We've analyzed war data for the last 8 weeks of wars. We've omitted the first test week for technical reasons. We've looked at a lot of data and explored it in many different ways. We cannot share everything but we can share the high level details. Plunder difference. We've shown the difference in plunder for each war in a week and charted the differences over time. In this example under 25% shows how many wars in this week resulted in a plunder difference below 25%. These are the best results, where the plunder scores are closest. You can see the last two weeks of wars have had the highest number of wars under 25% plunder difference. Almost all the wars in the last two weeks resulted in less than 100% difference. Comparatively we feel that the progress is net positive as a whole. We can also see that the average plunder has not decreased during this time. Which is important because it shows that the wars are not resulting less plunder despite the the smaller differences in scores. We can also see that the max plunder hasn't drastically increased either. Finally, we can see that in the wars with large plunder differences (blow outs), they are much smaller with the new system. So while there are still some blow outs, they are not as severe as the previous system. We feel these results are indicative of an overall positive trend in the current iteration of our war system. One question that arises though, is about war durations, types and KO times. Over the last two weeks we have had the same war setups and it's possible that the current system favours those war setups. In order to test against this theory we are setting up this weekends wars as all 2 hour classic wars. We'll analyze this data to ensure that overall results continue to be positive. Once we've concluded Week 10 of testing and analyzed the data we feel we'll have enough data and iterations to plan and announce Season 2. There are a lot of other changes we'd like to make to Season wars and we are listening to all feedback. Our first priority based on continuous feedback throughout the last season was to address the matchup system and blowouts. We feel we've made significant progress in doing so though we will continue to strive to improve upon it. We would greatly appreciate it if clan admins would post their thoughts and feedback. While we listen to all feedback, a unified voice from the leader of a clan is much appreciated. Thank you all for participating in these tests and contributing to the overall improvement of the game. You are brave warriors!
Love this data. Thank you for the share =) Going to read over a few more times..So what this shows is, week 8 pretty much was the best week, match up, and amount of decisive wins?
KaW Admin, I will still need time to digest the data, but would like to point out one detail. The "classic setup" as you put it was a 15 minute KO. This weeks test wars all have a 10 minute KO, so in a strict sense, they are not the "classic setup" from Season 1 wars. I'm sure there will be more suggestions and analysis of the data from myself and others, but just wanted to point out that fact in case you were going to use test war data and compare it with Season 1 war data. Cheers, Kage.
Devs I would prefer half 1h wars other half 2h. Why? Because 1h was fun, can go either way and only, ONLY one xtal which are getting pricy. Just my two cents
The reasons wars are closer is the better chance to sko. That just makes wars appear closer, but they aren't actually closer. With this weekends wars, u will likely see a larger plunder difference than during season 1 as you have pretty much eliminated the no match. It seems you also have spent too much time on bfe adjustments when bfa is a MUCH larger factor in wars. Untouchables will skew wars significantly. 1 top 50 can skew the score 20%.
"in the last two weeks, there has been the lowest plunder difference" Perhaps because half of kaw is in an osw?? Or at least a lot of the biggest clans.
so instead of fighting our guts out and getting a close score in an impossible match up we need to just lay down and not fight so you can see how poor your system is? FAIL!!! ps, i could have scraped that data and posted it in 10 minutes, where is the rest. like the stuff that matters per clan.
Thanks for answering non of my questions. Why has tgia been matched up against clans 60-100 CS more with more BFA and equal or better BFE. 1 hr war with advantage - CS and BFA is not as important. But 2 hr war with 10 minute timer ko. No a chance a clan can take in a clan that much bigger. So I ask again. Did u even bother looking into this? Thanks
This graph is annoying, you guys ever hear of a pie chart? Feel like I'm in a damn quarterly meeting.
Great Job Overall Devs. I like the charts and the statistical analysis. I like the 5 min man advantage in wars, can still time/track its just not as easy. And i been clamoring all pre-season to let us war with larger numbers. Its probably late in the game to allow larger numbers to war; but i still support and ask for it.
Thanks Devs. This is helpful data but want to point out a couple things that will hopefully help you going forward: 1. It has always been my observation that you have placed far too much emphasis on raw stats and not nearly enough on bfa. Having said this, I believe with the introduction of new lands and buildings it has helped reflect in raw stats, what was once a bfa difference. So therefore your algorithm can be more effective as bfa has effectively been converted to new lands and buildings. I would highly recommend introducing hf build upgrades as soon as possible as this will further serve to distinguish the conversion of bfa to raw stats and therefore help with better matchups. 2. I couldn't disagree more with the standard/no advantage KO setup. This has a few negative effects such as: (1) gives the 'losing side' very little chance to get back into the war because the opposition knows exactly when their shadow will appear at the doorstep and have 15 guns aiming at their heads-no chance. (2) it places a ridiculous burden on clans to find trackers whose only job is to sit and watch cc for the reds and constantly update their spreadsheet. It's overkill and frankly nobody should spend that much time on a game-especially if these wars will be the setup for season 2. 15-20 wars in a week....good luck keeping people on kaw for the long haul because you will have tremendous burnout and people will just quit warring and/or quit kaw altogether. 3. Place more emphasis on bfa. While I think introducing the upgrades on hf ASAP will mitigate some of this distortion, I still think you need to weigh more heavily the bfa of warriors when creating matchup. To me I never understood why you made this so complicated....you should just add up raw stats, bfa, bfe and get a sum of this. Maybe even the average or median as a means for you to create good matchups. 4. Come out much earlier than 20 hours before first war with war times. People have been preparing since Monday and using the times from last week as a basis for preparing rosters. 5. I agree with er-rudimentary that you will most certainly see larger gaps in plunder this war as a result of the standard 10 minute KO time. That's all I can think of for now but we do appreciate the transparency and insight into the statistics. And we appreciate your willingness to listen.