though. Furthermore, at level 13 (if the adventure reaches it) Rogue 7 and Rogue 8 are quite identical (both can be Rogue 8/Artificer 5). There are a lot of considerations to have but, from my point of view, are not as important as the one I touched in the main body.

I just realise that Two-weapon fighting might be doable with Returning Weapon. Making uses of the bonus action when not needed.

  • \$\begingroup\$ Thanks so much for this, and sorry it took so long to get back to you on it. I wonder—why not look at the 3/9 cases? Due to the artificer multiclass rounding up, you get to count as 2nd + 3rd = 5th level against the multiclass spellcaster chart, opening up 3rd level-spell slots. \$\endgroup\$
    – KRyan
    Commented Jun 17, 2019 at 19:12
  • \$\begingroup\$ @KRyan I guess I made a mistake with the math. I quickly calculated that 4/8 or 5/7 was enough to get 3rd level spell slot. I even included that in the Side notes. Now that I made the actual math, you are right. However, I have a mixed opinion on this. On one hand, 3rd spells slots are good (but not as good as having 3rd lvl spells). On the other hand, that ASI/feat might be as good as those 3rd in the long run. For example, an Elf Fey Teleportation is great for mobility in the fighter build (and is short rest), Or Magic Initiate for extra cantrips and a spell lvl 1 known and "slot". \$\endgroup\$
    – Chepelink
    Commented Jun 17, 2019 at 23:23
  • \$\begingroup\$ And magic initiate can be used for non wizard spells like Bless. And you'd be a back liner, thus easier to maintain concentration. Thus, that feat can be very beneficial in other ways. But, I think if you go for Rogue 8 it would be better to just go straight to rogue 9 for the extra 1d6 and extra feature. It would lock the race to variant human if you want max INT and sharpshooter. And I really think that Sharpshooter is needed for this build given the range of thrown weapon and avoiding cover. And you might want max INT for spells. \$\endgroup\$
    – Chepelink
    Commented Jun 17, 2019 at 23:31
  • \$\begingroup\$ I edited the answer to improve the answer and added the section of 3/9 builds. Given my math mistake, I can said that Rogue 7 is the best in terms of battle optimisation with all the goodies it has (and a solid skill monkey) . But, if looking for optimisation, the only race that fit the bill would be variant Human. Nothing wrong with playing one, but not as crazy as a flying axe-throwing birdman. \$\endgroup\$
    – Chepelink
    Commented Jun 18, 2019 at 0:41
  • 1
    \$\begingroup\$ Excellent, this seems to cover everything very nicely indeed. Thank you for all your work. \$\endgroup\$
    – KRyan
    Commented Jun 18, 2019 at 3:05

You must log in to answer this question.

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