Kha wrote:http://www.craftsage.com/probability.html

You are welcome.

This is so becoming my signature.

guess my 99 quals hade the same chance as mps today.

Edit: bottom one is where I started. For me, it's not even an discussion atm. 3rd time this happend.

Not sure, I think you're reading too much into it...

You had c. 200 crafts in those screen shots (if i'm reading them correctly).
You achieved 7 MPs (3.5%) which is pretty good assuming there is a 2% chance.
But as you say above, you believe there is an issue with streaks, not overall % chance.

To demonstrate some other possible streaks distributions for you I went to http://www.random.org.
I generated 6 lists of 200 random numbers between 1 and 100 (inclusive).
Using those 6 sets and your screenshots I counted the streaks of non MPs.

Here are the results:

*where you see an orange line, an MP was achieved.

I think it demonstrates that there is nothing too unusual about the results you show.
Within just 6 data sets we see long and short streaks, as we do within yours.
Random Set 1 seems remarkably similar.
Random Set 6 is probably the most evenly distributed, but as you can see looking at the other sets is not common to see that.

What do you think is so strange about the distribution you have shown there? Or perhaps i'm missing your point?

*Assumed a random number of 1 or 2 an MP result within the sets of random number.
*Happy to post the raw datasets if you like, or generate your own and do the same exercise. Takes a lot of space though.

Lurker wrote:Not sure, I think you're reading too much into it...

You had c. 200 crafts in those screen shots (if i'm reading them correctly).
You achieved 7 MPs (3.5%) which is pretty good assuming there is a 2% chance.
But as you say above, you believe there is an issue with streaks, not overall % chance.

To demonstrate some other possible streaks distributions for you I went to http://www.random.org.
I generated 6 lists of 200 random numbers between 1 and 100 (inclusive).
Using those 6 sets and your screenshots I counted the streaks of non MPs.

Here are the results:

*where you see an orange line, an MP was achieved.

I think it demonstrates that there is nothing too unusual about the results you show.
Within just 6 data sets we see long and short streaks, as we do within yours.
Random Set 1 seems remarkably similar.
Random Set 6 is probably the most evenly distributed, but as you can see looking at the other sets is not common to see that.

What do you think is so strange about the distribution you have shown there? Or perhaps i'm missing your point?

*Assumed a random number of 1 or 2 an MP result within the sets of random number.
*Happy to post the raw datasets if you like, or generate your own and do the same exercise. Takes a lot of space though.

How about the 99% quals?

For the record I don't keep track of my crafts (SC or otherwise) because that is just madness. I swear there are streaks, positioning matters, directional orientation can change it, Changing houses or going to the city and the list goes on. But I know this is just psychological to make myself feel better for statistical variations. I think randomnesss and probability have been covered already so lets not go into that again. Just remember each new craft is an independent roll and doesn't matter if it is your 1st attempt or 10,000th attempt.

The 99% quals? What is it you mean? The low quantity of them? It's fairly unlucky, i'll gant you that.

Because I'm bored and waiting on a realm switch....
I wrote a quick C# application that does the following:

1. Simulates 200 crafts based on the DoL pre 1.87 formula.

2. Runs some analysis of each set looking for:
a) long streaks ( >= 8 gems of the same quality in a row )
b) long gaps ( >= a string of crafts longer than 70 where we don't see a particular quality )
c) low occurrences (less than 8% of a particular quality crafted across the set of 200)

Here are the results from the first 30,000 sets of 200 crafts are here:
https://ibb.co/nfdhwQ

It seems that across any given 200 set...
- Getting as few as 12-14 gems of a particular quality being crafted is not too uncommon.
- Long stints of 70 or more where you don't see a particular quality occur fairly regularly.
- Streaks of 8 gems of the same quality are seen multiple times.

Roughly we see about 60 occurrences of these 'interesting' cases across the first 30,000 runs.
So every 1 in 500 sets of 200 crafts that happens on uthgard someone will experience one of the statistical edge cases.

It's worth noting that if i relax the reporting criteria even slightly there are so many results it just gets silly common.
e.g. Streaks of 7 happen extremely frequently

This was all far too much effort really! But i started and thought i may as well finish.
I'll leave it be now and let you draw your own conclusions.

* Item quality code from DoL https://github.com/Dawn-of-Light/DOLSharp/blob/master/GameServer/craft/AbstractCraftingSkill.cs
* Random Number Generator https://msdn.microsoft.com/en-us/library/system.random(v=vs.110).aspx
* You can see my code here: https://pastebin.com/ya2YMetX

Edit: Typo