Hello! My name is Nerosothe and I have once again emerged in February like the Groundhog of Lammania that I am to tell you if we are in for 2 more weeks of testing previews (spoilers: we are).
This aggressively oversized post will dive deep into the Arcane Trickster and how it plays both strength wise and general vibes. I have spent the better part of 8 hours on Lamm running through interactions with about 5 builds of Arcane trickster: 4 multiclasses and 1 pure life.
As per usual. I colour coat my comments so for easier skim reading:
Green denotes the subject I wish to address
Orange denotes any “problems” I see need to comment on
Red denotes core design problems or critical bug issues
Teal denotes my own suggestions.
Everything else is my yapping so delve into that with caution.
If you are a Dev or someone or just wish to skim: I Bold and Underline what I personally think as vital info/ data for this preview that I have gathered.
I also gotta say it every time: I am just a Bozo who yaps too much on a forum for a 20-year-old game. My opinions are just that: Opinions. I’m not here to tell the readers or any of the devs how to make an architype or how to objectively balance a game. I just wanna share my two cents in a constructive way in case that is of any help or to give those who don’t have the time to check out the preview a glance into what’s going down.
So, if you want the quick and short of it: Arcane Trickster design wise is amazing but needs AT LEAST one more preview to iron out bugs and some numbers in my opinion.
-On hit dice for spells scaling off Sneak attack dice, A machine gun caster playstyle which has surprisingly decent dps in epics via magical missile spamming, amazing (for the most part) utility spells from the stolen spells system, and Good hybridization for caster and crossbow trees. These are a few of the things I’ll dive into this post. I really do like what’s here conceptionally, but some stuff needs some touch ups i.m.o.
What has me hesitant about gushing full out for this Architype is a series of bugged interactions:
the magehand spells feel rather horrible to use, Mixed magics doesn’t give caster levels based on level, and no sneak attack dice in Arcane trickster’s tree leads to a drought in sneak attack scaling while leveling, Really, I could blather about how fun it is to theory craft this Archetype for a solid thesis length post, but I have a bed I yearn to sleep in, so lets get into the details.
Mechanic, Thief Acrobat, and Arcane Trickster as the trees:
I think the exclusion of assassin is the best choice here. Core Rogue should get the highest sneak attack dice possible so to offer its own class identity as the pizza cutter class (all edge and no point. Yes, I am so funny, thank you). If Arcane trickster could dip Assassin to get 4-5 more dice you really would have no reason to play base rogue as it would be the best melee, sneak attacker, and now caster option for rogue. Acrobat still gives you the choice of a melee build, the inclusion of Shillelagh as a stolen spell also gives this a nice boost and tinge of flavor too, hats off to the team for that. Mechanic Inquisitive dipped with a little Arcane Trickster is a great ranged option too, but it also means that the few people who want to play pure GCB rogue stay as the original class and feel no pressure to switch.
I will say I would love to see 5 or maybe 10 RP/ MP somewhere late in the tree as you really do sack that stat by splashing into this tree.
Maybe Tricks on Tricks could have a martial multiselector to give 5/10 MP/RP when you cast a spell would fix this? If that’s too strong you could swap Pin Point to t3 and Put Tricks on Tricks into t4. That or the cores giving 3-5 MP/RP as neither Thief Acro nor Mechanic give MP/RP in their early cores
Arcane Trickster’s identity.
If we are being reductionist, Arcane trickster’s most defining features are its scaling imbue dice, Mage hand, stolen spells, and sneak attack dice on spells from Magical ambush.
Right off the top, Magical Ambush is great conceptually. Shadowdancer’s ability to add sneak attack damage to spells/ blasts is a super cool idea in epics but it was weird that it seemingly came out of nowhere and has no direct support from any classes. The caster side of Shadowdancer has long needed a tree which is truly meant to gain something from specc’ing it, so it’s great to see this in here.
The biggest issue with Magical Ambush is in heroics Arcane Trickster has little to no way to gain Sneak attack dice bar the auto granted dice from rogue levels. While this does get fixed in expics via being 12+ rogue with epic sneak attack and the 6 odd dice you get from shadowdancer, man are you fiending for some sort of source of sneak attack dice in heroics.
As I said above, I think Assassin should be the master of Sneak attack scaling, but the
Arcane trickster really needs 2 or 3 sneak attack dice nodes somewhere cause in heroics Magical Ambush his hitting for a measly 4d3 with 50% Spell power when you get it and it only climbs to 8d3 x~250% (which still is terrible) by pure rogue 15 when mobs have hp in the thousands (Magic missile spells amend this slightly at the cost of being poor damage themself).
