10% double is slightly better than 25% return. It has been calculated and acknowledged for long time.
SharpkillerFTW2026-01-05 04:21
Is it like slightly better to the point it’s noticeable or just slightly better but negligible?
swagl0rd420dstep2026-01-05 04:25
it's pretty negligible
25% will give you return on a blue book and 3 blue book = 1 gold book
If you use blue book as the standard comparison, you get 0.25 value
10% will get you 1 gold book = 3 value, so you get 0.3 value
25% potentially allows you to get "more refunds", so it's value is technically slightly higher than 0.25 but the diff between the 2 is very small anyways.
SharpkillerFTW2026-01-05 04:29
So unless i REALLY care about that small margin, it doesn’t really matter?
swagl0rd420dstep2026-01-05 04:32
if you don't really care about the small margin it doesn't really matter much, the 10% double is still the one that gets you better returns mathematically so for most people that doesn't care they would naturally take the 10% chance
SharpkillerFTW2026-01-05 04:35
Ahh I got it, normally I’d only use 10% for when I craft like 30+ otherwise just stick to 25%
Greypawz2026-01-05 05:19
I use the 25% one if I’m one off of crafting another book. Say I’m crafting purple books, but I have 29 blue books. If I just craft as is, I’ll get 9 books and have 2 left over. So I usually single craft purple books until I get the 25% blue refund, which lets me craft another purple, and then use the 10% from there.
ataquemosle2026-01-05 06:54
this is the correct answer
BuilderAura2026-01-05 07:10
oh that's clever.
I set it to Eula back when I got her on her first banner and just never looked back XD
IgniteTheBoard2026-01-05 05:01
10% is better by such a small margin that it is just not worth the risk. 25% of the lower rarity has high odds of triggering since 25%>10% that you'll just find so many times that the 10% trigger gave you nothing and 25% trigger gave you a small reward. When the 10% procs it gives a much better reward but it has such big variance that it either gives you nothing or it over procs and gives you too much which is a waste.
Pinkerino_Ace2026-01-05 05:40
Sorry, but you aren't making any sense mathematically. The whole point of 10% being marginally better is because over a large enough sample size, 10% procs gives more value than the 25%.
So saying its not worth the risk is so mathematically inaccurate.
IgniteTheBoard2026-01-05 05:50
You're not making the same books at all times. If you trigger 10% on book type a and fail it on book type b then the net would be better than it with the 25% proc due to the maths involved but the thing is you get too much return on the 10% proccing on book a hat it doesn't actually make up for losing the 10% on book b since book a and b are not convertible.
Pinkerino_Ace2026-01-05 06:00
What have making same books got to do with this? The odds isnt dependant on what books you are crafting. In genshin terms, the "pity" doesn't reset just because you decide to craft a different book midway.
IgniteTheBoard2026-01-05 06:13
Because the sample size of creating books is not very big. If you're making 10 books then the odds of the 10% not proccing at all is 35%. Which is very significant and when it does happen more than twice that luck is just wasted as there is just more variance when something happens 10% of the times and the other happening 25% of the times. And if the expectation value for both events is so ridiculously close it's better to just go for the event which has less variance
Pinkerino_Ace2026-01-05 06:30
I dont understand a single word you are saying.
The whole premise of probability is you need to have a large enough sample size and it will reduce variance of sample mean.
If you are crafting 10 books; it doesnt tell you much because there's always chance of random error.
If you played genshin for 5 years and crafted thousands of books; the 10% proc chance will on average give more value.
It just math and probability, not an opinion, so yes I can confidently say you are wrong.
QWERTYAF12412026-01-05 06:23
The difference between 0.25 and 0.3 is a 20% increase in expected returns. Approximately since any extra material that's not already at the highest tier can be further crafted with a slight chance to give a big more back. It ranges but 10% double will always be better in the long run. Not using it is just wasting free resources.
AxShepard2026-01-05 11:22
I usually use 25% for green and blue books if I'm short on 1 or 2 books for the next tier. For gold I usually use 10%. One time Layla felt generous and got 3 extra gold books for 10 crafted ones. I got pretty shit luck with RNG though so I don't trust that 10% over much.
PieBob8512026-01-05 13:23
You did note it, but the way it should be viewed is one gives 25%/3 (11/12 the cost) less cost while the other gives 10% more value. We need to set costs equal, so it's 12/11 vs 11/10 and 11/10 is clearly slightly higher (1.09091 vs 1.1)
babangelsin2026-01-05 04:26
It is 5% better returns on crafted materials on average. You don't need deeper calculations than that.
Gregamonster2026-01-05 04:27
On a large enough time scale, 25% for 1\3 of an extra book gets infinitely closer to 10% for an extra book, but never quite reaches it.
babangelsin2026-01-05 04:29
It doesn't. With increasing time (which I assume means more books being converted) the 5% difference between them becomes only more pronounced.
