Item Search
     
BG-Wiki Search
+ Reply to Thread
Page 1 of 2 1 2 LastLast
Results 1 to 20 of 35
  1. #1
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Dancer Abilities and Step Accuracy Testing

    Step Accuracy is based on melee Hit rate (shown below) and caps at 95%. The theory behind these tests is that I can tell how much "Step Accuracy" I have using the difference between my melee hit rate and my Step hit rate. Steps cannot be resisted, so I assume a binomial distribution works. This is a less ideal way to do it than checking monster defense and calculating my exact theoretical hit rate, but I was under quite a bit of pressure (from ze girlfreend) to go to bed when I was setting these tests up, so ya get what'cha get!

    Format is going to be:
    (Steps_Landed/Steps_Attempted) Step_Hit_Rate% +/- 95%_Confidence -- Normal_Hit_rate 95%_confidence :: Type of Step

    Case 1: No Step Acc gear
    (904/1916) 47.2% +/- 2.2% -- 41.4% +/- 0.9% :: Box Step
    (2247/3588 ) 62.6% +/- 1.6% -- 58.6% +/- 0.9% :: Box Step

    Looks like there's an innate +8-10 Step Acc on Box Step.

    Case 2: Choreia Earring
    (547/972) 56.3% +/-3.1% -- 48.5% +/- 1.1% :: Feather Step

    Hard to draw too much from this both due to the low sample size and step change, but I'd guess Choreia Earring is between 2 and 6 Step Accuracy.

    Case 3: AF1 Hands, AF2 feet, Choreia Earring
    (1274/1863) 68.4% +/- 2.1% -- 49.5% +/- 1.0% :: Box Step
    (1040/1779) 58.5% +/- 2.3% -- 41.8% +/- 1.4% :: Box Step

    With the maximum amount of Step Accuracy available to me, it looks like there's apprixmately a 15-20% Hit rate bonus (30-40 Step Accuracy all told). I'd bet it breaks up:
    10 Step Acc innate
    5 Step Acc from Choreia Earring
    10 Step Acc from Dancer's Bangles (AF1 hands)
    10 Step Acc from Etoile Toe Shoes (AF2 feet)

    Looks like I need to do a control test with Feather Step to confirm it gets the same hit rate bonus as box step, increase sample size, do testing at a higher starting hit rate to increase my statistical power, etc. but I've put the tentative conclusions up there.

  2. #2
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Feather Step Testing

    My trusty mule and I have been swinging at the same Hpmede for the last five hours, with 5/5 Feather Step up the whole time. The evasion of this Hpemde matches the predicted evasion of the highest level Hpmede in this thread, which means it probably has 72 AGI.

    Without AF3+2 feet:
    80 DEX (dDEX=8, 10% predicted crit rate): 2076/14371 = 14.09 +/- 0.56%

    With AF3+2 feet applying the buffs:
    86 DEX (dDEX=14, 11% predicted crit rate): 3077/17375 = 17.7% +/- 0.56% <-- wearing AF3+2 feet
    90 DEX (dDEX=18, 11% predicted crit rate): 1839/10104 = 18.2% +/- 0.75% <-- THF mule with clubs
    91 DEX (dDEX=19, 11% predicted crit rate): 1346/7519 = 17.9% +/- 0.86% <-- wearing AF3+2 feet

    Because I've spanned almost the whole dDEX plateau there, it seems unlikely to me that Feather Step is doing anything related to dDEX. Because Crit rate is already a percentage, I'm skeptical of suggestions that Feather Step might be multiplicative instead of additive, but I guess it could happen.

  3. #3
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    307 Evasion Hpmede
    Quickstep level 1 -> 309 Acc for "Low Evasion", 308 is neutral
    Quickstep level 2 -> 305 Acc for "Low Evasion", 304 is neutral
    Quickstep level 3 -> 301 Acc for "Low Evasion", 300 is neutral
    Quickstep level 4 -> 297 Acc for "Low Evasion", 296 is neutral
    Quickstep level 4 -> 293 Acc for "Low Evasion", 292 is neutral

    So the 8->4->4->4->4 pattern is correct

    Also:
    48:47 - Lethargic Daze 1 (1:00)
    49:03 - Presto Lethargic Daze 3 (1:30)
    49:20 - Lethargic Daze 4 (2:00)
    49:37 - Lethargic Daze 5 (2:30)
    50:38 - Presto Lethargic Daze 5 (3:00)
    51:45 - Lethargic Daze wears off (3 minutes from first use)

    Despite giving extra potency to the step, Presto grants no duration bonus.

