Item Search
     
BG-Wiki Search
+ Reply to Thread
Page 2 of 2 FirstFirst 1 2
Results 21 to 34 of 34

Thread: CraftyMath     submit to reddit submit to twitter

  1. #21
    Ridill
    Join Date
    Apr 2011
    Posts
    23,736
    BG Level
    10
    FFXI Server
    Bahamut

    Quote Originally Posted by Byrthnoth View Post
    I want to know at what point I can stop the T1 Coconut Rusk test. The competition is Kitron Macarons, which were recently revealed to "increase the rate of synthesis success by 7%."
    Never!!! Actually if it helps I have a small data set on Kitrons as well for the using the same T1 low level synth again side by side exact same conditions minus food. 1468 synths per character.

    Kitrons: NQ - 1354 HQ - 98 breaks - 16

    None: NQ - 1287 HQ -102 breaks - 81


    So yeah definitely pretty small gains in the HQ department

  2. #22
    BG Content
    Join Date
    Jul 2007
    Posts
    21,103
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    More or less doubled my sample size with the same conditions:

    Goldsmithing skill: 106-110 goldsmithing skill with stall (5), shades (1), and, for new data, apron (1) and torque (2) for 112-119 skill total
    Additional gear: Craftmaster's Ring and Artificer's Ring
    Recipe: Avarite + Wind Crystal (Level 81, so Tier 2)
    Area: Bastok Markets, no Signet, Sigil, Sanction, or Ionis

    Rusk:
    NQ 4491
    HQ1 1163
    HQ2 299
    HQ3 84


    No Rusk:
    NQ 4434
    HQ1 1150
    HQ2 266
    HQ3 101


    The BGwiki description of how crafting works implies there are 3 independent probabilities:
    1) Break or Success
    I could potentially reanalyze my data for break rates if people particularly care, but at the moment I am ignoring the first comparison because that's not supposed to be related to what Rusks do.

    2) NQ or HQ (varies based on HQ Tier, would be 1/4 for this synth)
    Rusk:
    25.61% (1546/6037)

    No Rusk:
    25.49% (1517/5951)

    Chi squared p=0.883

    The raw HQ rate is not significantly different between these samples.


    3) HQ1 (12/16), HQ2 (3/16), or HQ3 (1/16)
    Chi squared p=0.193

    The HQ distributions are not significantly different between Rusk and No Rusk either. If I include NQs in the stats and assume steps 2 and 3 are not independent, the difference in the distributions becomes (p=0.346).

  3. #23
    BG Content
    Join Date
    Jul 2007
    Posts
    21,103
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    More or less doubled my sample size with the same conditions:

    Goldsmithing skill: 106-110 goldsmithing skill with stall (5), shades (1), and, for new data, apron (1) and torque (2) for 112-119 skill total
    Additional gear: Craftmaster's Ring and Artificer's Ring
    Recipe: Avarite + Wind Crystal (Level 81, so Tier 2)
    Area: Bastok Markets, no Signet, Sigil, Sanction, or Ionis

    Rusk:
    NQ 5602
    HQ1 1480
    HQ2 372
    HQ3 100


    No Rusk:
    NQ 6066
    HQ1 1577
    HQ2 367
    HQ3 134


    The BGwiki description of how crafting works implies there are 3 independent probabilities:
    1) Break or Success
    I could potentially reanalyze some of my data for break rates if people particularly care, but at the moment I am ignoring the first comparison because that's not supposed to be related to what Rusks do.

    2) NQ or HQ (varies based on HQ Tier, would be 1/4 for this synth)
    Rusk:
    25.84% (1952/7554)

    No Rusk:
    25.52% (2078/8144)

    Chi squared p=0.641

    The raw HQ rate is not significantly different between these samples.


    3) HQ1 (12/16), HQ2 (3/16), or HQ3 (1/16)
    Chi squared p=0.128

    The HQ distributions are not significantly different between Rusk and No Rusk either. If I include NQs in the stats and assume steps 2 and 3 are not independent, the difference in the distributions becomes (p=0.227).





