QisaBluevixen Posted February 2, 2012 Share Posted February 2, 2012 (edited) Overview I had heard some people saying that Corrosive Microbes might not be working. I just respec'ed to Lethality so I decided to test out some of that tree's skills. Lethal Injectors - Increases the duration of Corrosive Dart by 3 seconds. Devouring Microbes - Increases the damage dealt by poison effects by [5%/10%/15%] on targets below 30% of max health. Corrosive Microbes - Corrosive Dart has a [12.5%/25%] chance to tick twice when it inflicts damage. I tested against level 49/50 turrets southwest of the Business District terminal on Corellia. They're nice because I can back out of range of them and just watch the ticks. I used Fraps to record each attack so I didn't have to write them down quickly (and because of a quirk with Corrosive Microbes [see that section]). Corrosive Dart usually does its damage in 5 ticks over 15 seconds. The ticks happen at 3/6/9/12/15 seconds. For a given target, each tick does one amount of damage for non-crits and a higher amount of damage for each crit (based on Crit Multiplier). For instance, on a level 50 target my Corrosive Dart did: 314 522 314 314 522 The 314s are non-crits and the 522s are crits. Incidentally, the ticks were the same on the 49 turrets as well as the 50s - not sure if it does the same damage regardless of level, but I suspect it would. I fired 18 Corrosive Darts at the level 50 turrets (108 ticks) and 8 at level 49 turrets (45 ticks - some died before the last tick). Lethal Injectors Lethal Injectors simply adds a 6th tick which does the same amount of damage as the other ticks and this was born out in every one of my test shots. Verdict: Working Devouring Microbes Since my Corrosive Dart ticks always do 314/522 damage per tick, Devouring Microbes 3 should make that 361/600 when the target is below 30% health. In all of my tests, my ticks below 30% health were 362/600. So it seems my 314 ticks are really something like 314.4 rounded down. Additionally, I never saw these larger ticks at health above 30% - even a fraction of a percent above. Verdict: Working Corrosive Microbes The description of this skill seems very straightforward. With level 2, on average we should expect to see 1 or 2 "double ticks" per Corrosive Dart. Obviously, sometimes you'd see more and sometimes you might not get any. Also, the description does not specify that the level of the target should make any difference to the frequency of double ticks. Before I decided to more formally test this skill, I was running around Ilum attacking mobs with just Corrosive Dart to see how it worked. For instance, did the double ticks appear simultaneously (yes) or was another tick added to the end (no)? At first, I thought it was completely broken; I didn't see any double ticks at all on the first 4 mobs I tried. Finally I saw one on the 5th one. So I went to my usual test area and played around before really testing and noticed that I seemed to get more double ticks on level 49 turrets than on level 50s. That's when I decided to test it more strictly. I used Fraps to record the tests because I noticed that rarely some double ticks did the same amount of damage each and the numbers nearly overlapped each other visually. I could only tell that there was a double tick by the amount of health the target lost. By recording the fights, I could do the necessary subtraction to find those "hidden" double ticks. In the end, of the 108 opportunities (i.e. ticks) on level 50 turrets, I only got 8 double ticks (7.4%). On the level 49 turrets, I got 6 double ticks on 45 opportunities (13.3%). Overall, that's 14 double ticks out of 153 opportunities - 9.2%, well below the advertised 25%. Is this enough to be statistically valid? Incidentally, I had another 24 shots in my test that didn't get recorded (forgot to hit Record) and none of them appeared to have a double tick, but I didn't include them in my results since I couldn't be sure without the video that I had no overlapping double ticks. Verdict: Working, but probably not as well as the tooltip suggests. Could use more testing by others! This test plus my other one - where I showed that Experimental Explosives didn't impact Corrosive Grenade - makes me wonder just how many of our skills are actually broken. Edited February 2, 2012 by QisaBluevixen Link to comment Share on other sites More sharing options...
TheDeacon Posted February 2, 2012 Share Posted February 2, 2012 Suppose that's to be expected of an early, one-point talent that doesn't display its numbers. Horrible news about Experimental Explosives and Corrosive Grenade, too.. uh, just rolled a Sniper going Lethal Grenadier w/ Cybertech, affectionately named Kaboom. Seems there're way too many punched out holes in thse builds right now! Link to comment Share on other sites More sharing options...
Zerikin_Loukbel Posted February 2, 2012 Share Posted February 2, 2012 Thanks for the work. Someone posted some numbers in another thread and it seems like the 25% double tick was happening at about half the rate it should have. What about devouring microbes with corrosive grenade and cull ticks? Link to comment Share on other sites More sharing options...
Toat Posted February 2, 2012 Share Posted February 2, 2012 Thank you for doing this. And yeah, Razor Rounds > Experimental Explosives for the Cull/Cluster Bombs spec. Link to comment Share on other sites More sharing options...
DestyOwn Posted February 2, 2012 Share Posted February 2, 2012 Are you sending any Tickets concerning Corrosive microbes, because this is a very big difference... Link to comment Share on other sites More sharing options...
QisaBluevixen Posted February 2, 2012 Author Share Posted February 2, 2012 Cull is nearly impossible to test without a combat log. Too many numbers flying around overlapping each other. I haven't submitted a bug yet for this or the other test I did. I really should get around to that. Link to comment Share on other sites More sharing options...
frdrake Posted February 2, 2012 Share Posted February 2, 2012 Thanks for your work on this. I appreciate the cleanliness with which you conduct your tests. Link to comment Share on other sites More sharing options...