  4. #4
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    http://www.bluegartr.com/threads/120188-Dancer-Abilities-and-Step-Accuracy-Testing

    Doh, Crit damage boost is the same. I was wearing Charis Feather earlier. Anyway, AF3+2 head appears to give a +20% damage boost to hits affected by Climactic Flourish.
    Twashtar (D48+13), Charis Feather and Trait (Crit damage+7.8~8%) = 259 damage max

    It is either +22% crit damage or +20% independent mod. My money is on independent mod.

    Also, Yunalesca on FFXIAH reported:
    Quote Originally Posted by Asura.Yunalaysca
    was just testing it, the base animation is different, but interesting thing is that it will have a Dinging sound and your hands will have an enlight affect animation. this affect is only evident before my rudra's dmg shows the appropriate dmg for criting. so you can use striking and if you dont hear the dinging sound use evisceration over rudra because your ws will not crit. it's weird that this ja has this change because the ja it'self does not posses the potential to crit but the change in animation is definately tied to your next hit being a critical hit. im gonna continue testing but im 99% sure that is the case, which is pretty huge if you ask me, no more guess work with rudra dmg.
    ^ Actually, doesn't seem to be the case for me. Perhaps they just changed the animation.

  5. #5
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Windows Update ended my test last night prematurely, but I still got over 100 steps off...

    16 Acc in gear (AF3+2 Head/Hands), 35 Acc from trait, 91 Acc from DEX (122 DEX w/ 2H weapon), 128 Acc from skill (/WHM) = 270 Acc
    vs.
    the highest level Hpmede, 307 Eva

    56% hit rate predicted, 64% normal hitrate observed in 606 swings (hurrr? wait what? That's +/- 4%, and it definitely checks high evasion)

    Anyway, 123/125 Presto'd Steps landed. Even assuming my hit rate is secretly 70% somehow and factoring in the innate +10 Step Acc, that still means that Presto is at least +40 Acc, and realistically it probably means it's at least +70.



    Next edition:
    2537/4897 (51.81% Hit rate) with 122/148 Presto'd Steps successful (82.43% +/- 6.1%)

    Considering the innate +10 Step Acc on steps, I'd say Presto is about 50 Step Acc.

  6. #6
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Halp prz~!

    Guys, I'm DNC/WHM using club against the highest level Hpmede (307 Eva), which I have verified since.

    Set
    Kitty rod/Lady bell/_/AF3
    AF3+2/Love/Suppa/Brutal
    AF3+2/AF3+2/Hoard/Thundersoul
    Aisance +1/Acerbic/Tumbler/AF3+2
    =122 DEX ( 77 -11, +56)
    132 Skill + 61 from DEX + 35 from Acc bonus III + 16 from gear (AF3+2 head/hands) = 244 Acc

    vs. 307 Evasion enemy should = 43% hit rate.
    Instead, I parsed: 2537/4897 = 51.81% (+/-1.4%) HR, so approximately hit rate +10% or Acc+20.

    Where is my math going wrong?

    My mule wore:

    Sjugy THF90/SCH1
    Lady Bell/Lady Bell
    Turban/Love/Suppa/Brutal
    DH+1/Homam/Hoard/Snow
    Atheling/Anguinus/Aurore/Ballerines
    = 86DEX (77 +12 - 4)
    200 + 43*.9 from skill + 43 from DEX + 28 from gear = 310 Acc

    Predicted hit rate of 76% against a 307 evasion enemy.
    He parsed: 3552/4797 = 74.05% (+/- 1.2%) HR

    Same enemy, same timeframe. How did I math wrong?

    Edit: Oh man, I might know the answer but it's going to be hilarious if it's true...

