Page 1 of 2

Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 10:22 am
by panino76
Hello folks,

In an attempt to create a simple emulator for the experimentation process of an item (weapon, armor, food and so on...) I started to understand how the initial percentage is calculated (that one capped at 30% at the start of an experimentation process). To do it, I started to collect some data using the experimentation of a clothing and armor crafting tool (where only CD does matter) this where the results with different resources (each line is a different resource obviously):

113 118 2%
141 147 3%
186 193 4%
196 204 4%
200 208 4%
205 213 4%
240 250 5%
358 372 8%
370 385 8%
372 387 8%
425 442 9%
436 453 10%
442 460 10%
447 465 10%
461 479 10%
483 502 11%
509 529 12%
555 577 13%
584 607 14%
595 619 15%
649 675 17%
654 680 17%
724 753 19%
756 786 20%
899 935 26%
903 939 26%

The first column is the CD on each resource, the second column is the CD modified (with the 4% bonus) the third is the percentage I get after the initial assembly. Then, to double my results, I respecced my expertise, eliminating the 4% bonus from it.

So, I started another experimentation session, using a recycled metal (CD 200), I was expecting a percentage of 4% (see in the first attempt both 193 and 204 were giving me an initial percentage of 4%, so, since 200 is in the middle I was expecting the same result) but the tool gave me an initial value of 3%. At this point I gave up.

What is your opinion? Do you know where I could find a breakdown of the experimentation percentages and that I could use?

Thanks anyone for reading.

Re: Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 11:54 am
by Sobuno
I thought the 4% bonus actually was a flat +40 bonus? That would make your 193 actually be 226 and the 204 actually be 236.

That would make sense as the your 200 would be between a 3% and a 4%

That's what I was told at some point at least

Re: Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 12:23 pm
by panino76
Sobuno wrote:I thought the 4% bonus actually was a flat +40 bonus? That would make your 193 actually be 226 and the 204 actually be 236.

That would make sense as the your 200 would be between a 3% and a 4%

That's what I was told at some point at least
I don't think it's a flat amount, at least in the expertise it's expressed in percentage values....and I hope it's not a flat amount...I should reenginner my worksheet in that case, every math is done calculating it as a 4% increase....I should ask to a dev.

Re: Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 12:37 pm
by Hodru
I have read of other people who through experimentation have also come to the conclusion that rather than actually do 4%, they added a flat amount to the values, so if you're at 4% they add +40. Thus, instead of needing 962 to cap a resource at 1000, you actually are able to cap it at 960.

Re: Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 1:09 pm
by Sobuno
I don't know if a developer has ever said anything about it, but Zimoon was the one who told me about it first

Re: Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 1:15 pm
by panino76
Sobuno wrote:I don't know if a developer has ever said anything about it, but Zimoon was the one who told me about it first
If that's the case I need to review all the maths, thanks Sabuno, I'll make more experiments after I am done with university exams (tomorrow I have another one :/)

Re: Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 1:28 pm
by Sobuno
panino76 wrote:
Sobuno wrote:I don't know if a developer has ever said anything about it, but Zimoon was the one who told me about it first
If that's the case I need to review all the maths, thanks Sabuno, I'll make more experiments after I am done with university exams (tomorrow I have another one :/)
I have an exam tomorrow too :mrgreen:

Re: Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 1:52 pm
by panino76
Nothing to do!

I've tried with the +40 theory, but in that case I have these results:

370 410 8%
372 412 8%

000 461 10%
425 465 9%
436 476 10%
442 482 10%
447 487 10%


(Green lines With the +4 expertise bonus, the read one without it, I was expecting something between 8% and 9%)

It's weird....

Re: Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 4:38 pm
by Hodru
Are you taking into account the grade of successes as well? Perhaps the red one in this instance was an amazing, and the greens were all greats?

Re: Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 4:48 pm
by panino76
Hodru wrote:Are you taking into account the grade of successes as well? Perhaps the red one in this instance was an amazing, and the greens were all greats?
Yes, I have verified the grade at each step, and tried the ones that doesn't fit 3 or 4 times. Always the same result.

Re: Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 5:22 pm
by Kexi
Are you including the bonus to crafting stages based on subcomponents?

