seabasstion Posted May 21, 2015 Share Posted May 21, 2015 (edited) I had this battle happen to me earlier Chief Minister Kappa of Delta ordered a ground attack upon your nation of PowerPrittGlue. The attack was an immense triumph. Chief Minister Kappa's forces lost 10,016 soldiers and 322 tanks, while your defenders lost 728 soldiers and 813 tanks. You used 21.11 tons of munitions and 20.21 tons of gasoline defending the attack. The attack destroyed 29.41 infrastructure in the city of Ejiao. Chief Minister Kappa stole $138,130.41 in the attack and destroyed 0 improvements. prior to this attack i had 728 soldiers and about 2k tanks my opponent had around 150k soldiers and >5k tanks i find it very odd that 728 soldiers of mine would be able to kill 10x my soldier count facing insurmountable odds and a triumphant defeat. i think what is happening is you are calculating that: my nation SHOULD have lost around 20000 soldiers, however i only had 728 soldiers; so that means i lost 728 soldiers my opponent SHOULD have lost around 10000 soldiers, and lost 10,000 soldiers because they had over 10,000 soldiers What i propose is: for the nation getting beaten to zero soldiers - take the ratio of their starting soldier count with the calculated death amount and apply that to the calculated death amount to the winning nation. i get that there could be something like the Alamo where our soldiers valiantly defend their city and kill en extra amount, but 10x is rather high and im observing this in multiple places. i think justifying the kill counts with this logic is a case of it not being a 'bug' but rather a 'feature' instead of allowing an unintended bug to carry out a logical feature without control, it would only take a few lines of code to actually control this with a 1-5x Alamo Multiplier i believe you do this in php so i put something together quick to show you my thought process <?php $WinnerSoldierStartCount = 150000; //starting soldier count prior to attack of winning nation $WinnerSoldierCasualtyCalc = 10000; //your current calculated result for winner casualties incurred $LoserSoldierStartCount = 1000; //starting soldier count prior to attack of losing nation $LoserSoldierCasualtyCalc = 20000; //your current calculated result for loser casualties incurred $AlamoRNG = rand(100,500); //1x to 5x the amount of kills the loser will kill when faced with an alamo situation $AlamoPercent = $AlamoRNG / 100; //get it in a percent if ($LoserSoldierCasualtyCalc > $LoserSoldierStartCount){ //Loser will have 0 soldiers under current method $LoserSoldierEndCount = max(0,$LoserSoldierStartCount - ($LoserSoldierCasualtyCalc * ($WinnerSoldierStartCount / $WinnerSoldierCasualtyCalc))); $WinnerSoldierEndCount = max(0,round($WinnerSoldierStartCount - ($LoserSoldierStartCount / $LoserSoldierCasualtyCalc) * $WinnerSoldierCasualtyCalc * $AlamoPercent,0)); } else{ //Loser will have more than zero soldiers under current method $LoserSoldierEndCount = round(max(0,$LoserSoldierStartCount - $LoserSoldierCasualtyCalc),0); //your normal process $WinnerSoldierEndCount = round(max(0, $WinnerSoldierStartCount - $WinnerSoldierCasualtyCalc),0); //your normal process } ?> here is what it looks like with some outputs to better visualize echo "Winners Soldier Start Count = "; echo $WinnerSoldierStartCount; echo "<br>"; echo "Losers Soldier Start Count = "; echo $LoserSoldierStartCount; echo "<br>"; echo "<br>"; echo "Winners Calculated Deaths = "; echo $WinnerSoldierCasualtyCalc; echo "<br>"; echo "Losers Calculated Deaths = "; echo $LoserSoldierCasualtyCalc; echo "<br>"; echo "<br>"; echo "Losers Death Ratio = "; echo $LoserSoldierStartCount / $LoserSoldierCasualtyCalc; echo "<br>"; echo "Loser Death Ratio Applied to Winners Calculated Deaths = "; echo ($LoserSoldierStartCount / $LoserSoldierCasualtyCalc) * $WinnerSoldierCasualtyCalc; echo "<br>"; echo "<br>"; echo "Alamo Multiplier = "; echo $AlamoPercent; echo "<br>"; echo "Alamo Multiplier x Ratio Winners Calculated Deaths = "; echo ($LoserSoldierStartCount / $LoserSoldierCasualtyCalc) * $WinnerSoldierCasualtyCalc * $AlamoPercent; echo "<br>"; echo "<br>"; echo "Losers New Soldier Count = " . $LoserSoldierEndCount; echo "<br>"; echo "Winners New Soldier Count = " . $WinnerSoldierEndCount; Edited May 21, 2015 by seabasstion Quote Link to comment Share on other sites More sharing options...