  7. #7
    First invited, last in the zone.
    Join Date
    Sep 2008
    Posts
    1,449
    BG Level
    6
    FFXI Server
    Lakshmi

    I wonder if they raised the evasion floor, or it's in some way related to mob level. Take off all your acc gear and see if your hit rate drops below 50%?

    No clue otherwise, as IIRC you already tested DNC doesn't get ACC bonus IV.

  8. #8
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Hmm... technically both predicted values are outside the 95% confidence interval.

    If I combine the earlier broken parse with my mule's current data (same gear) it becomes 4883/6572, 74.3% +/- 1.1%. I don't know though, I'm willing to accept that might be 76%. I'd really really rather this be a floor issue than a re-work of the hit rate system.

    My initial guess was that SE flubbed Presto implementation and it's +50~60 Accuracy instead of +50~60 Step Accuracy, but I'm not sure it could be responsible for this even if it were true. Yeah, it couldn't. It would have only gotten me up to 46% average even if it was +80 Accuracy. I only missed 26 steps (13 minutes of presto active) out of a 148 minute parse.

  9. #9
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    I suppose this deserves a post. I believe there is a 50% hit rate floor now, and will test it overnight. Same gear set as before, 10000+ hits has me at 50.30% hit rate. I will test with even less Acc gear overnight.

    74.77% +/- 0.85% hit rate for my mule with ~10000 swings as well. So probably 75%. Same exact gear as before, so he's actually 1% low.

    > I don't know, but it seems the relationship between hit rate and accuracy may have changed in a more complicated way than adding a new floor.

  10. #10
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    FFFFFFFFFFFFFFFFFFFFFFF I figured it out. Closed Position 5/5. Chriiiist.

    Nothing was changed.

    I knew it would end up being something stupid like that.

  11. #11
    First invited, last in the zone.
    Join Date
    Sep 2008
    Posts
    1,449
    BG Level
    6
    FFXI Server
    Lakshmi

    Quote Originally Posted by Byrthnoth View Post
    FFFFFFFFFFFFFFFFFFFFFFF I figured it out. Closed Position 5/5. Chriiiist.

    Nothing was changed.

    I knew it would end up being something stupid like that.
    *slow clap* XD


    Edit: And yeah, with my ~100 hit sample I can already totally bust a 50% hit floor cap (21%+/- 7%)

    still 20% almost certainly.

  12. #12
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    http://www.bluegartr.com/threads/120188-Dancer-Abilities-and-Step-Accuracy-Testing

    Thanks guys, sorry I herp'd and then derp'd.

    5/5 Feather Step - I did two parses with 5/5 Feather Step, AF3+2 shoes, and two characters beating the Hpmede (72 AGI). I was setting out to verify a dDEX plateau to make sure Feather Step wasn't interacting with dDEX somehow, and I'm convinced it wasn't:
    <table border="0" cellpadding="0" cellspacing="0" width="448"><colgroup><col style="width:48pt" span="7" width="64"> </colgroup><tbody><tr style="height:15.0pt" height="20"> <td style="height:15.0pt;width:48pt" height="20" width="64">dDEX</td> <td style="width:48pt" width="64">Hits</td> <td style="width:48pt" width="64">Crits</td> <td style="width:48pt" width="64">Rate</td> <td style="width:48pt" width="64">Confidence</td> <td style="width:48pt" width="64">Predicted</td> <td style="width:48pt" width="64">Difference</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">14</td> <td align="right">17375</td> <td align="right">3077</td> <td class="xl65" align="right">17.7%</td> <td class="xl65" align="right">0.6%</td> <td class="xl66" align="right">11%</td> <td class="xl67" align="right">6.7%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">18</td> <td align="right">6151</td> <td align="right">1141</td> <td class="xl65" align="right">18.5%</td> <td class="xl65" align="right">1.0%</td> <td class="xl66" align="right">11%</td> <td class="xl67" align="right">7.5%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">19</td> <td align="right">7519</td> <td align="right">1346</td> <td class="xl65" align="right">17.9%</td> <td class="xl65" align="right">0.9%</td> <td class="xl66" align="right">11%</td> <td class="xl67" align="right">6.9%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">18</td> <td align="right">3953</td> <td align="right">698</td> <td class="xl65" align="right">17.7%</td> <td class="xl65" align="right">1.2%</td> <td class="xl66" align="right">11%</td> <td class="xl67" align="right">6.7%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" height="20">Combined:</td> <td align="right">34998</td> <td align="right">6262</td> <td class="xl65" align="right">17.9%</td> <td class="xl65" align="right">0.4%</td> <td class="xl66" align="right">11%</td> <td class="xl67" align="right">6.9%</td> </tr> </tbody></table>With that in mind, I figured it was safe to combine them. I feel fairly confident that Feather Step from that dDEX plateau gives approximately +7% Crit rate with AF3+2 shoes.

