100 days of crafting daylies - [BUG] Offest for Tailoring

geophonic_ESO

Dear crafters, dear ZOS,

I have finally reached the number of 100 craft daylies on ONE character and would like to point out a bug and/or bias I have found with tailoring.
Hard data first:


one character, all professions maxed since pre-Homestead, all traits known, all achievements, all pre-Homestead motifs known, all Homestead motifs known within first 30 days of Homestead.

Now, as you can see Tailoring has a much lower chance at receiving a master writ, although it should (with morifs and all traits learned) share the same overall chance of getting one.
Another observation is that somehow Alchemy and Provisioning have the same chance at getting a writ, although Provisioning should benefit from recipe knowledge, while Alchemy is only dependant on skill and number of known reagents (which by the time you are doing writs should be at 100%). Similarly Enchanting, which should rely on a similar system to determine the chance for a master writ as alchemy, i.e. skill and known glyph components, has a higher chance at seeing a master writ.

I am overall not complaining about the total number of vouchers I have gathered, since I was able to buy all the desired goods I wanted so far.
I`d rather point out a flaw in the system that, in my opinion, should be considered a bug.

@ZOS_GinaBruno , @ZOS_JessicaFolsom


cheers

(disclaimer: no members of House Hlaalu have been harmed during the course of this study)

ninti

geophonic_ESO wrote: »

Dear crafters, dear ZOS,
Now, as you can see Tailoring has a much lower chance at receiving a master writ, although it should (with morifs and all traits learned) share the same overall chance of getting one.

Dude, it's just RNG, there is probably no bug. You don't have anywhere enough of a sample size to tell if there is an actual bug yet. No one does really.

Synfaer

My stats are similar BUT woodworking is the odd one out.
On my main crafter (all motifs/traits/achievements on all crafts) the drop rates for Master Writs are as follows;

From 100 daily writs;

Blacksmith 17% (684 WV)
Clothing 16% (416WV)
Woodworking 7% (123 WV)

For my other crafters (most achievments/6-8 traits/2-3 motif sets);

From 600 daily writs
Blacksmith 6% (1105 WV)
Clothing 7% (1292 WV)
Woodworking 9% (908 WV)

From these stats it is clear that my main crafter has(had) a significant advantage in terms of master writs.
I say 'had' because drops rates have been very lean for my main crafter this last four weeks.


For the consumable crafts, all my crafters have mostly the same stats (all that differs is the housing recipe achievements).

From 700 daily writs
Alchemy 15% (375 WV)
Enchanting 17% (513 WV)

Provisioning was borked for a substantial time and the recipe count on all my crafters varies, so the stats are meaningless.

geophonic_ESO

ninti wrote: »

geophonic_ESO wrote: »

Dear crafters, dear ZOS,
Now, as you can see Tailoring has a much lower chance at receiving a master writ, although it should (with morifs and all traits learned) share the same overall chance of getting one.

Dude, it's just RNG, there is probably no bug. You don't have anywhere enough of a sample size to tell if there is an actual bug yet. No one does really.

The sample size is large enough to see the bias from Tailoring.
Apparently you have no clue about the meaning of RNG and Statistical chances. This is not just getting unlucky at the rolled dice, it is a significant offset.
It is however quite interesting to see that i.e. Synfaer sees this offset with Woodworking.

InvitationNotFound

If I get this right, you did only 100 crafting dailies in each category so far? I'm sorry, but this is far away from being representative.

I've posted my stats a while ago here:
https://forums.elderscrollsonline.com/en/discussion/324880/master-writs-statistics

doing crafting dailies on 8 / 9 chars for 20 days -> 160 - 180 crafting dailies in each category. Still not representative.
I might update this statistic if people are interested even that I've stopped doing crafting dailies lately (lack of time... doing it with 8-9 chars will consume ~45 minutes each day), but the current data should be somewhere around 60 / 70 days -> 480 - 630 crafting dailies in each category (still more data is needed for this to be representative).

geophonic_ESO

InvitationNotFound wrote: »

If I get this right, you did only 100 crafting dailies in each category so far? I'm sorry, but this is far away from being representative.