I.e. Advanced Armor Cores grant an overall +400 during the crafting process.

I don't know if other crafted stuff gets this bonus, but weapons will soon as well with the new GU5 weaponsmithing revamp.

Re: Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 6:20 pm
by panino76
Kexi wrote:Are you including the bonus to crafting stages based on subcomponents?

I.e. Advanced Armor Cores grant an overall +400 during the crafting process.

I don't know if other crafted stuff gets this bonus, but weapons will soon as well with the new GU5 weaponsmithing revamp.
I used the most simple item I could craft, a clothing and armor crafting tool (like in every experiment, you start from simple cases and extend your results to more complecated environment) where it's required only one type of resource, and only one stat (CD). The other resource is a recycled chemical that doesn't impact in the experimentation process.

Re: Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 10:58 pm
by Savacc
OK, as I understand it, you are trying to test the results of the "Assembly Phase", and not "Experimentation". That is, you are taking various CD coppers and trying them to see what the initial "Assembly" result will be for a crafting tool.

I wish we hadnt lost the old .com site. There was some initial work by Lunariel, back in 2003, on Assembly. Lunariel came up with the formula that I think you are looking for. Unfortunately, I dont have it. Maybe Z does.

If you are going to replicate Lunariel's work, I would recommend you first do things without the expertise. Find the pattern, then add expertise. It was actually some people doing things like your experiment that discovered that the 4% expertise bonus wasnt really 4%. The latest info we have is that it is +40. It has been tested at higher levels, 960 plus expertise does equal 1000. Wether it is holding true all down the line hasnt been proved, as far as I know. I think we are all hoping that SOE is consistant and it will be +40 from top to bottom, but I dont think anyone knows for sure.

Re: Experimentation theory: how it works?

Posted: Mon Jun 16, 2008 11:14 pm
by panino76
Savacc wrote:OK, as I understand it, you are trying to test the results of the "Assembly Phase", and not "Experimentation". That is, you are taking various CD coppers and trying them to see what the initial "Assembly" result will be for a crafting tool.

I wish we hadnt lost the old .com site. There was some initial work by Lunariel, back in 2003, on Assembly. Lunariel came up with the formula that I think you are looking for. Unfortunately, I dont have it. Maybe Z does.

If you are going to replicate Lunariel's work, I would recommend you first do things without the expertise. Find the pattern, then add expertise. It was actually some people doing things like your experiment that discovered that the 4% expertise bonus wasnt really 4%. The latest info we have is that it is +40. It has been tested at higher levels, 960 plus expertise does equal 1000. Wether it is holding true all down the line hasnt been proved, as far as I know. I think we are all hoping that SOE is consistant and it will be +40 from top to bottom, but I dont think anyone knows for sure.

From the last attempts I made, I'm almost sure (and when I say almost it means around 98%) that it's not a flat +40 from 1 to 960.

There are too many inconsistencies with that value, maybe the flat amount changes passing some gates (like +20 from 1 to 200, +25 from 201 to 400 and so on....).

What you said is exactly what I'm doing, and at this point, I think what you suggested is the best way to proceed. I hope to find enough different resources to try with.

After I'll gues what's the formula, I'll try to understand how the bonus works. On a side note, I've asked to Loche an help, hope he can find some time to give me a brief description of the way the bonus is applied.

Re: Experimentation theory: how it works?

Posted: Tue Jun 17, 2008 2:15 pm
by Wasahi
I had a similar question regarding assembly/experimentation and the resource bonus, rather than satrt a new thread I figured I'd add it here.

Say an experimentation line used an average (CD/OQ) 50/50.
And you had a resource that had 972 CD and 930 OQ

When computing the average after the resource bonus, would the formula look like:

(1012 + 970)/2

Or would CD be capped before the average was taken and so the formula would look like:

(1000 + 970)/2

This is assuming that it is a strait +40 bonus for the resource quality expertise.

It sounds like any work done previously in this area may have been lost, but I figured I'd ask.

Wasahi

Edit: Disregard the above after pouring over the Begginers guide it seems the reource bonus is applied AFTER the averages are computed.