    2/5 Feather Step: For obvious reasons, this is a harder test to do. I wrote a MATLAB script that parses the chatlog and finds Step uses, Dazes wearing off, and Presto wearing off. It also assigns every chat log on the basis of whether or not it's a hit, miss, or crit and which character it's from. If you're interested in a copy, PM me. Anyway, this let me look at crit rates only between steps that landed and when Dazes wear off (and look at the Presto Acc boost, which we will get into later).
    I meritted my crit rate to level 5 before starting these tests, but my mule is still level 4. Anyway:
    <table style="width: 537px; height: 283px;" border="0" cellpadding="0" cellspacing="0"><colgroup><col style="width:48pt" span="7" width="64"> </colgroup><tbody><tr style="height:15.0pt" height="20"> <td style="height:15.0pt;width:48pt" height="20" width="64">dDEX</td> <td style="width:48pt" width="64">Hits</td> <td style="width:48pt" width="64">Crits</td> <td style="width:48pt" width="64">Crit rate</td> <td style="width:48pt" width="64">Confidence</td> <td style="width:48pt" width="64">Predicted</td> <td style="width:48pt" width="64">Difference</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">28</td> <td align="right">11610</td> <td align="right">2015</td> <td class="xl65" align="right">17.3%</td> <td class="xl65" align="right">0.7%</td> <td class="xl66" align="right">13%</td> <td class="xl67" align="right">4.3%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">13</td> <td align="right">12805</td> <td align="right">1827</td> <td class="xl65" align="right">14.3%</td> <td class="xl65" align="right">0.6%</td> <td class="xl66" align="right">11%</td> <td class="xl67" align="right">3.3%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">50</td> <td align="right">4684</td> <td align="right">1337</td> <td class="xl65" align="right">28.5%</td> <td class="xl65" align="right">1.1%</td> <td class="xl66" align="right">25%</td> <td class="xl67" align="right">3.5%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">13</td> <td align="right">6635</td> <td align="right">969</td> <td class="xl65" align="right">14.6%</td> <td class="xl65" align="right">0.8%</td> <td class="xl66" align="right">10%</td> <td class="xl67" align="right">3.6%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">46</td> <td align="right">918</td> <td align="right">213</td> <td class="xl65" align="right">23.2%</td> <td class="xl65" align="right">2.5%</td> <td class="xl66" align="right">21%</td> <td class="xl67" align="right">2.2%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">13</td> <td align="right">1566</td> <td align="right">249</td> <td class="xl65" align="right">15.9%</td> <td class="xl65" align="right">1.7%</td> <td class="xl66" align="right">11%</td> <td class="xl67" align="right">4.9%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">50</td> <td align="right">2011</td> <td align="right">550</td> <td class="xl65" align="right">27.3%</td> <td class="xl65" align="right">1.7%</td> <td class="xl66" align="right">25%</td> <td class="xl67" align="right">2.3%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">13</td> <td align="right">2659</td> <td align="right">406</td> <td class="xl65" align="right">15.3%</td> <td class="xl65" align="right">1.3%</td> <td class="xl66" align="right">11%</td> <td class="xl67" align="right">4.3%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">50</td> <td align="right">366</td> <td align="right">94</td> <td class="xl65" align="right">25.7%</td> <td class="xl65" align="right">4.0%</td> <td class="xl66" align="right">25%</td> <td class="xl67" align="right">0.7%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">13</td> <td align="right">1060</td> <td align="right">189</td> <td class="xl65" align="right">17.8%</td> <td class="xl65" align="right">2.1%</td> <td class="xl66" align="right">11%</td> <td class="xl67" align="right">6.8%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" height="20">Comb 13:</td> <td align="right">24725</td> <td align="right">3640</td> <td class="xl65" align="right">14.7%</td> <td class="xl65" align="right">0.4%</td> <td class="xl66" align="right">11%</td> <td class="xl67" align="right">3.7%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" height="20">Comb 50:</td> <td align="right">7061</td> <td align="right">1981</td> <td class="xl65" align="right">28.1%</td> <td class="xl65" align="right">1.0%</td> <td class="xl66" align="right">25%</td> <td class="xl67" align="right">3.1%</td> </tr> </tbody></table>So, as you can see, there is approximately a +3~4% crit rate.

