When you build a character in Genshin
Impact, there are generally four different areas of progression that you focus on:
Character Ascension, Weapon, Talents, and Artifacts. Each of these things cost Resin
to upgrade, whether it’s in the form of Leylines, Domains, or World Bosses. For most of these areas,
it’s very clear exactly how much stronger you get and how much Resin it costs to upgrade that build
path. For example, leveling your Skill Talent from 6 to 7 increases its damage by 7%, and it costs
roughly 80 Resin to farm the Talent Ascension materials. Ascending and leveling your Weapon
from 80 to 90 increases your damage output by approximately 9%, and it costs about 120 Resin for
a 4-Star Weapon and 180 Resin for a 5-Star Weapon to farm the Weapon Ascension Materials. While the
exact numbers might vary for different characters, weapons, or talents, we can more or less
deterministically say how much stronger we’re going to get from upgrading something from level
X to Y, and also estimate how much Resin it’s going to cost us. Of course, there is some RNG
involved with the Ascension Material drop rates, so you might get unlucky and have to run a domain
1 or 2 extra times. But in general, the variance isn’t so extreme that it strays too far from the
expected value. In this way, the character build process is like a series of contracts you make
with the game. The game says “give me 300 Resin, and in return, I will make your character 10%
stronger.” or whatever that number might be. This model allows us to make informed decisions
about where to invest our Resin first so that we can upgrade the things that give us the biggest
return on investment, and work our way down to the upgrades that are less Resin-efficient.
This applies for most of the 4 areas in our character build progression journey. But there is
one part of the build that is unfortunately not so predictable. And that is Artifact progression. If
progressing talents, weapons, and character levels are like a series of contracts, progressing your
Artifacts is more like spending a month inside of a casino, not knowing whether or not you’re going
to walk out with anything worthwhile. When you consider all the different variables that go into
building an Artifact, the sheer number of possible permutations that an Artifact can look like is
a monstrously large number. Because of this, no matter how much Resin you gamble away on an
Artifact Domain, there is really no good way to exactly predict how much stronger your character
will become after farming for however many weeks. And likewise, it’s not very easy to estimate
how much Resin we need to spend to increase our character’s damage output by a certain amount. To
help tackle this problem, I wrote a small program that can randomly generate and upgrade a 5-Star
Artifact, based on the probabilities that we know are used in the actual game. This allows
us to simulate the process of rolling lots of different Artifacts on a very large scale. Using
this as our building block, we can simulate the process of progressively upgrading a character’s
Artifacts in a way that’s very similar to how I actually do it in the game: Whenever I obtain an
Artifact, I first evaluate whether the Main Stat is something that’s desirable for my character.
Then, I look at the substats and see if there are at least 1 or 2 substats that are useful for the
character that I’m building. If it’s an Artifact that starts with only 3 substats, then I’ll first
upgrade it to +4 to reveal that 4th substat, and then further evaluate whether it’s worth
upgrading all the way to +20. If I do decide an Artifact is worth upgrading all the way, then
I finally evaluate whether it’s worth replacing my character’s current Artifact with it, by comparing
the expected damage output before and after equipping that new Artifact. By repeating this
process X number of times, we effectively simulate what our character’s stats might look like at
the end of farming X Artifacts. As an example, here’s what Wanderer’s Artifacts and stats might
look like after simulating this process with 280 random Artifacts, or roughly 4 weeks’ worth of
farming. One problem that we face though is that because Artifact generation is so random, if I run
the exact same 4-week simulation again, we can get a completely different set of Artifacts and a
very different looking stat sheet for Wanderer. This ends up creating some very inconsistent
results that are difficult to draw any meaningful conclusions from. One way we can mitigate this
variance is simply to repeat this simulation a bunch of times – in this case, I chose to run it
a thousand times – and then average out the stats we get over all of those iterations. Now that
we’ve set up the tools and methodology, we can try working out some of the problems that we posed
earlier in the video. First, I tried simulating Wanderer’s stats after 1 weeks’ worth of Artifact
farming, then 2 weeks of farming, then 3 weeks, and so on and so forth, up to 20 weeks. I chose
20 weeks as a kind of arbitrary maximum because I felt that that’s usually enough time to farm up
a decent set of Artifacts for a single character, but of course we could always go higher if we
wanted to. With this series of simulations, we can see how Wanderer’s stats get stronger
and stronger with each additional week that we spend farming. The number I’m most interested
in tracking is this “Expected Base Damage” stat. This is just a product of the total
attack, Crit multiplier, DMG % multiplier, and Reaction multiplier. This is basically an
aggregated stat that can tell us how “strong” our character is by combining our offensive stats
into a single number. And we can use this number to track how strong our overall build is at any
given time. Now, what we can do with this data is consider week 20 to be our target stat goal,
and then track how close we are to that goal on any given week. Now that we have this dataset,
we can make some pretty interesting observations. The first thing that immediately stood out to me
is that we achieve over 60% of our 20-week goal after just 1 week of farming. That quickly goes
up to over 70% after week 2, and almost 80% after week 3. The key takeaway here is that the majority
of our character’s power budget from Artifacts is unlocked in just the first few weeks of farming.