Princess Bubblegum Posted May 21, 2015 Share Posted May 21, 2015 (edited) I'm pretty sure it's the 800 2000 tanks that led to killing 10k soldiers. You lost 33248 power. He lost 22896 power. The battle was approximately 80k power vs 462k power. So 80k killed 22896. He lost around 7% of his strength. You lost 41% of yours. Here is an incomplete formula for battles and casualties: http://politicsandwar.com/forums/index.php?/topic/4806-battle-mechanic/#entry73245 You may be right about excess damage being dealt not transferred over to the other available ground unit once a unit type runs out, though I can't tell from what's been posted. Edited May 21, 2015 by Princess Bubblegum Quote Link to comment Share on other sites More sharing options...
seabasstion Posted May 21, 2015 Author Share Posted May 21, 2015 (edited) the same logic would apply to a battle i had earlier though that cant be explained by your reasoning Sweet Potato seabasstion of PowerPrittGlue ordered a ground attack upon the nation of Mandarin led by General McLeod. The attack was an immense triumph. Sweet Potato seabasstion's forces lost 20,801 soldiers and 668 tanks, while General McLeod's defenders lost 4,681 soldiers and 798 tanks. The attack destroyed 12.38 infrastructure in the city of Isle of Lewisand 0 improvements. Sweet Potato seabasstion stole $379,110.00 in the attack. using your 40 points per tank and 1 point per soldier metric above: I would have lost 47521 power while getting a triumphant victory - whereas the losing combatant would have lost 36601 power the attack immediately preceding this attack was this: Sweet Potato seabasstion of PowerPrittGlue ordered a ground attack upon the nation of Mandarin led by General McLeod. The attack was an immense triumph. Sweet Potato seabasstion's forces lost 21,200 soldiers and 607 tanks, while General McLeod's defenders lost 27,954 soldiers and 782 tanks. The attack destroyed 11.87 infrastructure in the city of Jacksonville and 0 improvements. Sweet Potato seabasstion stole $334,620.00 in the attack. youll notice that i lost 45480 power whereas they lost 59234 power. the battle figures prior to both were very similar as they were rebuying soldiers and tanks. this can be observed in my power lost metric (within 5% of each other) this attack i came out 13754 power ahead in the first attack listed i came out 10920 power behind this is a net change of 24674 power points (soldiers) the soldier kill count difference between these two attacks for the losing nation is 23273 these battles have a consistent scenario in that the winning nation had a large advantage in soldier count (near or over 100k) and over double the tanks against the defeated nation. im not sure of the range of the RNG used in these attacks but with this info in mind i would say that my original suspicion holds true in that when a defeated nation goes to zero soldiers, it tries to subtract a large number of soldiers - in this example 25k+, but as the defeated nation doesnt have that soldier count it sets it to zero since negative soldiers dont make sense but still applies the full kill amount to the aggressive nation if the aggressing nation did enough damage to inflict 20,000 more casualties than what is possible and only realizes a fraction of that damage, it doesnt make sense that the defeated nation should inflict the full calculated damage, hence the suggestion for a ratio damage/kill when setting a opponent to zero. i suppose this could help in balancing the war mechanic so the defeated nation could come back but i would be interested in seeing simulated