Architect Archive
Thread: Using Crafting Stations Some facts about the Functionality Rating
Greetings,
Months ago, we have seen popping up a strange value when we examine a crafting station, the Functionality Rating. Some Architects just like me were keen to find out how to make better stations. Then, some threads described pretty well the way to calculate it from the resources used. The Functionality Rating is a value between theorically -45 to +45. The station schematic (and also components) tends todescribe it asthe "Quality" of the station.
On Eclipse,bunch of Architects made mass production of high end stations with amazing ratings greater than 42 or 43FR. Those stations were sold for good prices. But alas, none of us could really guarantee anything about the results to expect from those stations. The worst is that I could rarely hear the same things from the Master Crafters using them, some would say it improves their experimentation results, others say they have more failures... I though we could have better feedbacks, find a consensus or even a correct theory. Alas I could not find the answers yet to my questions.
Twice I had serious doubts about the effects of the Functionality Rating and could not continue that business honestly withoutproofs. I decided then to make further tests on artisan items to figure out how acts this mysterious value. Those tests are the results of some grinding with more than 700 assemblies and more than 1100 experimentations. I had used several stations: -42FR -14FR +15FR +43FR. I choosed to experiment on CDEF Pistols. Assembly results were done with CDEF Pistol (Complexity 8) and Energy Distributors (Complexity 19).
You can see the results and the detailed tests following that link:
http://swg-cantina.com/pages/frating/frsum.html
The results are simply amazing. I maystatetwo facts:
- The Functionality Rating hasNO real effect on experimentation results
- High Functionality Rating (more than 30) reduces the risk of assembly failures.
I had some problems to admit the first item, but finally it explains why nobody could agree on FRating and Experimentationresults
Of course those tests are limited to two items in a single profession. It may probably be extended and generalized to some other crafting professions.
It is still hard to figure outwhich mathematical formula used the Devs to subtly mix item complexity, tool total effectiveness, station functionality rating, experimentation points andskill level for thecrafting result probabilities.I think there are still lot of work to find it out... except if the Devs and Designers give ussome clues in the meanwhile. 
Thank you for your interest and your feebacks,
Akoo Pa'anh - Eclipse - Master Architect/Master Artisan
Based on your first set of graphs, it does seem like higher end crafting stations reduce the likelyhood of critical failures. It's interesting to note how the curve shifts to the left with increased complexity. I would have thought that they would be reversed, having the item with the lower complexity item having a greater boost in the reduction of the crit fails with the high end stations. I'd like to see this test tried out on items with perhaps a higher (or lower) complexity level. Also, perhaps try some other station rating on the CDEF and EDs with values in the middle and see if they follow the results predicted by your graph.
Now in the second part, is what you are doing using 1/3/5 experimentation points at the same time? It would seem that you are right and there is only a small, non-significant deviation in the type of success that you achieve. It seems like the developers used a normal distribution to see what the outcome is. I'd like to have seen what would the result be is you had separated good and great to see if this was supported. I'd expect good to be the crest, and great to start the drop on the curve.
Still, great work =)
You deserve a Badge of Intellect for your hard work! One question I have is that it seems to me that the droid storage subcomponent has no value add to to the final functionality rating. Have you seen that as well?
I was surprised too. But I suspect the devs to have usedother factors than the only complexity. CDEF is a simpleweapon (Complexity 8), Energy Distributor (Complexity 19)a master artisan assembly. Maybe there are other modifiers to fit with the profession or the type of items.
Drem wrote:
Based on your first set of graphs, it does seem like higher end crafting stations reduce the likelyhood of critical failures. It's interesting to note how the curve shifts to the left with increased complexity. I would have thought that they would be reversed, having the item with the lower complexity item having a greater boost in the reduction of the crit fails with the high end stations.
Drem wrote:
Now in the second part, is what you are doing using 1/3/5 experimentation points at the same time? It would seem that you are right and there is only a small, non-significant deviation in the type of success that you achieve. It seems like the developers used a normal distribution to see what the outcome is. I'd like to have seen what would the result be is you had separated good and great to see if this was supported. I'd expect good to be the crest, and great to start the drop on the curve.
Still, great work =)
I had separated great and good at the begining, but I had only max 1 Good success, it is irrevelant of the real rate. I should make 5 times more test to figure out with a moderate accuracy what is the part of good successes in the experimentation :/
I had made two series of 3PTS (T2), because the first serie could let people think of an improvment with high rating. But the second serie was really at the opposite side and balanced the results.
I think thatone or two failures (or good success) on 60or 40 samples are irrevelant. That leads to a +/-3%error on the results (for 60 samples). That is basically the difference seen between the series.
Auri wrote:
You deserve a Badge of Intellect for your hard work! One question I have is that it seems to me that the droid storage subcomponent has no value add to to the final functionality rating. Have you seen that as well?
