Bio Engineer Archive
Thread: Undocumented features.
Hi, there are many portions of this new system which seem like they should best be described as undocumented features. I've started writing down everything that happens in each clone I make and running experiments every time I see something that seems a bit odd. Unfortunetly noting down 5 DNA samples, initial template, 10 experimentation states and then final combine has meant that I don't run through them nearly as quickly. This has limited the amount of information I've gathered and made it so I would really appreciate results from anyone else who is doing likewise. However here are some things I have noted to date. As I saidI'm hoping people will add to them.
Kinetic resist low on Dragonette and Giant Sand Beetle (GSB) combinations. Dragonettes have a kinetic resist of 25. Combine 5 of these together andyou get a final kinetic resist of 25.GSB's have a kinetic resist of 20. Combine 5 of these togetherand you get a final kinetic resist of 20. Combine 2 Dragonette sample (physique and prowess) with 3 GSB samples (mental/psycholoy/aggression) and you get a final kinetic resist of 16. Experimentation will help this in no way and it won't go up on final combine. I've run this experiment with all high quality DNA which also does not effect things. This means the final combine has a lower kinetic resist than any of its donors. The results do however seem to be repeatable as I've gotten the same result on 6 runs of this combination.
Dependability high on Dragonnette and Giant Sand Beetle (GSB) combinations. Dragonnettes have a dependability rating in the lowe 500's. GSB's have a dependability rating in the low 1000's. Using 2 DNA samples from Dragonettes (physique/prowess) and 3 from GSB's (mental/psychology/aggression) I get a dependability rating in the mid 1000's. (i.e. higher than any of the donors) Howeverevery attemptI've made toexperiment psychology has failed. (Usually I'm shown a 67-69% psychology with 3 bars available for experimentation) After the first failure the dependability rating will drop below 1000. I'm still get failures on any sub-sequent attempts to experiment psychology. To date I have gotten 6 failures on 6 attempts. To date my failure rate seems from my records to be slightly over 10% in general. (Despite showing as 0% on the side bar) This would seem to make this failure rate indicative of a problem.
Failure on experimentation seems to make non-experimented statistics go down. I'm not sure if there is a specific pattern but I've noted that failing on an experiment will have the following effect. One of the related stats will go up. On of the related stats will do down and one unrelated stat will go down. On one of my standard Drag/GSB combines. 5th experiment point in prowess I get a failure taking prowess from 33% to 30%. Dexterity went up frm 267 to 351. Endurance went down from 334 to 253. Intellect also went down from 326 to 256. On a related note the dropped stat does not become re-elgiable for experimentation. On a different Drag/GSB combine (2nd generation) I had maxed out prowess (50% 470/531) throuhg experimentation. On my 9th experiment point I got a failure in mental. (Intellect went up 276 to 281, cleverness went down 10 to 1, and endurance went down 531 to 461.) However on my tenth experimentation point I was not given an option putting a point into prowess to make up the ground I had lost.
2 generation DNA sample have a 0 cleverness. My first generation Drag/GSB generally have a cleverness around 300 and final mental stat of about 8000. Every DNA sample I have taken of them gives a cleverness rating of 0. The initial template usually has a single digit cleverness rating and after experimentation it is usually in the teens somewhere. It also seems that 2nd generation DNA samples cannot be higher quality than the finaly template they come from. Most of my templates are high quality. (It seems to take 5 very high quality samples to produce a very high template. 4 very high and a high produces a high quality template) I have never gotten a very high sample from a high quality template creature. However about ~1/3 of all wild creature samples I get are very high. (Like many result conditions in the crafting portions of the game the result seems to clump so its tough to be sure)
Diminishing returns on physique experimentation. The first point I spend experimenting physique frequently gives me a 5-7% increase. Sub-sequent experiments will only yield a 2-3% increase unless its immediatly following a failure in physique. I also not that the physique listed percentage will some times drop by 1 even though the stat hasn't changed. On a customer ordered statue I had fnished up with physique (11% 145/63). I then put a point into prowess. Now the pyshique read (10% 145/63)
Max drop in CL per generation. My Drag/GSB first generation all tend to be CL 25-27. To date all my 2nd generation Drag/GSB are CL 15 and contained at lower level CL of 25. My best to date was (HAM 8324/8540/5201 Armor None Resists 16/38/17/17/v/v/17/v Attack 2.39/0.2/70-75 Ranged/linding/stunning). My worst to date was (HAM 8360/7912/5632 Armor None Resists 16/38/17/17/v/v/17/v Attack 2.33/0.2/55-65 No ranged/Blinding/Stunnning) I'm surprised these where both the same CL. I would expect at least1 point ofCL difference. They both went into a dewbackskin, which has a CL 10 lower limit I believe.
