You are not logged in. Please login or register.
I understand that the current system may work for large number of units, but for a low number of units the regiments rely solely on luck to win or lose.
I would like to see a system that works equally well regardless of how many troops are sent.
Agreed by me.
Agreed!!
It's kinda weird to win a battle with 6 units left when the victim ran away with more then 30 units 
With small regiment size, the system behaves like it is biased towards the extremes. I am getting really lucky, or really unlucky... there does not appear to be much happening in the middle.
Yeah I had two in a row where my two cheap units fled the first round, and my witches..... well they fought 3v1 to the death.
THAT becomes expensive real quick.
Maybe small regiments get a bravery bonus? To push the odds in either direction.
No.
How does this RNG work? Can someone post the link to the documentation for the one in use?
random number generating? I know for a fact that is ridiculous hard stuff.
well, the RNG does work as intended last time I saw the test. It is human nature to remember the extremes, but with such a low number the extremes become the norm.
I'm specifically wondering what types of variance reduction is in play.
Outcome: Repelled
Attacking Units
Unit Amount died
Wolf Riders 55 12
Defending Units
Unit Amount died
Temple Sentinels 10 10
Dark Warriors 10 9
Wolf Riders 15 13
not funny
Outcome: Repelled
Attacking Units
Unit Amount died
Dark Warriors 70 15
Wolf Riders 35 4
Defending Units
Unit Amount died
Dark Warriors 16 6
Wolf Riders 25 10
I just want to express how utterly speechless I am at how poorly this is working. In the last three battles I have had, every one, I REPEAT, every one of my units but one fled asap. theirs did not. statistically, this is very, very unlikely. I mean 3x of killing over half his units? And its great that 1 unit is left that really never flees.... that one just gets decimated.
and now, I have killed well over 3/4 of his troops, in 3 battles while he was offline, and everyone of them has been a loss. Its useless to battle in this war game with such low numbers. It is only, and I really want to stress this, only luck that determines who wins.
Last edited by DoomPope (2010-06-01 06:46:34)
Outcome: Repelled
Attacking Units
Unit Amount died
Wolf Riders 29 3
Defending Units
Unit Amount died
Dark Warriors 2 0
Executioners 10 0
Wolf Riders 2 2
Temple Sentinels 10 0
Attacking special units
Sorcerer lvl 1 died
The ground trembles as both armies march towards each other. The peasants of Demo's run away screaming, while the armies of parados's village invade their lands.
The mages prepare for combat.
An attacking Wizard failed to cast a Fireball (90% success probability).
Suddenly, your captain notices that the enemies archers are getting into place. Quickly, he orders your archers to take aim and fire!
10 Temple Sentinels don't wait and kill 0 units of wolf (29 Wolf Riders).
Your captain orders your troops to rush towards the enemy troops.
wolf (29 Wolf Riders) kill 2 units from 2 Wolf Riders in a melee.
10 Executioners kill 2 units from wolf (28 Wolf Riders) in a melee.
2 Wolf Riders kill 0 units from wolf (29 Wolf Riders) in a melee.
10 Temple Sentinels kill 1 units from wolf (29 Wolf Riders) in a melee.
wolf (29 Wolf Riders) run away crying. I told you all of this would end in tears.
The bodies of 2 Wolf Riders lie shattered, in a pool of blood, on the battlefield.
A river of blood flows from the battlefield. While the sun sets, your captain looks over the battlefield, full of wounded and dead soldiers. The fight is over... but is the battle?
A level 1 Wizard died.
OK, someone correct me if I'm wrong, but they system currently draws a RNG every round. What if we temporarily go to drawing an RNG for each side but only for every battle and using that same pair of RNG throughout said battle. So that regiments on one side will withdraw at 65% or 80% or 5% while the other side withdraws at say 70% or whatever. Or... we draw a single RNG and both sides use that.
Personally, I don't mind the extremely random nature of battles (I've argued for random march times for about 2 years) but I don't think that the current random nature is "by design."
OK, someone correct me if I'm wrong, but they system currently draws a RNG every round. What if we temporarily go to drawing an RNG for each side but only for every battle and using that same pair of RNG throughout said battle. So that regiments on one side will withdraw at 65% or 80% or 5% while the other side withdraws at say 70% or whatever. Or... we draw a single RNG and both sides use that.
Personally, I don't mind the extremely random nature of battles (I've argued for random march times for about 2 years) but I don't think that the current random nature is "by design."
