I have BLM73 with both WHM & RDM subs on Asura, and would be happy to help collect data. if anyone has two ppl and need one more, PM me please^^
cheers ~
I have BLM73 with both WHM & RDM subs on Asura, and would be happy to help collect data. if anyone has two ppl and need one more, PM me please^^
cheers ~
Incredible work, nice testing (as a research scientist, I approve of the methods).
I have done some research myself concerning the inner "clock" of FFXI (in connection to crafting mostly) and am convinced that everything happens on the scale of a "tick" rather than seconds, although I don't have the proof yet. The big problem is ticks in FFXI are unequal, although they do follow a set pattern.
The pattern is directly related to RL seconds, in this manner:
1st tick = 2 seconds
2nd tick = 2 seconds
3rd tick = 3 seconds
4th tick = 2 seconds
5th tick = 3 seconds
For a total of 12 RL seconds per pattern. This then repeats ad infinitum. This might be common knowledge, but I was wondering if the VE values could possibly be viewed in this context. There is an inherent approximation factor since you are measuring time in an internet app, however, this ticking corresponds exactly to the RL clock and never varies. Maybe the ticking could be used as a precise way to measure VE values, with some sort of plugin?
Also another thing bugging me is how you mention that a mob chooses his next target only at specific times. So the mob is polling its hate list to choose the higest target, possibly at each tick? Assuming -60VE per RL second, the aforementioned ticking pattern, and assuming mobs choose a target only on each tick, this could possibly skew VE value testing by as much as 119 during small ticks and 179 during large. Unless of course, VE values are also only assigned at tick boundaries.
I'm probably talking out of my ass here, but I figured I'd mention it. I don't have the means to test this myself, but maybe someone else will.
There is that whole part where haste gear/spell affect recast you completely left out, which can upward double /DRK's enmity generation.Originally Posted by barber2006
Went out and did some my testing with another ls member (Mashu) and a dbox'd char a couple days ago so thought I'd add what we found. Some of what I will say here has basically already been said so just take it as added confirmation and nothing more.
I followed Kanican's testing setups as best I could with a few differences. Since my char is whm and I have -ENM merits it was somewhat of a pain to counter it with ENM+ gear as there isn't much a whm can wear to push ENM up. What I did instead is equiped the RDM char with -ENM. Thus almost all testing was done with both chars at ENM - 4 (not perfect but still good for basic testing). Also no windower so no distance plugin or any of that, I did my best to make sure both testing chars where same distance from mob (basically right up against it) for all testing.
Started with basic confirmation of hate balance (10 dispells from one char, then 10 dispels from other, then cure 1's for 0 to confirm that CE's were equal). This was more or less just to confirm to ourselves of the basics (always nice to see things with your own eyes). After that we did some logging/zoneing tests. Basically any time the person with hate zoned/logged while they had hate CE was set back to 0. If the person logged/zoned when they didn't have hate they retained full CE and would pull mob back with a single cure 1 for 0 hp.
This goes along with what someone said eariler about hate resetting only if mob looks for you but doesn't find you. Didn't try, but the following would help confirm that statement (unsure if they tested this or was their general observation)
- First player gets atleast 2 CE more then player 2
- 2nd player uses provoke and a Cure 1 for 0 HP
- First player zones and waits for the provoke VE to wear off
- First player zones back and does a Cure 1 for 0 HP on player 2 (does mob turn?)
As we continued to test things we did notice one thing.
- Player A and Player B both dispel increaseing their CE higher then puller and equal to eachother
- Puller casts something that has low CE and high enough VE to pull hate (bar's are enough when added to the initial puller hate)
In these cases when the pullers VE drops sufficiently that they lose hate the mob will turn back on the other players. The player the mob turns back on alternates. If the mob was attacking player A, then after the puller casts the bar spell and loses hate the mob ends up attacking Player B. This alternate targeting was mentioned in the resting hate stuff posted not long ago but I think it is an important effect. Once we saw this we often used this effect as an added confirmation that hate was balanced between the 2 testers. This also makes me wonder if all actions also produce a minimum of 1 VE. If that was the case, rather then mobs switching targets in cases of a tie, they are acutallly switching targets becuase CE's are even but the person who did last action aslo has 1 VE.
One thing you can use this target switching effect for is to narrow down VE's down to exact vaules with the following type of setup.
- Player A and B builds CE to 1800 or 1801 CE
- Player C (can't use puller unless you reset their hate) uses provke
- Does mob switch target (between Player A and B)
This helps avoid the issues that Kanican described when VE testing.
Now in the case that you must exceed other player's total to pull hate (rather then tie) then a target switch would occur only when player A and B have 1800CE. This is becuase the initial provoke is 1 CE and 1800 VE). If the switch also occurs at 1801CE then you would expect that hate is pulled on ties. Ofcourse this goes under the assumption that provoke VE is indeed 1800. Again I never actually performed this test it's just a proposed testing method.
Next we did a very quick rough look at pulling hate. Just used a known VE hate (barspell I think) and timed hate switch. Not an exact number ofcourse but it seemed to be around 240CE or so. Do we have a correctly measured value for this yet? Would help in testing if we know this exact value.
