Forum Settings
       
This Forum is Read Only

Log Parse Results! Effect of New NM Gear Stat Bonus on 30MRDFollow

#1 Apr 25 2011 at 4:02 PM Rating: Excellent
*
56 posts
Note: This test is part of a compilation of tests. The goals of these tests and the scopes are defined in the following thread. This thread is a subthread of the master thread:
TEST CENTRAL


========= I. PURPOSE

It was easy to do a quick parsing test with all the new NM gear on my 30 MRD (TEST1), then take off ALL gear except my weapon and see if it affected my average damage (TEST2). See the spreadsheet for full details.

Note this is the first time I have ever done an explicit test on just the effects of the gear itself. Also note that because this was a quick test with approximately 300 trials of Light Swing, that is not enough to conclude anything definitive on the role of the Accuracy stat. But we can make a conclusion on the change in Attack/STR.


========= II. EXPERIMENT CONSTANTS, VARIABLES. EXPERIMENT RESULTS
The spreadsheet and definitions within should be self-explanatory.
Note the spreadsheet has multiple pages.
https://spreadsheets.google.com/spreadsheet/ccc?key=tUfQdbQE8srrkMThpSvfxrg#gid=2

Due to the fact the experience values capped out and the mob rank was not differentiable, only the average of ALL mob ranks in the spreadsheet is shown. The randomness due to what rank mobs spawn should not affect data accuracy much as shown by my latest DEX test on these mobs.


========= III. LIGHT SWING CONCLUSIONS

  • 1. Effect of Gear Attack/STR: ATK: 121(VS)110 STR: 89(VS)80
    • Conclusion:
    • Decreasing STR & Attack on gear simultaneously will result in a decrease in average damage and average critical damage.
    • Data Citation:
    • The change above in ATK and STR resulted in a decrease from 72.36 to 63.66 (-12.02%) for non-critical damage. The change above in ATK and STR resulted in a decrease from 183.39 to 165.88 (-13.01%) for critical damage.
    • Discussion:


  • 1. Effect of Gear Accuracy: ACC:131(VS)120
    • Conclusion:
    • Not enough trials to conclude if decrease in Accuracy results in a decrease in the incidence of misses.
    • Data Citation:
    • However 300 trials is a decent count. It is worth noting that as expected the change in ACC resulted in the Miss Rate going from 28.62% to 32.98%, a change of 4.36%, and a change ratio of +15.23%
    • Discussion:
    • I was not intending to test accuracy at all in this experiment, because I could not get enough data samples before ranking up.



    ========= IV. SKULL SUNDER II CONCLUSIONS

  • 1. Effect of Gear Attack/STR: ATK: 121(VS)110 STR: 89(VS)80
    • Conclusion:
    • Decreasing STR & Attack on gear simultaneously will result in a decrease in average damage and average critical damage.
    • Data Citation:
    • The change above in ATK and STR resulted in a decrease from 183.39 to 165.88 (-9.55%)
    • Discussion:


  • 1. Effect of Gear Accuracy: ACC:131(VS)120
    • Conclusion:
    • Not enough trials to conclude if decrease in Accuracy results in a decrease in the incidence of misses.
    • Data Citation:
    • Discussion:


    ========= V. TRUNKSPLITTER II CONCLUSIONS

  • 1. Effect of Gear Attack/STR: ATK: 121(VS)110 STR: 89(VS)80
    • Conclusion:
    • Inconclusive due to low number of trials.
    • Data Citation:
    • The change above in ATK and STR resulted in an unexpected increase from 187.89 to 203.50 (+8.31%) for non-critical damage. This is due to the randomness of mob spawn ranks and by the few number of trials.
    • Discussion:


  • 1. Effect of Gear Accuracy: ACC:131(VS)120
    • Conclusion:
    • Not enough trials to conclude if decrease in Accuracy results in a decrease in the incidence of misses.
    • Data Citation:
    • Discussion:


    Edited, May 1st 2011 2:55am by DeusBahamut
    ____________________________
    Deus Diabolus, Gstar Raw, Besaid Server
    Leader of Lifestream LS
    http://lifestreamls.enjin.com
    #2 Apr 25 2011 at 4:26 PM Rating: Good
    ***
    1,636 posts
    This is one of the better "experiments" I've seen done so far. It seems that somehow you're test 2 you managed to isolate the same level of mob (or only that data was posted), if you can repeat that you can get some more meaningful data. I'm curious as to how your limited testing showed that no gear did more trunksplitter damage than with gear though, probably just not enough trials, however it does suggest that there may be a problem somewhere.
    ____________________________


    #3 Apr 25 2011 at 4:37 PM Rating: Good
    *
    56 posts
    Correct, you beat me to writing it myself above lol. The rank is automatically assigned based on the exp values I get from the mob when it dies. It may actually be that the EXP values are not indicative of the mob rank in this test. As you can see the EXP numbers scale weird. I will have to go back there and observe the exp values again and compare to the visual rank that is now fortunately available after the patch.

    Regardless, this is even more reason to have a larger data set in case the mob rank cannot be differentiated. Then you can opt to the ALL totals for conclusions.



    Edited, Apr 25th 2011 6:39pm by DeusBahamut
    ____________________________
    Deus Diabolus, Gstar Raw, Besaid Server
    Leader of Lifestream LS
    http://lifestreamls.enjin.com
    #4 Apr 25 2011 at 4:41 PM Rating: Good
    ***
    1,636 posts
    glad to see you're definitely into more tests, thats the hardest thing to get through to people who insist on trying to prove something, more testing is always better. Really, this type of thing should be handed at level cap, and same gear, however reallocate all your str down to zero, while keeping every other stat the same, which sounds like an awful thing to go through.
    ____________________________


    #5 Apr 25 2011 at 4:54 PM Rating: Good
    *
    56 posts
    More testing is not only good but required to put anything in perspective. There are so many different scenarios like the one you just proposed. I prefer to choose grind scenarios with stats people typically would pick so it is more beneficial.

    I might do Rank 50 end game scenario parsing once there is something challenging enough to parse against, and enough gears with good stats to provide incentive to optimize your character at all. Currently there is none for end game, nor is there incentive to do a test at Rank 50 if I'm not getting SP (with perhaps the exception of Great Buffalo hmm) It was easy to do a test while I am already ranking up MRD. All I had to do was write down my stats and copy the log data. Anyone can do that. And I can run my parser on it.

    Edited, Apr 29th 2011 4:49pm by DeusBahamut
    ____________________________
    Deus Diabolus, Gstar Raw, Besaid Server
    Leader of Lifestream LS
    http://lifestreamls.enjin.com
    #6 Apr 28 2011 at 11:26 AM Rating: Good
    *
    56 posts
    No longer relevant

    Edited, May 1st 2011 2:47am by DeusBahamut
    ____________________________
    Deus Diabolus, Gstar Raw, Besaid Server
    Leader of Lifestream LS
    http://lifestreamls.enjin.com
    #7 Apr 29 2011 at 1:03 PM Rating: Good
    *
    56 posts
    No longer relevant

    Edited, May 1st 2011 2:47am by DeusBahamut
    ____________________________
    Deus Diabolus, Gstar Raw, Besaid Server
    Leader of Lifestream LS
    http://lifestreamls.enjin.com
    #8 May 01 2011 at 12:49 AM Rating: Good
    *
    56 posts
    I have reposted the spreadsheet data after going through it very carefully to ensure that the in-differentiable mob ranks did not have a significant impact on the damage averages. I am considering a new test on the effect of Attack and Accuracy bonuses, and measure those effects individually.

    Edited, May 1st 2011 2:55am by DeusBahamut
    ____________________________
    Deus Diabolus, Gstar Raw, Besaid Server
    Leader of Lifestream LS
    http://lifestreamls.enjin.com
    This forum is read only
    This Forum is Read Only!
    Recent Visitors: 19 All times are in CST
    Anonymous Guests (19)