I've posted my stats a while ago here:
https://forums.elderscrollsonline.com/en/discussion/324880/master-writs-statistics

doing crafting dailies on 8 / 9 chars for 20 days -> 160 - 180 crafting dailies in each category. Still not representative.
I might update this statistic if people are interested even that I've stopped doing crafting dailies lately (lack of time... doing it with 8-9 chars will consume ~45 minutes each day), but the current data should be somewhere around 60 / 70 days -> 480 - 630 crafting dailies in each category (still more data is needed for this to be representative).

I have seen that data and it is great.
However I tend to disagree that doing writs on X number of crafters, which all have a different chance at getting a writ (through motif knowledge etc.), is a more representative sample size. In fact, what you are doing is mixing statistics for crafters at different master writ chances, that´s like rolling a D4, D6, D10, and D20 dice a hundred times and try to figure out the chance of rolling a 1, and say this is valid for any individual dice.
As long as we have no equation for how motifs, skill, traits, etc. are weighted into the overall writ chance it is much better to do statistics for a single crafter rather than a group of crafters. Or we need to combine the data for crafters with the exact same amount of traits, motifs etc.
That being saif I think we need keep monitoring this. Let´s talk again after another 50/100 days and see what data we have by then.

Synfaer

I agree with @geophonic_ESO
There is a significant difference in master writ drop rate for crafters depending on what motifs, traits they have learnt and crafting achievements completed.
The sample size of 100 is also way to small. This can be seen in the lower than expected data for tailoring drop rate of the OP and my wood working drop rate.

For my other crafters (most achievments/6-8 traits/2-3 motif sets);

From 600 daily writs
Blacksmith 6% (1105 WV)
Clothing 7% (1292 WV)
Woodworking 9% (908 WV)

This larger data set makes a bit of sense, as my crafters will have more traits learnt on wood working as there are less to learn, so you would expect a slightly better drop rate.

helediron

From about 5500 daily writs amount and voucher totals and their relative percentages:

So, eventually the gear writs drop roughly equally.
Woodworking might be a bit ahead because of quicker trait research.

EDIT: The percentages are relative percentages, not drop rates.

zaria

helediron wrote: »

From about 5500 daily writs amount and voucher totals and their relative percentages:

So, eventually the gear writs drop roughly equally.
Woodworking might be a bit ahead because of quicker trait research.

Woodworking is also nice because of the cheap gold temper

Synfaer

helediron wrote: »

From about 5500 daily writs amount and voucher totals and their relative percentages:

So, eventually the gear writs drop roughly equally.
Woodworking might be a bit ahead because of quicker trait research.

What I really want to know is how you achieved a 30% drop rate for enchanting master writs...
I'm getting nowhere near that with 7 crafters, all with max achievements (except housing recipe achievement).
...or am i reading your data wrong?

Queo

Question:

Are you guys using an add on to track this, or spread sheet you made?

STEVIL

geophonic_ESO wrote: »

InvitationNotFound wrote: »

If I get this right, you did only 100 crafting dailies in each category so far? I'm sorry, but this is far away from being representative.

I've posted my stats a while ago here:
https://forums.elderscrollsonline.com/en/discussion/324880/master-writs-statistics

doing crafting dailies on 8 / 9 chars for 20 days -> 160 - 180 crafting dailies in each category. Still not representative.
I might update this statistic if people are interested even that I've stopped doing crafting dailies lately (lack of time... doing it with 8-9 chars will consume ~45 minutes each day), but the current data should be somewhere around 60 / 70 days -> 480 - 630 crafting dailies in each category (still more data is needed for this to be representative).

I have seen that data and it is great.
However I tend to disagree that doing writs on X number of crafters, which all have a different chance at getting a writ (through motif knowledge etc.), is a more representative sample size. In fact, what you are doing is mixing statistics for crafters at different master writ chances, that´s like rolling a D4, D6, D10, and D20 dice a hundred times and try to figure out the chance of rolling a 1, and say this is valid for any individual dice.
As long as we have no equation for how motifs, skill, traits, etc. are weighted into the overall writ chance it is much better to do statistics for a single crafter rather than a group of crafters. Or we need to combine the data for crafters with the exact same amount of traits, motifs etc.
That being saif I think we need keep monitoring this. Let´s talk again after another 50/100 days and see what data we have by then.