Anyways after that testing the WHM in me wanted to test cures a bit, the ENM-4 did make getting good numbers a bit more annoying but they should be resonably close.
Note: first ratio is with tested CE, the second is my calculated base CE (if my ENM had been 0). The hp of person getting cured was about 1200 I believe (I didn't think to write it down at the time).
Cure 1 -> 24CE/36HP -> 25/36 -> 0.694CE/HP
Cure 1 -> 22CE/32HP -> 23/32 -> 0.719CE/HP
Cure 4 -> 299CE/429HP -> 312/429 -> 0.727CE/HP
Cure 1's numbers are so small they are subject to bad rounding errors but the ratios between cure 1 and 4 are fairly close to eachother so the cures 1-4 might use a set ratio. We also did some basic VE testing by timing hate change. VE numbers for cures are quite large, but one odd thing (and I think someone mentioned this before) is that cure 4's CE:VE ratio seemed higher then cure 1's. Cure1's CE:VE seemed to be about 4-5:1 and cure 4's seemed to be up around 6-7:1. When testing VE from cure 4 I was surprised how long it took for the mob switched targets. Our roughly calculated VE from the 429HP cure was about 2000.
Another quick test showed that using echo drops seems to produce 1CE (unsure how much VE it was a fairly small amount).
Then we switched the RDM to a PLD to test for possible VE cap (and wanted to see the sentinel effect). First thing we did is tried a basic sentinel and timed VE. Came out to 1800, but unknown if the 100ENM is actually affecting sentinels VE or not. To test this would need a COR to reset abilities so it could be used 2 times in a row. Anyways we then tested sentinel and provoke and timed that. Since 2nd player had about 300 CE (5 seconds of VE decay) the expected time before hate switch was 55 seconds (with no sentinel bonus) and 85 (with sentinel bonus). It turned out to be 85 seconds helping to confirm the sentinel bonus. Next test was an attempt to find a VE cap (I was expecting 10k to be a cap). To do this we simply had PLD spam high VE abilities with low CE.
WHM -> dispell (-4ENM) -> 307CE
PLD -> Sentinel -> Invincible -> Provke -> Flash (1800 + 3600 + 14400 + 2400?) -> 22200VE. I think flash is around 1200 we had did basic tests but I just can't remember now.
If there was a cap at 10000VE you would expect a hate turn in slightly less then 166 seconds (2mins 46 seconds). If there's no cap it would be slightly less then 6mins 10 seconds. So off we went and spammed the PLD abilities. Ya the PLD held hate for a long time, but never turned at 2min 46 seconds. For whatever reason after about 5.5 mins I came to the realization that we had completly messed up this test. Flash's CE had been also doubled thus the mob was never going to be turning back. I threw in a quick dispel at about 5:50 or so and the mob did turn. At that time the PLD should still have had about 20 seconds worth of VE (about 1200) so the dispel shoudn't have turned the mob. So it does seem that there probably is a cap (I still think 10k) but I messed up in my testing. I tried to use cor 2hr to get pld 2hr back but no luck. It was late so I didn't try this again but it is a very very easy test to perform.
Anyways that's about all I got, thanks very much for your hard work in doing these Enmity tests, loved reading your journal posts. Hope people find my ramblings useful.
/sigh
Here we go:
-----------------------------------------------
http://wiki.ffxiclopedia.org/wiki/Talk:Enmity
I would like to state that a major part of the reason we'd like to discourage much outside use and citation of a lot of this info at this moment in time is because the testing is very very incomplete; and at this point we would like to keep it within BG because we trust the community here in being able to help us continue tests. There have been so many similar problems and testings and equations worked out that started here, in this community, because it's a player base that puts a lot of emphasis on knowing this sort of information.Originally Posted by FFXIClopedia Users
When we are more complete in results, I do plan on typing up a Wiki article but right now we're tracking everything on Kae's blog. At this point he wants to keep it there so that information remains in one place and corrections can be made etc, and bad information doesn't get copied somewhere and then not corrected. I'm sure everyone can appreciate that.
Regarding Clopedia vs BG issues, I guess that will come up more later on. It's not an issue I wish to even happen; it's not being "juvenile" in trying to withhold information but more of a keeping a tight leash so that things don't get bastardized and perhaps misconstrued in a medium where we cannot correct it. I already made an account on allakhazam to enter the thread that was started there and was very very quickly spiraling out of control and completely missing the point, and there is no way we could possibly keep on top of that on every forum and in every information guide.
Thank you, though, for at this point in time, respecting Kaeko's wishes.
I knew I wasnt going crazy.. I would always look at the game clock and wonder if it was ticking out of sync, or what. Good to know it's on purpose and there is some sort of repeat cycle and not random.Originally Posted by Freysi
1 Vana'Diel day is exact 3456 RL seconds. That'll make each game minute tick 2.4 seconds, which follows your rounded tick pattern.
So in case I missed something/misunderstood, there have been no tests yet on if enm+/- gear effects VE the same way it does CE?
