Hello & Welcome to our community. Is this your first visit? Register
  • News & Announcements
      As a counter measurement to Ad spam, newly registered users are required to create a thread in the Introduction Forum before allowing to be able to reply and create threads.
      Please allow some time for posts to be approved. We thank you for your patience on this matter.
Bookmark Thread
Results 1 to 6 of 6
  1. #1
    Moderator Leyana's Avatar
    Join Date
    Apr 2014
    Posts
    38,227
    Rep
    702
    Mentioned
    4764 Post(s)

    Testing OASW and Fatigue

    The theory that "OASW is unaffected by fatigue" was brought up again and for the life of me I couldn't find a credible source showing support or disproving that statement. So I decided to do a little experiment of my own.

    The ship: Lv20 Shimushu with Type 4 Sonar + Type 3 Depth Charge
    The map: 1-5-A

    Shimushu was selected because she has a low level to achieve OASW which mitigates one factor affecting the standard accuracy formula in used. 1-5 was the testing grounds because it's an easy to reach sub node. The major assumption here is that ASW and by extension OASW uses the same accuracy formula as shelling.

    Now the accuracy formula is stated to be:
    0.03 + (0.9 + sqrt(ShipLevel - 1)/50 + 1.5 * sqrt(Luck)/100 + EquipACC/100 + UpgradeBonus * sqrt(Star#)/100) * FormationMod * FatigueMod
    So based off that, we can expect the following results:


    So as you can see, if fatigue is not a factor in OASW accuracy, we can expect Shimushu to hit 97% of the time (because hit rate is capped at 97% in game).

    The experiment:
    Shimushu was sent on cancelled sorties until she was in red fatigue. She was then sent to 1-5 to hit up the first node. The sortie was then cancelled after the OASW phase so that she would not gain levels and mess up the experiment.

    The result:
    The observed hit rate was 61% which is much lower than the expected 97% but still higher than the expected 50%.

    The conclusion:
    Fatigue does affect OASW. The likely reason why people report successful OASW at low morale is because of the other factors in the accuracy formula. Namely: Levels, luck equipment upgrade, and formation modifier (line abreast gives a 1.2 multiplier).

    EDIT: Test 2 at full morale

    Result: Observed hit rate was 96%.

    Conclusion: OASW is indeed affected by morale.

    Dataset:
    [+] Spoiler

    [+] Spoiler
    [img]https://i.imgur.com/q4qzLf1.png[/ims]


    4 out of 0 members found this post helpful.
    Last edited by Leyana; 06-29-2017 at 02:47 AM.

  2. #2
    Community Staff Jifaru's Avatar
    Join Date
    Apr 2014
    Posts
    2,149
    Rep
    234
    Mentioned
    417 Post(s)
    Some basic data analysis:

    Using a one-tailed T-test (testing if the observed mean is significantly different from the theoretical mean in one direction) assuming that the mean is 0.97, a score of 1 was assigned to successful hits and a score of 0 was assigned to misses. Out of a sample of 100 tests, we would expect a mean value of 0.97. The inputs were 61 instances of 1 and 39 instances of 0.

    The T-value is -7.314966. The P-Value is < 0.00001. There is no way in fuckin hell OASW isn't affected by morale


    Last edited by Jifaru; 05-19-2017 at 07:03 AM.

  3. #3
    Commander Greenie's Avatar
    Join Date
    Jul 2014
    Posts
    374
    Rep
    7
    Mentioned
    14 Post(s)
    Had to look up student's t-test as a reminder (I rarely encounter it), and yeah, pretty convincing to say the least.

    Results are as expected, but it needed to be tested, I guess.

    Thanks for your hard work




  4. #4
    Admiral goesto11's Avatar
    Join Date
    Oct 2014
    Server
    Paramushir Anchorage
    Posts
    4,499
    Rep
    167
    Mentioned
    1361 Post(s)
    @Leyana: Thanks for doing this. Good to know.


    Thanks to Arkayda for the signature.


  5. #5
    Moderator Leyana's Avatar
    Join Date
    Apr 2014
    Posts
    38,227
    Rep
    702
    Mentioned
    4764 Post(s)
    So I actually went and did another test due to feedback. Apparently the listed accuracy formula isn't that trusted so here's the real results with the same setup but with full morale.

    [+] Spoiler


    It resulted in a 96% hit rate.



  6. #6
    Captain kazenorin's Avatar
    Join Date
    Feb 2016
    IGN
    かぜばね
    Server
    Hashirajima Anchorage
    Posts
    677
    Rep
    2
    Mentioned
    190 Post(s)
    Explains why my Lv110+ Asashio K2D couldn't consistently hit the subs at Spring E5 with red morale even with reasonably good gears


    Got stuck and cannot proceed, then you, my first battleship passionately came: Kongou 126/155, 2016-02-28
    LSC Journey | Taihou 2016-06-16 | Yamato 2016-09-03 | Musashi 2016-09-08 | Bismarck 2016-09-10

 

 

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
All times are GMT. The time now is 02:24 AM.
Powered by vBulletin® Version 4.2.2
Copyright © 2017 vBulletin Solutions, Inc. All rights reserved.
Resources saved on this page: MySQL 7.41%
vBulletin Optimisation provided by vB Optimise v2.6.3 (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
User Alert System provided by Advanced User Tagging (Lite) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Thread / Post Bookmarks provided by Thread / Post Bookmarking v1.1.1pl1 (Free) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd. Runs best on HiVelocity Hosting.
Live Threads provided by AJAX Threads (Lite) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
vBulletin Skin By: PurevB.com