Well, the key is this, if character a has the 3 equips running daily and character b has the 3 equips running daily, add in C and D each with 3... they might indeed have different "levels of success" between each other, but if we are assuming that whatever the % is it will apply on a given character equally to all 3 equips on that character - then the RATIOs between smith/wood/tailor should stay the same and so more is better.

if we do not accept that for any one character the odds of smith/wood/tailor is the same if the character is doing tier-10 on all of them, then your own data has no meaning. maybe your tailor has a lower chance because of your internals, not a flaw in the system.

i have not noticed a difference between smith/wood/cloth on my various crafters anywhere near your numbers show. I wont pretend to have it counted as tightly as you though. So mine is more drawn across a lot more writs over a lot more characters and general perception. if it was half, i know i would have noticed it.

helediron

Queo wrote: »

Question:

Are you guys using an add on to track this, or spread sheet you made?

I am using spreadaheets for statistics with manual entry of numbers. I know there are addons to record these, but manul entry is simple and works even when the game crashes (when addons would lose the data).

helediron

Regarding to collect stats with one or multiple crafters, i played a bit with the numbers i have collected. First we have to take time perspective and how ZOS has changed drop rates. First weeks after Homestead master writs greatly favoured true master crafters. In late February ZOS implemented some kind of caps that reduced drop rates of true master crafters to roughly same levels as good crafters. A good crafter is e.g. 8traits and a dozen epic motifs known.

Everything below is about amount of master gear writs. It's not about amount of vouchers because they are a second RNG on top of first, and act way too chaotically in these sample sizes. It's not about consumable writs since every crafter can get top drop rates easily in alchemy and enchanting. Provisioning is insignificant.

After February i can't find significant difference between 8trait and 9trait crafters. First weeks after Homestead the best crafters got twice better results than good ones - after february the difference is less than RNG fluctuation.

There is a slight difference in amount of master writs dropping between 8trait and 6..7trait crafters. The 6..7trait crafters get a slightly less, but it's still not significant.

Explanation how i try to figure out this "significance":

• If i divide the crafters into two groups, better and worse. E.g. this month the better have 8.8% drop rate of a single gear writ and worse have 6.9% drop rate (relatively 56% vs. 44%). So there is a little difference, showing that one trait more increases drop rate 2%. BUT
• If i check which group wins a day, it's 50%/50%. No difference from knowing one trait more.
• If i take one group and divide results into three day groups like day 1,2,3,1,2,3,... i should get three equal results (33%,33%,33%) in large sample. But here i get 50%,20%,30%, which shows that if a crafter is competing against themselves, i get artificial difference from having equal amount of traits. This is just because of RNG. Subsamples of equal crafters show bigger difference than assumed difference between trait knowledge.

We can safely group together numbers from true master crafters and good crafters. I myself limit this assumption of "good" crafters at least 7 traits learned, have done all reasonably reachable achievements and have at least a dozen epic motifs learned. I don't have any crafters below those criterias running top level writs, so i can't say anything how good/bad they would be.

disintegr8

I agree with OP on clothing voucher drops, my stats - posted in the below thread - show my characters getting a combined total of 38% (smithing) and 50% (woodworking) of the voucher return from pretty close to the same number of master writs.

https://forums.elderscrollsonline.com/en/discussion/344736/drop-rate-of-master-writs#latest

helediron

Synfaer wrote: »

helediron wrote: »

From about 5500 daily writs amount and voucher totals and their relative percentages:

So, eventually the gear writs drop roughly equally.
Woodworking might be a bit ahead because of quicker trait research.

What I really want to know is how you achieved a 30% drop rate for enchanting master writs...
I'm getting nowhere near that with 7 crafters, all with max achievements (except housing recipe achievement).
...or am i reading your data wrong?

Sorry, they are not drop rates, but relative percentages. Of all dropped master writs, 30% were enchanting. If i got 100 master writs, 30 of them were enchanting.

