Jump to content

perpixel

Verified Tanker [EU]
  • Content Count

    115
  • Joined

  • Last visited

Reputation Activity

  1. Upvote
    perpixel got a reaction from cccpivan in Motorized Vehicle Porn Thread   
    i'm loving the difference between the EU players and the NA players
  2. Downvote
    perpixel reacted to Garbad in WN8 breakthrough   
    *you're
  3. Upvote
    perpixel reacted to Praetor77 in ⟪WN8⟫ Development / Resources   
    Filtered the data from Phalynx as it was a HUGE dataset, picked at random 500k battles from tier 9 tanks (500k rows, although many tanks were in the same battles, does that make sense?). I realized normalizing scale is not necessary for inputing the data to Eureqa since the logistic function used to predict models for binary classification already squashes the result of each variable to a 0,1 range. Xelos, if you can do something similar with your variables before doing the logit regression that would be dandy! Otherwise, your resulting coefficients have to at least be multiplied by the mean for each variable, right?   Then I plugged all of that into Eureqa and ran 1.5e10 formula evaluations using AUC error and fitting damage, track_damage, radio_damage, kills, defense, spots, teamdamage to a binary win formula using a logistic function. (did not use cap, since it will be a strong predictor but only due to the fact that you can only get cap points while winning. Causality is completely reversed for this  stat, so no sense in including it).   Some preliminary results:     So, Gloating´s prediction was much more accurate than surfer´s. Sorry!
  4. Upvote
    perpixel got a reaction from Expendable_Lad in Dog thread   
    Here's mine he's about 10 now. The picture is a year or 2 old so he looks a bit older now. He's really one of those gentle giants
     

  5. Upvote
    perpixel got a reaction from Nope in Dog thread   
    Here's mine he's about 10 now. The picture is a year or 2 old so he looks a bit older now. He's really one of those gentle giants
     

  6. Upvote
    perpixel got a reaction from Hidhorse in Holy BTC   
    Nope no bitcoins i did invest in some fruit company some years ago. Never gotten around to check if those are worth something now. should check it out. If i could just remember the name. I believe it was pear or apple or something. 
  7. Upvote
    perpixel got a reaction from emperorguy in How broken is WOT new rating system?   
    8k battles in 60 days. require 11 hours a day every day of non-stop tanking. given an average battle time of 5 minutes.
  8. Upvote
    perpixel got a reaction from FlamingTier in How broken is WOT new rating system?   
    8k battles in 60 days. require 11 hours a day every day of non-stop tanking. given an average battle time of 5 minutes.
  9. Upvote
    perpixel got a reaction from Dlur in 2nd campaign promo   
    Don't know what to make of it yet.. but here it is..
     
    http://www.youtube.com/watch?v=Rc-_LN4emCk
     
     
    Some more prime info from WoT eu portal:
     
    Clan Wars: Second Campaign
    The Campaign will be launched on 5th November and consist of three stages:
    1st Stage: Revolution Epoch (Tier I-VI) 2nd Stage: Train Robbery (Tier I-VIII) 3rd Stage: Gold Rush (Tier I-X)  
    2nd Global Map
    The Second Campaign will take place on one of two separate Global Maps, which will be available for all players soon. The second map will work like a standard Global Map:
    Tank locking disabled Fog of War enabled Spies/counter-spies enabled Revolts enabled Ransacks enabled A clan cannot fight on both maps simultaneously but can move between them as much as it wants.
     
    Victory Points and Fame Points
    Victory Points will be acquired from provinces and special tasks by a clan as a collective. These points will be used to determine the winner of each stage and later – of the entire Campaign. Victory Points will also be added each turn as a 1/24thpart of the daily Victory Points’ income. The winner of the stage will be determined by the amount of Clan Victory Points earned within that stage. In the case of Victory Points being equal, the number of winners may be increased.
     
    Fame Points, on the other hand, will be earned by clan members individually. They will be based on the amount of experience received in the Campaign’s battles, as well as on the completion of secondary tasks. A player’s Fame Points will equal the total amount of the team’s Fame Points divided by the number of team members and multiplied by special rates. These points will be the basis for the player’s individual ranking.  Fame cannot be lost – gathered points will remain with the player until the Campaign’s end.
     
     
    Special Tank
    The top 10,000 players who earn the largest amount of Fame Points in CW battles will be credited with a special Campaign tank, VK 72.01! In the case of players receiving an equal amount of Fame Points, the number of winners can be increased.
     
    In-game Gold
    The top 30 clans with the largest amount of Victory Points will receive in-game gold:
    1st place – 1,000,000 gold to the clan treasury 2nd place – 500,000 gold to the clan treasury 3rd place – 300,000 gold to the clan treasury 4th-30th place – 100,000 gold to the clan treasury * Please note that the number of clans rewarded may vary in different regions depending on the number of clan wars participants and the general server population.
  10. Upvote
    perpixel got a reaction from Derpinator2013 in Are good platoons bad for your stats?   
    So i was having a discussion about platooning and their effect on your wn7. In my opinion it wouldn't make a a lot of difference there are some pro's countering the con's. and only on real extremes do the con's start getting big problems(for example a 3 embronic journey's in a hellcat platoon).
     
    solo
    pros:
    -More hp
    -more time to deal damage
     
    cons:
    -more team dependant
    -lower w/r
     
    platoons
    pros:
    -higher w/r
    -people having your back/saving you when you make a mistake
    -less bad players
     
    cons:
    -less time to do damage
    -more people fighting over the same damage pool.
     
     
    as i see it. in platoons its easier to reach a constant level. while solo you will get bigger ups and downs. eventually adding up to about the same.
    Is there a way to reliably test this? maybe someone has an idea how to statistically prove either point.
  11. Upvote
    perpixel got a reaction from GeeForcer in ⟪WN8⟫ Development / Resources   
    i actually think wn8 shouldn't take wins into account. you would need it as a separate parameter to determine people's skill same as avg. tier. Preator tried to explain this earlier.
     
    wn8 is a normalisation of the combination of statistics of your performance in the battles. the damage YOU do, the spots YOU do, the kills YOU make, etc. etc. the problem with winrating is that it isn't only your doing.
    Is someone a worse player if he got a pools medal and 10k damage in a tier 10 match but lost because of the other flank crumbled and allowed the enemy to surround him?
     
