Home  | Login  | Register  | Help  | Play 

Luck Calculations

 
Logged in as: Guest
  Printable Version
All Forums >> [Artix Entertainment Games] >> [EpicDuel] >> EpicDuel Balance >> Luck Calculations
Forum Login
Message << Older Topic   Newer Topic >>
2/7/2014 5:03:06   
Goony
Constructive!


What if they used percentages instead of comparing values?

Here is some code based on block chance, the advantage is that it would eliminate minimum and maximum chance ;)


AS3 code:

var attackerDex:Number = 0; // attackers dexterity
var defenderDex:Number = 0; // defenders dexterity
var attLuckyStrike:Number = 0;// Can be 0,2 or 7
var defNinjaReflexes:Number = 0; // Can be 0 or 4
var block:Boolean = ; //Can be true or false

function block():void
{
	totalDex = attackerDex + defenderDex;
	defenderPercent = defenderDex/totalDex; //This will give a number between 0 and 1
	attackerPercent = attackerDex/totalDex; //This will give a number between 0 and 1
	blockChance =((defenderPercent + (defNinjaReflexes)) / (attackerPercent + (attLuckyStrike/100)}) * 8; 
//8 is the modifier, so if there are no cores and dexterity was equal, then: 0.5/0.5 = 1*8 thus block chance is 8%
	blockRandom = Math.Random()*100
		if blockChance < blockRandom
		{
			block = true
		}
		else
		{
			block = false
		}
}


Any mathematicians here that may be able to prove that this would be worse than the current formula :p

Formula: 

Block Chance = 8% + (Defender's Dexterity - Attacker's Dexterity) / 2.5


< Message edited by Goony -- 2/7/2014 7:36:21 >
Epic  Post #: 1
2/7/2014 6:25:04   
DarkDevil
Member

doesn't the computer do multiply before addition.

i'd add another variable to add ninja reflex or luckystrike before its rolled.

wait a minute something doesn't seem correct.

the forumla is wrong , att or def percent are between 0 and 1 while reflex is at 4 so it breaks any formula.

best thing is to add or subtract ninja reflex or LS from the random roll instead.

also for block side , the block will just be dexterity ratio*8%

< Message edited by DarkDevil -- 2/7/2014 6:51:24 >
AQ Epic  Post #: 2
2/7/2014 6:46:19   
Goony
Constructive!


From my understanding of the current code, the ninja reflex and lucky strike core % is added after the % chance is calculated. But in my calculation it is added during the chance % calculation and has less impact on the actual chance %.

Also your post reminded me that I need to adjust the cores from a percentage to a value between 0 and 1 ;)
(i.e. 4/100 or 7/100)


The roll is in both instances after the chance has been determined!

Edit, thanks Dark ;) Made some adjustments and added some brackets... Oops!

< Message edited by Goony -- 2/7/2014 6:54:22 >
Epic  Post #: 3
2/7/2014 6:54:57   
DarkDevil
Member

another point poped up , reflex or LS will be multiplied by the 8% rather than an absolute value , in other words it just adds 0.04 to the ratio of the defender but it doesn't add 4% block.

i think making reflex/LS added at end is better.

also it still only translates to dex ratio*8% for block , meaning if you had double his dex you only get 16% block.

< Message edited by DarkDevil -- 2/7/2014 7:00:27 >
AQ Epic  Post #: 4
2/7/2014 7:09:50   
Goony
Constructive!


quote:

also it still only translates to dex ratio*8% for block , meaning if you had double his dex you only get 16% block.


That's exactly the point, there is way too much luck in the battles...

The thing is that while someone with 80 dex vs 40 dex will have a 16% chance to block while the opponent will have 4%.

Whereas with the old formula the person with 80 dex would have 24% and the opponent would have the minimum 2%...

With the old formula it only took a 15 dex differential for the opponent to be reduced to the minimum chance, regardless of if they had 100 dex versus 115 dex... Too me that is wrong!
Epic  Post #: 5
2/7/2014 7:22:16   
DarkDevil
Member

its dex/total dex

and attacker value/defender.

so i have rolled this a little:

someone with 30 dex against someone with 60 dex has 4% of being blocked

and someone with 60 dex against someone with 30 dex has 16% of being blocked.

i think there is something wrong.

< Message edited by DarkDevil -- 2/7/2014 7:34:27 >
AQ Epic  Post #: 6
2/7/2014 7:32:58   
Goony
Constructive!


hmm, yeah it's probably broken :)

I wrote it as code since I was trying to work through the method. The dex% part is correct but I reversed the chance part to defender value/attacker value... lol

Edit: the figures you ran are correct, but try it with something like 120 dex versus 30 dex then the block chance would be 32% versus 2% block chance. The old method would have resulted in 35% versus 2%... So at the extremes it is closer to the old!

But when both have high dex like 100 versus 80 the it would be 11% versus 7%. While old method would be 16% versus 2%...

Make your own mind up about that ;)





< Message edited by Goony -- 2/7/2014 7:48:49 >
Epic  Post #: 7
2/7/2014 8:35:35   
King FrostLich
Member

Do a pseudo-random distribution coding and decrease or adjust the max chance for every luck factor. In this way, you get a near actual chance of the % the luck factor does than a fixed % to happen.
Epic  Post #: 8
2/7/2014 11:06:23   
  Exploding Penguin
Moderator


I would actually like this concept more than what it is right now. I was thinking of making crit, deflection, and block chances increase at an exponential rate to punish no-support builds and support more balanced stats. However this might overpower casters in a way that they reach the block or deflection cap almost always even against decently tanky opponents.

