Hey there, I was wondering if you could take a look at something for me and let me know what you think. A Korean Barb player just let me know that somebody in Korea had done some testing and determined that the Frenzy breakpoints listed on Reddit, and thus on D3planner as well, are wrong.
I took a look and the work seems pretty solid, but I wanted another pair of eyes on it. I know this seems like a topic of fairly narrow interest, but if this is wrong, who knows what other speed-based breakpoints may be erroneous as well.
The page with the analysis of Frenzy rate is here: http://www.inven.co.kr/board/diablo3/2737/604576
Itās in Koreanā¦ I ran it through Google Translate, which was somewhat
helpful. Thanks in advance for your helpā¦ feel free to draft in anybody else you think might be interested enough to care, and who has a good knowledge of game mechanics.
My big issue here is that they donāt show their gear. The entire basis for his claim is that the first bunch of tests with what is supposedly a 1.4 sword show the base FPA for Frenzy is a different value than supposed.
Butā¦ all his discovered breakpoints would match up perfectly with 57.1446 if he was using a 1.4*1.06 sword.
To me, without him showing his gear, the more parsimonious explanation is that he mixed up some video files and he was actually using the wrong sword. Whatās the simpler explanation? That the base FPA of a skill changes (very conveniently by an amount that is exactly the same as a possible attack speed roll on a weapon, not 1.0542, not 1.06138, but exactly 1.06) depending on the weapon attack speed? Or that he messed up and was comparing the wrong videos.
Itās easy enough to make that mistake, I did something similar when I was doing the vault testing and had to retest a bunch of stuff.
He does a crapton of testing with a single sword (supposedly 1.4) to suggest that the base FPA is 53.91 (=57.1446/1.06). When instead all of those results could be explained by him accidentally using a 1.4*1.06 sword, and all those results would be consistent with 57.1446 base fpa.
The entirety of his claim is based on a series of tests with a single weapon and then varying attack speed on gear. The few tests he did with other weapons seem to be consistent with 57.1446. Seems like the simpler explanation is that he messed up and used the wrong sword and/or analyzed the wrong files?
Itās easy to do if you donāt keep them organized. Iāve made that mistake before. I had like 50 videos for my vault testing. Itās easy to mix them up. So I started quickly opening my sheet stats at the start of each video so that I always know what setup I used for each video.
I could be wrong, but Iād say thereās a simpler explanation than a varying base FPA and breakpoint tables. Without further testing and without showing his gear, his test doesnāt seem convincing enough. Thereās a simpler explanation: itās just that he used a 1.4*1.06 sword instead of a 1.4 sword. He could be right, but it just seems weird to me. Without further testing, Iād lean more towards that his data could be explained by him making a mistake with the sword.
Thanks so much for taking the time to look that info over. And you make an excellent point about the possibility for error. I had pretty much decided that duplicity was unlikely (since it seems like an awful lot of work to do, just to make it seem like the breakpoints for a single D3 skill are slightly different).
But, having made that assumption, I was then focused on looking for mathematical problems, so I neglected to consider the possibility of good old-fashioned human error.
Iāll write back and see if I can get further clarification. Thanks again.
Ok, I ran a test, and Iām pretty sure the guy screwed up somewhere. Specifically, I evaluated the claim that by going with 3 7% AS rolls on gear + 3% from Enchantress, you can get to a 7 Frame frenzy (without WOTB active), as opposed to the 8 that is shown in the old table, and on D3planner. If true, this should boost your damage by 14.28% ( (1/7) / (1/8) ). And it should take you 12.5% less time to kill a given target (or, looking at it the other way around, it should take 14.28% more time to kill the target, if you are talking about subtracting that AS roll).
I got my damage down to a manageable level, and went to fight the 2nd phase Belial, since he has consistent health in each instance (in this case, 7.458 Trillion), and summons no adds. Both tests use the exact same gear, with the only difference being that in test 1, I have an extra 7% AS roll (on my BoM), whereas in test 2, this is rolled to vitality.
Kill time in test 1 was 1ā56" (116 seconds).
If the āKoreanā breakpoint is correct, kill time in test 2 should be about 116 * 1.1428 = 132.56 seconds.
But, the actual kill time in test 2 was 1ā58" (118 seconds), only a 1.7% variance, which I think is just because of randomized damage rolls.
So, if he got that wrong, all the rest of it is probably wrong too.
Hereās the video evidence, take a look and see if I bungled anything.
Test 1:
Test 2:
Edit: Ugh, I just realized that I donāt show my Compass Rose in Test 1. It is exactly the same as in Test 2 (and you should be able to confirm this by looking at the sheet APS and sheet ASI.)