    Presto Step Accuracy: Testing Presto'd Step Accuracy is particularly annoying. Once every 30-31 seconds is max speed, and once every minute is what I did for the sake of the above 2/5 Feather Step test. In the end, I got approximately +30% hit rate on Steps, which (accounting for the normal +5%) brings us to +50 Step Accuracy from Presto. Just so you can see what I'm working with, here's a table of all my Presto'd Data:
    <table border="0" cellpadding="0" cellspacing="0" width="425"><colgroup><col style="width:48pt" span="3" width="64"> <col style="mso-width-source:userset;mso-width-alt:1499;width:31pt" width="41"> <col style="width:48pt" span="3" width="64"> </colgroup><tbody><tr style="height:15.0pt" height="20"> <td style="height:15.0pt;width:48pt" height="20" width="64">Hits</td> <td style="width:48pt" width="64">Misses</td> <td style="width:48pt" width="64">Hit rate</td> <td style="width:31pt" width="41">Steps</td> <td style="width:48pt" width="64">Step Hits</td> <td style="width:48pt" width="64">Step Rate</td> <td style="width:48pt" width="64">Confidence</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">13703
    </td> <td align="right">8750</td> <td class="xl65" align="right">61.0%</td> <td align="right">617</td> <td align="right">552</td> <td class="xl65" align="right">89.4%</td> <td class="xl65" align="right">2.4%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">6138</td> <td align="right">6027</td> <td class="xl65" align="right">50.5%</td> <td align="right">355</td> <td align="right">284</td> <td class="xl65" align="right">80.0%</td> <td class="xl65" align="right">4.2%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">1222</td> <td align="right">1674</td> <td class="xl65" align="right">42.2%</td> <td align="right">85</td> <td align="right">67</td> <td class="xl65" align="right">78.8%</td> <td class="xl65" align="right">8.7%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">2537</td> <td align="right">2360</td> <td class="xl65" align="right">51.8%</td> <td align="right">148</td> <td align="right">122</td> <td class="xl65" align="right">82.4%</td> <td class="xl65" align="right">6.1%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">388</td> <td align="right">218</td> <td class="xl65" align="right">64.0%</td> <td align="right">125</td> <td align="right">123</td> <td class="xl65" align="right">98.4%</td> <td class="xl65" align="right">2.2%</td> </tr> </tbody></table>
    Choriea Earring Step Accuracy: This was just sort of a throw away test that I did while doing other stuff, Saber Dance testing before bothering to quest daggers. I'm pretty sure it came out to +5 Accuracy though, on top of the native +10 Accuracy all Steps get.
    <table border="0" cellpadding="0" cellspacing="0" width="448"><colgroup><col style="width:48pt" span="7" width="64"> </colgroup><tbody><tr style="height:15.0pt" height="20"> <td style="height:15.0pt;width:48pt" height="20" width="64">Hits</td> <td style="width:48pt" width="64">Misses</td> <td style="width:48pt" width="64">Hit rate</td> <td style="width:48pt" width="64">Attempted</td> <td style="width:48pt" width="64">Landed</td> <td style="width:48pt" width="64">Step Rate</td> <td style="width:48pt" width="64">Diff</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">3113</td> <td align="right">4170</td> <td class="xl65" align="right">42.7%</td> <td align="right">1196</td> <td align="right">598</td> <td class="xl65" align="right">50.0%</td> <td class="xl65" align="right">7.3%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">3774</td> <td align="right">4006</td> <td class="xl65" align="right">48.5%</td> <td align="right">972</td> <td align="right">547</td> <td class="xl65" align="right">56.3%</td> <td class="xl65" align="right">7.8%</td> </tr> </tbody></table>
    Base Step Accuracy Bonus: Now that I've referenced it several times, I guess I should show my data that supports an innate +10 Accuracy on Steps.
    <table border="0" cellpadding="0" cellspacing="0" width="448"><colgroup><col style="width:48pt" span="7" width="64"> </colgroup><tbody><tr style="height:15.0pt" height="20"> <td style="height:15.0pt;width:48pt" height="20" width="64">Hits</td> <td style="width:48pt" width="64">Misses</td> <td style="width:48pt" width="64">Hit rate</td> <td style="width:48pt" width="64">Attempts</td> <td style="width:48pt" width="64">Lands</td> <td style="width:48pt" width="64">Step Rate</td> <td style="width:48pt" width="64">Diff</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">4021</td> <td align="right">2852</td> <td class="xl65" align="right">58.5%</td> <td align="right">2055</td> <td align="right">1290</td> <td class="xl65" align="right">62.8%</td> <td class="xl65" align="right">4.3%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">3034</td> <td align="right">2125</td> <td class="xl65" align="right">58.8%</td> <td align="right">1533</td> <td align="right">957</td> <td class="xl65" align="right">62.4%</td> <td class="xl65" align="right">3.6%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">934</td> <td align="right">659</td> <td class="xl65" align="right">58.6%</td> <td align="right">482</td> <td align="right">307</td> <td class="xl65" align="right">63.7%</td> <td class="xl65" align="right">5.1%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">1136</td> <td align="right">837</td> <td class="xl65" align="right">57.6%</td> <td align="right">603</td> <td align="right">382</td> <td class="xl65" align="right">63.3%</td> <td class="xl65" align="right">5.8%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">4434</td> <td align="right">6273</td> <td class="xl65" align="right">41.4%</td> <td align="right">1916</td> <td align="right">904</td> <td class="xl65" align="right">47.2%</td> <td class="xl65" align="right">5.8%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">2105</td> <td align="right">2911</td> <td class="xl65" align="right">42.0%</td> <td align="right">836</td> <td align="right">383</td> <td class="xl65" align="right">45.8%</td> <td class="xl65" align="right">3.8%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">1590</td> <td align="right">930</td> <td class="xl65" align="right">63.1%</td> <td align="right">622</td> <td align="right">419</td> <td class="xl65" align="right">67.4%</td> <td class="xl65" align="right">4.3%</td> </tr> </tbody></table>So it looks like a 4-5% Hit rate increase on Steps even without Step Accuracy gear. I approximated it at +10 Step Acc because that seems more reasonable than +8. Point is, Steps have a positive Hit rate bonus.