    Profit) This is unimportant on a theoretical level, but is a way to track the tangible outcome for myself. When made in bulk, NQ and HQ1 are valued at 2.4 and 3.8k respectively for all gemstones. HQ2s in bulk are only more valuable than their NPC value (3k) for Titanites, where Heliodors are worth about 40k each due to alchemists making Mallison Medallions for skillups. HQ3s vary between the different gemstones. Any avatear is worth at least 50k to me because some day SE could release neo-Brisingmen or Avatear earrings and I'll be damned if I'm not going to be ready. With a 10% HQ rate on Avarite rings (200k worth of Rhodium + Tear), Shivatears are worth 900k and Ifritear/Ramutears are worth 500k. I produce the other tears faster than I can sell HQ rings, so they are worth my floor price (50k). This really gives us four profiles:
    Shivite (2.4/3.8/3/900)
    Ramuite/Ifritite (2.4/3.8/3/500)
    Titanite (2.4/3.8/40/50)
    Garudite/Carbite/Fenrite/Leviatite (2.4/3.8/3/50)

    Just multiplying the probabilities that from my data by the values, I end up with the No Rusk distribution being more valuable overall. Its higher HQ3 rate makes it the unopposed champions of Shivites (+2,881gil/synth) and Ifritite/Ramuites (+1,594gil/synth), while the difference in the distribution is nearly irrelevant for Titanites (-7gil/synth) and Fenrite/Leviatite/Carbite/Garudites (+147gil/synth). At the same time, 7600 synths at 25 seconds per synth means I used ~100 Rusks (300k). I could have been using 100 Kitron Macarons to reduce my break rate by 4%, each of which would have paid for itself slightly (low value avarites) or substantially (high value avarites).

    So in terms of actual profit, the distributions encourage me to use Macarons over Rusks, but I'm going to try and hold out until I hit 10k in each sample (just because it's a nice, round number).

  4. #24
    BG Content
    Join Date
    Jul 2007
    Posts
    21,103
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    I hit 10k samples per condition and still have found no significant effect on HQ rate/tier/etc. This does not mean a difference does not exist, just that I cannot detect it in 20,000 synths. I will not be continuing this experiment, because it's annoying to track and my sample size is huge already.

    The effect of Kitron Macarons on break rate, however, is pronounced. If Macarons works as described in FFXI community lore then Kitrons may be overkill for a single craft synth and Coffee or Cherry+Ring (both of which are cheaper from moogles) should be enough to cap you.


    Personally, I'll be using Kitron Macarons for pretty much everything from here on out.


    Spoiler: show
    Goldsmithing skill: 106-110 goldsmithing skill with stall (5), shades (1), and, for new data, apron (1) and torque (2) for 112-119 skill total
    Additional gear: Craftmaster's Ring and Artificer's Ring
    Recipe: Avarite + Wind Crystal (Level 81, so Tier 2)
    Area: Bastok Markets, no Signet, Sigil, Sanction, or Ionis

    Rusk:
    NQ 7386
    HQ1 1983
    HQ2 488
    HQ3 143


    No Rusk:
    NQ 7466
    HQ1 1920
    HQ2 447
    HQ3 167


    The BGwiki description of how crafting works implies there are 3 independent probabilities:
    1) Break or Success
    I could potentially reanalyze some of my data for break rates if people particularly care, but at the moment I am ignoring the first comparison because that's not supposed to be related to what Rusks do.

    2) NQ or HQ (varies based on HQ Tier, would be 1/4 for this synth)
    Rusk:
    26.14% (2614/10000)

    No Rusk:
    25.34% (2534/10000)

    Chi squared p=0.196

    The raw HQ rate is not significantly different between these samples.


    3) HQ1 (12/16), HQ2 (3/16), or HQ3 (1/16)
    Chi squared p=0.180

    The HQ distributions are not significantly different between Rusk and No Rusk either. If I include NQs in the stats and assume steps 2 and 3 are not independent, the difference in the distributions becomes (p=0.164).

  5. #25
    E. Body
    Join Date
    Dec 2008
    Posts
    2,408
    BG Level
    7
    FFXI Server
    Quetzalcoatl

    It's possible that HQ rates don't stack the way we think, much akin to trying to wear two pieces of move speed gear.

    I think you probably should've done the test without the HQ ring to remove a potential variable.

