Bismarck First ;3

[Squallblade]

i had a 100 bonecrafter sign a wyvern helm for me, she told me if it HQs i can buy from her for like 2m, or she sells for 10m 0.o, she didnt HQ...i didnt have 2m anyways lol

[Septimus]

After nights where I HQ 4 out of 12 synths for a level 16 cooking item (when I am mich higher than 51 over the synth), I kind of think that tiers are less important than personal "luck". And by luck, I mean that some game days the randoms are not generated in your favor.

Also, it seems that some items are more HQable than others. When I was skilling up on Tonosama Rice Balls, I normally got at least one HQ on every stack of synths.

Just my 2gil.

[Fira]

Also, it seems that some items are more HQable than others.

Thats very true.

[bait]

My evidence is that the 11-30, 31-50, and 51+ range have constant hq rates regardless of actual skill within those brackets. Tests have been performed many times. Read Alla, look for Eruntalon's HQ Mystery Solved? thread, it contains a bunch of tests he and other people performed showing exactly where these teir breaks occured, which the data also shows that hq rate at 31 is ~= to hq rate at 50.

Extrapolate this conclusion to the 1-10 range.

OK, now I have given my evidence, show yours before talking out of your ass again.

all the things ive read and results people have are something like 24% hq 31 over and 26% or something at 50, then it jumps up at 51... im not saying that tiers dont exist im just saying 50 over is a better chance then 31 to HQ so its worth it to have support no matter... you showed no evidence against this and i am talking out my ass no more then you are

[aurik]

Don't you know? 2% difference is so small it isn't significant. Re-run the test and there's just as good a chance that the results will be flipped, with 26% hq at 31 over and 24% at 51 over, and then what will you conclude? Higher skill is harmful?

There's a concept you haven't considered. It's called margin of error. Since the results fall within the margin of error, we can't conclude with confidence that higher skill within the tier actually helps.

Don't worry, you'll learn all about this when you get to 11th grade math.

[Raiko]

Aurik you really try hard to be funny don't you?

[Cloude]

HQ rate has changed since last major update. SE screwed most NA crafters that are 100 or near 100. Tier is Tier, % HQ is % HQ, it's only experimental data. Most of it is on a little bit of luck.

[bait]

Don't you know? 2% difference is so small it isn't significant. Re-run the test and there's just as good a chance that the results will be flipped, with 26% hq at 31 over and 24% at 51 over, and then what will you conclude? Higher skill is harmful?

There's a concept you haven't considered. It's called margin of error. Since the results fall within the margin of error, we can't conclude with confidence that higher skill within the tier actually helps.

Don't worry, you'll learn all about this when you get to 11th grade math.

so we can conclude from our inconclusive data, that it has 0 effect. you're smart.

[Septimus]

Don't you know? 2% difference is so small it isn't significant. Re-run the test and there's just as good a chance that the results will be flipped, with 26% hq at 31 over and 24% at 51 over, and then what will you conclude? Higher skill is harmful?

There's a concept you haven't considered. It's called margin of error. Since the results fall within the margin of error, we can't conclude with confidence that higher skill within the tier actually helps.

Don't worry, you'll learn all about this when you get to 11th grade math.

Actually, a 2% difference can be significant with a high enough sample size. Don't worry; you'll learn about that when you get to freshman statistics.

[aurik]

24% vs 26% -- it's irrelevant. You've seen tracking of over 4000 syntheses? The experiments I've seen stop after at MOST a few 100.
http://en.wikipedia.org/wiki/Image:MarginoferrorViz.png

Here, read this.
http://www.h7.dion.ne.jp/~morikuma/memo.txt

[Raineer]

Don't you know? 2% difference is so small it isn't significant. Re-run the test and there's just as good a chance that the results will be flipped, with 26% hq at 31 over and 24% at 51 over, and then what will you conclude? Higher skill is harmful?

There's a concept you haven't considered. It's called margin of error. Since the results fall within the margin of error, we can't conclude with confidence that higher skill within the tier actually helps.

Don't worry, you'll learn all about this when you get to 11th grade math.

wow you're mean

[Septimus]

24% vs 26% -- it's irrelevant. You've seen tracking of over 4000 syntheses? The experiments I've seen stop after at MOST a few 100.
http://en.wikipedia.org/wiki/Image:MarginoferrorViz.png

Here, read this.
http://www.h7.dion.ne.jp/~morikuma/memo.txt

With a sample of one hundred you probably will not get statistical significance at all. (Maybe if you had a difference of 25% or more.) One hundred trials do not have enough power to say anything useful about the sample.

But I stand by what I said; doctors would go wild for a 2% effect if the experiment has the numbers to back it up.

That's great you know what a margin of error is and you can read Japanese, but I stare at boring columns of numbers for a living. You can put your e-penis back in now.

[aurik]

"Damn, he's correct! hmmmm....I know, I'll accuse him of waving around his e-penis!"

[Almaa]

my e-penis is bigger than your e-penis

[Septimus]

"Damn, he's correct! hmmmm....I know, I'll accuse him of waving around his e-penis!"

Actually, no. You said "2% difference is so small it isn't significant." I informed you that a 2% difference can be significant depending on the test.

You are trying to say that there is no difference, yet the tests you are basing this assumption off of have N's in the hundreds. ("The experiments I've seen stop after at MOST a few 100.") I am saying that you need trials in the tens of thousands before you can come to any solid conclusions about whether or not a 2% difference is significant or not.

[aurik]

You're taking me out of context, Septimus. I didn't announce to the world than in any possible situation 2% is not significant. That remark was made in response to:

all the things ive read and results people have are something like 24% hq 31 over and 26% or something at 50, then it jumps up at 51... im not saying that tiers dont exist im just saying 50 over is a better chance then 31 to HQ so its worth it to have support

Bait says in all the things he has read and results, 50 > 31 by 2%. My response is basically "Those results don't have the quanitity of data to support a low enough margin of error to make a 2% difference significant".

[Septimus]

You're taking me out of context, Septimus. I didn't announce to the world than in any possible situation 2% is not significant. That remark was made in response to:

all the things ive read and results people have are something like 24% hq 31 over and 26% or something at 50, then it jumps up at 51... im not saying that tiers dont exist im just saying 50 over is a better chance then 31 to HQ so its worth it to have support

Bait says in all the things he has read and results, 50 > 31 by 2%. My response is basically "Those results don't have the quanitity of data to support a low enough margin of error to make a 2% difference significant".

On the same vein, there is also not enough data to say that it is insignificant.

The only way to know for sure is to do hundreds of thousands of experiments (and to rule out other factors that may bias the results) or to kidnap a FFXI dev and make him talk. I like the latter option.

[Chumm]

Aurik is right btw, E-penis aside :wink: . Tiers are tiers, it's been tested to death. Skill means jack unless you break 11, 31, or 51. Period.

[Raineer]

Ok..so I'm noob... Since when was there a tier at 11? I thought 31 and 51 was it?

[Chumm]

10% at 11, 25% @ 31, 50% @ 51. Not like 11 is all that special, but for very high lvl synths it's important to break it when possible (10% on a lvl 95 synth is pretty good ^^ *EDIT* 94 for cloth )