    All gear Step Accuracy Bonus: I decided to put all my Step Acc gear on for some of the last Saber Dance tests. This is what resulted!
    <table border="0" cellpadding="0" cellspacing="0" width="448"><colgroup><col style="width:48pt" span="7" width="64"> </colgroup><tbody><tr style="height:15.0pt" height="20"> <td class="xl65" style="height:15.0pt;width:48pt" height="20" width="64">Hit</td> <td style="width:48pt" width="64">Missed</td> <td style="width:48pt" width="64">Hit rate</td> <td class="xl65" style="width:48pt" width="64">Attempted</td> <td class="xl65" style="width:48pt" width="64">Landed</td> <td style="width:48pt" width="64">Land rate</td> <td style="width:48pt" width="64">Confidence</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">3882</td> <td align="right">5406</td> <td class="xl65" align="right">41.8%</td> <td align="right">1779</td> <td align="right">1040</td> <td class="xl65" align="right">58.5%</td> <td class="xl65" align="right">2.3%</td> </tr> <tr style="height:15.0pt" height="20"> <td style="height:15.0pt" align="right" height="20">4863</td> <td align="right">4960</td> <td class="xl65" align="right">49.5%</td> <td align="right">1863</td> <td align="right">1274</td> <td class="xl65" align="right">68.4%</td> <td class="xl65" align="right">2.1%</td> </tr> </tbody></table>So it looks like approximately Hit rate +17~20% (+35-40 Acc). Since we already guestimated that Choreia was +5 Acc and the base rate increase was +10, we're left accounting for about 20 Accuracy. AF2 feet and AF1 hands are probably +10 Accuracy each. Obviously I should test them separately, but for an approximation I'm sure I could do worse.


    If you were going to take something away from my data so far, I would say:
    1) Bewildered Daze's first two levels are much more potent than its later levels. At the moment my theory with AF3+2 feet on is: (2%,2%,1%,1%,1%) up from (1%,1%,.5%,.5%,.5%)
    2) Step Accuracy gear is silly. As a Dancer, you never want to be in a situation where you don't have nearly-capped melee accuracy, and Steps already get a modest accuracy bonus. Carrying around extra gear just so you can land steps with Flash on is a little excessive. It's even dumber that it's a merit category, and dumber still that some people merit it. Ugh!
    3) (much the same as 2) Presto, innate bonus, and gear add up to about 85 Step Acc. If you're including some Acc gear you don't normally wear and have over capped hit rate normally, you will really have unfloored Step Accuracy with a blinding potion on. That's how much damned Step Acc we get!

    PS. I found tables. Life will never be the same.

  13. #13
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Desperate Flourish gives -10 Evasion, in additional to weighing the target down.

    307 Evasion target, put Desperate Flourish on.
    267 Acc = /checks neutral evasion
    266 Acc = /checks high evasion

  14. #14
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Thanks! I may be done testing for a little while until I finish leveling my NPC and either Ukon or Twashtar to 85 (still being lazy about killing the VNMs, and the 'dar/ASE patch and items stacking isn't really making me less lazy).

    Next test I do would be:
    dDEX=50, Dancer's Bangles and a target of 75% melee accuracy with Feather Step level 5.

    One interesting option for how Feather Step might work (given the weirdness in my data) is that "Lowers critical hit evasion" could actually give a second chance for crit to proc if the first chance fails. It would explain the seeming inverse relationship between base crit rate and seeming Feather Step potency. Testing at dDEX=50 with Step level 5 would let me see a larger version of this effect, if it exists, and keeping level 5 lets me spam steps and determine Dancer's Bangles Step Acc. Desperate Flourish's hit rate penalty would also be a nice thing to test, but I'd have to do all my analysis in MATLAB again instead of just reading KParser if I tried to do that test at the same time as these. I can stack it with the level 1 test without issue though, so I'll probably do that.

    Nevermind, I thought of a better way to test this. Nightfyre shows that the crit rate cap is now 100%. I'll get 94% crit rate from Atma (Dark Depths, Gnarled Horn, Razed Ruins) and stack Feather Step level 5 on some mobs. If I see a non-crit then I'll know that it's probably a separate chance to proc.

    Additional thought:
    This could also be used to approximate the crit hit bonus on WSs, eh?

  15. #15
    Groinlonger
    Join Date
    Oct 2006
    Posts
    2,964
    BG Level
    7
    FFXI Server
    Fenrir

    Much easier way to build an exclusion zone with certainty. For example, if you see non crits with a base crit rate of 95%, you know it's less than 5%.

  16. #16
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Put on GH, SQ, and Razed Ruins (85% crit rate). Sekka->WS a low-HP monster without any crit rate gear on. If you can't get a non-crit this way, the crit rate boost is close to 15% or higher than it. If you non-crit, put on AF3+2 head and a Crit+1 Wivre Gorget and repeat the test to see if the bonus is 10% or higher.

    Edit: yeah, what Mojo said. It wouldn't be fast going, but it could be easier than building TP on L0 monsters and WSing over and over with a fairly low power test (due to ~35% crit rate).

  17. #17
    Radsourceful

    Join Date
    Jul 2007
    Posts
    1,964
    BG Level
    6
    FFXI Server
    Bismarck

    Quote Originally Posted by Byrthnoth View Post
    Nevermind, I thought of a better way to test this. Nightfyre shows that the crit rate cap is now 100%. I'll get 94% crit rate from Atma (Dark Depths, Gnarled Horn, Razed Ruins) and stack Feather Step level 5 on some mobs. If I see a non-crit then I'll know that it's probably a separate chance to proc.
    Remember 5th crit merit, if applicable.