My drop rate in enchanting has been 18% this month. So if i do 100 enchanting regular writs, i get 18 master writs from them.

ninti

First weeks after Homestead master writs greatly favoured true master crafters. In late February ZOS implemented some kind of caps that reduced drop rates of true master crafters to roughly same levels as good crafters.

And how do you know that? I don't recall them saying anything to that effect in any of the patch notes, did I miss something?

have done all reasonably reachable achievements

Achievements have never been said to have any effect on master writ drops, just motifs and traits.

As for the grouping of different crafters together to find percentages for anything, I agree with those saying it is a lost cause. There are too many variables, and the odds are too low, and RNG sucks too much, for us to ever actually know these odds with any degree of certainty.

Hell, look at something a lot easier, Enchantment MW drop rate. Presumably everyone posting in this thread has all master enchanters, and we should all get the same rate, but look at the variance:

geophonic_ESO: 15%
Synfear: 16%
Me: 12%
helediron: 18%
disintegr8: 10.6%

So that is ALL just RNG. I have 8 characters doings writs, and they have all done almost 100 enchanting writs each since Homestead, and the variance can be seen quite clearly. I have one character that has a 19.2% enchanting MW drop rate, and I have another character that has a 4% drop rate. If we can't even show that rate, what hope do we have of figuring out the much more complicated equipment formulas?

Synfaer

Achievements have never been said to have any effect on master writ drops, just motifs and traits.

Achievement do influence master writ drop rates

See below for patch notes and the relevant quote from them;

https://forums.elderscrollsonline.com/en/discussion/318838/pc-mac-patch-notes-v2-7-5#latest

The frequency of Master Writ invites from standard Writ boxes is based on the overall associated tradeskill mastery possessed by the character.

This includes research, Motif knowledge, and achievement completion.

Synfaer

Queo wrote: »

Question:

Are you guys using an add on to track this, or spread sheet you made?

I just use a spreadsheet, doesn't take long to jot down drops/WV and then I update the spreadsheet weekly.

Synfaer

Hell, look at something a lot easier, Enchantment MW drop rate. Presumably everyone posting in this thread has all master enchanters, and we should all get the same rate, but look at the variance:

geophonic_ESO: 15%
Synfear: 16%
Me: 12%
helediron: 18%
disintegr8: 10.6%

@ninti
Check you enchanting achievements, this may be the difference for your low %. Achievements for enchanting contribute a lot to the drop rate (no motifs/research).
There is an achievement tied to crafting prismatic runes which some avoid due to cost of Hakeijo runes.

ninti

Synfaer wrote: »

Achievement do influence master writ drop rates

Well, kind of. Jessica made a more concrete statement about it: https://forums.elderscrollsonline.com/en/discussion/322313/official-discussion-thread-for-homestead-guide-master-writs/p2

Only Purple- and Gold-quality provisioning recipes and completed Motifs increase your chances of getting a Master Writ. Individual Motif Chapters do not increase your chances. Achievements are factored into Master Writ chances by way of learning associated Alchemy reagent traits and Enchantment glyph translations.

Qbiken

geophonic_ESO wrote: »

ninti wrote: »

geophonic_ESO wrote: »

Dear crafters, dear ZOS,
Now, as you can see Tailoring has a much lower chance at receiving a master writ, although it should (with morifs and all traits learned) share the same overall chance of getting one.

Dude, it's just RNG, there is probably no bug. You don't have anywhere enough of a sample size to tell if there is an actual bug yet. No one does really.

The sample size is large enough to see the bias from Tailoring.
Apparently you have no clue about the meaning of RNG and Statistical chances. This is not just getting unlucky at the rolled dice, it is a significant offset.
It is however quite interesting to see that i.e. Synfaer sees this offset with Woodworking.

To be honest, a sample size of 100 isn´t enough to prove anything (admire your dedication with writs though). Same thing with people trying to prove that RNG with maelstrom is skewed (trait destribution is fair, but weapontype that drops is skewed. Post made a few months ago summurizing over 1500+ drops from different players) and say that they did like 200 runs.....You need much, much bigger sample size if you want to prove that RNG is "bugged".