Gregamonster2026-01-05 04:42
Only if you don't account for the extra crafting instances you can make with the refunded materials.
With the refunded materials you get an infinite feedback loop that gets you increasingly close to them being even. But it never quite reaches it.
2. With the 10% option, sometimes you get an extra purple. On average it's 3 blues to 1.1 purple.
3. With the 25% option, sometimes you get refunded a blue. On average its 2.75 blues to 1 purple.
1.1 / 3 = 0.3667
1 / 2.75 = 0.3636
10% option is very slightly better.
GateauBaker2026-01-05 05:32
Still not quite, even taking that into account, summing the expected values of each conversion gives you an infinite geometric series that sums up to approximately .99176 of what the double product passive gives you and no more. Which is close for practical purposes, but it does not get you "infinitely closer on a large enough time scale".
For reference: The refund gives an expected value of x/(1-0.08333). And the double product gives an expected value of 1.1x. Where x is the initial number of product value.
refance2026-01-05 04:27
In 99 conversions you will get 33 converted books + 9 converted books or 24 unconverted books that can become 8 converted books. The difference is negligible
Shadourow2026-01-05 05:07
I think you've confused yourself here
Better do 100 conversions (300 blue books into 100 purple books)
Two cases :
\- 10% : 100 purple books + 10 bonus purple books
\- 25% : 100 purple books + 25 bonus blue books
\- - 25 blue books (8 conversions) → 8 purple books + 2 bonus blue book and 1 blue book left alone
\- - - 3 blue books → 1 purple book + 0.25 blue book
Conclusion :
For 100 conversions :
\- 10% gives 110 purple books
\- 25% gives 109 purple books + 0.25 blue books
Small difference indeed, but I'm not sure of what you mean here
>33 converted books + 9 converted books or 24 unconverted books that can become 8 converted books
QWERTYAF12412026-01-05 06:28
The difference is bigger when you use it on lower tier books. Because extra blue books from crafting blue books can then be crafted again. While getting green books from blue books and then coming green books is less valuable. Regardless, it's always better to go 10%. It's not even a gamble since it's just a passive talent. It either happens or doesn't.
The3nd0fT1me2026-01-05 10:38
Normal is 110 for 330 resources. (1 for 3)
10% extra gives 121 for 330 resources. (121=110*1.1)
25% refund leads to 120 for 330 resources. (330 = 360-360*0.25/3)
So the 10% option leads to one more product out of 330 resources.
heimdall17062026-01-05 16:38
So in the long run, always 10%, noted
cartercr2026-01-05 15:04
Let’s say you craft 10,000 materials. (You know, enough to where law of large numbers should really be affecting it.) Layla should have made 1,000 materials. Xingqiu, by contrast, will have made 2,500 of the lower ranked materials, which is only 833 materials of the same quality.
In the grand scheme of things it doesn’t matter too much, but it is technically a decent amount better.
BawsYannis2026-01-05 18:55
In reality its around 909 materials not 833, because crafting the 833 gets you mats to craft another 70 ish, those 70 gets you mats for like a few more
cherrycoco72026-01-08 15:58
But our conversions usually are not that large.
In my exp, I had more refunds of 25% than would convert to more golds than the 10% golds .
What me saying is , while probability of 25% refund and 10% double are good on paper, their likelihood hurt me.
177013-228922-42992026-01-05 15:46
if you have a lot of materials, 25% is better.
if you need 1 more to make 3, 25% is better.
otherwise go with 10%.
reason 25% ends up better than 30% (3x10%) is because you get them as the lower level book which you again craft into larger and thus can get another 25% every 3 procs, simple math
Axthen2026-01-05 15:49
The way my monkey brain sees it.
10% chance only needs to proc once to be more valuable than two 25% activations. And equal value with three 25% activations. However, three 25% activations is less common than one 10% activation
And the value scales.
If you get one gold extra, that's worth a lot more than just two purples refunded (but you would have to craft at least two golds to even have a chance at two purples being refunded.)
Count_Nick2026-01-05 04:48
I remembered a friend saying 25% is better but now 10% is better? Is it also faster for farming so that I don't need a quarter of a year to build a character's skills and levels?
ConciseSpy850672026-01-05 08:26
Well, since the lower tier materials are basically a third of the value of the next tier materials, that means the refund would need to be triple the value of the duplicate to break even, and 10x3 isn’t 25
CadetC2026-01-05 08:31
On paper, it is. In practice, it depends on pure luck.
MightySamMcClain2026-01-05 10:30
It's very slightly better when doing big numbers like 20+. If you are doing like 5 it's worse bc you usually get none
GinahyaKangaroo2026-01-05 14:48
Amused.. That's basic expected value math.
