9 Weeks of War Data - Analysis

Discussion in 'Wars' started by admin, Aug 8, 2013.

  1. Nice work war ppl and devs, this is all way more fun than EBs, fiddle the matchup or not, pt happy.
     
  2. At least in season one when you were blown out it was because you were outbuilt.

    And now we still have no answers on the lunacy you all call matches.
     
  3. LABEL YOUR AXES!!!! :idea:
     
  4. Hey Devs, thanks for your work in the EE wars so far, however, as a heavy-spending customer, I would like to provide a few salient loopholes in your war analysis. Some of which have been brought up by the previous replies, but it wouldn't hurt to reiterate.

    1) Using war plunder % differences as a yardstick to measure 'war improvement' is inaccurate. If you are comparing season 1's statistics with these test wars, obviously a 2 hour war would have a higher chance of having a greater plunder disparity, as compared to a 1 hr war. Furthermore, as brought up by many, the new advantaged ko system serves to mask the poor match-up that occurred. What you'll get is a 'false-positive'. (You think there's an improvement in match-up, but no!)

    In a nutshell, how do you even compare when season 1 were all 2 hour wars, and season 2 had many 1 hr wars? 1 hr wars have a narrower gap, and you failed (or happily-ignored) to take this into account.

    2) I'm a proponent for the old 'classic' wars. There are many who favour the advantaged ko-system and there's nothing wrong with that. But personally, there is really no fun spamming the mouse button for a few mins just to get in 1-2 hits, before the person self-kos. In the classic wars, with a fixed ko, more hits can get through, and you wont need to 'spam' clicks for minutes in a row.

    Perhaps a solution that has been previously brought up would be a 1 min advantage, (2 min max). 1-2 mins variability provides enough randomness to provide improved survivability for the person coming out of ko, yet not too boring for the attackers.

    3) We haven't tried 1 hr wars with fixed 8 mins ko time? Not sure if it'll work, but if the yardstick for war improvement is % plunder disparity, i dare say it'll do well.

    TL:DR, Having a 1 min advantage, (2 min max) on top of a fixed KO time (say 10 mins in 2 hr wars, 8 mins in 1 hr wars) could be a good fix to things.
     
  5. This reminds me of what my college professor always said "Statistics don't lie, staticians do". By just showing results means wars are closer but what's missing is because of what variable? Is it because you added dts/dtw and variable ko or because u are making better matchups, or both? Also what is missing is what % of wars were blowouts when clans total strength (including bfa/bfe/hit ratios or whatever measurement your using) are within 5%, 10%, 20% of each other etc. In other words, are wars closer because u making better matchups or simply because you changed the war mechanics? Is there a correlation between strength difference and blowouts? Are you saying based on your hit ratios of 75% of clan members being able to hit each other for matchups correlates to 25% differences in plunder therefore making wars closer? Where is the efficient frontier or sweet spot for hit ratios or differences in total strength in matchups?
     
  6. Its clear there are a lot of ppl here that are frustrated primarily with one thing: the matchup system playing too much of a part in deciding the outcomes of ee wars. Part of the frustration is due to the lack of transparency in the matchup process (we don't even get a clan strength rating during matchup anymore - even tho it was somewhat made up). If the war rosters are evenly matched and there's still a blowout then one can only complain about their OWN clans activity level (or lack thereof) and war strategy. Devs would be off the hook ;)

    I tend to agree that these adjustments being made by the devs while demonstrating a good faith effort to improve the war system falls far too short. Even after restricting roster sizes much too narrowly (prevents many clans from warring) the matchup system is still broken (allows too many mismatches based on overall clan strength and combined stats) effectively deciding the outcomes of wars. No Match was a step in the right direction but not a permanent fix. There will never be perfectly even wars of course but the current system still leaves much to be desired.

    Not until the matchup process is able to pull from a reserve pool of warriors to manipulate the combined stats and roster numbers (as many have suggested in the past) to even out the overall strength of two warring clans will the matchup system be complete IMHO. Until then we'll just keep on praying to the KAW gods that we don't get screwed over yet again by a bad matchup that decides the wars' outcome before it even gets started...
     
  7. Devs, are you studying the most important statistic of all? Participation rates
     
  8. Great work kaw universe!!
     
  9. When is Season 2?
     
  10. Thanks devs for the data.

    Since we agree with all posters before who already pointed out that the smaller difference in end plunder is most probably caused by the random KO system and not a better matchup system, we welcome this weekend tests. They should give more definitive information on the cause of the smaller difference of the end plunder.

    We will work with any system you present us with, be it random KO/set KO/KO with an advantage and 1 or 2 or even 4 hours war. Of course we have our favourites (set KO times or random with 1/2 minute window and a mix of 1 and 2 hour wars), but we will adapt as surely many other clans will.

    What we cannot adapt to is the bad matchups that will decide the outcome at the start. It starts there and so far we haven't seen much improvement.

    What would be much more interesting for us to know than differences in end plunder, is what the difference is between cs, BFA and BFE that this new 'hit ratio algorithm' allows for in the matchups. Because as stated before, most wars are decided at matchups simply looking at the difference between strenght of both teams warring.

    We'll try to participate as much as possible this weekend to provide you with enough data on this test. However, the willingness to participate is at an all time low.

    ༺Elysium
    ༻Isles of the Blessed
     
  11. Thanks devs :)

    Personally I'd like to see a mixture of war types in season two, would certainly make things more interesting. :)
     
  12. It is titled "9 weeks of war analysis," but they start off by saying it is 8? KK babay.
     
  13. The advantage ko allowed players to come out of ko at 8-13mins before getting attacked, which truely helps cause the tracker would have a hard time, so the player could inflict self ko, creating a lower plunder loss, rather than being timed in 10mins and simultaneously harrassed by 3+ people making self ko harder...soo I believe that the advantage ko was the reason for closer plunder differences
     
  14. @LegendaryScopez,
    Please do not spam or I will spam you. It is against the tou, as it is spam and very similar to posting "first" or "second."
     
  15. KaW_Admin,

    I will not presume to speak for everyone who has posted prior, but I will try to distill what (I think) has been said regarding set KO vs. Random/Advantage KO. This week will reveal whether that one variable merely had an effect on plunder outcomes or whether it demonstrates the strength of the underlying matchup system.

    Many of the posters above suspect that there will be more blowouts as a result of the set times. They argue that the matchups themselves are flawed and the random/advantage ko system only masked an already bad matchup by allowing the weaker team to try and avoid being sat on by the dominant team (the "Turtle" or "Whack-A-Mole" scenarios discussed in other threads).

    Conversely, if the plunder differential totals remain relatively consistent this week as compared to the last few, then it will underscore the fact that the matchups themselves were fairer regardless of the KO system employed.

    I look forward to seeing the results. (And, to my fellow KaWers, if I have misstated/oversimplified/totally screwed up the reasoning about random vs. set times--please feel free to point that out).

    Cheers,

    Kage.
     
  16. When is the last time anyone has seen a no match besides the super tiny clans?? There hasn't been one in weeks among the top 100. The disparity among the big clans is much greater than among the smalls. BFA plays a bigger factor than whether someone is 1m cs bigger than you at the same war roster spot. U can overcome that tiny difference with pots and eq, but when someone has 500m plus in bfa u won't be plundering them. BFE, with attk heavy eq, gives you an attk bonus over the defender. BFA gives u the most important thing, static def. if BFE was def based, it should be factored in more. But in the current system bfa is the overriding factor among the top tiered ee clans. The dtw/dts may be effective matching small clans with beginner builds but not in matching clans where everyone wars above 6-8m cs. In those wars bfa should come to the forefront in matchups.