You are not logged in.

Vedauwoo

Unregistered

41

Monday, January 26th 2015, 6:11pm

Legolas was for many battles and holy battles hero, double hero and triple hero - and this is not something supported by the Ratio Formula.
And why is that not supported by the ratio?

42

Tuesday, January 27th 2015, 2:10am

Because the ratio is not high - he said it in SB one day he calculated his ratio.
Just to give you an update: Fenrir - the strongest account in this server sold his account because - due to the battle script - he thought he was `raped` by his right of being hero in battles (for different things, wins or hitpoints). Right now the account is King Arthur. Mind you, he is 2nd in the world as level too. The turning point was a particular battle, when a player, 10 levels below him had like 150 wins and he had 40. And they fought on the same side. This is just one hw story about how the script works.

Vedauwoo

Unregistered

43

Tuesday, January 27th 2015, 7:04pm

If you want to know the script, then you still have to do the work.

Holy-War Combat Spading Project

And it is work. It means gathering the data reported in each fight. Falco and I worked on it for some time before he left the game. Our focus was how the script worked, what was the way it operated. It wasn't a secret and GAS changed things in the middle of the project. However, I don't believe the script changed that much.

Here is the way I believe it works.

The fight begins when you hit attack
  • A check is made to see if you can attack the player.
  • If he is safe, or too weak or some other reason for an attack, then no attack occurs and you are kicked out of the attack process.
  • If he can be attacked then you go to the fight script subroutine
The fight begins
  • Round 1
  • A random number between 6 and 10 is generated, this is to determine how many "moves" are to be made. Example below has 10 "moves". All but the last have no consequence.
XXXXX manages to counter the attack from Ved.
Ved runs up to XXXXX at full speed, stumbles and falls.
XXXXX successfully avoids the attack from Ved by rolling away sideways.
XXXXX charges at Ved, elegantly jumps over a bush, stumbles over a tree branch and clumsily falls to the ground.
Charging at full speed, Ved attacks his opponent, who is nonetheless able to defend himself.
XXXXX tries to come at Ved from behind, but his opponent notices him and evades the attack at the last moment.
The brave XXXXX manages to resist his opponent's ferocious attack.
Ved tries to come at XXXXX from behind, but his opponent notices him and evades the attack at the last moment.
Ved successfully avoids the attack from XXXXX by rolling away sideways.
Ved strikes at Proleter like crazy. XXXXX takes 19 HP of damage.
  • Each move a random number is created to determine if there is a hit. The actual math used by GAS is unknown, but basically it seems that the probability is in favor of the stronger player. So when you said that the ratio would not work in a 1 on 1 fight between a 2.6 ratio player and a 2.8 ratio player because it is a flip of a coin, I think you are wrong. I think it is the flip of an unfair coin, or that the 2.8 ratio player will win 60 and the 2.6 will win 40 in a hundred fights. It is statistical and based on random processes, and the stronger character wins more often than the weaker character (stronger in stats, and stronger in health), but you will see some random results, just like if the characters who are exactly the same, sometime one will win 5 times in a row. Just like flipping heads 5 times in a row, it happens.
  • If there is a hit, then the value of the hit is determined with a random number, but is based on the stats of the character. This is what Falco and I were seeking was the way the stats affected the amount of damage done. The damage is then placed into the text of a hit in the "move". This is the number 19 in line 10 of round 1 above.
  • After each move, the damage is added to each player and subtracted from their LP. If LP drops below a minimum, then the fight ends. If Damage is greater than a certain amount, then the fight ends. if the fight does not end, then another move is made. when the number of moves have been taken for the round, then the next round starts and the process repeats.
I believe this is the general layout of the fight script. The math and the amount of random variation are the unknowns. But it can be determined. It would take time. The more the variation (random number span), the more fights would have to be collected. We saw that there was clear differences in the way the script operates for HP, STR, ATT, DEF, AGI and STA. Certainly health is a major factor in the fight as well. The XP of the player is not quite as important. The higher the XP tends to have higher HP.
Tourneys are great for testing the script because Health is not so important. Everyone is put at full health at the beginning and the level variation is minimized by the brackets.
Battles are another matter. If you enter at half health, then you might have be able to have a lot of hit points but not so many fights. The fights will occur in a pattern set by the battle script, that only applies on the first cycle of fighting as after that, players are removed as they are knocked out. you can think of the battle order as 2 gears each with the number of teeth equal to the number of players. Suppose 5 players versus a hundred players, the 5 players would have to cycle 20 times before the pattern repeats, all the time a players are getting knocked out.

The battle fatigue script is another factor and quite frankly, it is stupid. GAS should have never done it. However, it can all be re-engineered if you gather enough data and do the math.
I just don't see it as worth the time.