Neospanner2026-01-05 21:09
It's amazing, looking at the comments under this response, how many people rationalize taking the provenly less-valuable option.
If you're going to be playing this game for any length of time, you're going to be making this gamble (assuming you have the characters to you need in order to be able to choose one or the other). Yes, if you take the 25% option, you'll see "wins" more often - but in the long run, you'll be getting less stuff overall.
No, it's not better to take the 25% option if you're doing only one book, or whatever. That one book is still part of the massive number of books you'll be collecting during your playthrough of the game, and every opportunity you have to make the 10% bet is more opportunities to win the 10% bet.
There's some kind of gambler's fallacy going on here, but I'm not sure what it is.
callmejamesx2026-01-05 04:19
not much to really gamble
10% double on avg is the best option in terms of returns
SharpkillerFTW2026-01-05 04:23
If that’s the case then I’ll probably start doing the 10%
I just haven’t seen it hit unless I craft 30+ while that 25% feels like 75% with the way it consistently hits
notthatjaded2026-01-05 04:30
It might hit more often but you're only getting refunded materials so it *has* to hit more often for you to be able to craft the same amount of materials you'd be getting from the "double" option. Of course, if you think you might also want to craft/have the lower level stuff in the future for some future character/weapon, it might not hurt to do it that way instead.
Christian71572026-01-05 15:08
My average must be skewed towards the 25% then, bc I've never seen the 10% trigger, like ever.
takenusername50012026-01-05 04:24
always
MyClientsBark2026-01-05 04:24
I do because the dopamine hit when it works feels good.
TheGungnirGuy2026-01-05 04:26
I switch at my fancy.
In my experience, either does whatever the hell they want. Sometimes I do an entire round of 100 and get nothing from the ten, but three rounds of three on the 25 gives me one a piece. Overall, they balance out in irrelevancy.
Evening_Parking26102026-01-05 04:31
I ask for returns when i still need some of the pre level up materials and use 10% when im feeling lazy and dont want to get that last whatever
-Illuzio-2026-01-05 04:34
i usually do the 25 percent when i craft a ton of things. and when im down to my last craft i do the 10 percent
T-280_SCV2026-01-05 04:35
I use Xingqiu until/unless I have a multiple of three, then all-in the remainder with Layla.
Holloway-Tape2026-01-05 04:44
This is super stupid mathematically but I use the 25% chance refund for common mats and the 10% chance for rarer ones. My rationale is that having common mats for base levels is nice to have on hand just in case and the dopamine hit of getting double rares makes me feel like a big man.
Nearby-Control-11192026-01-05 04:57
I gambled all the time. It FELT less than 10%
BeautyJester2026-01-05 05:04
Last i search it up, common consensus is 10% is more preferably
Attractive_Sock2026-01-05 05:26
always
DarKXkrules772026-01-05 05:26
More than slightly better, it's a whole 1/6 worse to use the 25%, expected return from 1000 tries it's 300 vs 250, so you should swap to the 10% If you want the better results
SecretSpectre112026-01-05 05:27
Law of large numbers, do it enough times and the sample mean will converge to the true mean. 10% is better on average. Since you're crafting thousands of materials it will be better in the long run. There really isn't much of a gamble here.
Sum4nJ2026-01-05 05:27
Only when the materials are divisible by 3. Gotta satiate my OCD.
Evodius__2026-01-05 05:28
Gambling only in the sense that the Probability theory evolved from gambling.
Ze_AwEsOmE_Hobo2026-01-05 05:40
If the talent books getting converted are divisible by 3, I use Layla. If they aren't, I use Xingqiu.
Sometimes I will go from 8 common books into 2 rare and 1 common from XQ's passive, then use the extra book to create another a rare with Layla, which occasionally gives me 2 rares.
worquinnprogress2026-01-05 07:22
This is the truth and what I do too !! 😤🤝
Hellcat7272026-01-05 05:42
Comparing to pull rate, 10% is too high to pass.
smartfalco2026-01-05 05:52
last i saw a graph on it, it was like laylas is 0.0027361% better or something (exaggerating)
bananaroll_2026-01-05 05:55
i always gamble that 10%
Ksarkin2026-01-05 05:56
I always do 10% buff unless I am very close to preferred quantity. You can always craft higher tier, but not disassemble to lower tier.
magnidwarf19002026-01-05 06:03
Yes
pc_player_yt2026-01-05 06:06
yeah because it just feels more satisfying when you see the bigger material drops. I always choose Sucrose over Dori when crafting these
QWERTYAF12412026-01-05 06:19
Everyone does. It gives slightly better returns.
