Effect of PIE and Level Difference on Phantom Dart and Damnation Accuracy: Parser Results

[DeusBahamut]

Major EDIT: Cleaned up the OP. I only kept his data sets with summaries of what's going on. I tried to keep as much of his own words as possible while shortening this post for your sanity. Some things from the OP are left out including some things regarding INT and Damage. Important conclusions are in bold. The OP's original full thread can be found on ZAM forums linked here.
~Kaeko


Purpose:
This test is strictly testing the effects of INT and PTY on Phantom Dart and Damnation Damage and Accuracy, versus Rank 55-59 Sphene Doblyn, at Thaumaturge Ranks 43 and 48.

Conclusions:
(1) PIE and dLVL do not effect the miss rate of Phantom Dart. Keeping in mind the small sample sizes, the estimated miss rate of phantom dart is static at 15%.
(2) As with Phantom Dart, the miss rate of Damnation appears fixed and roughly 20%. Again, keep in mind the limited sample size.

******************

Methodology and Parser:
Crabby Thread, by golfandsurf
DOWNLOAD CRABBY HERE
-- D/L at own risk -- etc. etc.
1) Download Crabby.rar from the link above and extract the EXE to your own directory.
2) Start FFXIV and log in using windowed mode.
3) Run Crabby.exe. Check if it has synced up by looking at the bottom of the window. It will say "Attached". If you need to resync, just double click the icon at the lower left of the window.
4) You can create separate window feeds for each chat type. Click "Settings > Chat Tabs" and select one or more types. They correspond to what you "Show" or "Hide" In-Game under "Configuration > Chat Log > Settings". Best of all, there is no limit to the history or number of log entries Crabby remembers.
5) Execute your test with the guidelines above, and copy out the timestamped entries into an external text editor. Please be aware that logging out of the game may crash Crabby and prevent you from copying out the logs. Be on the lookout for bugs and report in the Crabby Thread. Try a more stable version of Crabby if you experience problems. The newer features are not necessary for our testing purposes.
6) Parse these text files to determine your test conclusions. I have completed customized perl scripts which automatically evaluate your average non-critical damage, average critical damage, etc.

Great care was taken to ensure constants like weapon durability > 60%, ensuring consistent use of the same abilities and job traits, the same gear, the same point allocations, etc. My script even identifies and throws out attack data where the attack was done after the Coblyn readies Cluster Geyser, which results in subsequent hits doing upwards of 50% more damage due to a "broken" mob.

Limitations:
Please note the sample size and keep it in mind that the more samples, the closer the averages come to converging on the 'true' value before the game's random number modifier. Only with 1000 samples do we start to see consistency due to the randomness, especially with miss rates. The Rank 43 Tests are not as accurate with the miss rates, but are accurate with the damage averages.


Results:
TOTAL NUMBER OF HITS+CRITHITS AND (HITS+CRIT HITS+MISSES)
RANK 43 TEST 1 = 131 (151)
RANK 43 TEST 2 = 81 (93)
RANK 48 TEST 1 = 889 (1037)
RANK 48 TEST 2 = 1087 (1303)

Phantom Dart Results
Damnation Results

In the charts, the EXP column corresponds to experience points gained from Sphene Doblyn Rank 55, 56, 57, 58, 59, and "ALL" refers to the TOTAL statistics between all stated ranks. You can easily understand the significance of ranking up by comparing the average damage between mobs of higher and lower rank, and a difference of 5 THM ranks.

[Kaeko]

I feel like we may get more out of this with you posting some of your own testing rather than just outlining a ton of unknowns. May edit your original post in just a second so please don't be offended if I do.

Thank you for posting some relevant info.

[Takiwaki]

nvm

[Minwu84]

\؟مظ ةة

[Shifer]

This has been painfully obvious since beta, no idea how this is even near relevant.