Potential bug regarding triple attack and double attack

[MarkovChain]

According to this extended test (near bottom, click expand lolalla), double attacks and triple attacks prevent each other from proc'ing if the game rolled positively for both of them. For instance the test shows that 15% Double attack and 10% triple attack only gives 8.5% actual 2xrounds and and 13.5% dual rounds when you would expect

* 10% triple rounds and (100% -10%) x 15% = 13.5% double rounds if a double attack procs when the game rolls a triple attack AND a double attack
* 15% double rounds and (100% -15%) x 10% = 8.5% double rounds if a tripple attack procs when the game rolls a triple attack AND a double attack

This behavior is unexpected because having high values for DA means the amount of TA dimishes considerably, and conversely. As an exemple, 50% DA and 50% TA would mean that you are only actually getting 25% DA and 25% TA. A more exagerated example would be that 100% TA and 100% DA results in both double attacks and triple attacks to never happen.


Note : the sample size is large enough to cover +/- 0.5% variation in proc rate from the ingame algorithm.

[Darka]

There's a check order, Triple Attack checks before Double Attack, not at the same time. If TA rolls, DA will not on that hand/swing, DA will roll if TA fails. So having high TA diminishes DA only.

[Byrth]

He knows that. He's saying the test shows that having high DA and TA diminishes both of them. The question is whether the effect is statistically significant, which I don't have time to check right now.

[Karbuncle]

This seems like, if true, is a really big oversight by SE...

[Monchat]

There's a check order, Triple Attack checks before Double Attack, not at the same time. If TA rolls, DA will not on that hand/swing, DA will roll if TA fails. So having high TA diminishes DA only.

the test shows that there is no priority of TA over DA or the converse. If it were the case, with 10% DA and 15% TA, and TA was checked first, you would get 15% of your attack rounds with TAs, which is not the case.

[Byrth]

Bump for great justice and shortsighted programming practices.

Upon further examination at the OP's data, he's absolutely correct.
http://ffxi.allakhazam.com/forum.htm...57380216814143

Having more DA makes your TA less worthwhile. Having more TA makes your DA less worthwhile. Atma of the Apocalypse may be outdone by others for jobs with high DA.

[Arcon]

I still find this hard to believe. Sounds like an extreme oversight. I will have to test that myself. My WAR can get 38% DA rate, with Atma of the Lion, Atma of the Rapid Reptilian and Atma of the Apocalypse the tests shouldn't be too bad.

Three possible scenarios:
1. DA first
Expected DA rate: 38.0%
Expected TA rate: 19.6%

2. TA first
Expected DA rate: 27.6%
Expected TA rate: 30.0%

3. Both exclude each other
Expected DA rate: 38.0%
Expected TA rate: 30.0%

4. They cancel each other out
Expected DA rate: 27.6%
Expected TA rate: 19.6%

Should be easy enough to determine which one it is.. if this thread is correct, it should be option #4.

[MarkovChain]

The issue was coming from Kparser's algorithm to identify melee rounds and multitattacks. This was improved and it turns out that it is working as expected : TA is checked first then DA.

[vedder]

thank bloody god lol moment i heard this i was freaking out lol