Page 5 of 5 FirstFirst ... 345
Results 61 to 65 of 65

Thread: Accuracy of SoT and how to calculate enemy defense

  1. #61
    Post Demon Bijo's Avatar
    Join Date
    Jul 2008
    Posts
    1,771
    Quote Originally Posted by Ethan View Post
    Something still seems wrong there - the top line items (100% success) look right, at least for the first two (I don't know the third off the top of my head). But for the 50% with 0% random intel (halfer war scenario) we know it *ought* to occur when you have exactly 97% of target's defense. I assume this would be listed as 3% under, or maybe 3.093% "over" (as in target has 103.093% your sent MO). In neither description would it be 2.76, as I think your sheet listed.

    Also, the .01% and 0% cases should get *lower* as the intel uncertainty increases - the better the intel, the less lucky you can be in intel. Normally not relevant, since most assume they will be unlucky, but the fact that your numbers don't follow that trend suggests something is a bit fishy still.

    My formal probability is rusty, but I'll try to mash out the formal equation at some point soon. Intel variability range will be a parameter (good approach on that Bijo), not because I expect to have worse intel than 3%, but sometimes I have a tighter range due to taking multiple SoT... so intel error range will be unbound. Among other things I'd like it built into my custom SoT+SoM excel, so I can know what risks I'm taking exactly.
    i don't think there is anything wrong with it i checked and rechecked over and over again...

    http://thedragonportal.eu/1.png

    and here is the way i calculate the chance %
    i didn't wrote it that way at first
    but this is the shortest way to wrote it (and cleanest)
    i did it so that i can change even the random factor or any other factor :)

    http://thedragonportal.eu/chance.gif

    please tell me what do you think ^^

  2. #62
    Post Demon Bijo's Avatar
    Join Date
    Jul 2008
    Posts
    1,771
    here i made this just now for you

    http://thedragonportal.eu/chance.xls

    took me like 2 mins

    have fun!

  3. #63
    Post Demon Bijo's Avatar
    Join Date
    Jul 2008
    Posts
    1,771
    omg i just did it in one line!

    y (Intel Range %) : [1.00] [0.97] [0.75]
    x (Random Factor %) : [3.5]

    Chance = (100/x)*(1-2/(1+(Off/Def)*(y/0.97)))

    if ( Chance > 1 ) Chance = 1
    if ( Chance < -1 ) Chance = -1

    Chance = -25*(Chance)*(Chance)*(Chance)+75*(Chance)+50

    and i also fixed the xls file: http://thedragonportal.eu/chance.xls

  4. #64
    Post Fiend
    Join Date
    Feb 2011
    Posts
    237
    Your probabilities are incorrect for 3% and 25% range. You seem to simply be basing probability off of the worst case scenario assumption, which is the case where the intel is low by the max amount. The 50% mark on all 3 ranges should be the same, because the average of the variabilities in intel will be 0, presuming it is positive and negative equal amounts of the time. The difference would be that the graphs of probability against relative army strength would be stretched out more the more inaccuracy there was. If intel innaccuracy is 25%, there's a chance that you got 125% of their actual defense, and that your army would receive +3.5% while theirs received -3.5%. This would give you a victory so long as you sent 72.36% of what the intel said they had for defense. Thus your 0% chance isn't until you are sending 27.64% less than they "have".

  5. #65
    Post Demon Bijo's Avatar
    Join Date
    Jul 2008
    Posts
    1,771
    Quote Originally Posted by Hanoumatoi View Post
    Your probabilities are incorrect for 3% and 25% range. You seem to simply be basing probability off of the worst case scenario assumption, which is the case where the intel is low by the max amount. The 50% mark on all 3 ranges should be the same, because the average of the variabilities in intel will be 0, presuming it is positive and negative equal amounts of the time. The difference would be that the graphs of probability against relative army strength would be stretched out more the more inaccuracy there was. If intel innaccuracy is 25%, there's a chance that you got 125% of their actual defense, and that your army would receive +3.5% while theirs received -3.5%. This would give you a victory so long as you sent 72.36% of what the intel said they had for defense. Thus your 0% chance isn't until you are sending 27.64% less than they "have".
    thats not true

    X : [1.00] [0.97] [0.75]

    Chance = (100/3.5)*(1-2/(1+(Off/Def)*(X/0.97)))

    if ( Chance > 1 ) Chance = 1
    if ( Chance < -1 ) Chance = -1

    Chance = -25*(Chance)*(Chance)*(Chance)+75*(Chance)+50

    look also in the xls file: http://thedragonportal.eu/chance.xls

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •