Snowflake's Feed

06-26-2021 at 08:27 AM
Rate this Entry

ESI bug

the weird thing is everyone except me and john marks esi is calculaed correctly. I'm getting a little more than i deserve and hes getting a lot more than he deservves. Cant blame it on my multiple submisssions, since that would've aided me far more, and also theres no reason why multiple subs of mine would've helped john marks alone and noone else.

john's is the easiest to see is too high. he beats 3 people but has over 6 esi. you can never doulbe esi, only first place, if they beat everyone by 100% (so everyone else score zero, or first place getting infinity) can they get double and no more.

mine is a bit harder, but as rtm's bonus multiplier is roughly 2/3s (a litlte higher) and i'm roughly 2/3s his score (a little lower) my multipler should be 2/3*2/3=4/9 and i beat 4 people. so 4+ 4*4/9=4+16/9=5+7/9=5.77 meaning i've got a little more than a point extra what i deserve

trying to speculate how such a bug is possible before i report it in scoreboard errors thread cause i'm not sure they can do much about it until we figure this out better

ThanksJJT_Defender thanked this post
LikesJJT_Defender liked this post
  1. Rogerpoco's Avatar

    Crazy, you've figured out the ESI calculations that well, very interesting!

    Good luck getting it addressed(is prolly not a problem exclusive to this track), I think we are kinda walking around in empty hallways around here lately.

  2. Barthax's Avatar

    I've got ESI calculations of:

    Robert T Mruczek 10.2012
    Bill Gaholt 7.39
    Snowflake 5.84
    John Marks 3.61
    Garrett C. Holland 2.11
    Roger Edwin Blair 111 1.04
    Alan Hewston 0

    What's the timing on multiple submissions having been accepted? Perhaps it's due to get cleaned up in the next maintenance run.

  3. Snowflake's Avatar

    i didnt use a calculator but yeah those numbers look right, or at least very close to right. the timing is across about this last week. usually though the multiple sub errors would give much different bonus, like all the subs together added up. john gets an extra 3 points, i didnt submit underneath him 3 times.

    LikesBarthax liked this post
  4. swaggers's Avatar

    Did you just get a higher score? Most likely hasn't been fully recalculated.

  5. Snowflake's Avatar

    i did, i've had multiple submissions. still not sure why that would affect marks and noone else. also the old bug with multiple subs wouldnt do this, and merely not recalculating wouldnt either as a number like this would've never been correct. but of course yes i imagine the multiple submissions in some way did this i just cant figure out how

    john marks is the guy that i was immediatley under then had 3 subs over

  6. Snowflake's Avatar

    i do remember there was a bug with PSI (pre esi) days. psi just measured how many people you beat, so very easy. i was innaugral so i had zero psi and first place. someone beat me, that pushed me to second and gave me negative 1 psi. Its clear to me that rather than calculate psi from scratch it tried to be efficient be just calculating the change. it saw i used to be at first and went down to second and figured thast a one point drop, failing to take into account i didnt truly go down so much as new players came in ahead of me, so the number of people i beat (zero) remained constant

    i wonder if the algorith is back to trying to figure out esi changes and then apply those chanegs (and have a bug in that) as opposed to just recalculating esi from scratch each time

  7. stygian's Avatar

    Hi, Snowflake. Did you remember to calculate his and your 'Single cavern' performance ESI into the totals you see in the left column?

    ThanksSnowflake, bensweeneyonbass thanked this post
    LikesBarthax liked this post
  8. Snowflake's Avatar

    lol i did not. thx stygian. false alarm everyone.

Join us