    Edit: I see you listed 85% in another post, just a reminder

  18. #18
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Yeah, I goofed there. So used to 24%, but I like 25% even more! Either way, I will either have about a 4% non-crit rate or a 100% crit rate. I assume I can figure this out in 3-4 Turtles or Glade Wivres.

  19. #19
    xXNyteFyreXx420Sharingan
    Join Date
    May 2009
    Posts
    3,710
    BG Level
    7
    FFXI Server
    Fenrir

    Quote Originally Posted by Byrthnoth View Post
    Put on GH, SQ, and Razed Ruins (85% crit rate). Sekka->WS a low-HP monster without any crit rate gear on. If you can't get a non-crit this way, the crit rate boost is close to 15% or higher than it. If you non-crit, put on AF3+2 head and a Crit+1 Wivre Gorget and repeat the test to see if the bonus is 10% or higher.

    Edit: yeah, what Mojo said. It wouldn't be fast going, but it could be easier than building TP on L0 monsters and WSing over and over with a fairly low power test (due to ~35% crit rate).
    What kind of sample size would I be looking at to be reasonably certain of a given critrate bonus with this method? Was planning to test CDC after finishing Almace since I don't think there's been any testing on it yet.

  20. #20
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Quote Originally Posted by Nightfyre View Post
    What kind of sample size would I be looking at to be reasonably certain of a given critrate bonus with this method? Was planning to test CDC after finishing Almace since I don't think there's been any testing on it yet.
    It depends how sure you want to be. For instance, if the real crit rate is 99% (worst case scenario), your odds of getting 100 consecutive crits would be .99^100 = 37%. So if you got 100 consecutive non-crits, I think you could be 63% sure that your crit rate isn't really 99%. It should follow this table

    <table style="width: 497px; height: 134px;" border="0" cellpadding="0" cellspacing="0"><colgroup><col style="width:48pt" span="6" width="64"> </colgroup><tbody><tr style="height:15.0pt" height="20"> <td style="height:15.0pt;width:48pt" height="20" width="64">What if it's really...?</td> <td style="width:48pt" width="64">n=200</td> <td style="width:48pt" width="64">n=100</td> <td style="width:48pt" width="64">n=50</td> <td style="width:48pt" width="64">n=30</td> <td style="width:48pt" width="64">N for 95% confidence</td> </tr> <tr style="height:15.0pt" height="20"> <td class="xl63" style="height:15.0pt" align="right" height="20">99%</td> <td class="xl64" align="right">86.6%</td> <td class="xl64" align="right">63.4%</td> <td class="xl64" align="right">39.5%</td> <td class="xl64" align="right">26.0%</td> <td align="right">298</td> </tr> <tr style="height:15.0pt" height="20"> <td class="xl63" style="height:15.0pt" align="right" height="20">98%</td> <td class="xl64" align="right">98.2%</td> <td class="xl64" align="right">86.7%</td> <td class="xl64" align="right">63.6%</td> <td class="xl64" align="right">45.5%</td> <td align="right">148</td> </tr> <tr style="height:15.0pt" height="20"> <td class="xl63" style="height:15.0pt" align="right" height="20">97%</td> <td class="xl64" align="right">99.8%</td> <td class="xl64" align="right">95.2%</td> <td class="xl64" align="right">78.2%</td> <td class="xl64" align="right">59.9%</td> <td align="right">100</td> </tr> <tr style="height:15.0pt" height="20"> <td class="xl63" style="height:15.0pt" align="right" height="20">96%</td> <td class="xl64" align="right">100.0%</td> <td class="xl64" align="right">98.3%</td> <td class="xl64" align="right">87.0%</td> <td class="xl64" align="right">70.6%</td> <td align="right">73</td> </tr> <tr style="height:15.0pt" height="20"> <td class="xl63" style="height:15.0pt" align="right" height="20">95%</td> <td class="xl64" align="right">100.0%</td> <td class="xl64" align="right">99.4%</td> <td class="xl64" align="right">92.3%</td> <td class="xl64" align="right">78.5%</td> <td align="right">58</td> </tr> </tbody></table>
    Yeah, I used a table just because now I know I can.