Yeah, except as I mentioned before, the ticks are not equal in actual length. They are not 2.4 seconds each, the are in fact either 2 or 3 seconds. The 5 tick pattern is 12 seconds total and repeats itself exactly 288 times per in game day (24 hours * 60 ticks)/(pattern of 5 ticks) = 288. Also, 288*12 = 3456 seconds.Originally Posted by VZX
Originally Posted by Bagel
I guess depends what you mean about "same." Right in Kae's blog there are tests regarding +Enmity gear raising spikes on VE. Or if not there, I know I've mentioned it a few dozen times. >.>
Sorry I haven't updated or posted any replies, I was out all night and crashed.
I haven't read through all the PMs and posts here, but it seems like there is some really great data. The ones about the internal clock and VE cap are ones I definitely need to look further into. The cure curve one is pretty good as well! I'll try posting replies when I get the chance to read through everything.
Also to dispel any confusion because I wasn't as clear as I should have been... The test for Enmity Gear was "Effect of Enmity Gear on CE". It did not test anything about VE. We've done that but haven't posted it yet - Enmity Gear has the same effect on VE as well - spikes it but does not change the decay rate.
Making Mierin understand where you are coming is about as fruitless an endeavor as you're likely to find. It's small consolation, but I think that you folks are in the right. It's a shame that so much effort has to be expended for damage control, when it could be more effectively applied towards on-topic discussion and more research.
I guess I should say something on topic... I like hearing that +enmity impacts VE without touching the decay rate. That answers the age-old question "should I macro in +enmity for Provoke?", since people who lean on the voke button with +enmity will have net positive total Kaekos compared to those who don't, and are simply breaking even.
Interestingly, and I know I mentioned it briefly in the Alla thread.. and maybe here... Foe Sirvente slows the decay rate and doesn't affect the spike. :DOriginally Posted by Liselle
I personally have nothing against old WIki - Ashira is right, this is completely an issue of releasing bad information. Everyone is really excited and want to jump on those early numbers we've come up with, but the fact is this is not something intuitive that just "makes sense" to test. If you jump the gun and post numbers 99% of the people reading them WILL believe it unconditionally. This can be really bad for everyone since the information we've provided needs fine-tuning. There are a lot of mistakes that get back-corrected on there.Originally Posted by Ashira
When it is all/mostly completed and VERIFIED, then we can talk about what databases this goes on. I would not personally put the time in to do something then be an asshole and refuse to let people use it. There has to be 1 source to document mistakes - this is why the information needs to be limited to certain places right now.
I consider myself a BG forum member and a part of the BG community. In addition to getting some really great information for gameplay over the years, I also feel this was the place that originally popularized my LJ, so that is why I say BG can go ahead and use the information freely. This does not mean oldWiki shouldn't have access to the information in the long run though. The article written up there is fine in my opinion as it is only a summary and there's really little chance of anything written there being wrong. He asked for my permission actually then asked me to verify if it was ok. I said it was fine so it's not really an issue. I think a couple of the oldWIki mods just have it out for BG and immediately jumped on it when they say "BG" mentioned.
Hasn't stopped them, before...While it is sudden, it's a great disservice to the users to leave blatantly false and now proved to be unclear information on the article.
When it comes down to it, otherwiki bandwagoners and their admins like Mieren and Chrisjander are just full of themselves, blindly lashing out without any reason other than to spite. If they want to misuse this information and take it for granted, they will. Unfortunately, not much can be done about it other than just to express your displeasure. However, at least Chrisjander is making some sense. If they're going to end up using it at some point, whether you like it or not, they may as well credit you. You're lucky he was offering that much.
While some of the admins seemed a bit like they were lashing out at BG, the actual poster of the article has been very gracious about making sure nothing on there isn't "acceptable" to me. Unless another person or one of the more gung-ho admins like personally tries to edit that article, I don't see any harm being done regarding citations and "stealing" information at this point.Originally Posted by Khamsin
Well, true, Chrisjander in particular was very reasonable. So he probably didn't deserve what I said. However, certain other admins like Mieren and non-admins like Charitwo do seem to have an anti-BG sentiment when it comes to the wiki issue. They'll go out of their way to prevent or remove any sort of crediting to BG where possible.Originally Posted by Kaeko
Originally Posted by Khamsin
Khamsin, I think the hauberk in your avatar is the new color skin! IS IT PINK OR IS IT RED OR IS IT MAROON OR IS IT BURGUNDY OR IS IT WINE...
(if you don't get that, go download and listen to this week's ROCL NEWS NETWORK! :D)
Sorry, I just saw it and it was all pinky and it reminded me of that. :D
AFK!
Ok, fine, enough of the wiki issues. lolOriginally Posted by Ashira
Can't believe nobody caught that.When it comes down to it, otherwiki bandwagoners and their admins like Mieren and Chrisjander are just full of themselves, blindly lashing out without any reason other than to spite. If they want to misuse this information and take it for granted, they will. Unfortunately, not much can be done about it other than just to express your displeasure. However, at least Chrisjander is making some sense. If they're going to end up using it at some point, whether you like it or not, they may as well credit you. You're lucky he was offering that much.