VongQuocKhanh2026-01-05 06:26
I go with Yae cause sometimes I just need a few books but don’t feel like using resin on that specific domain
notokawaiiyo2026-01-05 06:46
My standard used to be 6/6/6 for all characters, so I needed more of the blue books than gold so I used to use the 25% return only when crafting gold. Now that I'm doing more 9s and 10s, I should use the 10% more.
wphxyx2026-01-05 06:52
What gamble? The 10% for double is just better, mathematically. Put yourself on Layla and forget about it: you'll be better off.
SinesterBrayn232026-01-05 07:15
I only do it when I either only have enough mats to synthesize one time, or I don’t believe I’ll get enough mats back (from Xinqiu) to get an additional synthesize.
Proper_C_nt2026-01-05 07:28
I always do it based on what I need. If i have enough mats to upgrade, ill go xingqiu for the refund(also if i have like 20 mats and i only need 1 more for another conversion), but if i have enough to get 5 higher mats and I NEED 6, ill go to the 10% (again, only if my lower mats need 2 or more to form another 3->1 conversion)
zickzi2026-01-05 07:28
If i craft lower than 30 items then i chose 25% chance. But higher than 30 i chose 10% double, work pretty well
ConciseSpy850672026-01-05 08:24
Technically it’s just straight up better, a 10% chance to get a dupe or a 25% chance to get a refund, but the conversion rate for lower tier to higher tier is 3-1 meaning it would need to be a 30% chance for a refund to make it even…but then you’d just use the refund every time unless you’re a true gambler
Adventurous-Risk59192026-01-05 08:25
I use it for the higher crafting
DrDerpyDerpDerp2026-01-05 08:31
when crafting purple 10% is always better
Alexasbc2026-01-05 08:36
i always use the 10% chance for extra units of the crafted material just because of how i manage resources.
i usually craft things to the next level when i have as much of the previous level as i need, so 10% chance triggering is way more valuable for my resin efficiency then the 25% chance
Chuck0062026-01-05 08:41
I only use Layla. Over time, it's much better value.
SHIN-YOKU2026-01-05 08:46
Depends on quantity for me, in large batches I find getting enough to craft again better.
trueHolyGiraffe2026-01-05 08:55
Lets convert 300,000 blues into 100,000 purples!
I'm naming a huge number because it doesn't matter if we do it all at once, or one-by-one, so in a sense, it may as well be your entire genshin-career, here's the table:
|How we'll convert|How we'll convert|End result (ON AVERAGE, OBVIOUSLY)|
|:-|:-|:-|
|Using the 10% talent|The conversion is pretty straight forward, just multiply 100,000 by 1.1.|**110,000 purples**|
|Using the 25% talent|For 300,000 blues I get 100,000 purples, and get an extra 25,000 blues on my first conversion. Let's convert these extras into purples! I'll get 8333 more purples (total 108,333) and have 1 spare along with 2,083 new blues! Let's convert these extra-extras! That's 694 more purples (total 109,027) with 2 spare blues, and 173 new blues. Let's convert these extra-extra-extras! That's 58 more purples (total 109,085) with 1 spare blue, and 14 new blues. Let's convert these extra-extra-extra-extras! You see where this is going, its an infinite sum formula. S≈4/11 x 300,000=109,090.909...|**109,090,909... purples**|
So yeah, I always pick the 10%, unless I'm down to like - 5 last piece of blue, and I only need 2 more purple,
then (and only then) I convert with the 25% because it saves me another run with a 25% chance of success rather than 10%.
EDIT: I'd like to point out the difference is SO SMALL, its negligible almost always.
Only in the REEEEEEEEEEEEEEEEEEEALLY long term, its a "loss" of about 0.83%, if you kept using the 25%, and it doesn't take into account the time, and needs of the player, sometimes YOU NEED more blue rather than purple, for example.
It BARELY matters.
_zacque2026-01-05 09:01
I prefer using 25% refund to build other characters. if I don’t have enough materials, I will gamble for 10% to get an extra gold material jajaja.
flippin_Cal2026-01-05 09:01
Rule of the thumb is if you have a lot of materials to craft then double but if you have less materials then refund
No-Piano-9872026-01-05 09:04
An easy way to think about it is expectation. If you make 100 purple books, you could expect 10 extra purples with the 10% chance to double or 25 extra blues with the 25% chance to refund a blue. 10 purples are worth 30 blues so the 10% chance to double is better.
sorarasyido2026-01-05 09:06
If the materials used has more than 50 pcs, I would pick Layla.
Any less is Xingqiu. Why 50? Just my personal sweet spot.
oncffch2026-01-05 09:18
Omg I once crafted like over 30 talent stuff with her, I got nothing so I decided to just always go with the bigger chance.
Ironexeption2026-01-05 09:39
Yes. Because i use Eula for it every time since i got her.
PaulStarhaven2026-01-05 09:46
Yes
esmelusina2026-01-05 09:57
They are basically the same, it’s better to pick whichever gives you better crafting efficiency.