What this amounts to is Magical ambush being horrendous damage let alone serviceable in heroics unless you can find 5+ sneak attack dice from a non rogue source. To point out how bad it can be: If you want to play a Thief Acrobat Arcane Trickster, you will get NO Sneak attack dice bar your level up dice unless you take the Thief Acrobat capstone. You will also only gain 3 more Imbue dice compared to base Acrobat+Assassin. Mechanic can avoid this issue if they go nonrepeating crossbows, but like man does it feel bad if the only heroic style of arcane trickster that has good sneak attack damage is one who doesn’t use spells. I want to emphasize that this ability is super, super cool conceptually but
please, in Cores 2 and 3, add 1 sneak attack dice to Arcane Trickster or put 1-2 in the bottom 3 tiers for some heroic scaling.
Okay I’ve yapped for long enough on Magical ambush, what about stolen spells?
Listen, stolen spells are flat out amazing.
Class identity? Immaculate. Need early healing? Covered. Wanna see many, many numbers on your screen? Magic missile, chain missile, force missile wahoo. Need a panic button? Boom, Radiant Forcefield. Wanna be a freak? Take Fire Seeds as a lvl 18 core SLA.
I jest, but really; bar like, Disintegrate, Neutralize Poison, and
Fire Seeds (Please put this one a core earlier), These cores are cool and useful.
“Wow. That was really concise regarding stolen spells. Does this mean the rest of the post will be as straightforward?”
HAHA, you’re cute for thinking that reader…
Mage hand Is too damn slow.
On one hand, these spells are prob the best thematic part of this tree, on the other hand (yes this is a pun, lets move on here),
it feels antithetical to play a sly magical thief who’s invisible hand servant takes whole seconds to order around. Mage Hand Spells cannot be quickened, casting animation is a ~1.5 seconds feels terrible to use in mid combat, I can cast almost three spells in the time it takes to use either slap or poke. This currently destroys the play loop for me and if these abilities keep their animation time, they will not be used by a majority of players similar to smoke bomb/ snap trap from Dark Hunter.
Please either change the animation of the mage hand spells, allow them to be quickened, or double its speed.
I really would love to use the mage hand spells in combat but right now they really take me out of the gameplay loop just to cast. I don’t think it’s a question of them being too weak. A pretty much guaranteed stun or blind with sneak attack immunity strip is always good utility, and Helping Hand is just a great panic button.
The insta kill from Presto is also neat but 30 second CD single target insta kills are just not worth pressing or putting points into when it is the only one you have. Again, not saying it’s a bad skill, but Go for the Kill is 20 second CD, and Death Attack is 12 seconds. It seems modern Insta kills are much shorter on the CD than this so
I don’t think bringing this down to 20 seconds CD is out of line.
Regarding Imbue dice
The 1d4+1 imbue is a neat little thing, it gives a solid option for Mechanic and Acrobat, and I’m sure there’s some martial build that will love this as their chip damage option (as that is what imbue damage is for the vast majority of builds). I think the coolest thing about the imbue is the t4 node that gives you imbues based on level. The fact that this is in the tree itself and not a core means you can stack with another class core like alch or EK for some fun. I will get around to seeing the viability on this.
Big Boi Bugs
Mixed Magics from T3 arcane trickster is not applying its levels correctly: At cap your CL is 19 ( on the character this was tested I was 12 AT/8 WIZ so 12 base + 5 from epic levels and +2 from legendary). WIZ spells were at CL 15 (8 wizard +7 as per above so again, Mixed magics isn’t working as intended.
*checks notes* wait is that the only bugs have written down? Well tbh I’m sure there’s more but most of my testing was just looking at how the new tree worked and some rough build sketches at 8, 15, 20, and 34. On that note: 12 ArcTrick 7 Wiz Archmage 1FVS AoV is damn good force caster. I wept when FVS archon didn’t get the added damage from Magical Ambush. And yes, for those who know me I indeed tried to make sneak attack warlock a thing with this and well; it’s, by definition, better than it was before (this was torture, enjoy getting racial completionist to make Acolyte of the Skin Arcane trickster just barely work).
For now though I think I’ll call it here. I will get more into the numbers when a more final build comes out.
Overall, I think this is a super cool Archetype with some much needed attention before wrapping it in a nice bow and shipping. Really excited to see what next preview brings and if you have any questions about what exactly I tested, how I do my testing, or just want clarification, feel free to dm me or msg me in the official discord
@Nerosothe.