Some CL10 pets cannot be given to non-ch's. A guild mate asked me to make him a "show pet" woolimander. Gathered some local junk DNA, combined and got CL13. OKresample, combine and get a CL 10. Tame it, train it name it, try to transfer and it will not transfer. Not sure what the problem is but grrrr. Buddy was in a hurry so I'm not even sure of the stats or samples. No idea where to start looking for what went wrong there.
Some resists effected by experimentation and final combine otehrs not. It seems like any non-vulnerable resist below 10-15 (?)will be pushed up by experimentation. Ifind that when all is said and done resists at 5 inthe initial template will be at 17 by then final creature. I find this very surprising when there are stats high than the initial which end up lower than the final (My standard Drag/GSB experiment has a Kinetic of 16% from inital to final creature. blast/heat/acid allstart at 5 but end up at 17.
Vulnerable can be over-ridden. 2 DNA samples from rancor (physique/prowess), 2 from ancient grauls (mental/aggression) and 1 from horned voritor lizard (psychology) . Forelectric resists this guves us (60/60/V/80/V) with a final resulting electric resist of 14. Armor showed up after 5 experiments in physiques (1 failure in there) but the resists never changed.
Kinetic resist low on Dragonette and Giant Sand Beetle (GSB) combinations. Dragonettes have a kinetic resist of 25. Combine 5 of these together andyou get a final kinetic resist of 25.GSB's have a kinetic resist of 20. Combine 5 of these togetherand you get a final kinetic resist of 20. Combine 2 Dragonette sample (physique and prowess) with 3 GSB samples (mental/psycholoy/aggression) and you get a final kinetic resist of 16. Experimentation will help this in no way and it won't go up on final combine. I've run this experiment with all high quality DNA which also does not effect things. This means the final combine has a lower kinetic resist than any of its donors. The results do however seem to be repeatable as I've gotten the same result on 6 runs of this combination.
I've found that you really need very high quality samples to get a perfect blending otherwise the final result is tainted on the template because it's a poor quality clone. Typically though if you had the GSB in two slots (mental and psych) then the dragonet in the other three you'd find the kinetics being much higher since those slots accound for around 80% of the resistances. One thing you might want to check is that while the GSB has effectiveness kinetic protection the Dragonet might have special kinetic protection. When blending the special protection overrides the effectivness type leading to the fact it's can't be experimented higehr, in my opinion this is a bug carried over from what appears to be the armorsmithing programming being dumped into this half arsed cloning system.
OK, Your right about the Dragonet being special kinetic protection. So maybe mixing and matching special with effective is bad. I've produced both Above Average templates and High quality templates. (I've never gotten atemplate belowAbove Average in the new system but I think thats because asmaster we don't seem to get DNA samples below Above Average)Both seemed to have the same listed stats. I haven't tried a very high quality template but I got the last DNA sample I needed for it last night. I'll probably give that experiment a try tonight.
Tal-N wrote:
I've found that you really need very high quality samples to get a perfect blending otherwise the final result is tainted on the template because it's a poor quality clone. Typically though if you had the GSB in two slots (mental and psych) then the dragonet in the other three you'd find the kinetics being much higher since those slots accound for around 80% of the resistances. One thing you might want to check is that while the GSB has effectiveness kinetic protection the Dragonet might have special kinetic protection. When blending the special protection overrides the effectivness type leading to the fact it's can't be experimented higehr, in my opinion this is a bug carried over from what appears to be the armorsmithing programming being dumped into this half arsed cloning system.
I need to dig up my formula post again... its due for a reworking to include new info... but I found that resists followed a very strict formula irregardless of the quality of the sample.
- Physique 40%
- Prowess 25%
- Mental 5%
- Psychological 5%
- Aggression 25%
This has a few holes in it.... vulnerabilities is one (not sure what the value of a vulnerability is... usually -50 to -200 works)... some strings of identical "effectiveness" clones bother it slightly, but these values typically rise to the prediction with experimentation in physique and someone brought up that the +1 to +4's that we see popping up are related to a fractional value of Fortitude / 10 (think it was 10 or was it 30?).