Do 25% until you hit a multiple of 3, then just use 10%.
FigTechnical80432026-01-05 10:04
I used Layla on quite a few because I never read, she actually does return quite well, but the more you create the more obvious it is. I'll often slog my guts out to get materials and she's sat there going "here's 4 spares" thanks Layla, I'll use them on the next one. Checks the next one...different book, whyyyyyyyy!!!???
jonnevituwu2026-01-05 10:08
Only when Im doing a lot at once
Narissis2026-01-05 10:11
I gamble whichever makes the most sense.
If I'm crafting a material that I know I will also need to use for another character, I use the refund chance so I get some back for future use.
If the character I'm working on is the only one I'll be building for the foreseeable future who uses the material, I use the bonus production chance because why get refunded materials I won't use?
jisooed2026-01-05 10:15
who in the world uses 25%
Kingpimpy2026-01-05 10:16
idc what others say i keep using mona over my albedo in those
albedo NEVER procs the shit 10%
tried it too often he never performs
CoolMintMC2026-01-05 20:17
Me too.
Same for all the other ones too.
CashOtherwise87232026-01-05 10:29
Yesterday i gambled for mavuika purple core thing Made 6 got 4 bonus 🙏🏻😏 ALWAYS GAMBLE LADIES AND GENTLEMANS
Independent-Can-1832026-01-05 10:33
I will never be able to understand the calculation on these things because the 25% chance to refund a "portion" of the crafting material is not clear at all. Is the refunded portion mean that you have a 25% chance to always refund 1 blue book back or can it be 2 blue books back as well?
If you have a 25% chance to get 1/3 gold back, it will result in 8.3333% of getting a gold. And if you have a 10% chance to get 1 gold back, it will result in 10% of getting a gold. The difference will be a 1.6666% advantage for the gold. This means that the 10% is better by 1.6% which is not that much.
5UP3RBG4M1NG2026-01-05 10:34
10% gives more mats so I do
Mundane-Ad-94012026-01-05 10:37
I always do Yae Miko no matter what
Adorable-Comfort-2802026-01-05 11:34
Def would if I had her
KataklysmGI2026-01-05 11:52
I use both depending on how many items I'm crafting. The more I craft, the better the 10% becomes (since it's 3 whole item's worth of "refund").
If you craft, say, 1000 3★ items, and the odds are perfect, that 10% double production becomes 100 extra 3★ items, which are equal to 300 2★. Meanwhile, the 25% refund would mean you get back 250 2★ items, or 83 3★ ones, so 17 less in total. Over time, this average converges to become true, but well, it's RNG at the end of the day, so sometimes you might not even reap the benefits from the 10%, especially on smaller crafting sessions.
ThatRandomCanadianV2026-01-05 12:00
Crafting >20 - refund
Crafting <20 - double project
MindingMyBusiness022026-01-05 12:20
It's gambling what do you think
Frosty-Soil16562026-01-05 12:48
Me every time
ramsdit2026-01-05 12:56
I switch between them depending on if I want more silver or gold books. For example if I have enough silver books to craft all the gold books I need I'll use Xingqiu as I would just rather have a surplus of silver books on my account that future characters I'm building can use than a lot of extra gold books that might never be used.
Real-Contest49142026-01-05 13:10
I use 25% is when I have remainders already, 10% is when it has no remainders to go all in.
Silver-Bet79092026-01-05 13:37
I do because Eula is the 10% chance and she was my main until Kinich…
hollyherring2026-01-05 13:39
I always do the 10%
One-Spare-7982026-01-05 14:10
Like a wise man said
ALL OR NOTHING!
Cosimov2026-01-05 14:15
10% is better on average, but since my luck is terrible I rarely ever see the effects of it.
So 25% for meeeeeee, especially for the weapon mats because that 10% never triggers lol
DasBleu2026-01-05 14:30
lol when it defaults and I am not paying attention. Otherwise I prefer to use Yae because she’s more fun.
WhoInvitedDJ2026-01-05 14:40
I situationally take the gamble. When I'm on 3/3 for what I'm trying to craft and getting 1 back wouldn't give me enough to craft another one, I'll take the chance on trying to double it.
cartercr2026-01-05 15:07
Depends on the situation really. If I’m crafting like three gold books when I need four, but I only have two purples then I’m using Xingqiu every time. I’d rather use the higher chance of getting what I need rather than the lower chance of getting more.
However if I’m needing to convert a lot of stuff and maybe still need to farm more then I’m fishing for as much value as possible and I’m going with Layla.
Edit: Yae Miko is also a fun one. I’ve actually used her to craft before because sometimes I don’t need one material, but I need a few of a different one, so I craft with her to fish for an extra couple of the one I need. Very very inconsistent, but when it hits you feel so big brain.
