[NEW MERCENARIES] Please note that all new forum users have to be approved before posting. This process can take up to 24 hours, and we appreciate your patience.

Add Damage vs ATT Limit

Tpyop0Tpyop0
Vindictus Vertreter: 1,655
Posts: 39
Mitglied
edited July 18, 2017 in General Discussion
I've been doing some testing using different configurations of Additional Damage and ATT Limit. The existing belief is that a multiplier of 6.25 is applied to the additional damage stat and none is applied to the ATT Limit stat when calculating the effective ATT/power. This doesn't appear to be true in the current version of the game according to my test results. I'd like to request some data points from Fionas with 1000 - 2000 attack limit and ATT/Bal cap: record the damage of ~25 non-crit, un-deflected Amaranth Kicks in Hero Havan with 10% Awakening enabled.

Link to google sheets: https://docs.google.com/spreadsheets/d/1REiSY-0oPi7qfA4Zr7Uqijy70kAOUJl4ewpD6_buQkA/edit?usp=sharing

I basically need some additional columns with different ATT Limit numbers. I can't test the 711 - 2299 range with my gear.
Current test results suggest that 3900 Add Damage + 2300 ATT Limit is weaker than 3800 Add Damage + 2710 ATT Limit. The multiplier for additional damage still appears to be 6.25. Please note that I'm not attempting to make any claims about the attack formula, I'm trying to understand it myself.

kls9EnigmaTaro

Kommentare

  • Tpyop0Tpyop0
    Vindictus Vertreter: 1,655
    Posts: 39
    Mitglied
    Bump, for science, for a dead game. A list of numbers is fine, I don't mean a video clip when I say recorded data. PM works too.
  • boddoleboddole
    Vindictus Vertreter: 1,410
    Posts: 123
    Mitglied
    edited July 10, 2017
    Ok, so I pawed though your data in Excel for a bit and came to some conclusions that I feel relatively confident in:

    1) I believe your hypothesis about Attack Limit Release having a modifier of 3.125 is accurate. Trying to account for both predicted and observed damage by changing the ALR modifier to 1.0 and changing the AD modifier to compensate creates a situation where some projections are quite accurate (within a dozen damage points), while others are completely wrong (off by thousands of points). This clearly indicates that there are at least two variables in play, and your 3.125 number for ALR describes the observed results perfectly. It's probably also no coincidence that this new modifier happens to be exactly half of the AD modifier.

    2) This change probably happened during Rise, from this late 2016 thread (http://forums.vindictus.nexon.net/discussion/1322/attack-limit-clarification) the numbers this person came out with works out pretty much as expected if the ALR modifier was 1.0.

    ->As someone pointed out to me, ALR hasn't been around all that long in KR before Rise (a year?). My assumption was that KR would notice any changes when ALR came out, and that information would filter over to us fairly quickly. Though as they pointed out, I can't say that for certain, and it is possible the numbers have been changed for a while and nobody noticed/cared enough to document it. But again, seeing how when Rise came out, both Tpyop0 as well as Shippu and his friend noticed something was different fairly quickly, I'd still lean on the side of saying this is a recent change (I would guess to keep it somewhat relevant against AD, otherwise ALR would be a complete joke, especially for how expensive it is to get).

    3) Your assertion that, "3900 Add Damage + 2300 ATT Limit is weaker than 3800 Add Damage + 2710 ATT Limit" is also correct (keeping in mind that you have to have enough ATT over the boss DEF to actually use all that ALR). Just like it is simple to understand how AD works (just multiply each point of it by 6.25 before final modification), ALR works almost the same (multiply each point of that by 3.125, then by 0.95 (assuming max balance) ). Meaning that in an optimal situation, each point of AD is worth ~2.1 ALR. So in your example you lose 100 AD points (~210 ALR), but gain 410 ALR - so everything works out like it should. Yay!

