Medic Archive
Thread: Crafting Experimentation Study (my attempt to confirm the dev's answer to our question)
Page 1 of 2
Obata
Sat Apr 24, 2004 10:47 am
#1
I experimented 100 times each with two different crafting stations and one crafting droid. All the experiments were performed with only the same crafting tool, same resources, same draft schematic (ABEC), and the same character (Master Doc) wearing +9 med experimentation clothes. I meant to take the clothes off for the study, but forgot to so I just left them on the whole time to keep that constant. I did not use any food or drink.
The numbers:
Station 1, rating: 43.68:
Amazing Success: 6
Great Success: 93
Critical Failure: 1
Station 2, rating: 26.55:
Amazing Success: 12
Great Success: 87
Critical Failure: 1
Droid Crafting Module:
Amazing Success: 3
Great Success: 92
Critical Failure: 5
These numbers certainly do not confirm the answer we just got from the devs on the effect of a crafting station's rating, though it does seem to support the statement that droid crafting modules are nuetral (essentialy a 0 rating). I'll do a set of 100 at a public crafting station next to see if it gives numbers that are similar to the droid's. This is enough to convince me to use my lower rated station when making buff pack schematics.
Zarlor
Sat Apr 24, 2004 3:48 pm
#2
I'll definitely pass that on.
You may also want to try with a negative rated station if you can get one. Then repeat the experiment using one of the stations, but using positive and negative rated Tools instead. That would help out a lot as well.
Thanks for the extra work!
Obata
Sat Apr 24, 2004 4:36 pm
#3
I was planning on doing the study with different tools as well. I'll have to wait for my guild's architect to come online to get a negative rated station though.
Obata
Sat Apr 24, 2004 7:13 pm
#4
I got more numbers. This time using the same tool (+14.79), resources, clothes etc. from the original post at a public crafting station.
Station 3, public:
Amazing Success: 10
Great Success: 89
Critical Failure: 1
I also tested a negative tool (-10.28) at the same +26.55 rated station as in the original post and got:
Amazing Success: 9
Great Success: 88
Critical Failure: 3
The results for the public station are very similar to those for the 26.55 station, though not quite as good. The numbers also indicate that the positive rated tool is better, though not by as much as one would expect.
Houdani
Mon Apr 26, 2004 7:05 pm
#5
I wonder if experimenting on an ABEC is the best choice. Even when you only put points into one category, there's still a chance that your experiment will affect the other attributes.
For example, if you put points into charges and click experiment and get a "great success" -- it's entirely possible that you really got an "amazing success" in charges but only a "success" in power. The average of these two would give you the "great success."
So, in my opinion, the best way to gauge the results of experimenting would be to do it on an item which has only one attribute ... such asan ALS, ASDS, or ACRDM ... or some other equivalent item.
Thanks for doing this. It's nice to see some people put in extra effort which benefits the community as a whole.
H.
Zarlor
Mon Apr 26, 2004 8:07 pm
#6
It's OK. Power and Charges are two seperated experimentation categories, so it wouldn;t have an impact in this case. Now if it were something like an Enahnce Pack, where Duration and Power are affected onthe same experimentation line then I can see where it may be a problem.
Obata
Mon Apr 26, 2004 8:10 pm
#7
Yeah, I agree with Zarlor on this one (not just because I don't want to repeat the work). It would be a problem if I experimented both lines at once, but I only experimented one point at a time so it could only affect one attribute or the other.
Houdani
Mon Apr 26, 2004 8:12 pm
#8
I guess I was thinking along the lines of when you have a critical failure and you lose stats across the board, even though you were only experimenting on one line. /shrug.
H.
Obata
Mon Apr 26, 2004 10:09 pm
#9
For the record, I stopped experimenting on a line if I got a critical failure that brought it to 0. I chose ABECs because I would be able to use all my points on them, and each line does only experiment on one stat. I'm rarely able to use all my points before maxxing a line, so it would have taken me much longer to do the study with CRDMs (for example).
Gooney
Tue Apr 27, 2004 7:26 am
#10
Your results seem to confirm my belief that the quality of tools and stations have little if any affect on the final product.
It still seems terribly random to me. I still get about 1 critical fail every crafting session while making enhance D's it doesnt seem to matter where I am doing it and or the quality of my crafting station.
IMO a critical failure out right at first combine is a lot easier to stomach than one in experimentation. I really really hate those....WTH am I supposed to do with a 770 Enhance D with 91 med use..../destroy item.
-Gooney
Obata
Tue Apr 27, 2004 8:07 am
#11
The data do support that a higher rated tool is better (33% more amazing successes and 1/3 the number of critical failures). As far a stations go, the droid is clearly the worst (I was very surprised by that), but the data do not confirm the statements from the devs that a higher rating gives a better chance at success.
Message Edited by Obata on 04-27-2004 11:08 AM
Gooney
Wed Apr 28, 2004 2:40 am
#12
I really think the "data" is too small a sample size to draw any conclusions. 100 isnt all that much statistically speaking.
While I do think that its interesting I dont think any conclusion can be made from it.
I make my enhance D's usually 6 at a time, I almost always have to make 9 ASDS's as I can almost always count on one critical fail at combine and at least 2 fails or critical fails during experimentation (resulting in trashing the ehance). When doing such small batches I havent noticed that quality of tools or stations makes any difference. I still get critical fails occasionally. During experimentation on Enhances one failure makes a pretty big difference.
I have noticed that some places have a higher incidence of "Amazing" experimentation results as well as fewer failures. Normaly in a research city. But thats more of a gut feeling than anything else.
-Gooney
Obata
Wed Apr 28, 2004 8:58 am
#13
100 samples at each point is enough. I do this kind of thing everyday. A typical process validation uses less data than this for each stage. If you don't see a trend with 100 samplings at each point, then there is no trend to find.
I can give one piece of advise in making enhance packs, use bespin port. I haven't noticed a higher rate of amazing successes with it, but I have noticed a drastic decrease in failures.
Page 1 of 2