44

Wednesday, January 28th 2015, 1:51am

The problem with battles - at least from what I checked and had the chance to check same level chars in same battle on same side reports - you have big chances to not have the same opponents all the report. That means, the opposing side is split it in `columns`and depending where the `cut`is done, you might get tough guys or not. When they started to change the script, from what I`ve seen, was column vs column and the next column kick first the survivals from the first one, before engaging the new column. Now it is changed again. You say 60-40, I say flip a coin because you will not know on what side of the coin you fall when you bet.
And the fatigue has something to do with the number of fights won in a sequence - after that you start losing when you should not. Has nothing to do with stats, but with the random linked to a number of wins in row that is related to the number of participants. Maybe it doesn`t make sense, but this is what I could gather about the battles.

45

Wednesday, January 28th 2015, 2:07am

Add the life percentage too. If your life is lower (for a normal char) you have chances to win against stronger and healthier chars with you getting 0 damages. German forum was full of reports of super tixxers getting kicked 12k to 0 by almost dead chars. They play with this `fatigue`sometimes during one month. And fatigue was asked by normal players who stood no chances in battles against merc tixxers. One thing was when tixxing was ending at lvl 50 - you could buy one stat with 30EU, and other thing is now.

46

Wednesday, January 28th 2015, 3:44am


  • Each move a random number is created to determine if there is a hit. The actual math used by GAS is unknown, but basically it seems that the probability is in favor of the stronger player. So when you said that the ratio would not work in a 1 on 1 fight between a 2.6 ratio player and a 2.8 ratio player because it is a flip of a coin, I think you are wrong. I think it is the flip of an unfair coin, or that the 2.8 ratio player will win 60 and the 2.6 will win 40 in a hundred fights. It is statistical and based on random processes, and the stronger character wins more often than the weaker character (stronger in stats, and stronger in health), but you will see some random results, just like if the characters who are exactly the same, sometime one will win 5 times in a row. Just like flipping heads 5 times in a row, it happens.
  • If there is a hit, then the value of the hit is determined with a random number, but is based on the stats of the character. This is what Falco and I were seeking was the way the stats affected the amount of damage done. The damage is then placed into the text of a hit in the "move". This is the number 19 in line 10 of round 1 above.
  • After each move, the damage is added to each player and subtracted from their LP. If LP drops below a minimum, then the fight ends. If Damage is greater than a certain amount, then the fight ends. if the fight does not end, then another move is made. when the number of moves have been taken for the round, then the next round starts and the process repeats.
I believe this is the general layout of the fight script. The math and the amount of random variation are the unknowns. But it can be determined. It would take time. The more the variation (random number span), the more fights would have to be collected. We saw that there was clear differences in the way the script operates for HP, STR, ATT, DEF, AGI and STA. Certainly health is a major factor in the fight as well. The XP of the player is not quite as important. The higher the XP tends to have higher HP.


That line bold there, is not what is happening in this :

X charges at Zed, elegantly jumps over a bush, stumbles over a tree branch and clumsily falls to the ground.

The attacking X aims and hits Zed. His opponent is hit 968 HP and screams in pain.

X manages to resist his opponent's ferocious attack.

X strikes at Zed like crazy. Zed takes 808 HP of damage.

Zed strikes out, but misses X.

X strikes out, but misses Zed.

The attacking X aims and hits Zed. His opponent is hit 955 HP and screams in pain.

The brave Zed manages to resist his opponent's ferocious attack.

X falls for a feint and cannot dodge the blow aimed at him by Zed (1953 HP).


Round 2

The brave X manages to resist his opponent's ferocious attack.

X aims and hits Zed. His opponent is hit 855 HP and screams in pain.

X starts an attack against Zed and hits him. His victim screams in pain (919 HP).

The brave Zed manages to resist his opponent's ferocious attack.

X runs up to Zed at full speed, stumbles and falls.

Zed tries to come at X from behind, but his opponent notices him and evades the attack at the last moment.

X is severely hit by Zed. (Damage: 1653 Hitpoints)


Round 3

Zed attacks X with full force but misses his shot.

X strikes out, but misses Zed.

Charging at full speed, X attacks his opponent, who is nonetheless able to defend himself.

X attacks Zed with full force but misses his shot.

Zed attacks X with full force but misses his shot.

Zed manages to counter the attack from X

Zed starts an attack against X and hits him. His victim screams in pain (2043 HP).

X runs up to Zed at full speed, stumbles and falls.


Even if Zed won, he was hit 5 times and X had the upper hand in hits even if in defense. Saved by HP (that was why I said 3 instead 2.5 in ratio formula)
I can say that stat difference is above 500 (just what I can see) and a 7 dif in levels, but it is not a factor. I still think is not only the numbers for stats that is giving a win or a loss, but the way you develop the char. That is why I was not quite buying what Falco was trying to say long time ago. I realized that chars with visible stats STR AGi are having hard time to win against almost equal stats and are at par with same development. And chars with big ATT and DEF are wining more often to the almost equal. But this is nothing proved, just an assumption based on try-outs. For me it makes more sense to find a formula based on percentage for each stat, than trying to go for pure numbers.

47

Wednesday, January 28th 2015, 3:59am

PS In this particular case, it was STR AGI against almost equal.

Rate this thread