Bio Engineer Archive
Thread: Are the High health, High Kinetic pets really abberations A surprising analysis
I did some thinking, and some math, and came up with something that really surprised me. Please take a thorough look at this and see if it makes you think twice about these creatures
The Cowardly Gurreck used to be "top dog" among the non-CH pets (or top cat, if you prefer). It had 3200 HAM across the board, something around 15% resists to kinetic and energy, and I think I remember pre-patch ones as doign about 200 damage, as a CL 15.
The patch lowered the maximum CL to 10, and some time back it was suggested that BE's were expected to be able to exceed a wild creature's abilities by about 50%. Taking this into consideration, we should then be able to make a CL 10 creature with the stats of a CL 15 "wild" creature.
OK - Now we come to "stat migration". A pet with 3200 Health, 3200 Action, and 3200 Mind has then a total pool of 9600 to work with for stats. Take a rough one-third of the stats as a minimum, which is approximately what a player must keep of any given stat pool in most cases, though it may be less. That means a CL 10 must retain , for example, 1100 points in any stat pool, with the others shifted to another pool, as is the case with players. That gives us about 2100 from each stat pool to migrate, or a maximum of 7400 health (3200+2100+2100), 1100 Action, 1100 Mind as a Maximum Health template for a BE CL 10 creature, if the stat migration is allowed to function the same as it is for a player.
Damage and Resists are also figured into the CL. We know this empiricly, though we do not have a formula for it that gives any repeatable accuracy yet. If the Cowardly Gurreck has, as an example, 10% Kinetic, 10% Energy, 5% Blast, 5% Electricity resists, (these are the current Coward's stats, I do not have the pre-patch ones), we might consider migrating them to 30% Kinetic. If we go to vulnerability, say that "gives back" 5% to the pool, so we might add as much as another 30% to the pool, making a 60% Kinetic with all other resists vulnerable an actual possibility.
Damage is difficult to manage in the calculation. A current Cowardly Gurreck does 150-160 damage. Dropping that to 90 might concievably allow a boost of 25% to some stats, figuring half of the amount lost in damage as a boost.
So, with all of this as possible, are we still sure the 9K/1K/1K, 60% Kinetic, 75 damage pet is the total abberation that we have been assuming it to be?
I am not arguing right or wrong in this one - I am simply positing a theory that might actually make us consider whether our "gut feel" that these high health, high kinetic pets are really as far off as we have been assuming that they are. For myself, thinking this through has given me a different perspective on them, and I need to re-think my take on how they affect the game. I am quite sure the developers did not expect the results in this manner, but I also doubt they had the time and capacity to stretch the limits the way the be community has done.
So , to sum up, simply:
Cowardly Gurreck - 3200 Health, 3200 Action, 3200 Mind. 10% Kinetic, 10% Energy, 5% Blast, 5% Electricity, 160 Damage .31 to-hit
Stat migration towards Health - 7400 Health, 1100 Action, 1100 Mind (3200+2100+2100 into Health )
Resist Migration 60% Kinetic, Vulnerable to all others (10+10+5+5+20 for the vulnerabilities)
Damage Migration - 100 damage at .21 to-hit (30% speed reduction, 30% damage reduction)
Possible boost toHAM pool based on the damage / speed reduction of 30% - 8510 health, 1210 action, 1210 Mind
So a BE creature with 8500 Health, 1200 Action, 1200 mind, 60% Kinetic resist, 100 damage at .21 to-hit becomes a possibility, as the mathematical equivalent of a Cowardly Gurreck.
I'm not saying this is the answer , but it does provide something to consider.
LLan
Message Edited by ArthurDentOnBria on 05-18-2004 12:12 PM
Remember - I am not arguing FOR these creatures as acceptable, merely pointing out the math that produces a very different view of them than the "gut feel" we seem to share on the high-health creatures.
As for what the devs have said, forgive me, Talthazar, but after months of discussion and living with the badly out-of-control pet checks and CL determinations, whatever the devs may have said that long ago seems at best tangential to the problem, and relegated to irrelevance by their inaction. I wish it were otherwise, but the appearance is hard to shake after this long. If it walks like a duck, and quacks like a duck, it's a kaadu.
