Quadrakill Issues
Old 01-18-2019, 08:33 PM   #1
Epic Scholar
 
Huggles's Avatar
 
Huggles is offline
Join Date: Aug 2010
Posts: 4,845
Default Quadrakill Issues

While testing average damage on piercers in Traveler's Way, I noticed that I wasn't getting many Quadrakills. Actually, I wasn't getting any. Not a single one.

150 kills in without any Quadrakills, I decided to report.

Test conditions:
1H Piercer in main hand
No items in off hand
No Pets equipped
Nibbles has 43 Quadrakill


Test process:
Enter combat with T. Way mob
Hit attack
Exit
Repeat

As the mob starts at full HP every time, I should be getting the maximum Quadrakill chance. However, even at 43 Quadrakill (supposedly a 4.3% chance), I did not see any quads.

I then did a little more testing, this time with nearly identical test conditions, the only difference being a shield in the off-hand. Again, not a single quad.

__

I changed tack and decided to try to test overall quad rate by dueling someone while using the Staff of the Masquerade, as this should easily show quads due to the way the proc works.

I ended up getting 2 quads from 163 swings against a target who was full HP nearly the entire time thanks to regen. That's a 1.2% rate of success, rather than the expected 4.3%.

I then tested further and got 1 proc on 72 attempts, making is 3 for 235, or 1.28%. Then 1/40, total 4/275, 1.45%.

I will continue to test as much as I am able, but something looks screwy.

If you think I need to show more proof before you investigate, I can do that, but it'd save me hours of time if you were to look over the code.
__________________
Quote:
Originally Posted by Glitchless
If they put a nice icon on you it's beneficial if they put a mean one on the other guy it's detrimental.
Quote:
Originally Posted by Glitchless
The player was banned for using an exploit repeatedly without reporting it. There's only 1 proper way to profit off an exploit: report it.
Quote:
Originally Posted by Huggles
Old Delay/(1+haste%) = New Delay

For slow effects, use the slow % but as a negative.

Just repeat for multiple effects.

DO NOT multiply by 1+slow% to get your new delay. 1*1.4 is not the same as 1/0.6 and your answer will be wrong.
  Reply With Quote

Old 01-19-2019, 02:58 AM   #2
Administrator
 
Glitchless's Avatar
 
Glitchless is offline
Join Date: May 2008
Posts: 8,569
Default

I double checked and the % being applied appears to be correct.

If you're counting swings you're not going to get an accurate % because quads can't proc off of the additional swings from double attacks, only the first.

http://forums.nodiatis.com/forums/sh...ad.php_t=18108
__________________
Computing the probability that at least one of the following events will occur:
P(a or b ... or z) = 1 - P(!a and !b ... and !z)
Probability
  Reply With Quote

Old 01-19-2019, 09:24 AM   #3
Epic Scholar
 
Huggles's Avatar
 
Huggles is offline
Join Date: Aug 2010
Posts: 4,845
Default

By counting swings, I meant I was checking at each delay whether there were at least 4 attacks, as would be indicated by four 'Cursed' texts and the accompanying damage splooshes.

I used weighted to slow down the attacks to make sure previous attacks were not getting mixed in to make it easier to distinguish.

___

Here's a video of testing this morning in T Way with the piercer. 243 fights (give or take a few due to miscounting) and not a single quad:

https://youtu.be/gSi4Q2qFKgw

Watch at 2x speed to save your sanity.
___

Possible causes:

Quad is somehow having a bad interaction with measuring HP of mobs on opener. Low level mobs are instantly found as having 0 HP and thus don't proc quad. This could also cause an issue where HP is being measured after the first hit and thus causing actual quad chance to be lower in all cases.

Doubleattack chance is somehow preventing quad from activating as often as it should. The math doesn't seem to add up for this one to be the culprit though.
__________________
Quote:
Originally Posted by Glitchless
If they put a nice icon on you it's beneficial if they put a mean one on the other guy it's detrimental.
Quote:
Originally Posted by Glitchless
The player was banned for using an exploit repeatedly without reporting it. There's only 1 proper way to profit off an exploit: report it.
Quote:
Originally Posted by Huggles
Old Delay/(1+haste%) = New Delay

For slow effects, use the slow % but as a negative.

Just repeat for multiple effects.

