

Hit Registration Problems. Can Any One Confirm?
#1
Posted 25 August 2015 - 01:13 PM
just now i had a match where all on my side got wiped, many had hitreg problems (writing that in chat) and the other got many one shot deads!
So can more people confirm that?
If so this is a matter for a priority patch.
I only register hits from back, which are registered in my front CT. Maybe it is vice versa for fast kills, that they are registered in back instead of front.
Regards
#2
Posted 25 August 2015 - 01:18 PM

So don't just bother to understand MWO hit mechanic
Atlas: hit it upfront in that sweetspot to destroy his rear armor
CTF: hit it randomly to dmg his arm on the other side
AWS is self explanatory (watch the distance aquisitor)
Edited by DuoAngel, 25 August 2015 - 01:21 PM.
#3
Posted 25 August 2015 - 01:30 PM
DuoAngel, on 25 August 2015 - 01:18 PM, said:

So don't just bother to understand MWO hit mechanic
Atlas: hit it upfront in that sweetspot to destroy his rear armor
CTF: hit it randomly to dmg his arm on the other side
AWS is self explanatory (watch the distance aquisitor)
Dude, damage transfers through destroyed components. So shooting that Cataphract in the right torso - which has been destroyed - will mean that 20% of that damage (if I recall the percentage correctly) will be applied to the next component, moving from the limbs to side torsi and side torsi to center torso.
Think about it: if damage didn't transfer from a destroyed component to the next extant innermost component you could just use your destroyed side torsi as a total damage shield - which would break the game, because your aim would be true and it would be a hit but no damage would be applied to the mech. Instead we have the system that we use now, where a hit to a destroyed component partly transfers the damage that would be applied to said component to the next component in (meaning moving from limbs to center torso). That way you can still shield yourself partly with a destroyed side torso or whatever, but yet you'll still take some damage, which is as it should be.
Edit: Perhaps a better way to put it is that a registered hit has to do damage for the game to function on the most basic level. If you hit a component that is already destroyed, you obviously can't apply any damage because the component doesn't exist anymore to begin with. If there was no damage applied at all as a result, a mech would become progressively more invulnerable as it lost components - the exact opposite of what any sane person would reasonably expect to happen. Thus there is a mechanic where a percentage of the damage that a destroyed component would take is transferred to the next innermost component.
Edited by grendeldog, 25 August 2015 - 02:33 PM.
#4
Posted 25 August 2015 - 01:37 PM
grendeldog, on 25 August 2015 - 01:30 PM, said:
Think about it: if damage didn't transfer from a destroyed component to the next extant innermost component you could just use your destroyed side torsi as a total damage shield - which would break the game, because your aim would be true and it would be a hit but no damage would be applied to the mech. Instead we have the system that we use now, where a hit to a destroyed component partly transfers the damage that would be applied to said component to the next component in (meaning moving from limbs to center torso). That way you can still shield yourself partly with a destroyed side torso or whatever, but yet you'll still take some damage, which is as it should be.
Have you even read my post?

Dmg gets thru side torso to CT. You got it right. But this dmg NEVER gets further to the other side of the mech, not to mention that far arm when ST attached to it takes zero dmg in process. And that's how it works...how it supposed to work.
#5
Posted 25 August 2015 - 01:54 PM
DuoAngel, on 25 August 2015 - 01:18 PM, said:

So don't just bother to understand MWO hit mechanic
Atlas: hit it upfront in that sweetspot to destroy his rear armor
CTF: hit it randomly to dmg his arm on the other side
AWS is self explanatory (watch the distance aquisitor)
You didn't understand what i mean.
The enemy was behind me and hitting me, but my front torso got red!
That was for sure, cause i was hitting a DW from behind!
How hits should function i know!
So if you have anything else to help, go on. Else i don't think you are helping to find what i am reading since last patch.
#6
Posted 25 August 2015 - 02:30 PM
IF
armour_value == 0
{ check_hit_location; check_for critical:
apply_damage_to_component;
}
ELSE
apply_damage_to_armour}
Now, it seems that MWO is built in such a way that many checks happen simultaneously (asynchronous(?)), and as I have it it is inherent in this architecture that checks will be missed. Effectively Hit Reg thinks there is no component or armour in the way of the bullet, which is actually simulated as passing through the internal structure of the mech cycle by cycle.
Since I am anything but an expert, any computer experts feel free to correct me, I will take no offense.
The main thing to take away that Hit Reg will always swallow a small part of the damage, and damage will pass through intervening components. It depends on the degree of damage loss which is acceptable.
However, it seems that Laser Hit Reg at least has improved hugely, as I now return damage numbers twice that of 4 months ago. Where I would do 150 damage in a 4 LL Flame, it is as of recently 300 or so quite often. Whether this is due to me upgrading my system, a ping improved by about 50 ms from 330 down to 280, and/or PGI's Hitreg I cannot tell.
Fact is they actually squashed a few bugs in Hitreg and mentioned it in the patch notes about a month back.
However, according to Karl Berg, it seems the way PGI maintains the MWO code makes it possible for changes to be inadvertently reverted. Any emergency hot fixes by hand are eliminated and/or not possible in between patch iterations either. Remember how sometimes bugs in closed Beta would sometimes re appear down the line? Had to do in part with keeping the code base as consistent and maintainable as possible, with no manual fiddling by engineers allowed on a case by case basis. Also a large part of the code is auto generated, and you do not want humans to poke around too much.
However, this does place an onus on PGI to check if that happened.
Edited by Hammerhai, 25 August 2015 - 02:38 PM.
#7
Posted 25 August 2015 - 02:37 PM
#8
Posted 25 August 2015 - 07:52 PM
#9
Posted 25 August 2015 - 08:55 PM
Myself i had several immobile Mechs i shot several times in the back and was killed when they woke up, turned around and finished me off as i was already close to overheating in the attempt to get them to their expected death.
Everything as usual...

#10
Posted 26 August 2015 - 01:47 AM
#11
Posted 27 August 2015 - 01:27 PM
1 user(s) are reading this topic
0 members, 1 guests, 0 anonymous users