QisaBluevixen Posted February 2, 2012 Author Share Posted February 2, 2012 FYI, I created a ticket in-game for this issue. Link to comment Share on other sites More sharing options...
Tibbel Posted February 3, 2012 Share Posted February 3, 2012 (edited) Thanks for this testing! Is this enough to be statistically valid? Well, since you asked... This test is an attribute test, which means each data point is either a TRUE or a FALSE. Let's set our null hypothesis to say: "The proc rate is 25%." The null hypothesis is the claim we are trying to test. We can never confirm it through testing, only reject it or fail to reject it with a certain confidence. Confidence is typically denoted as γ or as (1 - α) and is a value between 0 and 1. The resulting data from the test leads to a p-value, which is the chance that the actual conditions are at least as extreme as the observed. (In this case, it's the chance that you would get the test results you did if the proc rate is actually 25%.) Like confidence, the p-value is also a number between 0 and 1. It's dependent on the hypothesis, the observed data, and the sample size. (Look up the math if you care about it -- I used Minitab.) If p < α, then we reject the null hyposthesis (i.e. the results are statistically significant); otherwise, we fail to reject it. So, for your first test, your null hypothesis is: "The proc rate of 2/2 CM is 25%." Let's say we want to be 95% confident, so α = 0.05. 14 ticks out of 153 trials results in a p-value of: p < 0.001. Since p < α, we can reject the null hypothesis with a confidence of 95%. In other words: YES, your test results are enough for us to say the proc rate is not 25%. ----------------------------------------------------------------------------------------- You also conducted a second test, which tested the null hypothesis: "The proc rate is the same, whether the target is level 49 or level 50." Let's shoot for the same 95% confidence target, so α = 0.05. Using your data... Lv 50: 8 procs in 108 opportunities Lv 49: 6 procs in 45 opportunities The resulting p-value is: p = 0.355. Since p > α, we fail to reject the null hypothesis. In other words, NO, we cannot say from these results that the proc rate is different for level 49 vs level 50. Hope that helps. Oh, by the way: LFL in Alderaan Places This has been my personal guild status note since launch. You know... they say great minds think alike. Edited February 3, 2012 by Tibbel Link to comment Share on other sites More sharing options...
QisaBluevixen Posted February 3, 2012 Author Share Posted February 3, 2012 Well, there you go then. Statistically valid. Thanks Tibbel! I got a response back to my in-game ticket thanking me for my valuable input. So I presume that means it's in their bug database anyhow. Maybe I'm being optomistic though. BTW "LFL in Alderaan Places" is my guild name. Link to comment Share on other sites More sharing options...
Vermalech Posted February 3, 2012 Share Posted February 3, 2012 One question I would have to ask the op is that, Does Lethal Dose help I.probe? Or did they meant to say poisons? Kinda like experimental explosives. Link to comment Share on other sites More sharing options...
Ayestes Posted February 3, 2012 Share Posted February 3, 2012 (edited) I actually really dislike Corrosive Microbes in it's current form anyway. Enhancing RNG elements with more RNG (I'm looking at this with Lethal Injectors) creates erratic and extremely undependable results. I'd be so much happier with a talent that simply changed it so that it ticked every 2.0 seconds / 1.5 seconds with a -15% Damage / -25% Damage dealt per tick. Assuming it ticked over the full duration still, the damage would remain about the same (slightly greater then it is now) but it'd be a lot more reliable in terms of tick frequency and helping out Lethal Injectors. Edited February 3, 2012 by Ayestes Link to comment Share on other sites More sharing options...
QisaBluevixen Posted February 3, 2012 Author Share Posted February 3, 2012 One question I would have to ask the op is that, Does Lethal Dose help I.probe? Or did they meant to say poisons? Kinda like experimental explosives. I haven't yet tested Lethal Dose with Interrogation Probe. Testing crit chance is such a pain! I'd have to do hundreds of attacks before I was satisfied the test was statistically valid. And the term "periodic effects" is so vague. Corrosive Dart and Grenade are clearly periodic. Interrogation Probe really is too, so I'd assume it should be impacted. But what about Orbital Strike? For that matter, what about Series of Shots or Suppressive Fire or Cull? An argument could be made that those are periodic, but I doubt they're included. But would I be surprised to find something working that shouldn't? Ha! Link to comment Share on other sites More sharing options...
JoyProtocol Posted February 15, 2012 Share Posted February 15, 2012 (edited) Just wanted to thank you for awesome research, and bump this for greater education. From my experience with MMO's, a "period effect" is something that applies a debuff to the target that causes damage at intervals. Meaning that Interrogation Probe would be one, and Series of Shots wouldn't be. Edited February 15, 2012 by JoyProtocol Link to comment Share on other sites More sharing options...
Stranden Posted February 16, 2012 Share Posted February 16, 2012 Gonna bump this too. Guess one should send it in to the dev team as well.. Thanks for the research! Link to comment Share on other sites More sharing options...
QisaBluevixen Posted February 16, 2012 Author Share Posted February 16, 2012 I've submitted all results from my tests to Bioware. They assure me they have forwarded it to the right people, but that's probably just a form letter. Link to comment Share on other sites More sharing options...
GeorgZoeller Posted March 17, 2012 Share Posted March 17, 2012 Gonna bump this too. Guess one should send it in to the dev team as well.. Thanks for the research! We're aware and Game Update 1.2 is resolving this. Link to comment Share on other sites More sharing options...
Recommended Posts