But %s are, IMO, the way to go.
Epic  Post #: 9
2/8/2014 1:12:04   
edwardvulture
Member

So a 20 Dex advantage would mean 0% chance of getting blocked?
AQ DF MQ  Post #: 10
2/8/2014 6:58:26   
Goony
Constructive!


Theoretically, there is a way that having a 20 dex advantage which would give you 0% chance of being blocked:

If you had 20 dex and your opponent had 0... But, it is your percentage of the total dexterity that determines your block chance.


So, if you both have 40 dexterity the block chance for both players would be 8%

If you had double his dexterity, say 80 versus 40 you would have double the original block chance 8x2=16% block chance and he would have half of the original 8/2=4% block chance.

There is a maximum and minimum chance, 100% max and 0% min... Although, it is not based on stat difference but the ratio of total dexterity between the 2 players...

< Message edited by Goony -- 2/8/2014 7:06:25 >
Epic  Post #: 11
2/8/2014 11:39:59   
edwardvulture
Member

ah... so you're basing it on ratios instead of stat differences.
AQ DF MQ  Post #: 12
2/8/2014 12:40:23   
toopygoo
Member

i like this better than what we have now. Investment in a stat should give full benefits in its main sub-numerical ability. if you hate guns, why shouldnt you get to
deflect all the time? :P


Approved :D

_____________________________

AQW Epic  Post #: 13
2/8/2014 12:59:25   
  Exploding Penguin
Moderator


I can't say I approve of the exact formula, but I can actually provide an example of how this would probably work better:

So many people running around with 17 support builds. Let's say I have 51 as a focus build. That's 34 more support, which averages about 9% crit chance (actually is high 8%, but I don't even remember my own formulas these days). I find a 9% crit chance very wimpy and not at all fair given that I have triple the opponent's support! Shouldn't that mean I get a much higher crit chance than sub-10%? Some bias involved, but that's kind of the idea I get from the current RNG system.
Epic  Post #: 14
2/8/2014 13:23:18   
toopygoo
Member

but even if thats correct, crits are debunked by blocks and deflects, and thus can be countered. It is basically a way of making high HP/strength builds less lucky overall.

_____________________________

AQW Epic  Post #: 15
2/8/2014 13:33:43   
Mother1
Member

@ toopygoo

which is also an issue. Critical have always been outshined by deflections when they are put together. Even more so since they consistently been doing so. Blocks and deflections also need a rework outside of calculations (Meaning how much they take away from actual damage as well) because as it is now, anyone not using that build against high dex/tech builds are hard pressed dealing with them especially with a crit deflection.

EXP say you get one of these and the with the crit you alone you would have gotten 400. however add the deflection to this and it goes down to 200. The crit only adds 30-50 more damage to any attack thanks to the changes to crits while a deflection takes away 50% of the over all damage. In this situation the deflection takes away far more damage than a critical could ever give.

Another issue I have with this is that if we change it to be this away any debuff build will get a massive buff. Meaning someone uses smoke and is able to take away enough dex to make the % go down to 0 (meaning no chance of blocking) that right there spells abuse. The same can be said of malf for tech in deflections. for this to happen debuffs will need a serious rework as well so this can't happen otherwise it will just spawn a new breed of builds that's soul purpose is to abuse this.
Epic  Post #: 16
2/8/2014 13:52:06   
toopygoo
Member

@ mother1
which i forgot to ask: would it be possible to make these debuff skills, purely shield based instead of stats? i think that altering them gives strength builds another advantage, and thats the las thing we need. Anyone can recover a shield, but stats, only certain classes can compensate when they are debuffed. i would rework the whole thing into affecting just damage taken from debuffs.
AQW Epic  Post #: 17
2/8/2014 20:38:45   
Goony
Constructive!


Hmm, will try to look at a better solution using ratios, while they will basically work for block the method may not work as well for other calculations like critical s since the modifier will not give enough variation...

I think that this discussion highlighted some issues that I had overlooked. The main one being you cannot divide by 0... So, the calculation would fail :/

Edit: What about a combination of both :P

8 + (( 0.0001 + defDexRatio ) / ( 0.0001 + attDexRatio )) * (0.0001 + (( defDex - attDex) / 5 )) + ( defNinjaCore - attLuckCore ))

< Message edited by Goony -- 2/8/2014 21:47:07 >
Epic  Post #: 18
2/9/2014 8:28:14   
Jacob [Back to ED]
Member

That's what I really like to happen, thanks for reminding us my friend!
Epic  Post #: 19
2/9/2014 8:40:17   
DarkDevil
Member

it still breaks when opponnent has 0 dex either.
also to add , ninja reflex and lucky strike gets squashed in between , so i could suggest moving it to the begining of the formula added or subtracted from the 8.

< Message edited by DarkDevil -- 2/9/2014 8:47:36 >
AQ Epic  Post #: 20
2/10/2014 6:25:04   
dfo99
Member
 

in theory, the max block chance possible is 39%, but probably is more than it
Post #: 21
Page:   [1]
All Forums >> [Artix Entertainment Games] >> [EpicDuel] >> EpicDuel Balance >> Luck Calculations
Jump to:






Icon Legend
New Messages No New Messages
Hot Topic w/ New Messages Hot Topic w/o New Messages
Locked w/ New Messages Locked w/o New Messages
 Post New Thread
 Reply to Message
 Post New Poll
 Submit Vote
 Delete My Own Post
 Delete My Own Thread
 Rate Posts




Forum Content Copyright © 2018 Artix Entertainment, LLC.

"AdventureQuest", "DragonFable", "MechQuest", "EpicDuel", "BattleOn.com", "AdventureQuest Worlds", "Artix Entertainment"
and all game character names are either trademarks or registered trademarks of Artix Entertainment, LLC. All rights are reserved.
PRIVACY POLICY


Forum Software © ASPPlayground.NET Advanced Edition