Fancy-Listen-52422026-01-05 15:24
just do double when it's the low star product then when u get to purples or gold u do the 25% refund
Suspendrz2026-01-05 15:50
Double is like 3 of the refund.
Double is 3 times better and refund is 2.5 times more likely.
3>2.5
Gullible-Turnip-67222026-01-05 15:53
I do 25% when I only need 1 more mat to fuse, otherwise always 10%
JrButton2026-01-05 15:56
You're going for the other one just because it's a bigger number? That's big brain reasoning right there lol
Statistically (altho only marginally) the 10% is better
Fish_On_A_Piano2026-01-05 16:11
Always; as has been said it's just better on the numbers l, but also just to avoid the nuisance of having to craft multiple times with the lower tier book refunds
Chakolatechip2026-01-05 16:12
Yae and Layla are the only ones worth using.
reeeekin2026-01-05 16:12
If I craft a bigger amount, i Use the 25% to return some materials. If I Craft just a few, I gamble the 10%. Makes sense to me this way.
DaymD2026-01-05 16:33
The real question is who doesn't always prefer the 10% ?
Andrew120902026-01-05 16:37
10% is better because Xingqiu hates me and the 25% chance feels like 5%
neat-NEAT2026-01-05 16:38
Only really do the 25% if I'm 5/3 on my last craft.
PantheraAuroris2026-01-05 16:39
Always, because the math says so. 10% for double is just better.
Watsons-Butler2026-01-05 16:41
Mathematically, in the long run, the 10% is the better option.
BrandNewKitten2026-01-05 16:41
For me:
- Lower tier mats get the 25%
- Higher tier mats get the 10%
- Extra mats get Yae Miko’d when I am building a specific character.
Gabbyxo972026-01-05 16:58
I do. I think - but I'm no math-genius - in the end it doesn't matter that much. A slightly higher chance for slightly less. The idea of getting one extra feels nicer and more useful than getting one back. But again, no math-genius so I don't know much about probabilities and such
ce1es2026-01-05 17:02
I take Miko's 25% because I love the chaos that lady provides, not because math.
the_tygram2026-01-05 17:07
Statistically it's better. In order for the 25% chance to refund materials to be the same benefit as the 10% bonus you'd need it to proc 3 times (since that's how many materials it takes to craft again). This means it would need to happen 3x mor than the 10% bonus for them to be equal (i.e. have a proc rate of 30%). Since it only have a proc rate of 25% (2.5x the 10% bonus rate), in the grand scheme of all your crafting it gives less reward.
Conscious-Log93942026-01-05 17:22
ALL THE TIME! TO THIS DAY!!!
minobi2026-01-05 17:22
To be on par with 10% it should be 30%. So 25% bonus is 17% less effective than 10% bonus.
Cieneo2026-01-05 17:22
Craft 60 times, you'll get:
10%: 66 materials, so 6 extra.
25%: 60 materials + 15 refunds = 65 materials, but you have to *spend more mora to craft the 5 additional materials* (will most likely get 1 additional refund).
So no, 10% is not just slightly better, it is slightly better AND cheaper.
jasmsaurus2026-01-05 17:48
Sometimes I get really lucky and sucrose and Layla go off
Admirable_Fig11412026-01-05 17:54
If I still need the lower tier drops I do the refund if only need the highest tier stuff I try to go for the 10% chance to double. Ultimately I believe the 10% is ever so slightly more efficient plus u don’t need to spend mora to craft up again on the refunded books but the difference is pretty negligible iirc.
six_seasons2026-01-05 18:16
Nah 30% all day
aCozyCloud2026-01-05 18:31
I do the 25% if I'm one away from a multiple of 3 to not have any leftover, then do the 10% to maximize chances of doubling. Works pretty well !
Echos_light2026-01-05 20:17
Only when I’m mass crafting
Rare-Cat-45722026-01-05 20:33
When I have a lot of craft I use 25%, when I craft 3 item or less I try 10%
Anyway... considering my luck, it's not 25 and 10% but rather 10 and 5%...
CoolMintMC2026-01-05 20:34
I use Xingqui & Yae Miko because I like getting the passive to proc more often.
I also looked it up & apparently for 'Weapon Ascension Materials" Mona gives the 25%, whereas 4 characters share the 10% version. Those being Albedo, Ayaka, Alhaitham & Wriothesley. (I didn't know about Ayaka as I don't have her)
vorator_2026-01-05 20:43
yes obviously
Luc4r1o2026-01-05 20:54
I use 25% if the mat is not divisible by 3
Let's say 32/3 then I use 25%
If it's 33/3 then I use 10% all the way
slayaechan2026-01-05 20:57
yes, I do it when I have a lot to make and I get good returns.
FRICKENDEKU2026-01-05 22:21
I do. Just in case I ever get that 10%. I don't get it often, but I get it.