And if you understand how Artifacts work, this actually isn’t very surprising. Most of an
Artifact’s power comes from its Main Stat. And for most sets, the set effects are also a huge
part of what makes the Artifact desirable in the first place. RNG is still a big factor in
finding set pieces with the right Main Stats, but the randomness is still relatively mild
compared to finding the right substats. It’s completely reasonable to expect that we can find
these desirable Main Stats within the first few weeks of farming and unlock most of that power
budget that the Artifacts have to offer. The second thing that stood out to me is that at the
halfway point in this simulation – or after 10 weeks of farming – we are already 95% of the way
to our 20-week benchmark. Another way of framing this is that it takes just as long to achieve
the last 5% of our goal as it did to reach the first 95%. This really puts into perspective how
difficult it is to improve your Artifacts after you get past a certain point. Once you’ve already
hit these major stat benchmarks, trying to improve your stats any more just comes down to making
these tiny micro-optimization that cost you a ton of Resin and barely increase your damage. Bringing
this back to the question of Resin efficiency, we can already start to formulate a strategy from
what we’ve learned so far: When you start building Artifacts for a new character, I recommend
starting off by spending about 2 weeks’ worth of Resin to farm your Artifacts. And in this
time, you’re just looking for pieces that have the right Main Stat for your character while also
finishing a 4-piece set if that’s something that your character needs. The pieces that you find in
those first 2 weeks are probably not going to be great Artifacts. But based on the data that’s
generated by this simulation, even with these far-from-perfect Artifacts, as long as you have
the right Main Stats and you upgrade them to +20, you can still expect to unlock about 70% of the
total power budget that your Artifacts have to offer. And given the relatively small amount
of Resin that’s used in those 2 weeks, this is a really good deal – this is a very efficient
usage of your Resin. The second thing that we learned is that Artifact farming quickly becomes
very Resin-inefficient when you’re just trying to improve your substats. Up to Week 5, it’s actually
not too bad; gaining 10, 6, 3.5% more damage for a week’s worth of Resin is actually not a bad deal.
But once you start going into Week 6 and beyond, you’re looking at only 1 to 2% damage increases
for each additional week that you spend farming, and that’s really just not a very good rate of
return for the Resin that you’re investing. Your Resin is much better spent on other parts of
your build first. Get your Weapon to 90, get your Talents up to 8 or 9, level your character to
90. All of these things are more Resin-efficient, in terms of increasing your damage, than trying
to improve your Artifacts beyond the 5-week benchmark. It’s only after you’ve done all of
these things that I would continue farming more Artifacts. From this point, I usually aim for
stats that are roughly around this 9 to 10-week benchmark. Of course, there’s nothing stopping
you from farming beyond this point. You can only get stronger from farming more Artifacts. But the
key point that I want you to take away from this, is that after approximately this 10-week
benchmark, it’s gonna be really tough to make major improvements on your Artifacts.
Most weeks, you will probably walk away with no new good Artifacts. On the weeks that you do,
you’re probably going to increase your damage by 1 – maybe 2 – percent. So if you’re in the process
of building new characters or multiple characters, really do stop and think whether there are
other places that you could spend your Resin on, before you get this deep into Artifact farming. Up
to this point, we’ve done all of these simulations and analysis while making a very big assumption.
And that is that we only care about ATK, Crit Rate, and Crit Damage. And I actually picked
Wanderer in this simulation very deliberately because he is a character that doesn’t care too
much about having a lot of ER or EM, and he really just wants those raw, offensive stats to maximize
his Normal and Charged Attack damage, which is where he gets a majority of his damage output
from. And in this sense, Wanderer is actually very picky about what he considers a good Artifact.
If it doesn’t have all 3 ATK, Crit Rate, and Crit DMG, he’s probably going to want to keep looking
for a better piece. Now, on the very opposite end of this spectrum, we have another damage dealer
who is much better at utilizing all kinds of different stats. And that is none other than our
Spiral Abyss poster child, Xiangling. The key difference with Xiangling is that ER and EM are
actually quite valuable stats on her that directly contribute to her damage output, on top of the
traditional offensive stats like ATK and Crit. And this means that Xiangling is much less picky about
what she considers a decent Artifact. If we’re farming Artifacts for Wanderer and we come across
a piece that looks like this: We’re probably going to throw it away – or at least try to replace it
very soon – because it doesn’t really offer a lot of value to Wanderer, other than that Main Stat.
