Curse

Lordhelmos · Oct 26, 2006, 07:14 AM // 07:14

Rather than complaining about the nightmare after effects of the new patch, as the mods say we should be thinking on ways to fix these and offer positive suggestions.

1.) For the kiting and movement problems I have sent a bug report to NCsoft detailing the problem. Its quite a big deal, I assume they will fix it.

2.) For aggro problems and mob behavior there is definitely a problem with running past mobs and getting hit by invisible swords and hammers. The new patch seems to have a massive latency issue and alot of things are out of synch. For example the running while casting stops you and then you cast halfway and its canceled. It's like the game is too slow to keep up with events going on. Yes it sounds like lag but everyone is experiencing it. Maybe its part of the network, because I see the same effects when I ran through Ran Musu from Tsumei and monster attacks that were done a long time ago started hitting me out of nowhere, even though the mob was miles away.

~Now as we all know UW/SS and farming with AoE and mobbing is moot. It's been killed by the misbehavior of the A.I. The A.I. has been plagued by several issues:

-Godly interrupts. Not to human at all. I would say tone the interrupts down a little and make them more sloppy. higher level monsters with better reflexes of course.

-Now seemingly random scattering. Before mobs would herd over one character and "holding aggro" was the staple of PvE. Anet claims they want to make the mobs as "human" as possible but I just don't see it in the monsters we face. How would the mobs operate? We are all used to the herd and conquer method of playing but now with the coming of dervish this way is no longer a possibility for Anet. Keeping mobs how they were will turn the game into a giant farm fest with the new skills and dervish. Perhaps Anet dug themselves into a hole with this class, this is not going to be fun to fix.

-UW is now dead with the new AI, but I think its good SV/55 is no longer monopolizing the place. Let's face it, guild wars was stagnating with players only allowing certain builds to enter specific areas once they found out they worked. A certain degree is needed to keep cookie cutters from dominating the game. This is a move for the greater good of the community, so UW farmers get over it and play. This game will not boil into a farm fest.

******

Now back to the issue at hand: Improvements.

-We want an auction house and stylists. Nuff said.

~MOB BEHAVIOR~

Ok so we don't want guildwars boiling down into a herd fest. These are vicious monsters not cattle we can herd for gold. However by making the monsters completely random as they are, classes like tanks lose their value and casters will be frustrating to play. As one poster said, whats the point in tanking if there is no aggro to take?

~Improving the mobs:

An important part of the monster AI is targeting. This is the first hurdle I will shoot for. Who should the monster target? Well as of recently, mobs target players based on AL and DP. Mobs will always go for the weakest link, or what they believe is the weakest link. This definitely is a primitive way for the AI to think. The AI should use more strategy in target choice.

I surmize that the AI should target on a skill and class basis. For example a monster using backfire will have a high priority on using it on a ele or monk. Lets say the monster casts the spell, the %'s on what target the monster chooses should be based on the player and class:

For example: the player's party is 1 War, 1 ele, 1 monk. The monster is lvl 24 and has a 50 % chance of backfiring the monk and switching target to him, 30% for the ele, and 20% for the tank. Lower monsters will have poorer choices such as 50% on the tank and 25% on each caster. This goes for sins, higher sins will prioritize healers and lower ones armored targets.

So where does the tank come into play? There is a threat factor, for example a warrior wailing on a shadow war that attacks a monk increases the chance the monster will change target to him by increasing his threat factor with damage. If the warrior intercepts the enemy and deals enough damage or body blocks (or knocks the enemy down/uses snares or hamstring) the AI will check its threat factor and change target to the warrior. Of course higher level enemies will be harder to persuade.

What im saying is that mobs should not just be stuck on one target but choose an initial target and change targets based on threat factors from the player party and the skills the enemy itself is using.

~AOE~

Of course enemies shouldnt stay in AoE's like idiots, but at the same time not be smart enough to avoid all AoE's. defensive skills should be triggered by AoE's first. For example, an enemy stuck in an MS with high life should use something like balanced stance or RoF first before attempting to run. If the AoE is super massive the enemy flee's based on a % of damage its taking and a random factor. Lower enemies will be dumber and stay in AoE's longer. There should be a internal comparison % of how much life the creature is losing from AoE and a % chance it will run. The higher the threat of death, the higher the monster's % is to flee. There should also be an desperation factor in which a monster with high damage being stacked on it activates a train of skills in a clever order to try to take down a player as it is dying (Berzerk mode).

So dumping an AoE on a monster, several factors should come into play. If the monster is taking heavy AoE damage, it has a higher % to run. But also it has an internal factor that sometimes desperation mode can be triggered and it goes crazy on the nearest character, adding some form of randomization. Also higher monsters use defensive skills in response to AoE before trying to run, like ward against elements or stability. Those skills should factor into AoE as well. The monster shouldn't just activate a ward based on proximity and run based from AoE based on how much damage over time it's taking. I mean players have predicted that the monster will run from an AoE if it is taking area damage more than 3 times per second..... thats just too stiff.

I propose that there shoudl be two rolling %'s in the AI. One for how much damage the creature is taking, (its recording its damage over time from area attacks). It compares this % to another random cycle that decides if it will run or not. This second % is lower for higher level monsters, causing smarter mobs to scatter if they are taking too much damage. But these numbers should never be set in stone, there should always be a chance that the monster stays in the entire AoE.

For example: 1 MS is dealing 50 dmg per second to a 600 life monster. Internally it has a random number between 40-20% of its max HP that it can take in one second. If the MS aoe attack breaches this random number it rolls a for a 30-60% (done randomly) chance to run.

On a higher level monster with 700 HP and 2 MS pouring down on it for 100 dmg a second. It internally has a random number of 30-15% of its max hp it can lose before it triggers the run script, in which it has a 50-75% chance of running.

In both scenarios there should always be the possibility that the monster will stay in the AoE, also in both cases when the monster takes 50% damage and is under AoE it may have a 30-50% chance based on level to go into rampage mode and attack regardless of AoE. The rampage would overpower the other 2 rolls.

Well those are suggestions, hope that gets us on a better track so the mods wont spend the whole night shutting down threads.

Arcador · Oct 26, 2006, 07:30 AM // 07:30

1) and 2) are majors. I hope it is fixed easily and soon.