First off I really appreciate the work. I have always wondered about this. you expected 9.1% and got 8.1%. Pretty close in my book tbh. Not enough difference imo to really call it unfair. It’s actually better than what I thought 600 was.
1 Like
Makes sense that heavy has worse rolls than light, have to make it hard as possible for tanks to get a good set, asmodeum is used most is very recipe and so is the most expensive and hardest to come by gotta keep you grinding the tread wheel for that one in a thousand roll that is actually a good.
I never understood why we can’t choose the perks for bags and tools, letting only the GS be random. Those arent game changing for PvP, there is no point in make the so expensive to make.
How we are still fighting for BASIC THINGS in the PTR 18 months later is so mind-boggling.
1 Like
Right, but at what point do you say the difference is big enough to be significant? That’s why we have to apply the binomial distribution to see the probability of the results by random chance.
So it tells us that the odds of getting 349/4300 or fewer (382 expected, -33) at 1/11 has a 1.2% of occurring by pure random chance, which is below the 5% error threshhold.
The important thing here is the sample size, since that helps reduce random deviation. At 1/11 the odds of getting 4/100 are actually higher than getting 349/4300, even though the former has a larger % difference from the expected.
Sorry dude, but these tests aren’t representative of the crafting community.
No one crafts without selecting max azoth.
Failure to do this values all this data pretty useless.
Can guarantee that the results would have been better if you did this. Don’t know how much better. But they would have been better.
Azoth only affects perk slots, not GS. You can still get non-gold 600s.
1 Like
My bad did not know this.
My point still stands though.
Most people are crafting only 595-600 and chasing Legos.
Going 590-600 and counting 600s is a bit redundant - no one crafts like this.
Yes, the point was to test if 600 is equally weighted across the window i.e. 1/11 or if it has lower odds than <600.
595-600 can’t be tested until next PTR goes live.
1 Like
Bags require dramatically less azoth, and the 600 chances for those seem even more abysmal, I’ll be doing these tests on the PTR with those to see what the odds actually are.
What is the point inknowing the odds of 590-600 crafting is what I want to know.
It doesn’t have any application, because everyone does 595-600.
It’s like me rolling 575-585 and counting how many 585s I get, just irrelevant data.
While I do not always agree with everything you post, this is some EXCELLENLY SUPERB research done to clear up confusion about crafting odds. Thank you for your efforts, they are GREATLY appreciated! 
1 Like
Did you, perhaps, miss this part of his post?
Or maybe this part?
I disagree. I have crafted 100x 595-600 Bags across two characters on two servers and the results were 15 legendary bags so just 1.67% shy of expected values. Worth noting I only made 3 out of 30 attempts in my first batch, followed up by 7 out of 30 in my second batch and finally 5 out of 40 in my 3rd batch.
Similarly, out of 400x 595-600 gathering tools (100 of each type) I’ve ended up with 61 legendary tools or 15.25% - again 1.4% shy of the expected values.
Always crafting with max Azoth for what it’s worth - and that means spending the extra 15 or w/e it is of wasted Azoth to guarantee a 4th perk even though I am forcing one with a perk modifier because who knows.
I don’t have nearly as many 595-600 armor crafts logged but of the 54 I have crafted (mostly with golden scarabs) I’ve made 9 legendary items which curiously enough is exactly 16.67% or 1 in 6. 
Because it is useful to know if crafting is uniformly distributed across all possible gearscores or if 600/max is lower weighted, and if so, what the actual crafting distribution looks like.
Look I welcome the research, you put alot of time into this but it doesn’t really help anyone.
If only you managed to do it on 595-600 
That’s fine if you do not think it is useful for you. I don’t think you’re getting the point of the post.
basically his results showed you can roll anywhere from 1-8 on a six sided die…that’s not unfair?
pretty sure anyone who disagrees is missing the point.
Love your post, love your OTHER post, and love how you stand up to everyone in here who needs schooling on confidence intervals. As someone else put it, „doing gods work“, keep up the good shit 
This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.