But if we’re farming for Xiangling and we find this piece, we would be pretty happy to see this.
The EM directly amplifies her Vaporize reactions, and the ER improves the uptime on her Burst,
which is where the vast majority of her damage comes from. So this is a really good Artifact
for Xiangling, even though it’s not the typical high-CV Artifact that we’re usually looking for.
Because she’s able to utilize a lot of different substats effectively, she has a much bigger pool
of viable Artifacts to pick from. That having been said, let’s see what the stat progression looks
like when we re-run the 20-week simulation with Xiangling instead. What we see here is that the
overall trend is still very similar – we still have this logarithmic curve where most of our
progress is front-loaded in the first few weeks, we hit that 95% mark right around Week 10
again, and we very slowly approach the 20-week benchmark with each subsequent week. But when we
compare this to Wanderer’s stat progression, we immediately see one key difference: the progress
in the early weeks of farming is noticeably better for Xiangling. After Week 1, we are 66% of the way
to our 20-week goal, compared to 60 for Wanderer. After Week 2, we are 76% of the way there,
compared to 71. And this is directly related to Xiangling being able to utilize other stats,
instead of just ATK and Crit. Even if you’re not finding Artifacts with perfect substats in
those first 2 weeks, there’s a much higher chance that you’ll find at least a semi-decent
Artifact that still boosts Xiangling’s damage, because she has a much bigger pool of substats
she can work with. The key takeaway here is that it’s generally less Resin efficient to build
characters that depend on just ATK and Crit – at least for the first few weeks. If you’re building
a character like Wanderer, you should expect that his damage will ramp up a little slower in the
early stages of Artifact farming, because you’ll be looking for very specific pieces with a smaller
pool of acceptable substats, which will naturally take longer to find. Some other characters that
have similar build styles to this are Itto and, to some extent, Neuvillette. On the other hand,
characters whose damage scales off of several different substats ramp up slightly faster. This
means that you can get them off the ground and at least somewhat working in just the first couple
weeks of Artifact farming. Aside from Xiangling, characters like Alhaitham and Cyno are on this end
of the spectrum. These are characters that utilize a lot of different stats to compound their damage,
and it makes it a lot easier to find – not perfect Artifacts – but at least workable Artifacts. Most
damage dealers fall somewhere in between these two extremes. Most characters have a desirable substat
pool consisting of one primary-scaling stat like ATK or HP, then Crit Rate and Crit DMG, and
finally EM or ER, or sometimes both. Aside from the difference in desired substats, one other
major difference between farming for Xiangling and Wanderer is that Xiangling uses the Emblem of
Severed Fate set, which is a set that’s available from the Artifact Strongbox. This means that
even if we find an Artifact that’s not desirable, we can recycle it to give us another chance
at finding something better. One thing that newer players might not know about the Artifact
Strongbox is that it actually has a significantly higher chance of generating a 4-line Artifact
compared to Artifact Domains. Roughly 1 in 5 Artifacts you get from Domain runs will start with
4 substats. This rate is increased to about 1 in 3 for Artifacts you get from the Strongbox. So not
only does using the Strongbox give you a second or third chance at rolling a desirable Artifact, but
it also statistically produces better Artifacts. When you compound these two benefits together,
we end up cutting our farming time by a pretty significant amount. Let’s see what our build
progression for Xiangling looks like when I enable the usage of the Strongbox. This means that after
we’re done farming and upgrading our Artifacts, we recycle all the Artifacts that we didn’t
upgrade, and look for new Artifacts to upgrade out of those that come out of the Strongbox.
Then, we keep repeating this process until there are no more Artifacts left to recycle. And
this is what we get: At the end of the 20 weeks, we’re about 2.4% stronger than if we hadn’t used
the Strongbox. And this alone doesn’t sound very impressive. But rather than looking at how much
stronger we got, I’m a lot more interested – from a Resin Efficiency perspective – in finding out
how much faster we hit the same stat benchmarks compared to when we weren’t using the Strongbox.
In the original simulation we ran with Xiangling, we achieved about 90% of our 20-week goal in
about 6 weeks. If we compare that against the run where we did use the Strongbox, we’re now able
to achieve that somewhere in between 4 and 5 weeks instead. Let’s just call that 4.5 weeks. If we’re
aiming for that 10-week benchmark like I mentioned earlier – or about 95% of the power of our
20-week goal – the Strongbox allows us to hit this benchmark somewhere between Week 7 and 8 instead.
Again, let’s just call that 7.5 weeks. In total, we save about 2.5 weeks worth of Resin on farming
Artifacts. This is 2.5 weeks worth of Resin we can use to start building a new character sooner;
or work on other parts of our build; maybe even clear the Spiral Abyss that much sooner. The
point is, being able to obtain extra Artifacts through the Strongbox cuts down our farming time
by a really big amount. If we look at Week 14 of our “Strongbox-assisted” run, we’ve already
hit the 20-week benchmark from the original simulation where we weren’t using the Strongbox.
