To Hold and Be Held. A Sad Game Issue.

[Chai]

I was being snarky with my response. I should be more careful, sarcasm doesn't translate well to text too well.

Sorry, I glossed over the sarcasm indicator™

[Llewndyn]

This is a byproduct of the U5 physics check modifications. In order to fix it they would have to reverse the order of operations for mob physics.

If a mob began an action, or the system commanded it to begin an action (even if it doesnt look like the action was physically initiated yet) the mob gets to take that full action. Just like when TWFing and your main hand hits a mob if you get an offhand proc for that main hand attack, it too is considered an action that completes "before" any other state on you can take its hold. Its also considered to be "within range" because the main attack was also "within range". Go into a module with hobgoblins and kill all but one, then try dodging their sword attacks, which have a huge wind up. If it started that wind up when you were in range, you are considered to be "in range" when the swing happens, even if you are accross the room by that time.

If a mob with archer / caster AI is about to move and you stun it, it will slide down the hall to where it was going to move to. If it was casting, the spell goes off, even if it died before that point in some cases. If you trip an ogre as its two stepping backward before its triple smash attack, then everyone runs up to beat on it, they all get triple cleave smashed, by an ogre on its back.

The lowering of the number of physics checks was done to reduce lag, but the side effect of that is immersion shattering in a game where success is based heavily on mastering moving combat and order of operations of ability use. A fix would entail reversing the rules in the code for "if action initiates, then it automatically completes full action, regardless of other circumstances" as well as the rule for "if physics check indicated [in range] at beginning of single action (which can be a multiple attack sequence) then [in range] is true for entire action."

That is the most helpful, well thought out, illuminating response I have seen on these forums to date. -1 for your insolence!

(I didn't really neg rep, but awesome response, it makes a little sense this way. I don't like it, but it makes sense!)

[TheDjinnFor]

This is a byproduct of the U5 physics check modifications. In order to fix it they would have to reverse the order of operations for mob physics.

Summary for the technologically impaired:

It reduces lag.

/signed on the fix by the way, I'd rather crazy lag than crappy combat.

[GoRinNoSho]

boolean mobilityImpaired;

if(mobilityImpaired == true) notify/interrupt the behavior thread.

That will suspend the use of graph theory (np hard) as mobs are still reactionary when scooching. That is assuming that the extra move segments are generated by the monster on course, otherwise they are preplanning/cacheing the next refinement move and trashing several graph calculations normally. So thus the held monster doesn't trash the calculations and iterates across all refinements causing figure question mark behavior.