Emx77 -> 3.4 beta is more favorable to defenders (8/12/2010 7:35:06 PM)
|
Have anybody noticed that version 3.4 is more favorable to defender? I have performed a couple of test to see whether this is subjective observation or not. Test description: I ordered two infantry battalions with mortar Coy to attack one fortified infantry battalion, and saved this setup. You may see this at following screenshot: [image]http://img37.imagefra.me/img/img37/8/8/12/emx77/f_w05um_d238388.jpg[/image] Then I executed attack, recorded results and repeated this process 20 times both with version 3.2 and with version 3.4. The results are shown in next table: [image]http://img38.imagefra.me/img/img38/8/8/12/emx77/f_ww60wm_1de3b56.jpg[/image] From the table above we can see that attacker losses are on average 14% higher and defender losses are on average 40% less in v3.4. The second difference is rather big one. Although mean Attacker Loss is somewhat higher in v3.4, Independent samples t-test calculated in SPSS 17.0 statistic package doesn't showed this difference to be statisticly different between two versions. At same time, Independent samples t-test showed that there is statisticly significant difference between v3.2 and v3.4 in respect to Defender Loss. This result objectively supports my hypothesis that v3.4 is more favorable to defenders compared to v3.2. Also, here are results which shows how many times defender is retreated, unfortified or divided into subunits after attack: [image]http://img02.imagefra.me/img/img02/8/8/12/emx77/f_ww60xm_6f5e20c.jpg[/image] From tables above we can see that in v3.4 defender is less prone to retreat, to be unfortified and to be divided after attack, which further supports hypothesis about v3.4 being more favorable to defenders. Question is how this will affect existing scenarios balance?
|
|
|
|