results the amount of casualties i would have incurred if i had only send 21,000 soldiers in (with the same number of tanks) and seeing if i would have still killed all their soldiers while still incurring 20801 soldier losses. im confident we wouldnt see this high of a kill count as i did something similar (to another nation) after i observed this phenomenon You ordered a ground attack upon the nation of United Canada led by TheJokeMan. The attack was a Pyrrhic victory. Your forces lost 4,681 soldiers and 43 tanks, while TheJokeMan's defenders lost 18,240 soldiers and 0 tanks. You used 41.00 tons of munitions and 40.00 tons of gasoline executing the attack. The attack destroyed 10.96 infrastructure in the city of Montreal. You stole $47,020.00 in the attack and destroyed 0 improvements 6401 power lost (winner) 18240 power lost (loser) back calculating the gas we know i used 4000 tanks. that leaves 1 munitions left over which equals 5000 soldiers. this was a phyrric victory (needless to say NOT triumphant) and i only lost 4681 soldiers. it doesnt make sense then that a TRIUMPHANT victory over another opponent with 4x less soldiers would kill 4x more soldiers (16 times more powerful). if i were to come in with 20,000 soldiers would i still only lose 4681 soldiers? probably not it would be expected to lose a few more since there are more targets but how many? 2x the amount? 3x? if i were to come in with 120,000 soldiers should i expect to lose 5x the amount of soldiers and still win the battle than what i would have otherwise lost? under the current model it shows that from a damage standpoint, it is actually better to send in fewer soldiers which will accomplish roughly the same damage to the defeating nation (meaning all their soldiers were killed). i know these arent perfect examples as one nation had tanks and the other didnt but it goes to show that a ratio kill (what i think is happening) isnt very logical is this 16x figure because of tanks? i guess but if we could somehow get apples to apples data on battles like this we would know for sure: battle 1) attacking soldiers: 100,000 attacking tanks: 5000 defending soldiers: 5000 defending tanks: 1000 battle 2) attacking soldiers: 15,000 attacking tanks: 5000 defending soldiers: 5000 defending tanks: 1000 if battle 1 is capable of delivering over 20,000 deaths to the attacking nation in a triumphant victory (which i think it would) than it would stand to reason that battle 2 should result in a defeat for the attacking nation. however, im highly confident both would be triumphant victories if we simulated them today, which does not make sense. the defending soldiers are either around to kill the attacking soldiers or they are not. unless there is a lot of friendly fire, ghosts, or reincarnation this high soldier kill count cannot be accurately resolved Edited May 21, 2015 by seabasstion Quote Link to comment Share on other sites More sharing options...
Phiney Posted May 21, 2015 Share Posted May 21, 2015 tanks are worth 27 soldiers that have munitions, incase that helps. Quote Link to comment Share on other sites More sharing options...
seabasstion Posted May 21, 2015 Author Share Posted May 21, 2015 it's more of a conceptual problem though phiney. to summarize my wordy response above does it make sense that: 100k Soldiers + 5k tanks against 5k soldiers + 2k tanks = triumphant victory with 21k deaths when 20k soldiers + 5k tanks against 5k soldiers + 2k tanks = triumphant victory with [some death figure i dont know] Quote Link to comment Share on other sites More sharing options...
