PvP - High attack speed vs HP Charm - Charm sometimes will not tick at 50% HP

Zsw - Dreamweaver

It is necessary for another character with 3.33 to 5.0 attack per second to attack you in order to reproduce this glitch.
It is necessary for your character to have a Guardian Charm to reproduce this glitch.
Your character must be able to die to this other character in roughly 5 hits IF you were not charmed.

1. Use any character that fits the above requirement.
2. Have the other character attack you at around 3.33 to 5.0 attacks per second.

RESULT:
Your character will die without your guardian charm ticking.

If you observe your damage log however, you can see that your guardian charm should have activated, but yet it did not.

Screenshot:


Please refer to the damage log. The fight I conducted included barb buffs. This puts my hp at roughly 15,600. The last 5 hits conducted on me killed me without my guardian charm ticking disregarding the fact that I even had deaden nerves activated. My charm should have restored me to full hp after the third hit. Yet, I was able to die without my hp charm ticking.

Fissile - Archosaur

Same can happen if you are being attacked by multiple people. Archers also have a skill combination that can attack too fast for the charm to work (though it is only 3 hits I think).

Adroit - Lost City

Fissile - Archosaur wrote: »

Same can happen if you are being attacked by multiple people. Archers also have a skill combination that can attack too fast for the charm to work (though it is only 3 hits I think).

To add to that, wizards can do this with the right timing, it happens a lot in sutra combo with d.pyro (damage is delayed) and wotp (instant damage) where they seem to hit together and bypass charms.

Oups_Dead - Harshlands

wow that's so old since pwi first came i guess... but i kinda like it another way 2 charm bypass . last 2 charm bypasses i got that way were on full +12 Jod archer r9rr and on a full vit stone barb r9rr full +10
so i'd like to add that u do not have to be squishy to be charm bypassed although both targets died in 3 aps hits.
also first hit u do from stealth does not show on damage log unless it's a skill with a long cast time or u amp before u hit it from stealth (mire/poison) sometimes more than 1 hit is not shown
idk if it depends on lag or it is a glitch but i think both cases cannot be fixed since charm needs a fraction of sec to heal u and if u receive a lethal hit before that fraction of sec then maybe u deserve to die.
also according to my experience u cannot be charm bypassed with deaden on that is EXTREMELY STRANGE AND UNLIKELY .
it happens when 2 successful hits r given and r enough to kill u. ur log says three 3ks in a row deaden should have activated on second hit leaving enough time for a charm tick ... strange

Zsw - Dreamweaver

Oups_Dead - Harshlands wrote: »

wow that's so old since pwi first came i guess... but i kinda like it another way 2 charm bypass . last 2 charm bypasses i got that way were on full +12 Jod archer r9rr and on a full vit stone barb r9rr full +10

If I were to guess, i'd say its caused by a delay in the way PWI registers damage, so taking multiple hits in a short period of time causes the game to not be able to react quick enough. It would explain why the glitch is consistent with wizard sutra combo and also occurs with getting ganked.


To reproduce the sutra combo glitch:
It is necessary for another Wizard to attack you in order to reproduce this glitch.
It is necessary for your character to have a Guardian Charm to reproduce this glitch.
Your character must be able to die to this other character in exactly 2 hits IF you were not charmed.

1. Use any character that fits the above requirement.
2. Have the wizard use the following skills in order without delays: Essential Sutra -> Divine Pyrogram -> Will of the Phoenix

RESULT:
Your character will die without your guardian charm ticking.

Oups_Dead - Harshlands wrote: »

also according to my experience u cannot be charm bypassed with deaden on that is EXTREMELY STRANGE AND UNLIKELY .
it happens when 2 successful hits r given and r enough to kill u. ur log says three 3ks in a row deaden should have activated on second hit leaving enough time for a charm tick ... strange

That's what I thought too, but I definitely had it on and it still went through =/
My HP was close to full, so charm should have ticked on the second or third hit. If not, then deaden should have activated on the 4th or 5th hit and hp charm should have saved me there. But neither case happened.

Oups_Dead - Harshlands

wow man that's not a small lag then must be a huge one for it being delayed for over 2 hits :P must have gotten worse with timeb:shocked

Zanryu - Dreamweaver

It also happens if you take two or more large hits in succession (within the same second), though one hit would allow for a charm tick, your charm doesn't have time to register it and you get bypassed. I've had it happen to me a few times. Would be nice if our charms would function properly.

Arawin - Raging Tide

I've had this happen a lot, especially in large scale PK or TW. It's really irritating.



I'm guessing based on what's been said the "charm tick" is only calculated at specific intervals, say every .5 seconds. That would describe the exact symptoms listed here. (Large damage numbers coming at the same time = charm bypassed.)