I dont mind random either, but a coinflip is removing most of the strategy. In my last 6 battles, 2/3 of my regiments fled as soon as they crossed the 10% line and the other side did not. ONE battle, my troops did not flee the first round, 93% victory. Its all chance.
I would rather see more variables each with smaller amount of chance. It would help give the illusion of true random, while still not making it a coinflip.
Hellikin wrote:OK, someone correct me if I'm wrong, but they system currently draws a RNG every round. What if we temporarily go to drawing an RNG for each side but only for every battle and using that same pair of RNG throughout said battle. So that regiments on one side will withdraw at 65% or 80% or 5% while the other side withdraws at say 70% or whatever. Or... we draw a single RNG and both sides use that.
Personally, I don't mind the extremely random nature of battles (I've argued for random march times for about 2 years) but I don't think that the current random nature is "by design."
I dont mind random either, but a coinflip is removing most of the strategy. In my last 6 battles, 2/3 of my regiments fled as soon as they crossed the 10% line and the other side did not. ONE battle, my troops did not flee the first round, 93% victory. Its all chance.
I would rather see more variables each with smaller amount of chance. It would help give the illusion of true random, while still not making it a coinflip.
What if running away was based on what the troops have to fight and not how many allies are left?
What if running away was based on what the troops have to fight and not how many allies are left?
What I was thinking is having more variables.
ie
The base chance to flee is the % of troops lost. If a unit loses 10% of its troops, it has a 10% chance of fleeing.
If a unit is being flanked (attacked by 2 units) it has an additional 5% chance to flee.
If it is being flanked on both sides (attacked by 3 units) it has an additional 15% chance to flee.
If a unit has killed more troops than it has lost in a given round, it has a 5% lower chance to flee.
ect, ect.
Voldtekt wrote:What if running away was based on what the troops have to fight and not how many allies are left?
What I was thinking is having more variables.
ie
The base chance to flee is the % of troops lost. If a unit loses 10% of its troops, it has a 10% chance of fleeing.
If a unit is being flanked (attacked by 2 units) it has an additional 5% chance to flee.
If it is being flanked on both sides (attacked by 3 units) it has an additional 15% chance to flee.
If a unit has killed more troops than it has lost in a given round, it has a 5% lower chance to flee.
ect, ect.
But why would we raise the chances to flee when they're already too high?
But why would we raise the chances to flee when they're already too high?
It actually much lower. With what I posted in order to have the same base 50% we have now, a unit would need to lose half their troops.
Its also just meant as a framework. Dividing the base in half losing 10% of a unit would give it a base flee rate of 5%. At that rate, a 50% flee rate would be achieved when everyone in the unit was killed.
a regiment that loses more troops per round, has higher probability of fleeing after 10%. regiment losing 1 troop per round really shouldnt be afraid....
a regiment that loses more troops per round, has higher probability of fleeing after 10%. regiment losing 1 troop per round really shouldnt be afraid....
Exactly.
The system would be much more robust if the actual RNG was downplayed but many variables acted together.
Uhm, I am added to this list. I murdered this guy, yet was repelled....
Date: 01.06.2010 23:50
Outcome: Repelled
Attacking Units
Unit Amount died
Gu'al Cavalry 64 15
Defending Units
Unit Amount died
Horsemen 108 47
6 Knights saw a spider and jumped to location 2.
10 Citadel Guards moved from location 5 to 4.
Green Munchers (19 Gu'al Cavalry) kill 2 units from 6 Knights in a melee.
6 Knights kill 0 units from Green Munchers (19 Gu'al Cavalry) in a melee.
10 Citadel Guards kill 1 units from Green Munchers (19 Gu'al Cavalry) in a melee.
6 Knights run away crying. I told you all of this would end in tears.
10 Citadel Guards ran towards location 3.
Green Munchers (18 Gu'al Cavalry) kill 2 units from 10 Citadel Guards in a melee.
10 Citadel Guards kill 1 units from Green Munchers (17 Gu'al Cavalry) in a melee.
Green Munchers (17 Gu'al Cavalry) run away crying. I told you all of this would end in tears.
A river of blood flows from the battlefield. While the sun sets, your captain looks over the battlefield, full of wounded and dead soldiers. The fight is over... but is the battle?
A level 3 Wizard died.
I just watched a defending unit lose 5-10% of its units 4 turns in a row, against 3 of my units, and win.
Fortunately I didn't lose more than 15% of any one regiment tho I suppose.
Powered by PunBB, supported by Informer Technologies, Inc.
Currently installed 10 official extensions. Copyright © 2003–2009 PunBB.