Princess Bubblegum Posted May 21, 2015 Share Posted May 21, 2015 (edited) im not sure of the range of the RNG used in these attacks It's a range of 60%, or 40% to 100%: AttSolCasRand = MAX(RANDBETWEEN((Armed Soldiers * 1.75 + Unarmed Soldiers)*.4,(Armed Soldiers * 1.75 + Unarmed Soldiers)),1)DefSolCasRand = MAX(RANDBETWEEN((Armed Soldiers * 1.75 + Unarmed Soldiers)*.4,(Armed Soldiers * 1.75 + Unarmed Soldiers)),1) AttTankCasRand = MAX(RANDBETWEEN((Tanks * 40 * .4),(Tanks * 40)),1) DefTankCasRand = MAX(RANDBETWEEN((Tanks * 40 * .4),(Tanks * 40)),1) So casualties can swing wildly for both the attacker and the defender with varying ratios between the two from battle to battle. but with this info in mind i would say that my original suspicion holds true in that when a defeated nation goes to zero soldiers, it tries to subtract a large number of soldiers - in this example 25k+, but as the defeated nation doesnt have that soldier count it sets it to zero since negative soldiers dont make sense but still applies the full kill amount to the aggressive nation But once again, I still can't say for sure if this may be happening since we don't have access to the complete formula. In the scenario where all soldiers are killed, the excess kills should be transferred over to tanks in some amount. Is it though? I can't answer that. it's more of a conceptual problem though phiney. to summarize my wordy response above does it make sense that: 100k Soldiers + 5k tanks against 5k soldiers + 2k tanks = triumphant victory with 21k deaths when 20k soldiers + 5k tanks against 5k soldiers + 2k tanks = triumphant victory with [some death figure i dont know] Part of the problem may be that you seem to be approaching this under the assumption that you need soldiers to win the battle. You can initiate a battle with 5000 tanks and 50 soldiers and still come out the victor because soldiers and tanks are both sufficient on their own to win. Tanks are effectively their own soldiers. Only 50 actual soldiers are required to initiate a battle, not 1 soldier per tank. Edited May 21, 2015 by Princess Bubblegum Quote Link to comment Share on other sites More sharing options...
seabasstion Posted May 21, 2015 Author Share Posted May 21, 2015 (edited) there has to be more than this or i am not fully understanding the implementation the floor of the rand in AttSolCasRand for 100k soldier (assuming 100% armed soldiers) is = (Armed Soldiers * 1.75 + Unarmed Soldiers)*.4 =(100000 * 1.75 + 0)*.4 = 175000 * .4 = 70,000 the ceiling of the rand in AttSolCasRand for 100k soldiers is =(100000 * 1.75 + 0) = 175000 the rand will pick a number between that (70000 being the lowest possible) and then choose the max between that figure and 1, presumably when there are zero soldiers on hand there may be some problems with having a variable set to 0 is this formula for a soldier strength in battle and not actually the casualty number? the nomenclature suggests casualty but there has to be another component for how this is used (or it is the wrong function) as this equation calculates to a minimum of 70,000 deaths when attacking with 100,000 soldiers and we know this isnt observed Edited May 21, 2015 by seabasstion Quote Link to comment Share on other sites More sharing options...
seabasstion Posted May 21, 2015 Author Share Posted May 21, 2015 (edited) But once again, I still can't say for sure if this may be happening since we don't have access to the complete formula. In the scenario where all soldiers are killed, the excess kills should be transferred over to tanks in some amount. Is it though? I can't answer that. it isnt though. case in point : both of these attacks i initiated which were strikingly similar in setup aside from a low number of defending soldiers by the enemy. attacking tanks/soldiers i believe were the same (max i think) and defending tanks were close to the original attack (maybe a few less) and soldiers as i said notably lower. Sweet Potato seabasstion of PowerPrittGlue ordered a ground attack upon the nation of Mandarin led by General McLeod. The attack was an immense triumph. Sweet Potato seabasstion's forces lost 21,200 soldiers and 607 tanks, while General McLeod's defenders lost 27,954 soldiers and 782 tanks. The attack destroyed 11.87 infrastructure in the city of Jacksonville and 0 improvements. Sweet Potato seabasstion stole $334,620.00 in the attack. Sweet Potato seabasstion of PowerPrittGlue ordered a ground attack upon the nation of Mandarin led by General McLeod. The attack was an immense triumph. Sweet Potato seabasstion's forces lost 20,801 soldiers and 668 tanks, while General McLeod's defenders lost 4,681 soldiers and 798 tanks. The attack destroyed 12.38 infrastructure in the city of Isle of Lewisand 0 improvements. Sweet Potato seabasstion stole $379,110.00 in the attack. if it were actually the case and it were transferred, roughly 20,000 additional soldiers should have been killed if it mimicked the first attack. since there wasnt 20000 additional soldiers this equates to 20,000 additional power. lets give the current model the best case scenario and take 40% of the 20,000 figure meaning the first attack happened to be at 1 and the second at .4 and would be the lowest amount of 'power' that could be transferred this is 8000 power. 8000 power / 40 = 200 tanks bare minimum, 200 more tanks should have been destroyed in the second attack compared to the first. however there were only 16. this represents 640 power that was 'recouped' compared to the first which can be interpreted as 3.2% (well outside the 40% lower bound). this assumes the tank factor was the same. lets say the tank factor in the first example was 1 and the second attack was .4. that would still mean that it would be expected a bare minimum loss of 80 tanks. this figure of 16 is still over 4x lower than that. i dont remember his exact tank difference from attack 1 to attack 2 but im positive it wasnt 4x lower i think it all comes back to this: I would have lost 47521 power while getting a triumphant victory - whereas the losing combatant would have lost 36601 power that in itself seems undeniably wrong Edited May 21, 2015 by seabasstion Quote Link to comment Share on other sites More sharing options...