DO NOT multiply by 1+slow% to get your new delay. 1*1.4 is not the same as 1/0.6 and your answer will be wrong.
  Reply With Quote

Old 01-19-2019, 10:33 AM   #4
Epic Scholar
 
Huggles's Avatar
 
Huggles is offline
Join Date: Aug 2010
Posts: 4,845
Default

Tested with Staff of the Masquerade in T. Way as well. No quads.
__________________
Quote:
Originally Posted by Glitchless
If they put a nice icon on you it's beneficial if they put a mean one on the other guy it's detrimental.
Quote:
Originally Posted by Glitchless
The player was banned for using an exploit repeatedly without reporting it. There's only 1 proper way to profit off an exploit: report it.
Quote:
Originally Posted by Huggles
Old Delay/(1+haste%) = New Delay

For slow effects, use the slow % but as a negative.

Just repeat for multiple effects.

DO NOT multiply by 1+slow% to get your new delay. 1*1.4 is not the same as 1/0.6 and your answer will be wrong.
  Reply With Quote

Old 01-19-2019, 03:47 PM   #5
Epic Scholar
 
Huggles's Avatar
 
Huggles is offline
Join Date: Aug 2010
Posts: 4,845
Default

Quote:
Originally Posted by Huggles
...

Doubleattack chance is somehow preventing quad from activating as often as it should. The math doesn't seem to add up for this one to be the culprit though.
Actually, this is looking close to my results on the duel testing. 4.3% of doubleattacks (4.3% of 25.6% is 1.1%) is quite close to my quad rate of around 1.45% and would explain why I wasn't seeing any 5x hits. However it still doesn't explain the issue with openers unless there are two simultaneous issues.

Hopefully I can get some people to help test this when I get home.
__________________
Quote:
Originally Posted by Glitchless
If they put a nice icon on you it's beneficial if they put a mean one on the other guy it's detrimental.
Quote:
Originally Posted by Glitchless
The player was banned for using an exploit repeatedly without reporting it. There's only 1 proper way to profit off an exploit: report it.
Quote:
Originally Posted by Huggles
Old Delay/(1+haste%) = New Delay

For slow effects, use the slow % but as a negative.

Just repeat for multiple effects.

DO NOT multiply by 1+slow% to get your new delay. 1*1.4 is not the same as 1/0.6 and your answer will be wrong.
  Reply With Quote

Old 01-19-2019, 05:02 PM   #6
Administrator
 
Glitchless's Avatar
 
Glitchless is offline
Join Date: May 2008
Posts: 8,569
Default

I'm going to add battle text for quads, tris, and doubles so it's clearly apparent when they occur.

I'm also adding an extra update check for health percentage in case there was an issue with openers vs. newly spawned creatures.
__________________
Computing the probability that at least one of the following events will occur:
P(a or b ... or z) = 1 - P(!a and !b ... and !z)
Probability
  Reply With Quote

Old 01-20-2019, 10:44 AM   #7
Epic Scholar
 
Huggles's Avatar
 
Huggles is offline
Join Date: Aug 2010
Posts: 4,845
Default

https://youtu.be/Ans1-u0O9z0

Still no luck on mobs that are getting 1 hit. I'll try testing overall proc rate next.
__________________
Quote:
Originally Posted by Glitchless
If they put a nice icon on you it's beneficial if they put a mean one on the other guy it's detrimental.
Quote:
Originally Posted by Glitchless
The player was banned for using an exploit repeatedly without reporting it. There's only 1 proper way to profit off an exploit: report it.
Quote:
Originally Posted by Huggles
Old Delay/(1+haste%) = New Delay

For slow effects, use the slow % but as a negative.

Just repeat for multiple effects.

DO NOT multiply by 1+slow% to get your new delay. 1*1.4 is not the same as 1/0.6 and your answer will be wrong.
  Reply With Quote

Old 01-20-2019, 03:47 PM   #8
Administrator
 
Glitchless's Avatar
 
Glitchless is offline
Join Date: May 2008
Posts: 8,569
Default

Ok that finally makes sense then. It's based on the HP of the target after the first attack and double attacks actually go through, so if the target is 1 shot then it has 0% health and no chance of quadra proc.
__________________
Computing the probability that at least one of the following events will occur:
P(a or b ... or z) = 1 - P(!a and !b ... and !z)
Probability
  Reply With Quote

Old 01-20-2019, 04:47 PM   #9
Administrator
 
Glitchless's Avatar
 
Glitchless is offline
Join Date: May 2008
Posts: 8,569
Default

I can definitely see the point being made that the check should happen prior to any swings landing. This bug report is a win for you, sir. 5 TCs for the time spent testing and will do the following:
-Removing Quad/Trishot/Double spam Apparently people like it.
-Perform the Quadra check on the HP % prior to any swings
__________________
Computing the probability that at least one of the following events will occur:
P(a or b ... or z) = 1 - P(!a and !b ... and !z)
Probability
  Reply With Quote
Reply



All times are GMT -5. The time now is 11:04 PM
Boards live since 05-21-2008