Was just talking to someone about Quality of the DNA sample why this caught my eye. I honestly don't think the Quality has any bearing on the clone other than the Quality relating an average value / spread for each of the 10 characteristics with minimal overlapping of stat ranges.
Sumorex wrote:
I need to dig up my formula post again... its due for a reworking to include new info... but I found that resists followed a very strict formula irregardless of the quality of the sample.
- Physique 40%
- Prowess 25%
- Mental 5%
- Psychological 5%
- Aggression 25%
Hi Sumorex, It was actually your formula (or one very like it that I was trying to validate) Howver if you look at my data you see.
25 *0.4 + 25 * 0.25 + 20 * 0.05 + 20 * 0.05 + 20 * 0.25 = 23.25 for an expected result.
However on 6 run I've gotten a 16% kinetic resist each time, notthe23/24 expected result. However as Tal-N has pointed out earlier maybe your formula works on all special protection or all effective protectionbut not any combinations of those two.
It could be tied up in there.. special vs effective.
Sorry if I'm missing points, I'm battling the flu and ain't at my sharpest today 
*shakes his fist at Kel* Give us documentation **edit**!
One thing to point out is that in actuality when you first have ANY effectivness resistance they start of low initially and need to be experimented up to the amount they should be. Typically what you find is that they start at 50% of their 'potential. Kahmurrith snappers should be 30% but they initially start at 15%. The merek deaths death should be 10% with light armor but starts around 30% (since it's a grand total of 50%+10%)
Therefore the effectiveness for the giant sand beetle would initially be recognised as 10 rather than 20. The dragonet having special kinetic protection of 25 would indeed be 25 still. So in reality the math isn't going to be (20+25)/2 resulting in what should be roughly 22% to 23%. The actual math is (10+25)/2 resulting in EXACLY what you'd expect as a ballpark figure as proven in those 6 tests... 17% to 18%. Admittedly you got 16% however we're not talking about an exact 50/50 merging here because of how the slots are weighted.
So be warned folks, when mixing special protections and effectiveness resists your actually getting only around 50% of the effectiveness value leading to lower resistances.. however given the CL miscalculations it's probably a pay off anyway
Not sure why people insist on one staring Tal-N for saying what most of are thinking (*****). Anyway tried my standard experiment using all Very High DNA and got a Verh High quality template.Kineticwas 16 like always. On the positive note the final combine gave me 1 higher blast/heat/acid than usual (18 instead of 17). While that could just be the vagaries of the final combine I think it might suggest a benefit to all VH DNA samples. This seems to rule the quality effecting carry though of armor theory.
The other theory (Tal-N's modification to Sumorex) was the that in mixed combines the specials come through at full value but prevent experiment in that trait and the effective comes through at 1/2 value. My normal experiment line gives the following.
25 * 0.4 + 25 * 0.25 + 20 * 0.5 * 0.05 + 20 * 0.5 * 0.05 + 20 * 0.5 * 0.25 = 19.75
The actual result was 16 which indicates the predicted value is a little off. I also tried reversing the sample donors on a second experiment.GSB for physique/prowess and Dragonet for mental/pshychology/aggression.Expectedresults forthis formula would be
20 * 0.4 * 0.5+ 20 * 0.25 * 0.5 + 25 * 0.05 + 25 * 0.05 + 25 * 0.25 = 15.25
The actual result was 8.While the formula does predict a lower result it does not predict one that much lower.
If we assume that NONE of the effectiveness gets through at all then we get for my normal experiment.
25 * 0.4 + 25 * 0.25 = 16.25
Pretty well spot on. Trying again with the second experiment we get.
25 * 0.05 + 25 * 0.05 + 25 * 0.25 = 8.75.
Once again pretty well spot on. I think the lesson here is that Sumorex's basic formula is spot on but that you should never mix special with effectiveness because you will lose ALL of the benefit of the effective protection.
That owuld explain a lot. I was trying to get a high kinetic, high stun resist creature. But the stuns were coming out at less than half of what they were supposed to be.This must be why.
Why in the hell would the game need to distiguish two types of the same resist?
If indeed the special protection is over-riding the effectiveness completely then it's pretty much doing exactly what happens with player armor.
Next batch of mixed special/effect resists clone I work with I'll pay attention to this. Looks dead on though.