    *typo fixed: should have been 'lose' instead of 'gain' and 'was' instead of 'is'.
  • Tpyop0Tpyop0
    Vindictus Vertreter: 1,655
    Posts: 39
    Mitglied
    Thanks for the input. There is strong evidence that the ATT Limit multiplier is no longer 1, assuming that the effective damage formula wasn't completely overhauled and it's not sheer coincidence that my numbers are matching up fairly closely. The numbers for the baseline and config #4 matches up closely with a multiplier of 1 because the ATT Limit for both is 300.

    I still would like to get some more data points since all numbers are subject to sampling error caused by RNG damage rolls. Plugging in any number from 2.5 to 3.5 for the ATT Limit multiplier value gives me decent matches, and I'd like to get a better number. 3.0 is currently the closest match. It would have been good if the Gremlin dumplings were still usable.
  • _Kale_Kale
    Vindictus Vertreter: 1,310
    Posts: 34
    Mitglied
    Tpyop0 wrote: »
    Thanks for the input. There is strong evidence that the ATT Limit multiplier is no longer 1
    Just curious, was the ALR multiplier definitely 1 prior to Rise? Could you point me to some data?

  • boddoleboddole
    Vindictus Vertreter: 1,410
    Posts: 123
    Mitglied
    Do you happen to know what the last known attack multiplier was for the attack you used?
  • Tpyop0Tpyop0
    Vindictus Vertreter: 1,655
    Posts: 39
    Mitglied
    _Kale wrote: »
    Tpyop0 wrote: »
    Thanks for the input. There is strong evidence that the ATT Limit multiplier is no longer 1
    Just curious, was the ALR multiplier definitely 1 prior to Rise? Could you point me to some data?

    https://vindictus-hd.blogspot.com/2014/06/formula-common-formulas.html It's from 2014, when people cared about such things. None of this is my work. The original website in which the data comes from appears to be down.


    boddole wrote: »
    Do you happen to know what the last known attack multiplier was for the attack you used?

    The skill multiplier of Amaranth Kick? I don't know. But it shouldn't matter since skill multiplier should be applied after the effective damage/power and I use only one skill for my comparisons.


  • boddoleboddole
    Vindictus Vertreter: 1,410
    Posts: 123
    Mitglied
    " it shouldn't matter"
    -Yeah, you're right, either trying to 'align' the multipliers across all sample points or the predicted damage shouldn't matter in the end.

    Any chance you could get, say ~100 data points per gear setup? I'd be curious to see how the numbers move. Currently, the standard deviation in damage per hit compared against the set average runs from ~0.65% to 1.37%. Also, is there a data input error in setup #5? Noticed the 39.5k hit, which is waaay off for the rest of the samples (which makes the st.dev/set average over 15%...).
  • Tpyop0Tpyop0
    Vindictus Vertreter: 1,655
    Posts: 39
    Mitglied
    edited July 10, 2017
    boddole wrote: »
    "
    Also, is there a data input error in setup #5? Noticed the 39.5k hit, which is waaay off for the rest of the samples (which makes the st.dev/set average over 15%...).

    That's a good catch. That's the crit number. We can just take it out. This makes config #2 the only sample set in which the two bases deviate from each other by more than 2% so I'll run that again when I get a chance. I'll get burnt out if I have to kick Havan in the crotch for 100 more times per config set.
  • boddoleboddole
    Vindictus Vertreter: 1,410
    Posts: 123
    Mitglied
    edited July 10, 2017
    "I'll get burnt out if I have to kick Havan in the crotch for 100 more times per config set."
    -Been there, did a study on the damage formula / ein lacher a while back sampling multiple gear setups against multiple bosses usually 95+ samples per run...it does get a bit soul crushing...

