|
Post by Leonid Govorov on Oct 30, 2017 23:24:35 GMT
20 more trials. PL triggered 9 times... Probability (if something wrong with medal) is only 24%. So did it trigger more then supposed to?
|
|
|
Post by stoic on Oct 30, 2017 23:57:49 GMT
20 more trials. PL triggered 9 times... Probability (if something wrong with medal) is only 24%. So did it trigger more then supposed to? Well, yes, significantly... if the theory that there is a problem with medal is correct...
|
|
|
Post by stoic on Oct 31, 2017 0:05:42 GMT
Anyway I think that more tests are necessary...
|
|
|
Post by kanue on Oct 31, 2017 5:36:39 GMT
Before I tried testing the skill in an invasion, I had done it in Axis 3 Pacific campaign and noticed that pz leader rarely triggered (that's why I did with invasion instead). So now I test the trigger chance with this Axis 3 Pacific normal again. Here's what I did, - Put a general on the starting heavy tank. - move 2 grids up left. - attack the double stack light tank next to bunker. - restart the mission - repeat 15 times Here's the result - pz leader not trigger: 79 83 89 91 83 86 86 79 91 77 84 (avg = 84.36, N = 11) - pz leader trigger: 140 152 155 130 (avg = 144.25, N = 4) - Dmg diff = 59.89 or 70.98% boost. Trigger chance = 26.67% - Using binomial distribution, given pz leader trigger properly at 60%, the chance that it would trigger 4 out of 15 times is minuscule at 0.9347660%. Given that stoic 's info, I guess WC4's RNG is not so random. Maybe it's because of the seed number for RNG. If that so, it would explain why my experiment done on the same round of the same campaign produces the same low chance of triggering. These are all my guess tho.
|
|
|
Post by stoic on Oct 31, 2017 5:53:08 GMT
Before I tried testing the skill in an invasion, I had done it in Axis 3 Pacific campaign and noticed that pz leader rarely triggered (that's why I did with invasion instead). So now I test the trigger chance with this Axis 3 Pacific normal again. Here's what I did, - Put a general on the starting heavy tank. - move 2 grids up left. - attack the double stack light tank next to bunker. - restart the mission - repeat 15 times Here's the result - pz leader not trigger: 79 83 89 91 83 86 86 79 91 77 84 (avg = 84.36, N = 11) - pz leader trigger: 140 152 155 130 (avg = 144.25, N = 4) - Dmg diff = 59.89 or 70.98% boost. Trigger chance = 26.67% - Using binomial distribution, given pz leader trigger properly at 60%, the chance that it would trigger 4 out of 15 times is minuscule at 0.9347660%. Given that stoic 's info, I guess WC4's RNG is not so random. Maybe it's because of the seed number for RNG. If that so, it would explain why my experiment done on the same round of the same campaign produces the same low chance of triggering. These are all my guess tho. Did the same, PL triggered only 3 times out of 15... I moved my tank also close to bunker (using plain), result is 4 times out of 15 it is significantly worse than in Battle of Kursk.
|
|
|
Post by stoic on Oct 31, 2017 6:17:48 GMT
Just out of curiosity replayed 3 times 3 first rounds of battle of Kursk. PL triggerd 4 times (twice in attack, twice in defense), 5 times(2-3) and 4 times(1-3)respectively. It is early to say something deffinitely, but more tests for PL in defense are also of high importance ...
|
|
|
Post by Leonid Govorov on Oct 31, 2017 6:56:23 GMT
OK this is getting weird... First the damage inflicted was a wonder, now the percentage of triggering is.
|
|
|
Post by stoic on Oct 31, 2017 7:01:32 GMT
OK this is getting weird... First the damage inflicted was a wonder, now the percentage of triggering is. Well, it is always the same when you start to dig deeper
|
|
|
Post by Leonid Govorov on Oct 31, 2017 7:08:04 GMT
OK this is getting weird... First the damage inflicted was a wonder, now the percentage of triggering is. Well, it is always the same when you start to dig deeper I am not sure which will come first, either me dying of exhaustion or us discovering the secret lol
|
|
|
Post by stoic on Oct 31, 2017 14:01:42 GMT
|
|
|
Post by andrei on Oct 31, 2017 15:14:23 GMT
You mean base unit's attack is responsible for the PL increase effect? It is all about type of the unit?
|
|
|
Post by stoic on Oct 31, 2017 15:28:01 GMT
You mean base unit's attack is responsible for the PL increase effect? It is all about type of the unit? I have no other explanation... Have you ?
|
|
|
Post by andrei on Oct 31, 2017 15:46:03 GMT
You mean base unit's attack is responsible for the PL increase effect? It is all about type of the unit? I have no other explanation... Have you ? It doesn't explain why in my testing Manstein on double medium tank had 15-20% increase and in soonerjbd testing Manstein had 50-65% increase. The units and generals are the same There should be a very complicated formula with coefficients for different terms. I still think that the main factor is unit's attack.. BUT I also suppose there is also something about the defence of the enemy unit You are attacking. Otherwise I can't explain such a huge difference in different testing. The only thing we were not specifying for each testing comparison is unit we are attacking. Moreover some units (due to tech tree) have bonus damage to this or that enemy unit (like tank against infantry and infantry against tanks) which could also influence a lot.
|
|
|
Post by Max Otto von Stierlitz on Oct 31, 2017 15:49:34 GMT
Just checked with Nimitz (lv2 fleet leader), made 50 shots to level 1 ship. There were 14% critical shots (not 15% as written just matter of luck I think). Increase was 29.5%. Number looks quite good (I mean 30% looks ok, not like some 23 or 17). I think the problem may be in level defence (which was in wc3 and I think is in wc4 too). For example, your average base damage to first level unit is 100. Let's assume critical hit gives 30% extra damage, so it is 130 with critical hit. But if we attack unit, which lowers damage, for example, by 20, than if those 20 points are deducted after critical hit is applied, it is 80 base damage and 110 critical hit, which is 37.5% increase. Also there may be some trick with terrain. For example, terrain defence bonus may affect only base damage or critical hit may totally ignore terrain defence bonus. And these results look so good because I attacked first level unit and water has no terrain bonuses. I think there's something like this. Anyone knows the effect of unit's level?
I will do some more investigations (with Nimitz as besides him I have only Timoshenko) later and check the results on some other level units.
|
|
|
Post by stoic on Oct 31, 2017 16:43:32 GMT
Just checked with Nimitz (lv2 fleet leader), made 50 shots to level 1 ship. There were 14% critical shots (not 15% as written just matter of luck I think). Increase was 29.5%. Number looks quite good (I mean 30% looks ok, not like some 23 or 17). I think the problem may be in level defence (which was in wc3 and I think is in wc4 too). For example, your average base damage to first level unit is 100. Let's assume critical hit gives 30% extra damage, so it is 130 with critical hit. But if we attack unit, which lowers damage, for example, by 20, than if those 20 points are deducted after critical hit is applied, it is 80 base damage and 110 critical hit, which is 37.5% increase. Also there may be some trick with terrain. For example, terrain defence bonus may affect only base damage or critical hit may totally ignore terrain defence bonus. And these results look so good because I attacked first level unit and water has no terrain bonuses. I think there's something like this. Anyone knows the effect of unit's level? I will do some more investigations (with Nimitz as besides him I have only Timoshenko) later and check the results on some other level units. That's a difference between lvl 1 and lvl 2 triple stacked heavy tank. I don't have Military Academy yet, so I can't produce lvl 3 tanks...
|
|