Princess Bubblegum Posted May 21, 2015 Share Posted May 21, 2015 (edited) is this formula for a soldier strength in battle and not actually the casualty number? the nomenclature suggests casualty but there has to be another component for how this is used (or it is the wrong function) as this equation calculates to a minimum of 70,000 deaths when attacking with 100,000 soldiers and we know this isnt observed Yes, from my understanding those numbers are then divided by constants to determine casualties. bare minimum, 200 more tanks should have been destroyed in the second attack compared to the first. however there were only 16. this represents 640 power that was 'recouped' compared to the first which can be interpreted as 3.2% (well outside the 40% lower bound). And this might be getting to something I've actually suspected in the past, that casualties may not be determined by combining the power of tanks and soldiers, but by applying the full force of the RNG power twice, once to tanks and once to soldiers. If that is the case, as I've suspected before, then that is a very large detriment to tanks (or more specifically a detriment to using both tanks and soldiers), since every time both tanks and soldiers are used, the player is facing potentially twice the amount of power loss just for the privilege of using tanks. And indeed, one of the consequences of such a formula would be massive casualties for the person going up against a small number of soldiers and a large number of tanks, since tanks do large amounts of damage and take relatively fewer amounts of casualties. The tank power would be applied twice, once to enemy soldiers and once to enemy tanks. The obverse would not be true. Both enemy soldiers and enemy tanks would each do damage only once to tanks instead of twice to tanks when there are 0 soldiers. The enemy would have double the casualty ratio comparatively. Edited May 21, 2015 by Princess Bubblegum Quote Link to comment Share on other sites More sharing options...
seabasstion Posted May 21, 2015 Author Share Posted May 21, 2015 (edited) thats a really clever observation if that is true. i can understand the rationale a bit though if the intended purpose was one day ground soldiers perform better and tank squad happens to flaunder, both perform poorly, tanks well soldiers poorly etc. looking at the three above attacks: attack 1: 10016 soldiers ; 322 tanks (defender kill set to 0) attack 2: 21200 soldiers ; 607 tanks (defender kill NOT set to 0) attack 3: 20801 soldiers ; 668 tanks (defender kill set to 0) assigning 40 soldiers per tank we get damage ratios of attack 1: 10016 : 12880 = 0.7777 attack 2: 21200 : 24280 = 0.8731 attack 3: 20801 : 26720 = 0.7785 assigning 27 soldiers per tank (since they are armed) we get damage ratios of attack 1: 10016 : 8694 = 1.152 attack 2: 21200 : 16389 = 1.2936 attack 3: 20801 : 18036 = 1.1533 i know the sample size is very low but i find it interesting that both the defender to 0 situations have a ratio very similar to each other (probably an observed rounding error if i had to take a guess). ill look for more data points of people going to zero soldiers in the morning. ill even buy 5000 or so for myself and hopefully ill get another ground attack on my nation to see if i get this 0.77n ratio again. Edited May 21, 2015 by seabasstion Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.