The problem lies in that the mathematical equivalent is not the functional equivalent, and that is where the problem is as well. 60% Kinetic resist is FAR more useful than 10% Kinetic, 10% Energy, 5% heat, 5% cold, 10% acid, 10% stun , 5% blast and 5% restraint, though both add up to the same number. Without pretty careful "weighting", any numeric-based system can be played to its own disadvantage, and this is a very clear case.
The difficulty is in the reinforcement being given to a particular set of actions or equation. If the "acceptable" non-CH pet is consistantly nerfed to uselessness by the pet checks, the only ones we will see made are the "monsters", because they are what is adjudicated by the system as usable. The only alternative is to stop making any Non-CH pets, which leaves those that care about their effect on the game ineffectual, and those who see the pets as what works in control.
BTW, the 50% number I used comes from a mention attributed to Holochron I believe, in which he was describing the BE as being able to make creatures half again as powerful as their wild versions, so I used it as a place to start in theory, and it does seem to support the results we get. Scaling the creatures up to 70 may or may not play out, but what I am concerned with is the "uber" non-CH pet we are concerned about at this time. (Just trying to clarify the scope of the discussion)
LLan
It seems that we are mostly on the same page, but part of our problem is that no one has ever said how much stronger a BE pet should be than a wild pet. Holo told us BE pets would be stronger, but he never said half again nor has anyone given us a definitive "BE pets will be xx stronger than wild pets." Unless you see it coming from a red name post, take it with a grain of salt.
Forgive me, Talthazar, but lately, when I see it coming from a "red-name" post, I take it with not a grain of salt, but with a dash of cayenne pepper, as too much of it seems like hot air. 
I also agree that the non-CH level was lowered to prevent the "pet wars" that were common at the time, not to boost the BE's relevance in the game; I'm just using the CG as a good example of a CL 15 for the sake of comparison, in part because of its familiarity. I would love to see someone do an analysis in the same manner I did here on the rock beetle as an example; I was a master CH before the changes so I was not much into the rock beetles myself, and lack the familiarity.
LLan
Akelan,
I have long thought it was just a matter of the maths underlying the CL calculation. I have spent a lot of time recently making repeat pets, but altering as few stats as possible to see if I can figure out how the CL calc is made. And I am making some progress which I will share. Everyone please note that these are preliminary results and tested against a few select "types" of template. They may be wildly wrong, they are most definitely mildly wrong .
What I wanted to work out was why a 3k HAM, 38 Kinetic, 8 Energy, 8 Blast creature was level 16, whereas a 9k 60 kinetic creature was 10 - was there any sense. Yes there is and there is even a formula which delivers +/- 1 CL to a range of creatures I have tested.
In all its splendour the best fit formula I have is (and remember this is not yet quite right and might be very very wrong)
0.25*Hardiness + 0.25*Dex + 0.25*Intellect + 2.5*Cleverness + 0.5*Courage + 1.5*Power + 4*Kinetic resist + 8*Energy resist + sumof(1* the otherresists)
Add 1170 (this is the fudge factor)
Divide by 100 and round (I tend to round up).
If a resist is a vulnerability deem it to be -100.
To note
- cleverness (the "to hit" chance has a huge impact - more so than power).
- A vulnerability decreases the sum hugely - vulnerability to energy is -8 levels!
- If fortitude is over 500 add about 8 (still working on this) for the armour.
Some examples. The first creature I mentioned has dna sample of the 10 stats of, say,
204, 21, 197, 330, 103, 66, 729, 336, 68, 90 this gives us 3.7k/3.3k/2.1k HAM 0.23 to hit 80 -90 damage
Resists are 29 kinetic 8 energy 8 blast rest vulnerables
Calculate
Stat impact = 204*.25 + 197*.25 + 103*0.25 + 66*2.5 + 334*0.5 + 1.5*90 = 593
Resist impact = KInetic = 4*29 + Energy (6*8) + blast (8*1) = 172
5 Vulnerabilities = -500
Total = 160 add fudge factor of 1170 = 1435, creature level of 15 (or maybe 14)
Take a monster... HAM of 9+k and kinetic resists. Mine have dna like this..
579, 475, 390, 295, 188, 45, 760, 472, 26, 177 This gives you a 9.k, 6k, 4k 2.2 hit chance 150-160 damage 60 kinetic resist monster.