On average, we’re hitting the same stat benchmarks approximately 1.4 times faster as a result of
using the Strongbox. Another way to think about this is that the Strongbox multiplies our Resin
efficiency by about 1.4x, which is a huge factor. Because of this, using the Strongbox is the single
most impactful thing you can do to improve your Resin efficiency when farming Artifacts. Please
do not feed your +0 5-Star Artifacts for Artifact experience. Hold onto them, run them through the
Artifact Strongbox for other sets that you need, and you will save yourself weeks or months worth
of Resin over time. I know some newer players might be hesitant to use the Strongbox because
it feels like you’re losing a lot of Artifact experience by not feeding those 5-Stars instead.
But please trust me when I say this – the Resin that you save from using the Strongbox is going
to be infinitely more valuable than the experience that those 5-Stars would be worth. Let’s just
consider the scarcity of the two resources. You come across Artifact experience in a lot of
different ways as you play through this game – you can get it from your Teapot, you can get it
through exploration, from the 4-star and 3-star Artifacts that you get from the Domain runs.
Over time, you will very naturally accumulate an abundance of Artifact experience. Resin, on
the other hand, not so much. Extra Resin is very difficult to come by and that is intended by the
game’s design. Resin is one of the two premium resources in this game that ultimately gate your
progress – the other one of course being Fates, or more generally, Primogems. At a very basic
level, Fates allow you to expand your character roster and Resin allows you to strengthen your
character roster. So given that Resin is one of these two premium “progression-gating” resources
as I like to call it, it’s really important to use every tool that you can to make the most of
your Resin. And the Artifact Strongbox is exactly one of these tools that can really add a lot of
additional value to the Resin that you spend. One last thing I wanted to look at is what this
build progression looks like for a character that cares about only one stat. A prime example that
many players already know about is Kuki Shinobu, when she’s built as a Hyperbloom carry, who only
cares about maximizing her EM. And this is what her stat progression looks like. What we see is
that her progression is even more front-loaded than Xiangling’s. After just 1 week of farming,
we’ve already achieved 74% of the damage we would output at Week 20, and by Week 4 we’re already
at 90%. Understanding why this happens is kind of related to the first major lesson that we learned,
and it’s that the Artifact Main Stat makes up the bulk of its stat budget. And as it happens, EM is
available as a Main Stat on 3 different pieces: the Sands, Goblet and Circlet. Once you find
these 3 pieces with that EM Main Stat on them, there really isn’t that much more room to improve
the build. The only way you can get more EM is from the Substats of the Flower and Plume, but
those are relatively small increases in EM that are going to take a much longer time. So it’s a
combination of the fact that Kuki only needs EM and that she doesn’t need to spend extra time
looking for better Substats on these 3 pieces that allows her to ramp up her damage so quickly.
In general, characters that have mono-stat builds like this are extremely Resin-efficient to
build. Zhongli, most Anemo supports, and the damage-augmenting family of supports are some
other characters that fall under this category. These types of characters are really valuable to
have in a new player’s roster because they add a lot of value to your teams without requiring a
large Resin investment to start being effective. This is a big part of the reason why characters
like Zhongli and Kazuha are used on so many different teams by a lot of players. Of course,
the main reason is that they have very strong, generalized kits that allow them to fit into many
different teams. But cost to build and time to find viable Artifacts is an often overlooked, but
important factor, especially if you’re a newer player. If you are a veteran player and you’ve
been playing Genshin for a long time, you probably don’t care much about Resin optimization anymore.
You probably have a lot of Weapons already at 90, an overabundance of Ascension materials, and
lots of +20 Artifacts that are just waiting for the right character to come and pick them up.
But if you’re a relatively new player who’s just starting to get into the mid or late game,
I’m sure you can already feel that Resin is an extremely limiting factor to your Characters’
progress. And starting at AR45, the Resin that you need to spend on Artifact farming is a big part
of that challenge. With that having been said, I hope this video and the Key Takeaways that we
covered gave you a better idea of how you should approach Artifact farming in a way that allows
you to get the most out of what limited Resin you have. For a lot of people watching, there’s
a good chance that you’re already familiar with the concepts that we discussed. “Prioritize Main
Stats, focus on Substats later” – not exactly a novel idea considering how long theorycrafting
in this game has been around. But even if you are a veteran player, I hope the visualizations of
the stat progression and the ability to quantify just exactly how much stronger we’re getting
per Resin spent was still a valuable takeaway. At least for me, these were the kind of findings
that motivated me to write this program and make this video in the first place. Regardless
of how far you are into Genshin, I hope you learned at least one new thing from watching this
video. As always, thank you so much for watching.