  6. #26
    Old Odin
    Join Date
    Oct 2006
    Posts
    6,198
    BG Level
    8
    FFXI Server
    Asura

    CraftyMath

    Quote Originally Posted by urat View Post
    It's possible that HQ rates don't stack the way we think, much akin to trying to wear two pieces of move speed gear.

    I think you probably should've done the test without the HQ ring to remove a potential variable.
    there are samples with no ring and just macarons which state the same thing: the food provides no increased HQ rate

    Now the question is if it provides a boost to desynthing? maybe not HQ ratebut success rate? since desynth have a different success rate calculation?


    Gesendet von iPhone mit Tapatalk

  7. #27
    BG Content
    Join Date
    Jul 2007
    Posts
    21,103
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    I suspect that Rusks give a decreased material loss rate, and the large disparity between Macarons (increased skillup rate, decreased break rate) and Rusks (decreased material loss rate) is just SE's twisted idea of balance.

  8. #28
    Old Odin
    Join Date
    Oct 2006
    Posts
    6,198
    BG Level
    8
    FFXI Server
    Asura

    Quote Originally Posted by Byrthnoth View Post
    I suspect that Rusks give a decreased material loss rate, and the large disparity between Macarons (increased skillup rate, decreased break rate) and Rusks (decreased material loss rate) is just SE's twisted idea of balance.
    well decressed materials lost makesit good for desynths, since the success rate there is shit compared to normalcrafting


    Gesendet von iPhone mit Tapatalk

  9. #29
    A. Body
    Join Date
    Jun 2007
    Posts
    4,445
    BG Level
    7
    FFXIV Character
    Sir Taint
    FFXIV Server
    Gilgamesh
    FFXI Server
    Cerberus

    Necro bump. I didn't want to start a new thread.

    Based on a the few post on the last page, the HQ rings do in fact add about 1% HQ rate? Any test with both HQ+1% rings?

  10. #30
    Ridill
    Join Date
    Apr 2011
    Posts
    23,736
    BG Level
    10
    FFXI Server
    Bahamut

    Not sure if that was really determined. Looks like it did something but not sure an exact value got nailed down and certainly looked lower than 1% additive. Also to note that I didn't really notice before is it looked like rusk helped without hq rings but didn't with them. So like the success rate the hq rates might have caps based on tiers and you might be capping with a single hq ring

  11. #31
    Old Odin
    Join Date
    Oct 2006
    Posts
    6,198
    BG Level
    8
    FFXI Server
    Asura

    Quote Originally Posted by dasva View Post
    Not sure if that was really determined. Looks like it did something but not sure an exact value got nailed down and certainly looked lower than 1% additive. Also to note that I didn't really notice before is it looked like rusk helped without hq rings but didn't with them. So like the success rate the hq rates might have caps based on tiers and you might be capping with a single hq ring
    didnt byrth somewhere calculate that citron macarons while not providing more HQ rate, in the end still produce more HQ stuff because of the lowered fail synths compared to rusks??? (when Crafting above your level)

  12. #32
    Ridill
    Join Date
    Apr 2011
    Posts
    23,736
    BG Level
    10
    FFXI Server
    Bahamut

    Quote Originally Posted by Damane View Post
    didnt byrth somewhere calculate that citron macarons while not providing more HQ rate, in the end still produce more HQ stuff because of the lowered fail synths compared to rusks??? (when Crafting above your level)
    Well it was more like putting numbers to an idea of how it could work based on some assumptions.

    The assumptions though that aren't proven/have some contradictory info. Like it assumes rusks have zero effect on hq rate which his test showed but mine showed one though we had different test parameters that could of caused it I'm banking it on the hq ring capping you because that is such an SE thing to do so it might just be a matter of swapping rings and such though that said rings wont cap you success rate so should still be in favor of kitron. It also assumes kitrons will increase nqs and hqs equally while the smallish (1468 synths x2) test at the top of this page suggest otherwise.

  13. #33
    Ridill
    Join Date
    Apr 2011
    Posts
    23,736
    BG Level
    10
    FFXI Server
    Bahamut

    So anyone working on new rates? I've started working on T1 again not done but so far seems promising

  14. #34
    Melee Summoner
    Join Date
    Feb 2014
    Posts
    45
    BG Level
    1
    FFXI Server
    Asura

    I actually just came to start a new thread since all this old data is now useless.