I can see 2 simple solutions:

1) Check for charm tick after every single hit. (Ex: hit 1 -> check charm -> hit 2 -> check charm -> hit 3 -> check charm, and so on... even if all hits come at the exact same, impose an ordering on them and make sure to always check charm.)

This may be slightly more computationally expensive than the current implementation and produce excess lag. In that case see solution 2. (Though this could also be a "charm buff")

2)Always check for charm tick before someone DIES. Basically "If character health is less than zero, check if charm tick available." This would prevent any deaths without charm tick, and not add very much computational overhead. However it would buff charms quite a bit and remove 1 shots if charmed.

Bhavyy - Raging Tide

Arawin - Raging Tide wrote: »

I've had this happen a lot, especially in large scale PK or TW. It's really irritating.



I'm guessing based on what's been said the "charm tick" is only calculated at specific intervals, say every .5 seconds. That would describe the exact symptoms listed here. (Large damage numbers coming at the same time = charm bypassed.)

I can see 2 simple solutions:

1) Check for charm tick after every single hit. (Ex: hit 1 -> check charm -> hit 2 -> check charm -> hit 3 -> check charm, and so on... even if all hits come at the exact same, impose an ordering on them and make sure to always check charm.)

This may be slightly more computationally expensive than the current implementation and produce excess lag. In that case see solution 2. (Though this could also be a "charm buff")

2)Always check for charm tick before someone DIES. Basically "If character health is less than zero, check if charm tick available." This would prevent any deaths without charm tick, and not add very much computational overhead. However it would buff charms quite a bit and remove 1 shots if charmed.

I like the idea of the second one except it would just be impossible to kill barbs with deaden nerves and double the pool of hp for the charm tick to work through. Same with sins and their deaden nerves.
I'm not really sure why this happens but yes I don't think the interval needs to be below 0.3 since archers can do it too with 0.9 int

Zsw - Dreamweaver

Arawin - Raging Tide wrote: »

2)Always check for charm tick before someone DIES. Basically "If character health is less than zero, check if charm tick available." This would prevent any deaths without charm tick, and not add very much computational overhead. However it would buff charms quite a bit and remove 1 shots if charmed.

I don't think this method would work since that would prevent even proper charm bypasses.

I have no knowledge on how the PWI server actually calculates damage, so I will be making some guesses.

It shouldn't really be computationally expensive to check after every hit if your current hp is lower than 50% of your max hp.

So I would guess that the problem resides in the function that calculates damage being called multiple times simultaneously. The function simply checks if current hp - damage taken > 50% of max hp, if so, decrease hp by damage taken. For each hit, it is true that current hp - damage taken is greater than 50% of your max hp, and no charm tick is necessary. However, when 5 calls to that function is added up altogether, it causes a death.

So the function that calculates damage would have to only process one request at a time, as opposed to 5 of them simultaneously. Although that would surely lag if you're trying to process 5 hits individually in under 1 second.

I would also guess that the function to determine whether or not your character has died is not tied to the function that calculates damage, or else it should be able to pick up this problem when it sees that your hp has reached zero when your current hp is still greater than damage you took. The death check function most likely is called between intervals and simply check if hp > 0.

In which case, it may be necessary for a separate function to exist to ensure that if your current hp is higher than the damage you last took, your charm must tick before you die. Lets call it is_charm_bypassed. The function would have to be called every time damage is calculated. The function would store both your current hp and the damage you just took. When the game detects that you die, the function to calculate death would have to called is_charm_bypassed. If your current hp is greater than the damage you just took, tick your charm instead of dying. Otherwise, you are dead.

This would mean that if 5 hits were calculated simultaneously, non of those hits should meet the condition that damage taken > current hp. Therefore, you should recieve a charm tick for those hits. If you were to be 1 shot or properly charm bypassed, then the last hit you took would satisfy the condition that damage taken > current hp, and you should die properly.

Fissile - Archosaur

My guess is it is far simpler than that. I would think that a charm tick is determined on the client side and latency between the client and the server lead to the sever determining that you are dead before it receives the charm tick notice from the client.

This would also explain why it seems to happen more often in TW where you typically experience more latency.

Aeroboi - Heavens Tear

Fissile - Archosaur wrote: »

My guess is it is far simpler than that. I would think that a charm tick is determined on the client side and latency between the client and the server lead to the sever determining that you are dead before it receives the charm tick notice from the client.

This would also explain why it seems to happen more often in TW where you typically experience more latency.

Except the fact that a client can disconnect from the server, but that Toon can remain online charm ticking.

Ive also had large lag spike and was kept alive by my charm until I warped back upto speed.

Zsw - Dreamweaver

If it was handled client sided, I think it would be very risky as it could be subject to exploits.