Doctor Archive

Okay, get a coffee, this is a number crunch post, so it might take some reading. I thought after 9 months of making stuff in this game, I knew everything there was to know, but I am guessing I've been a bit blaze about it.

Here is what happened:

Made 1 set of buffpacks (Set#1), 4 days ago:

ABEC 22c 15p
SDS 70
CRDM 65
Avian (Eog.)
Reactive (Pento.)

I spent my 12pts like this:
Charges: (1 amazing)
Med Use: (1 great)
Power: (10 great/amazing to max)

= Buffpack 913 power, 32 charge, 89 med use

Then made another set 4 days later (Set#2):

ABEC 22c 15p
SDS 70
CRDM 65
Avian (Eog.)
Reactive (Pento.)

spent points the same way, *exactly*

= Buffpack 913 power, 31 charge, 89 med use

So, something must be different, I hear you say.

Well, the gas and Avian and the CRDMwere *identical*.
The SDS and ABEC components I build have *identical* stats but were made with different resources as follows:

In Set#1 (32 Charges end result)
SDS: Same Oats, Dolovite (Xiroo) with OQ994 got same power = 70
ABEC: Tat Fiber (Aspep.) with OQ837 UT927, got same result = 22c 15p

In Set#2 (31 Charges end result)
SDS: Same Oats, Dolovite (Aunemio) with OQ956, got same power = 70
ABEC: Tat Fiber (Osath) with OQ752 UT829, got same result = 22c 15p

In both cases, the initial "build" resulted in 30 charges. Spending ages waiting for 1 amazing success, on the first build gave me 32 charges, but on the second, still 1 amazing success only gave me 31. In both cases, spending a second point would have raised it to 33 charges.

I think we have only 2 options and this is where I want opinions (and feedback from the Correspondant would be lovely), either:

1) Fraction Theory

That the "stats" on our components are displaying only full numbers, but are actually carrying forward PARTIAL numbers, or fractions, which have a noticeable affect on the final product. Eg. The 22c 15p ABEC's I listed above are actually something like: 22.1c and 22.7c

2) Component Resource Quality affects Final Build

That the resource quality of the component make actually DOES affect the final build (something I have always believed NOT to be true).

My opinion is on (1). Any comments, opinions, findings?

If it is (1), I want a change. I want to be able to see the fractions.

Thoughts anyone?

---

Sarkyn wrote:

1) Fraction Theory

That the "stats" on our components are displaying only full numbers, but are actually carrying forward PARTIAL numbers, or fractions, which have a noticeable affect on the final product. Eg. The 22c 15p ABEC's I listed above are actually something like: 22.1c and 22.7c

---

i hope is that one on carrying partial numbers but displaying full numbers.

I maded one long factory run on buff packs and get in the 920 power 31 charges

in the next factory run i maded exactly same packs on exact resources but changed the hervibore on the advanced chemicals, and lost one point in power in that ones, 65 power the first ones and 64 power the last ones, but.... get the same 920 power packs 31 charges at end.

Great and amazing sucesses do not always add the same amount dots when you get one.

Try a few dozen ABECs. You'll notice that 2 great exp in one run does not equal 2 great exp in another 20 ABECs later.

In short, a great sucess will get you 68% most of the time. Every now an again, that same great sucess with the same resources will get you 63% or 72%. (68% is just an example, your percentage will vary)

Blue-Quick
Gorath Server

---

Blue-Quick wrote:
Great and amazing sucesses do not always add the same amount dots when you get one.

Try a few dozen ABECs. You'll notice that 2 great exp in one run does not equal 2 great exp in another 20 ABECs later.

In short, a great sucess will get you 68% most of the time. Every now an again, that same great sucess with the same resources will get you 63% or 72%. (68% is just an example, your percentage will vary)

---

That is because, the %-based system of experimentation hides the amount below the decimal point. Effectively, you are not getting 7% each time, and never have been. You get a figure somewhere between 7 and 8%, which is then rounded (for DISPLAY) but not for use. Thus the tendency with even "great" successes to get 7,7,8,7,7,8.