    So its an indication of thing you do in the battle. And all those variables should add up to a higher winratio. Then when you actually want to know if he is sacrificing his wins to get to those stats you can choose to look at his winrate and determine if this is actually the case.
     
    It shouldn't be a 1 stat tell all situation. you will need a couple to determine someone's skill. But those should be a lot more precise. And i believe the current route is the correct way to precisely normalize a players contribution to a battle. whether he wins or loses.
  12. Upvote
    perpixel got a reaction from Zartok in ⟪WN8⟫ Development / Resources   
    That would work.
     
    "see as wn rating as boobs. wn7 are easily enlarged fake breasts while wn8 are more natural. The only difference is that we changed the cup rating so a wn7 D-cup is now more of a C-cup. And by keeping all the breasts natural the real big ones stand out more."
  13. Upvote
    perpixel got a reaction from Praetor77 in ⟪WN8⟫ Development / Resources   
    That would work.
     
    "see as wn rating as boobs. wn7 are easily enlarged fake breasts while wn8 are more natural. The only difference is that we changed the cup rating so a wn7 D-cup is now more of a C-cup. And by keeping all the breasts natural the real big ones stand out more."
  14. Upvote
    perpixel got a reaction from Rexxie in ⟪WN8⟫ Development / Resources   
    That would work.
     
    "see as wn rating as boobs. wn7 are easily enlarged fake breasts while wn8 are more natural. The only difference is that we changed the cup rating so a wn7 D-cup is now more of a C-cup. And by keeping all the breasts natural the real big ones stand out more."
  15. Upvote
    perpixel got a reaction from Folterknecht in ⟪WN8⟫ Development / Resources   
    That would work.
     
    "see as wn rating as boobs. wn7 are easily enlarged fake breasts while wn8 are more natural. The only difference is that we changed the cup rating so a wn7 D-cup is now more of a C-cup. And by keeping all the breasts natural the real big ones stand out more."
  16. Upvote
    perpixel got a reaction from ViktorKitov in ⟪WN8⟫ Development / Resources   
    That would work.
     
    "see as wn rating as boobs. wn7 are easily enlarged fake breasts while wn8 are more natural. The only difference is that we changed the cup rating so a wn7 D-cup is now more of a C-cup. And by keeping all the breasts natural the real big ones stand out more."
  17. Upvote
    perpixel got a reaction from boom_town in How do you manage your crews?   
    The ideal would be only grinding one line per nation. and use one crew per nation. This will give you the best possible results. Though this means selling all your finished tanks and only keeping tier 10's. Then just switch the crew whenever the situation asks for it.
    Problem is this is pretty much fun ruining
  18. Downvote
    perpixel reacted to bjshnog in ⟪WN8⟫ Development / Resources   
    No.
  19. Upvote
    perpixel got a reaction from Praetor77 in ⟪WN8⟫ Development / Resources   
    of those people with more than 10000 games. is there a way to extract groups that played more than an x% in en certain tank type or have an avg tier below certain levels? this way you can make certain the avg wn8 of these groups is what it is supposed to be.(silence those arty players and show it will level out now. And make sure sealclubbers aren't rewarded to much)
  20. Upvote
    perpixel got a reaction from SumiXam in Hellcat Appreciation Thread   
    just felt like a junior Embryonic journey 
     
    Victory! Battle: Malinovka 13-9-2013 23:28:21 Vehicle: M18 Hellcat Experience received: 2.988 Credits received: 67.906 Battle Achievements: Top Gun, Radley-Walters's Medal   and the screens:


     
  21. Upvote
    perpixel got a reaction from ProfessionalProGamerer in An outcry   
    TBH. only an uptrend here. almost a 500 wn7 difference with my overall.. i am secretly stealing your skill while you sleep miros
  22. Upvote
    perpixel got a reaction from Private_Miros in An outcry   
    TBH. only an uptrend here. almost a 500 wn7 difference with my overall.. i am secretly stealing your skill while you sleep miros
  23. Upvote
    perpixel got a reaction from GoodEyeSniper14 in EVERYONE HATES THE UNICUM CLANS! (something interesting I found)   
    jingles gets carried A LOT because he's always playing with unicums because of his name. hence the high W/R
  24. Upvote
    perpixel got a reaction from Praetor77 in EVERYONE HATES THE UNICUM CLANS! (something interesting I found)   
    i do believe there is some sort of stats inflation going on. I am seeing a lot more high stats people lately (and must admit my stats are going up faster in i suppose they should.) Could also be just being around better players is manipulating my point of view.
  25. Upvote
    perpixel got a reaction from LokiSnake in EVERYONE HATES THE UNICUM CLANS! (something interesting I found)   
    jingles gets carried A LOT because he's always playing with unicums because of his name. hence the high W/R
×
×
  • Create New...