Doctor Archive
Thread: Experimentation on Buff Packs? Are you joking me?
Page 2 of 2
toothlessviper
Tue Feb 01, 2005 8:04 pm
#14
i get the best average when i exp 3 3 then2 1 1 1.... and yes get used to the fails you seem to get more than armor and or ws.... shug im just used to it... welcome to the club...
Lunariel
Wed Feb 02, 2005 5:28 pm
#15
Gwig13 wrote:
I have been a long time weaponsmith and dabbled in armor crafting and have recently tried making buffs. I have 11 experimentation points but the experimentation on buff packs seems COMPLETELY screwed. First of all, I get many more fails and good successes than I ever do in weapons or armor. Secondly, even if I get all great successes it still takes all 11 points to get max effectiveness. In weapons and armor, great successes generally add 6-7% whereas in buff crafting it seems that I'm lucky if I get 3%. The results aren't even that consistent. I was just wondering if this was an accepted issue or it is being corrected. Has it always been like this? Thx for any input.
The reason for this is a hidden feature with experimentation.
Whenever there are more than one attribute on a single experimentation line, it is represented internally as one percentage for each attribute. The percentage you see is just an average of all of them. The hidden feature is that there is a compensation system for multiple percentages. Normally we don't notice this unless the same experimentation line has different formulas.
The compensation system takes part of the 7% advance (for great success) from the highest percentage and gives it to the lowest percentage.Unfortunately the compensation continues after the lowest one has been maxed. Since the lowest one can't do anything with this bonus, it is just thrown away.
Let's take an example with a buff pack which maxes out duration at 60% and power at 100%. If you experiment one point at a time they will both end up fairly close to each other just below 60%. But once you pass 60% the duration is capped and you get less and less each time for the power. At 90% you will only get 7% * 0.8 = 5.6% for a great success on the power. And the average will only increase by 5.6% / 2 = 2.8% showing a 2% or 3% increase in the total...
- The experimentation system is set up so that you can always reach maximum with 10 points when experimenting on a single property. With multiple properties you may need more.
- When you have multiple properties the experiment result you get is a bit misleading. A separate roll is done on each property and the result you see is the average rounded down. If you get great success on the duration and amazing on the power, it will report 'Great Success' back to you. You will only see 'Amazing Success' if both are amazing.
- If you have problems getting great successes at master, you shouldmake sure you are in aResearch City. That does wonder for you experimentation. Or you could try and find some Bespin Port. Using less points each time will also help out.
- Because of the compensation, the relative difference is largest just before starting to experiment. The worst thing you can do is to use all points at start. You may risk to overshoot past the lowest cap. The optimal way to experiment is to first get close to lowest cap with single or multiple points. And then try to use as many points as possible each time.
Lunariel
Ledao
Wed Feb 02, 2005 6:02 pm
#16
Lunariel wrote:
- The experimentation system is set up so that you can always reach maximum with 10 points when experimenting on a single property. With multiple properties you may need more.
Lunariel
Very nice explanation of enhance experimentation... The part I've quoted isn't the "whole truth" though. 
The experimentation system is actually set up so that you can max out any single line with 9 points, though typically at least 7 of those points have to be amazing successes. (29% start, + 8% per point * 9 points = 101; 29% + 8% per point * 7 points + 7% per point * 2 points = 99)
This applies to enhances just like everything else, although if the duration is low enough it becomes impossible.
In my experience (and no, I haven't worked the numbers out precisely, but they're pretty close...), any enhance pack with a maximum duration over about 9.6k seconds can be maxed out with 10 points, and any pack with a max duration over 10.6k or 10.8k seconds can be maxed out with 9 points, although it can be quite difficult... (Even a pack with a max duration in the 12k second range might take me -- a +25xp doc in a research city with +12 bespin port -- 20 or 30 tries to max out with 9 points...)
And, of course, this ties back in to experimentation point usage -- if you know how many points need to have amazing successes to max out the property in question with 1 fewer point than usual, then you can make much more informed choices about how you experiment than "I use 2-2-1-1-1" or what have you.
Lunariel
Wed Feb 02, 2005 7:06 pm
#17
Ledao wrote:
Lunariel wrote:
- The experimentation system is set up so that you can always reach maximum with 10 points when experimenting on a single property. With multiple properties you may need more.
Lunariel
Very nice explanation of enhance experimentation... The part I've quoted isn't the "whole truth" though.
The experimentation system is actually set up so that you can max out any single line with 9 points, though typically at least 7 of those points have to be amazing successes. (29% start, + 8% per point * 9 points = 101; 29% + 8% per point * 7 points + 7% per point * 2 points = 99)
Thanks for clarifying it. Imeant to say10 points of great successes. Btw, amazing success is 8.05%.
Lunariel
Page 2 of 2