    Also, in case you haven't already done this - the st.dev between predicted damage / a theoretical ability multiplier / and observed damage is ~1.06% using 3.125. Of course the margin of error for each testing set also averages around 1%, so aside from using bigger sample sets, I don't know what can be done to make more accurate predictions.
  • HallyHally
    Vindictus Vertreter: 1,985
    Posts: 245
    Mitglied
    by a 3.125 multiplier.. how do it work? For example, if you have 1k more attack than the attack cap, and 500 limit release. Does it mean you have 1k extra effective attack, and cannot make full use of that limit release; or does it mean you get the full benefit.
  • gogoshicagogoshica
    Vindictus Vertreter: 500
    Posts: 39
    Mitglied
    edited July 10, 2017
    ATT limit multipier better not be just 1,considering how insanely difficult it is to get that.Takes less money to buy a +13 and enjoy its additional damage than to get capped on att.limit.
  • LeucosieLeucosie
    Vindictus Vertreter: 2,105
    Posts: 164
    Mitglied
    I can help when I get home in the evening ( i'm from eu ) :)
  • Tpyop0Tpyop0
    Vindictus Vertreter: 1,655
    Posts: 39
    Mitglied
    Hally wrote: »
    by a 3.125 multiplier.. how do it work? For example, if you have 1k more attack than the attack cap, and 500 limit release. Does it mean you have 1k extra effective attack, and cannot make full use of that limit release; or does it mean you get the full benefit.

    It's a good question. I did some testing instead of going to bed... I tested 3 configurations on Hero Havan, which has a Battle ATT Cap of 23,000:

    1) 23,021 ATT, 3300 Add Damage, 2400 ATT Limit
    2) 25,451 ATT, 3300 Add Damage, 2400 ATT Limit
    3) 28,424 ATT, 3300 Add Damage, 2400 ATT Limit

    The Additional Damage and ATT Limit values are same for all 3 configurations. Bal is at 90. Configurations 2 and 3 did the same damage. Configuration 1 did much less damage. You should get the full benefit of 500 ATT Limit if you are 500 ATT above the Battle ATT Cap. These 3 configurations are #8, #9 and #10 on the spreadsheet.
  • SamalenkoSamalenko
    Vindictus Vertreter: 1,475
    Posts: 221
    Mitglied
    edited July 11, 2017
    300 limit release


    500 limit release


    A noticeable at least 2.5% difference . There is no way that Limit release has no multiplier after RISE.
  • RadeRade
    Vindictus Vertreter: 1,020
    Posts: 39
    Mitglied
    What are the implications of an attack limit release multiplier on battles without attack caps? Does it just do nothing, is there some arbitrary value it uses to scale the damage, or is that benefit just always present?
  • LeucosieLeucosie
    Vindictus Vertreter: 2,105
    Posts: 164
    Mitglied
    As promised

    316te00.jpg

    28453
    28313
    28220
    28360
    28453
    28964
    28406
    28592
    28174
    28313
    28360
    28592
    28638
    28638
    28731
    28546
    28453
    28453
    28592
    28731
    28592
    28731
    28731
    28638
    28685
    28917
    28546
    28871
    28546
    28360
    28406

    hope it helps

    Delteros
  • Tpyop0Tpyop0
    Vindictus Vertreter: 1,655
    Posts: 39
    Mitglied
    Rade wrote: »
    What are the implications of an attack limit release multiplier on battles without attack caps? Does it just do nothing, is there some arbitrary value it uses to scale the damage, or is that benefit just always present?
    I don't expect any impacts since the multiplier should just be applied to a number 0. I don't plan on testing uncapped numbers.


    Leucosie wrote: »
    As promised


    hope it helps

    Thanks. I appreciate the help.




    I'm going to wrap up this set of experiments with: "There is a multiplier for attack limit and it's around 3.0". Anywhere from 2.8 to 3.2 seems to give close enough matches. Anyone that can afford ATT Limit will have high enhancements as well anyway. Make no mistake that the Additional Damage stat will have the biggest impact on the damage dealt by far, and its impact is big enough to allow end-game characters to ignore most fight mechanics.



    KingRichRiceGhengisJohn