Starting from the same point (just crafted) and getting the same success type (amazing) yields the same consistent result in my tests last night.

On set#1, it led to 32 charges.

On set#2, it led to 31 charges.

Soo - try the same exact crafting with same resources, only this time drink Bespin Port. I want to check something out.

---

Mahabret wrote:

I think it's #1, fractions that don't show. I've seen this happen with charges AND power (I do experiment some on power in my ABEC's).

Guess I need to start tracking the final % result in the ABEC names.

---

I think, if this happens (and the evidence seems to support it everywhere I look) that it's happening in the entire crafting system.

That would mean weaponsmiths, armoursmiths, even chef's must be seeing oddities like this, and probably more thana few of them have noticed it as well.

I'm going to start plotting %'s in my component names as well. We've got no choice really, if SOE won't give us all the numbers they're using, we have to record them ourselves.

Does anyone else feel this should be shown?

I thinkat leastcomponents should have shown more decimals. Crafting is the only thing you can use them for and the decimals matter there.But you can easily run the calculations to find the decimals yourself.

The final values are calculated like this:

ABEC Power: 5+25*Percentage

ABEC Charges: 5+20*Percentage

ACRDM Power: 10+65*Percentage

ASDS Power: 10+65*Percentage

And the max Percentages like this:

ABEC Power: (Wheat_PE+Wheat_OQ+Fiber_OQ)/3000

ABEC Charges: (Wheat_OQ+Fiber_OQ+Fiber_UT)/3000

ACRDM Power: (Herb_PE+C4_PE+Herb_OQ*2+C4_OQ*2)/6000

ASDS Power: (Oats_PE+Oats_OQ+Dolovite_OQ)/3000

If you experiment a line to the max, you can simply replace the percentage in the value formula with the maximum percentage. If you are only spending a few points, it gets a bit more complicated. Then you have to use this formula instead:

Initial Percentage = Max * ( Max + 100% ) * 15%

Add 7% for each great success

Add 8.05% for each amazing success.

An example: You make an ABEC with 90% max on power and 89.52% max on charges. Then you max out the charges with 9 points and gets 2 great and 1 amazing for the power. The exact power percentage is then 90%*190%*15%+2*7%+8.05% = 47.7%. And the charges percentage is of course 89.52%. This gives the final value of 16.925 power and 22.904 charges. These fractions will be added to the final buffpack values.

There is a bug where all the boxes fills out before you reach max. It happensoccasionally when you get close to the maximum without reaching it. This one can be hard to spot and will mess up your calculations.

If you do any of these calculations, remember to preserve every single decimal during the steps. I recommend doing them in a spreadsheet.

---

Sarkyn wrote:

That is because, the %-based system of experimentation hides the amount below the decimal point. Effectively, you are not getting 7% each time, and never have been. You get a figure somewhere between 7 and 8%, which is then rounded (for DISPLAY) but not for use. Thus the tendency with even "great" successes to get 7,7,8,7,7,8.

---

Great success always adds exactly 7% unless you are experimenting on a line with multiple properties. Like duration and power on the same line for buff packs.

Multiple properties has a lot of weirdness. Internally each property has its own percentage and its own experiment roll. And if the two percentages are not equal, parts of the gains from the highest will be taken away and added to the lowest one instead. So instead of both getting 7%, one may gain 8% while the other gain 6%. However the shown average should normally increase by exactly 7% if both are great success.

Trouble is the roll you get reported is the average of both of them and truncated like anything else in the game. So a great plus amazing success will be reported as great. Then the shown percentage will normally increase by 7.525 which may make it look like 8 percent were added if the decimals were high enough. On buffpacks you can normally observe the increase in either power or duration was more than expected when this happens. It will only report an amazing success if both rolls on the power and duration are amazing.

Multiple properties also has a bug/feature with the redistribution. It continues to redistribute to the lowest one after you have maxed it out, essentially wasting your precious experiment points. If one of them has maxed out, you will notice you get less and less each time on the higher one. This is why you have trouble reaching max power when you make packs using low DR.