Cl level?
579*.25 +390*.25 + 188*.25 + 45*2.5 + 472*.5 + 1.5*177 = 903.25
Kinetic 60 = 240
Energy Vuln = -800
Rest Vuln = -600
Fudge factor = 1170
gives: 9.1 - this creature actually had a cl of 9.
I have tested it with a variety of creatures normally with vulnerables to everything except kin and energy so I am not sure about the "1" factor for other armous.
But its held up quite well. Anyway give it a shot see if it helps. I have a spreadsheet which does all this, but its based on someonelse's spreadsheet. If they give permission I will make it availble.
Zardo
ZardoZap -
That looks as close as anything I have seen to working, and I am willing to help test if you would like. I can also write this into the "DNA Sequencer" program I have written, which takes the stats from each sample in the slots used and uses the formulae from Mirrl and others to predict a template result, to a very good level of accuracy. Given an effective formula for CL, this would be quite easy to extend into a program that you alt-tab to, enter your samples, and get a live prediction of the base template and maximum, including the CL. With a prediction on the effect of an experimentation point (which I do have in the current version), the program could even predict the effect on the CL and stats if you experiment here or there before doing so.
I currently show predicted Health, Action, Mind, Armor, Speed, To-Hit and Damage, and the resist percentages; adding CL to it would not be at all difficult to do as a calculation from the template stats, using your formula. IF it works out, we could enter the DNA stats in each slot, calculate the resulting template, do "virtual" or "what-if" experimentation, and see a good prediction of our results as to the CL/HAM/Damage that we are all interested in knowing.
LLan
Just looking at your analysis again, the to-hit making a large difference would logically have a large effect if the pet's damage as seen as DPS (damage per second). The Damage x speed calculation is easy to see, but the to-hit chanse makes a bigger difference as the percentage of that damage that is actually "delivered" to the target.
say for ease of example, 100 damage, 2.00 speed. Creature hits for up to 100 damage every 2 seconds, or 50 dps raw score.
use a to-hit of .25 and you have 25% of that damage, or 12.5 DPS being delivered to the target.
Use a to-hit of .30 and you get 15 DPS delivered, .35 brings it to 17.5 DPS, and so on. You would have to cut the attack speed to 1.0 to make as much difference as a .5 over a .25 to hit, so a little change in that makes for a large final impact on the damage output of the pet in combat. which would certainly be reflected in its CL.
Just some math to back up the empirical understanding of the to-hit's value. (Forgive me, WWWAAAYYY too much time parsing logs in Everquest to determine damage output for taking down raid-level mobs and keeping a combat area clear.)
LLan
Hey Gang,
I hate to be the bee in your Coke, especially since I was so excited about this formula, but after plugging in about 30 sets of stats with known results, I found some discrepancies. Maybe they will be helpful?
Basically, your formula works really well for figuring out the high Kinetic Monsters at level 9 or 10. However, the first generation of creatures (the ones you sample from to make those monsters) always come out as about half the level the game puts them at.
From what I can tell, maybe there is a sliding scale on the energy resist vulnerability value based on HAM, or maybe based on level of the creature used? What I found is that if I plugged in the stats for these sub-component creatures, I could get your formula to work if I changed the energy resist to -20 or -30 (x8) as opposed to -100 (x8). Then it was hitting very close to the right level for all the subcomponent creatures and not for the High Kinetic monsters. When I plugged -100 back in for those, they worked.
It could also just be some other stat change that is not taken into consideration by your formula, but honestly, when there is both a Kinetic and Energy value to put in (most of my subcreatures are Energy vulnerable, so only had a few examples of these) it seems to be at least mostly accurate.
Anyway, hope this helps narrow it down. I could feed you some of the stats I have if you are interested. Typically they were creatures with Kinetic and Blast resists in the teens with energy vulnerability. Could be that you found the bug too. Maybe Energy resist is supposed to be calculated x8 for the class if there is a positive number in there, but only supposed to be a grand total of -100 (not -100 * 8) if it is vulnerable...... Just a thought......
Anyway, thanks for the work on this people!
Message Edited by GotharMarath on 05-21-2004 02:32 AM