Kronman5902026-01-05 23:25
25% until im at a multiple of 3 then 10% the rest
FicklePangolin17172026-01-06 01:43
I always do 25% Xinqui from brown to silver, and 10% from Eula from Silver to Gold. I only used Yae Miko if needed books for several characters I was building. But now I only build one at a time.
Commentaires
25% will give you return on a blue book and 3 blue book = 1 gold book
If you use blue book as the standard comparison, you get 0.25 value
10% will get you 1 gold book = 3 value, so you get 0.3 value
25% potentially allows you to get "more refunds", so it's value is technically slightly higher than 0.25 but the diff between the 2 is very small anyways.
I set it to Eula back when I got her on her first banner and just never looked back XD
So saying its not worth the risk is so mathematically inaccurate.
The whole premise of probability is you need to have a large enough sample size and it will reduce variance of sample mean.
If you are crafting 10 books; it doesnt tell you much because there's always chance of random error.
If you played genshin for 5 years and crafted thousands of books; the 10% proc chance will on average give more value.
It just math and probability, not an opinion, so yes I can confidently say you are wrong.
With the refunded materials you get an infinite feedback loop that gets you increasingly close to them being even. But it never quite reaches it.
You're right,
[https://www.wolframalpha.com/input?i=sum+of+0.25\*\*n](https://www.wolframalpha.com/input?i=sum+of+0.25**n)
I guess I'll go for the 25% for lower variance now
edit : wait a second, 33% isn't 30%, am I confusing myself right now
edit2 : I guess my brain is fried, can't do the maths rn, but just simply doing an example case is easy enough, 10% is indeed better
[https://www.reddit.com/r/Genshin\_Impact/comments/1q4adr4/comment/nxrc49x/?utm\_source=share&utm\_medium=web3x&utm\_name=web3xcss&utm\_term=1&utm\_content=share\_button](https://www.reddit.com/r/Genshin_Impact/comments/1q4adr4/comment/nxrc49x/?utm_source=share&utm_medium=web3x&utm_name=web3xcss&utm_term=1&utm_content=share_button)
1. Baseline conversion is 3 blues to 1 purple.
2. With the 10% option, sometimes you get an extra purple. On average it's 3 blues to 1.1 purple.
3. With the 25% option, sometimes you get refunded a blue. On average its 2.75 blues to 1 purple.
1.1 / 3 = 0.3667
1 / 2.75 = 0.3636
10% option is very slightly better.
For reference: The refund gives an expected value of x/(1-0.08333). And the double product gives an expected value of 1.1x. Where x is the initial number of product value.
Better do 100 conversions (300 blue books into 100 purple books)
Two cases :
\- 10% : 100 purple books + 10 bonus purple books
\- 25% : 100 purple books + 25 bonus blue books
\- - 25 blue books (8 conversions) → 8 purple books + 2 bonus blue book and 1 blue book left alone
\- - - 3 blue books → 1 purple book + 0.25 blue book
Conclusion :
For 100 conversions :
\- 10% gives 110 purple books
\- 25% gives 109 purple books + 0.25 blue books
Small difference indeed, but I'm not sure of what you mean here
>33 converted books + 9 converted books or 24 unconverted books that can become 8 converted books
10% extra gives 121 for 330 resources. (121=110*1.1)
25% refund leads to 120 for 330 resources. (330 = 360-360*0.25/3)
So the 10% option leads to one more product out of 330 resources.
In the grand scheme of things it doesn’t matter too much, but it is technically a decent amount better.
In my exp, I had more refunds of 25% than would convert to more golds than the 10% golds .
What me saying is , while probability of 25% refund and 10% double are good on paper, their likelihood hurt me.
if you need 1 more to make 3, 25% is better.
otherwise go with 10%.
reason 25% ends up better than 30% (3x10%) is because you get them as the lower level book which you again craft into larger and thus can get another 25% every 3 procs, simple math
10% chance only needs to proc once to be more valuable than two 25% activations. And equal value with three 25% activations. However, three 25% activations is less common than one 10% activation
And the value scales.
If you get one gold extra, that's worth a lot more than just two purples refunded (but you would have to craft at least two golds to even have a chance at two purples being refunded.)
If you're going to be playing this game for any length of time, you're going to be making this gamble (assuming you have the characters to you need in order to be able to choose one or the other). Yes, if you take the 25% option, you'll see "wins" more often - but in the long run, you'll be getting less stuff overall.
No, it's not better to take the 25% option if you're doing only one book, or whatever. That one book is still part of the massive number of books you'll be collecting during your playthrough of the game, and every opportunity you have to make the 10% bet is more opportunities to win the 10% bet.
There's some kind of gambler's fallacy going on here, but I'm not sure what it is.
