Pikeman Archive
Thread: Tests of defense affect on hit chance
Page 1 of 1
Bermag
Tue Jul 26, 2005 4:01 pm
#1
I made some test trying to find out how CoB and defense in general will affect chance to hit.
Target was my Pikeman/Doc/Swordsman which have 210 ranged defense and CoB (+250 defense). Attacker was my BH/rifleman with 315 ranged accuracy and +5 accuracy on rifle (using sac reduction pup) using ranged shot.
I shot 400 shot in each test case. Note: One thing to think about when testing. Not even 100 shot is enough to test correctly. I had a large variation when I compared 100 series of shots (for example with no COB/Pikatta it was as low as 4 misses and one was 10 misses). Might need as many as 1000 shots to get some statistical correct results.
With no CoB/ no Pikatta:24 misses out of 400 = 6 %
With no CoB / Pikatta: 31 misses out of 400 = 7.75%
With COB / no Pikatta: 71 misses out of 400 = 17.75%
With CoB / Pikatta: 73 misses out of 400 = 18.25%
The difference between using Pikatta or not when having COB active is so small that it might just be random results.
Note: When taking Pikatta I did re-equipweapon (on target) after taking food.
One thing is sure however, having CoB active made a big difference in damage taken, mostly because of damage mitigation. When I had no CoB running I had to switch and use bacta shot to not get incapped, when cob was active bacta infusion was enough to keep health up (and did not go lower than 2/3).
Bermag
Tue Jul 26, 2005 4:06 pm
#2
There are two things I would like to test as well. How much difference will improved cob have and what happens if you increase accuracy? However the test I did took all my gaming time today already lol
Anyone willing to do some more tests?
Bermag
Wed Jul 27, 2005 2:08 am
#4
Note: Pikatta was 50% (not sure if it was pre or post cu made though)
As you can see the effects of Pikatta is very low. I have heard people saying that Pikatta is more useful when you have very high defense running improved aura/imp CoB, but my test results show a different picture. The effect is actually lower when you have improved your defenses with CoB. The improvement you see with Pikatta both with and without CoB might be to small to be evidence of it working or not (might still be to few test results).
I have a theory about Pikatta. It is NOT a dodge chance (i.e. 50% Pikatta would let you dodge 50% of the time), but it was not like that pre-cu either. Pre-cu I think it actually raised your defenses with 50% (and it used the actual defenses as base, not the defense cap. I have a friend that tested that and having defenses over the cap helped a lot when using Pikatta). It mightbe like that still and my theory is that Pikatta is working on the skill defenses without taking CoB or other temporary enhancements (maybe also including attachments). It might also be that it only looks at ranged/melee and not using weapon specific defenses (Polearm defenses in this case). Knowing how development works (and especially how SOE do their work lol) it might be that devs have forget about the new defenses
If it only looks at ranged defense that would be an increase of 70 ranged defense, that fit the increase in miss rate pretty well. However, keep in mind that just a few more misses would change the % misses so it is hard to make any sure conclusions (for example 5 more misses would have increased the miss rate to 9% when using Pikatta with no CoB). When using Pikatta with CoB the increase of defense is probably to low to make any significant impact (530 defenses instead of 460).
Also, Pikatta do not increase the defenses when it comes to damage mitigation (which it should not do so that part is workign as intended).
My pretty high accuracy might affect the results and making it count for less though.
I urge everyone interested in exploring this to make your own tests with other numbers for accuracy/defenses. Keep in mind though that not only 100 shots is enough, I saw much variation when doing series of 100 shots. I think that you actually would need at least 1000 shots until you get results which are statistically significant. But there is a limit how much I am willing to test hehe. I am thinking of trying some kind of automated testing (using a macro to shoot 500 shots).
Also keep in mind when testing, keep everything else the same. Only test one variable at a time. Don't change anything else or it might affect results. If you use a sac reduction powerup then you need to use the same type of powerup for all testing. Use the same special, weapon, distance to target and posture so there are no other thing affecting test results (even if you "know" it should not affect results).
Page 1 of 1