Do you really not see a difference between measuring a computer controlled timing effect and looking at statistical variances across a population of people? Really? The only error in this measurment is human timing on the start and stop of the spell, possible in-game rounding, and lag. More trials = less error, sure. But the numbers are consistent enough to show that 100 runs is enough to factor out those errors.

I'm not sure what analogy to human testing would work, but it isn't yours. That's ridiculous.

20 casts is probably enough to find break points, but you'll have more mistakes in the exact, per cast numbers. I would not use it to catch exact casting speeds, but I think its plenty to find a dex break point.

isn't that the test they did to test the link between vaccinations and autism? People believed that one....and still do.

 

-----signature-----

Yes i have spent 2 years posting in here exclusivly. My point still stands however, if you want a proper test you need to do more than 10-20 samples and then multiply them with whatever.
My 5000dex cast was 2s and then i took that times 2000 and my cast time is 2s wtf?!? ...

Anyway it does not matter to me what so ever. They aparantly changed a lot of things somewhere becuse the old tests is aparantly invalid now and they where done for real and not 10 cast multiplied with 100 to get a good sample pool.

you again, didn't you fail enough in this thread?

EDIT: Was to offensive so i will write it in a more polite manner... Hello random american with a two digit IQ

Not sure how anyone still believes that the casting cap with 10% is 386? There is VERY noticeable cast speed differences between 391 and 392 dex when casting the same spell. Since I have not done the tests myself I could never say for 100% that they are correct, but I know for sure that 386 IS NOT the final break point for dex with 10% speed.

Simply without using any programs and sitting there with 386 dex (hell even 391 dex) and then comparing to 392 there is a very noticeable difference.

 

-----signature-----

Jocke, you are a fool.

If you did any dex break point test you'd see how simple, and easy, it is to figure out.

20 casts is plenty, when the difference between test samples is over 2.0 seconds with + - .2 seconds margin of error. Guess your reflexes are a bit slow though.

 

-----signature-----

i am from germany -> you fail -> as usual.

Aww cute Jocke is getting all defensive and resorting to 5th grader name calling


And FYI 10-20 casts is perfectly fine for tests like this dude. They aren't testing damage variance or something that you need a huge sample size from to eliminate any variance. They are simply finding major drops in cast times. Not like 1-2 more casts over the span of an hour, like 10 more casts in the span of 2 minutes.

 

-----signature-----

Just to be clear...

A 1000 nukes e.g. isn't something you need to press a qbind 1000 times to do so the number really doesn't matter in that sense, quick and dirty 20 casts will go a long way towards finding the break ofc.

I tend to do that many not just for the cast speed, but also variance on darkcarver baseline and so on. Last time I did it was when I kept getting told that wp9 was broken (which I run on runie) so wanted proof hat it wasn't (it isn't)

 

-----signature-----