^^ needs sticky

 

-----signature-----

bump

 

-----signature-----

edit: deleted because I responded to a 1,5 year old post

 

-----signature-----

its still one of the most relevant posts for today's game. I used it because i was respeccing RA's on my Druid and wanted to know how much to get. Currently I went for 368 I think, which gives me enough for a better casting speed on buff shears as well as some of my heals I believe. I haven't done any testing but I trust the testing others have done, and just from looking at it I see a noticable improvement in casting speed between 368 and taking off a piece of equipment with 16 dex on it. It's not just in my head that I can see this, there is a 100% distinguishable different in casting animation, as when using 368 dex, the casting animation very often cuts off shorter than what it used to be.

 

-----signature-----

In case you haven't noticed or it hasn't been posted (or washed away with the traffic), Charplan now has a tab under the Stats screen called "DEX-Break" that will list break points & speeds! Gotta love Charplan!


It's possible that Mythic did some re-coding, but I personally think it's always been this way but was not discovered because it's not intuitive to our internal biases of how the world works. This is a digital world though, and for programmers this will make a lot of sense. Integer math, pure and simple. Values get truncated not rounded, which naturally leads to break-points rather then a smooth progression.

Maybe a single calculation was changed in the combat resolution algorithm to use integer-division instead of floats, and that alone can introduce these break points. In fact, now that you mention performance boosts Mythic mentioned, considering the amount of combat calculations a DAoC server has to make it is very reasonable to say that changing over to only integer-based math could lead to significant performance improvements over using floats. (If they were using floats before that is, which is what we don't know.)

As I said though, I can also believe that good ole fashioned human tester bias made us blind to seeing these break-points that have always been there.

 

-----signature-----

i think this is possible. plus some of the old tests where done on goa eu-servers. those where very laggy and had a strange server setup iirc. testing was so unreliable, that even huge testrows with 1k+ casts would deliver different results in different zones. (!)

 

-----signature-----

I highly doubt that. Back when 250 000 people played, I'm sure people did hundreds and hundreds of hours of testing. Mind you, things obviously wouldn't have been nearly set in stone until sometime after ToA items were nerfed (brought back into reality) with 10 instead of 25% cast speed.

Once again, i remind people that just weeks before Slajzer came out with this info, there was a hotfix, or patch notes I forget which one, that specifically said they made some behind the scenes changes to decrease the lag in NF. This might not be related at all, but the testing with cast speed is 100% clear now, and I think there's no way that everyone just had it wrong all along.

 

-----signature-----

"ou don't know what you're talking about. LOL @ trying to use statistics on something that has no variance. That's like you going to the store and buying a candy bar and telling the cashier you don't believe that the price is correct because the sample size is too small.

Then, even if there is some variance (say from random lag or some unknown wacky function) you have no idea what it is so your attempt to sound smart and say what the CI of the original tests were is 100% full of fail."

The fallacy in the above statement is that although the true value may be a constant, errors in measurement accumulate to make statistical variance a must in any experimental measurement of an unknown value. That being said:

Perhaps the test would be more informative if it were done in two sets:

Set A as already measured,
and Set B with the 10 percent cast speed bonus removed.


It is quite possible we have two tangential slopes that are useful for understanding the root(s) of the measured results. Or then again, it could be that 10% cast speed is constant in comparison to the break point tiers that the data describes.

If we accept hypothetically that Dex times Spell Cast Speed is the root calculation, with additional unknown modifiers to create a desired slope or a desired curve, then the question becomes this:

Are these results which originally were seen as a "smooth slope" turned into a "staircase' by amplifying the results of each progression along the Dex axis with 10 percent cast speed bonus?

At any rate, any harmless testing done is useful, even if only shows we are more in the dark than we previously thought. That is the true spirit of science.

 

-----signature-----

Slajzer is that possible that you could do one more test with minstrels AE mez (5 sec) cast time. I would like to see how much it is needed dex & cast speed on minstrel

My account has been closed for over a year. Identifying a breakpoint at the dex range you're interested in is done quickly though.

Say you want to find a breakpoint at 320-350 dex for example.

Test 350 dex
Test 320 dex

10 casts is enough to notice a difference if you time it properly

>>>>
Same cast speed = try a higher/lower number
Different cast speeds = Try ~335 dex

>>>>
335 is the same as 320 but slower than 350 = try with ~343 dex


Keep going until you find the breakpoint.
Then you should probably confirm the breakpoint with a higher number of casts.

 

-----signature-----