Mind Wrack tooltip incorrect. See inside for details.
I’m not convinced there’s an actual cap yet (sample size is still very tiny), but this definitely needs further investigation.
If I find the time later on I’ll run some tests on the golems, too.
Indeed. It’s why I’m hesitating to call any firm conclusions other than the fact that tooltip is wrong, and there’s something wiggy going on. Further higher power tests demonstrate a lock around 200% for the 3 illusions wheras I’m hitting 207-209% with less power.
You know, the tooltip indicates that a 3-illusion shatter is not as strong as 3x 1-illusion shatters.
What’s this thread about?
Also, I got 4 crits on Mind Wrack last night for a total of 8,000 damage. Wait till I get more gearz ~
It’s about the fact that the tooltip inaccurately portrays the relative amounts of damage. If you read the post on my site you’ll note that I indicate the tooltips ratio at 100%/135%/175% when in reality the actual damage is 100%/155%/200+%(roughly)
I’ve been finding out that a lot of our tool tips are inaccurate, so it doesn’t surprise me that Mind Wrack is off as well. However, I’m not convinced that there’s any diminishing returns on higher power. The results are “close enough” to the 100%/150%/200% that we’ve been seeing since BWE3.
You can find that old thread on GW2Guru here:
http://www.guildwars2guru.com/topic/43027-gigantic-consolidated-mesmer-changes-thread/page__st__90#entry1630406
As a general note, I don’t mind the damage being higher than the tooltips. Just saying.
As a minor sidenote: based on the logic of the scaling to “multiple” illusions, the 100/135/175 seems counter-intuitive. A 100/145/175 would make more sense.
Generally, most of the tooltips are a bit off, so. Shrug?
(I mean you’ve seen the iWarden tooltip damage right? Even if you x12 it, it’s still rofl.)
“Close enough” is not a particularly useful term when you’re looking at data that is demonstrating a possible diminishing returns. You say you’re not convinced and that’s fine. I haven’t provided remotely enough data to convince anyone yet including myself.
However the data is what it is. One importance is in 8 sets of data with 15 events in each I not once averaged below 152% for the 2 illusion shatter. That is statistically significant and strongly suggests the ratio is not 150% but higher, more likely around the 155% mark. Simple common sense says that if the true value was 150% I’d be getting both above and below that mark. But only above? That gets exponentially less likely as you do more tests and I did quite a few. Still possible? of course but really not very likely at this point.
The 200% mark again is true for higher data. In 4 sets of high end power with 15 events in each I never deviated more than 2% from 200%. However in 4 sets of low end power I deviated by a massive 7-9%. This again is statistically significant.
They’re not close enough. They’re clearly demonstrating different numbers. The question isn’t whether or not the data is correct. The question is whether or not there’s enough data for it to have been wildly skewed.
I’ve seen the thread you linked and it is part of why I started to do some actual testing. There is no real data that demonstrates any of the conclusions drawn there. Whether or not the data actually exists is not known to me, but it’s not public and when conclusions differ from my own data it is not very helpful in its current state. Not only that but the thread is relating on a set of variables that my data doesn’t cover, specifically in relation to Illusionary Persona and so it’s just not a very good comparison at all.
TL:DR? Their conclusions don’t appear to have any numbers behind them, and contradict my data in front of me. Doesn’t mean I’m right and their wrong. But it means it needs a closer look.
As a minor sidenote: based on the logic of the scaling to “multiple” illusions, the 100/135/175 seems counter-intuitive. A 100/145/175 would make more sense.
Generally, most of the tooltips are a bit off, so. Shrug?
The tooltips are only off in real terms, but not usually in mechanical terms. The numbers on tooltips need to be understood as relational figures rather than real ones. The 88 damage on Cry of Frustration doesn’t mean 88 damage. It’s a value which represents a more complicated set of math. Simply put 88 damage would be 88% of the value of 100 damage and this is why it’s important because on Mind Wrack the actual tooltip values are -wrong- in proportion.
As for counter-intuitive? Yes. But that’s the ratio they put so… shrug.
Well, my Cry of Frustration doesn’t do “88%” of much of anything, since it crits for 600, hitting for around 250-300. I’m not sure what stat I have anywhere that is “300-350”.
It’s clearly only a representation of scaling, however the representations are all inaccurate at any detailed level.
This is very clear when you compare Phantasm tooltips. iWarlock has some sort of absurdly low tooltip, even though it hits for the same ballpark iB/iS. iS’s tooltip is strangely low. iB’s tooltip is actually higher. And, they are incorrect relative to each other as well. And then you compare a Phantasm tooltip to a regular attack (like Sword4 counterattack), and they are completely off.
So, all that said, I don’t see the relevance of the detailed values or relations of the tooltips. They’re erroneous across the board to one extent or another.
It’s significant enough to say the tool tip is wrong (which we knew), but the data set is way too small to make the claim that there’s diminishing returns as power goes up. The skill is random to begin with, so a variation of 10% would not be the least bit surprising to me.
There’s also things in your analysis that suggest problems with your math. You claim a p<.05 result, but didn’t even bother to report the standard deviations. You report some of your damage results to 6-7 significant digits, when there’s clearly not that much precision to the results. You don’t even mention if you assumed a normal or uniform distribution in your calculations.
I think I’d rather ask Anet to alter the damage listed on tooltips, so that it makes more sense in general, as per what Easymode was saying, than ask them to make this one tooltip make sense within itself.
That being said, it’s nice to see someone actually doing tests on something instead of ranting about how weak an ability is or how OP an ability is without any data on the subject.
Leader of The Harbingers of Serendipity [LIFE] : Fort Aspenwood