No you are wrong, the contribution of Droid Storage Compartment is of +/-5 FRto the final rating.
You can calculate it from the resource stat: OQ / 100 - 5 (using maximum experimentation)
I use an excel sheet which calculate the Functionaly Rating with an accuracy of 0.01FR from the material stat list.
I have not created the formula, I pickedit up from another thread in this forum (+43 Crafting Station...). I have madeit more accurated. According to those adjustments a station using materials with OQ 1000, CD 1000 can only reach 44.53FR. Those materials will probably never exist all in the same galaxy. But it explains why nobody could make 44FR station yet. I think I shall produce soon stations better than 43.70FR as I have all the resources required.
Akoo Pa'anh - Eclipse.
That's some good hard data you have there. I actually had started doing that youself, but now I don't have to! Yay!
Have you considered trying this as a Novice? It's possible that because you were Master you had already "maxxed out" your experimentation potential. Just a thought.
_-ArchAngel-_ wrote:
That's some good hard data you have there. I actually had started doing that youself, but now I don't have to! Yay!
Yes, I have considered to test it with my other character (only novice artisan with almost no skills). I have to find some patience to start more test batches
But I guess that the results would be extremelly interesting
JaatoWaals wrote:
I would be interested to see whether there is an effect when you use a droid crafting station (one built into a droid by a DE) instead. I've had numerous customers tell me that they get better results (fewer failures and more 'great successes') when using a droid crafting station as opposed to even a 40-rating architect-made station. I've noticed a similar effect- critical failures go way down when I'm using my droid as opposed to the station in my house (which has a 40-rating)
I have ordered a new droid from my favorite Master DE. I have also provided high quality materials to maximize the Mechanism Quality -Eatsob Copper (CD 961 / OQ 950) and Biba Gas (OQ 988). The DE excepts to get a value of 5 for the Mechanism. No idea if this value is another Functionality Rating.
I shall add a serie of tests with the Droid Station.
This is great work, however, there is a rather large omission from your results. Your Stanard deviation. Even with a 1100 datapoints, there is still error in results due to random numbers. My guess, based on your experimentation results is that your standard deviation is somewhere between 2-4%, which would mean that you have to give some leeway (up to 4%) for any of your results.
IE, it could mean that in actuality there is no effect on assembly in making the crafting stations.
More datapoints would be great.
-Jubey, Master Architect
Ashii wrote:
This is great work, however, there is a rather large omission from your results. Your Stanard deviation. Even with a 1100 datapoints, there is still error in results due to random numbers. My guess, based on your experimentation results is that your standard deviation is somewhere between 2-4%, which would mean that you have to give some leeway (up to 4%) for any of your results.
IE, it could mean that in actuality there is no effect on assembly in making the crafting stations.
More datapoints would be great.
-Jubey, Master Architect
Indeed, I had the same feedback. I shall add it.
Lakys wrote:
I was surprised too. But I suspect the devs to have usedother factors than the only complexity. CDEF is a simpleweapon (Complexity 8), Energy Distributor (Complexity 19)a master artisan assembly. Maybe there are other modifiers to fit with the profession or the type of items.
What could it be then? The only two stats you gain as you go up the Artisan tree are Artian Assembly and Artisan Experimentation. Both of these remain constant throughout your data, therefore your data will not show it. You could perhaps get a novice to perhaps go through the same run of experiments as see what the resulting graph is. I would expect the shift to be upwards, thus raising the number of failures, but not much else.
I had separated great and good at the begining, but I had only max 1 Good success, it is irrevelant of the real rate. I should make 5 times more test to figure out with a moderate accuracy what is the part of good successes in the experimentation :/
I had made two series of 3PTS (T2), because the first serie could let people think of an improvment with high rating. But the second serie was really at the opposite side and balanced the results.
I think thatone or two failures (or good success) on 60or 40 samples are irrevelant. That leads to a +/-3%error on the results (for 60 samples). That is basically the difference seen between the series.
I've read somewhere (can't remeber exactly; perhaps instruction manual? Website?) that good is actually at the bottom of the results, therefore making average better than good. Since I'm a droid engineer more so than anything else, I rarely look at the end result to see which is better. Good better than average makes more sense, but nobody said this games was logical
. If this where so, then good would fit in the bell curve. It still stands though that station does not help with quality, but only with failure rate.
I did my tests on four different stations. I can't remember the exact FR but I was in the neighborhood of -7, 0, +14 and +43. Not only did I not notice a difference in crit fails when crafting using the same materials (I wasn't really testing the stations so much as trying to figure out how crit fails work for Architects) but I found that I could nearly recreate an 80% crit fail rate on the +43 station just as easily as the -7 station.
By using the absolute worst quality ore with the lowest Malleability and Unit Toughness stats I found I could crit fail a structure module in about 8 of ten times on any station. I switched to some great quality Carbonate ore and my crit fail dropped to more like 1 in 100 and remained the same on each station.