10% double on avg is the best option in terms of returns
I just haven’t seen it hit unless I craft 30+ while that 25% feels like 75% with the way it consistently hits
In my experience, either does whatever the hell they want. Sometimes I do an entire round of 100 and get nothing from the ten, but three rounds of three on the 25 gives me one a piece. Overall, they balance out in irrelevancy.
Sometimes I will go from 8 common books into 2 rare and 1 common from XQ's passive, then use the extra book to create another a rare with Layla, which occasionally gives me 2 rares.
i usually craft things to the next level when i have as much of the previous level as i need, so 10% chance triggering is way more valuable for my resin efficiency then the 25% chance
I'm naming a huge number because it doesn't matter if we do it all at once, or one-by-one, so in a sense, it may as well be your entire genshin-career, here's the table:
|How we'll convert|How we'll convert|End result (ON AVERAGE, OBVIOUSLY)|
|:-|:-|:-|
|Using the 10% talent|The conversion is pretty straight forward, just multiply 100,000 by 1.1.|**110,000 purples**|
|Using the 25% talent|For 300,000 blues I get 100,000 purples, and get an extra 25,000 blues on my first conversion. Let's convert these extras into purples! I'll get 8333 more purples (total 108,333) and have 1 spare along with 2,083 new blues! Let's convert these extra-extras! That's 694 more purples (total 109,027) with 2 spare blues, and 173 new blues. Let's convert these extra-extra-extras! That's 58 more purples (total 109,085) with 1 spare blue, and 14 new blues. Let's convert these extra-extra-extra-extras! You see where this is going, its an infinite sum formula. S≈4/11 x 300,000=109,090.909...|**109,090,909... purples**|
So yeah, I always pick the 10%, unless I'm down to like - 5 last piece of blue, and I only need 2 more purple,
then (and only then) I convert with the 25% because it saves me another run with a 25% chance of success rather than 10%.
EDIT: I'd like to point out the difference is SO SMALL, its negligible almost always.
Only in the REEEEEEEEEEEEEEEEEEEALLY long term, its a "loss" of about 0.83%, if you kept using the 25%, and it doesn't take into account the time, and needs of the player, sometimes YOU NEED more blue rather than purple, for example.
It BARELY matters.
Any less is Xingqiu. Why 50? Just my personal sweet spot.
Do 25% until you hit a multiple of 3, then just use 10%.
If I'm crafting a material that I know I will also need to use for another character, I use the refund chance so I get some back for future use.
If the character I'm working on is the only one I'll be building for the foreseeable future who uses the material, I use the bonus production chance because why get refunded materials I won't use?
albedo NEVER procs the shit 10%
tried it too often he never performs
Same for all the other ones too.
If you have a 25% chance to get 1/3 gold back, it will result in 8.3333% of getting a gold. And if you have a 10% chance to get 1 gold back, it will result in 10% of getting a gold. The difference will be a 1.6666% advantage for the gold. This means that the 10% is better by 1.6% which is not that much.
If you craft, say, 1000 3★ items, and the odds are perfect, that 10% double production becomes 100 extra 3★ items, which are equal to 300 2★. Meanwhile, the 25% refund would mean you get back 250 2★ items, or 83 3★ ones, so 17 less in total. Over time, this average converges to become true, but well, it's RNG at the end of the day, so sometimes you might not even reap the benefits from the 10%, especially on smaller crafting sessions.
Crafting <20 - double project
ALL OR NOTHING!
So 25% for meeeeeee, especially for the weapon mats because that 10% never triggers lol
However if I’m needing to convert a lot of stuff and maybe still need to farm more then I’m fishing for as much value as possible and I’m going with Layla.
Edit: Yae Miko is also a fun one. I’ve actually used her to craft before because sometimes I don’t need one material, but I need a few of a different one, so I craft with her to fish for an extra couple of the one I need. Very very inconsistent, but when it hits you feel so big brain.
Double is 3 times better and refund is 2.5 times more likely.
3>2.5
Statistically (altho only marginally) the 10% is better
- Lower tier mats get the 25%
- Higher tier mats get the 10%
- Extra mats get Yae Miko’d when I am building a specific character.
10%: 66 materials, so 6 extra.
25%: 60 materials + 15 refunds = 65 materials, but you have to *spend more mora to craft the 5 additional materials* (will most likely get 1 additional refund).
So no, 10% is not just slightly better, it is slightly better AND cheaper.
Anyway... considering my luck, it's not 25 and 10% but rather 10 and 5%...
I also looked it up & apparently for 'Weapon Ascension Materials" Mona gives the 25%, whereas 4 characters share the 10% version. Those being Albedo, Ayaka, Alhaitham & Wriothesley. (I didn't know about Ayaka as I don't have her)
Let's say 32/3 then I use 25%
If it's 33/3 then I use 10% all the way