And the % of master writs you got for tailoring is very similiar to alchemy and provisioning so wouldn´t say that this is bugged.

helediron

ninti wrote: »

First weeks after Homestead master writs greatly favoured true master crafters. In late February ZOS implemented some kind of caps that reduced drop rates of true master crafters to roughly same levels as good crafters.

And how do you know that? I don't recall them saying anything to that effect in any of the patch notes, did I miss something?

have done all reasonably reachable achievements

Achievements have never been said to have any effect on master writ drops, just motifs and traits.

As for the grouping of different crafters together to find percentages for anything, I agree with those saying it is a lost cause. There are too many variables, and the odds are too low, and RNG sucks too much, for us to ever actually know these odds with any degree of certainty.

Hell, look at something a lot easier, Enchantment MW drop rate. Presumably everyone posting in this thread has all master enchanters, and we should all get the same rate, but look at the variance:

geophonic_ESO: 15%
Synfear: 16%
Me: 12%
helediron: 18%
disintegr8: 10.6%

So that is ALL just RNG. I have 8 characters doings writs, and they have all done almost 100 enchanting writs each since Homestead, and the variance can be seen quite clearly. I have one character that has a 19.2% enchanting MW drop rate, and I have another character that has a 4% drop rate. If we can't even show that rate, what hope do we have of figuring out the much more complicated equipment formulas?

My enchanting drop rates per each crafter varied between 6% and 31% this month. Every crafter has done every pre-homestead achievement in enchanting. They all should therefore have equal (and highest possible) chances, assuming ZOS kept formulas unchanged. Nevertheless drop rates keep bunnyhopping, sigh...

Synfaer

ninti wrote: »

Synfaer wrote: »

Achievement do influence master writ drop rates

Well, kind of. Jessica made a more concrete statement about it: https://forums.elderscrollsonline.com/en/discussion/322313/official-discussion-thread-for-homestead-guide-master-writs/p2

Only Purple- and Gold-quality provisioning recipes and completed Motifs increase your chances of getting a Master Writ. Individual Motif Chapters do not increase your chances. Achievements are factored into Master Writ chances by way of learning associated Alchemy reagent traits and Enchantment glyph translations.

Not exactly a comprehensive statement.
You could be right but exclusion of information does not prove anything. However, if she stated that achievements do not count for clothing, blacksmith and woodworking master writ drop chance....

geophonic_ESO

Qbiken wrote: »

geophonic_ESO wrote: »

ninti wrote: »

geophonic_ESO wrote: »

Dear crafters, dear ZOS,
Now, as you can see Tailoring has a much lower chance at receiving a master writ, although it should (with morifs and all traits learned) share the same overall chance of getting one.

Dude, it's just RNG, there is probably no bug. You don't have anywhere enough of a sample size to tell if there is an actual bug yet. No one does really.

The sample size is large enough to see the bias from Tailoring.
Apparently you have no clue about the meaning of RNG and Statistical chances. This is not just getting unlucky at the rolled dice, it is a significant offset.
It is however quite interesting to see that i.e. Synfaer sees this offset with Woodworking.

To be honest, a sample size of 100 isn´t enough to prove anything (admire your dedication with writs though). Same thing with people trying to prove that RNG with maelstrom is skewed (trait destribution is fair, but weapontype that drops is skewed. Post made a few months ago summurizing over 1500+ drops from different players) and say that they did like 200 runs.....You need much, much bigger sample size if you want to prove that RNG is "bugged".

And the % of master writs you got for tailoring is very similiar to alchemy and provisioning so wouldn´t say that this is bugged.

correct, I agree to this now too. 100 is a good start but far from being representative.
I was just a little angry because everytime we try to discuss these things someone jumps in and explains everything with just RNG.
That´s why I said we need to pool the data for crafters which supposedly have the same (or at least a very very similar) drop chance, and then see if we can look through the variance of randomly generated data sets by increasing the sample size.
However I often see all these comparisons of so many different crafters at so many different "levels" (i.e. chances to drop a writ), which have sample sizes maybe large enough for being close to representative, but are mixed data sets (as if you just drop your box of dices on the floor containing every possible x-sided dice available).