Character sheet - click here!Specializations have been implemented.
There will invariably be mistakes, and the below specializations have not been implemented. Most masteries also have not been implemented. Still, test it out.
Enlightened
Expertise
Sentinel Aura
Dominion
Rejuvenated
Durable
Relief
Savage
Bulwark
Calculated offense/defense values may vary at most 1%, because of client imprecision. (Did you know that Justice gear actually gives +147.4 offense, not +147 offense? I take this into consideration, BTW).
http://co-forum.perfectworld.com/showthread.php?t=250771
The purpose of this project is to essentially derive all in-game character sheet calculations (such as crit chance, severity, dodge chance, avoidance) from scratch using character values (stats, equipment).
How to use:
1. Copy the sheet to your local computer / google account.
2. Open up your character sheet and your equipment.
3. Put in values from your character sheet into the yellow highlighted boxes.
4. Watch the gray boxes update.
5. Tweak your values to run simulations of what your character will look like with that new crit chance mod or more offense.
Many thanks to
ayonachan and flowcyto for the initial data compile.
----
Version historyalpha
0.001 Initial version pushed to public.
0.002 Added some info and version history, backend work on deriving stats from chosen specializations, added "helpful" notes
0.003 Backend, user indicated Primary and Secondary stats
0.1001 Extensive work on filling in the spec trees.
The various colors in the calculations sheet mean:
Green - calculations for this spec are being done.
Yellow - calculations for the spec are/will not be factored in, or don't have enough data. See comments.
White - Haven't got to this spec yet.
Main page has not been updated.
0.1002 Experiment using Google API query language to get arbitrary misc stats.
0.11 Interface for stat trees, does not work yet
0.2 Specializations now work. Still alpha, because I can't guarantee that anything is remotely even correct.
0.21 All Equipment stats now separated into lists
0.22 Melee critical bonus displayed (so overpower actually does something.)
(This was already calculated, but was not shown in the interface.)
----
How you can help:
Collect more data on the following specializations:
How do these scale?
Enlightened
Expertise
How much?
Sentinel Aura
Dominion
Verify:
Rejuvenated
How do these scale?
Durable
Relief
Savage
Need more datapoints for regression:
Bulwark
Correct any errors.
Feedback!
Comments
No problem! Bit of a WIP so it might change, but it works for now.
One way you can help is to provide data on all the specializations that scale non-linearly (anything under here without a specific number per stat associated) in the form of 2 column tab-delimited lists. I can use those to do further regression fitting - I'm assuming most of these curves look the same, so this should go relatively quickly. It would help further if you could refer to the character sheet to get more precision (e.g. STR's brutality reports tooltip values in whole integers, but crit severity in the character sheet is reported as tenths).
- Be safe and have fun, champs - for science!
That Dork In The Suspenders, signing out.
WARNING: Not An Actual Internet Reviewer
Get the Forums Enhancement Extension!
Download it to your computer, or google docs.
The sheet was made read only because it's a public resource. I'll check if google docs can support a simple form or app.
#didthemath
Another way to look at it is that for each point of increased offense, you actually get
1. +0.25 Defense
2. +1.25 Offense
and for each point of defense,
1. +1.25 Offense
2. +1.25 Defense
This is without other equipment modifying specializations.
Cheers.
"Without data, you're just another person with an opinion." -- W. Edwards Deming
I use a combination of GraphPad for verification along with Eureqa and ZunZun.com for discovery. I do use R (through RStudio) for my real work (bioinformatics, molecular biology), but generally do not like it very much. I do use R to make awesome heatmaps, but I really can't think of anything in CO that requires a heatmap to explain ... lol.
tbc in my next reply
- Be safe and have fun, champs - for science!
Pretty sure Moment of Glory and Follow Through mimic 6th Sense and Brutality/Deadly Aim, but I can spot-check that later.
Imprecision strikes again. I used the char sheet for severity values >10%, but most the other stuff can't be tracked on the char sheet (ex. melee_ranged cost discount) or had problems w/ my gear (crit% w/ the crit already on my offense gear) so I had to settle w/ the tooltip limitations in places.
- Be safe and have fun, champs - for science!
and spot-checking moment of glory (data matches 6th Sense values):
That's prob it from me for a bit concerning this topic. Let me know if ya need anything else in the future.
- Be safe and have fun, champs - for science!
STR - brutality / DEX - deadly aim:
STR - brutality:
%SeverityBonus = (Rank/2) * 1.113 * sqrt(SSS)
DEX - deadly aim:
%SeverityBonus = (Rank/3) * 1.113 * sqrt(SSS)
Shown are maxed specs:
EGO - follow through:
%SeverityBonus = (Rank/3) * 1.573 * sqrt(EGO)
Shown are maxed specs:
EGO follow through (red) vs STR/DEX crit severity specializations (blue): (all specializations maxed)
1 point of EGO is worth about 1.413 points of SSS (which is almost exactly sqrt(2)) when it comes to these particular specializations. Funny how elegant that is.
%MeleeCrit = (Rank) * 1/(0.231 + 9306*(STR)^(-2))
At rank 1:
STR - physical peak:
%CostReduction = (Rank) * 1/(0.057 + 9430*(SSS)^(-2))
At rank 1:
EGO - Sixth sense:
%CritBonus = (Rank) * 1/(0.1912 + 31074*(SSS)^(-2))
At rank 1:
EGO - Insight:
%CostReduction = (Rank) * 1/(0.0582 + 2322*(SSS)^(-2))
At rank 1:
(And no, my software doesn't spit out nice numbers like 1 and -2 (for the exponent) automatically. Some of this is based on my personal judgment (parsimony, duh). I refit everything afterwards, of course.)
Stats that need further calculations:
I'll try to get to the ones in the 2nd list after I do some more activation delay testing. If anyone else wants to jump in and help w/ 'em, though, be my guest.
- Be safe and have fun, champs - for science!
<I> Energy Shield's Laser Knight adv is 25% final/multiplicative DmgRes:
(battle instance testing)
Sample hits:
@ 0 Defense/DmgRes: 152 taken w/ 203 base (74.88%); 815 taken w/ 1088 base (74.91%)
@ ~28.5% Def/DmgRes:
hit #1: 675 taken, 1155 base = 58.44% taken
`-> 28.5% DmgRes = 1/(1+0.285) = 77.821% taken (before LK)
`-> 1155 * 0.77821 = 898.83 taken * 75% = 674.122 ~ 675 taken
hit #2: 294 taken, 503 base = 58.45% taken
`-> 503 * 0.77821 = 391.44 taken * 75% = 293.58 ~ 294 taken (confirmed)
<II> Ebon Void's VD adv adds 10% DmgRes (not final) per stack:
(laser testing)
1500 base hit
test 1: Blocking w/ R1 EV (250%, 0 added Def/DmgRes)
test 2: No blocking w/ 10x VD and 0 added Def/DmgRes:
(should add 100%, or the same as halving dmg at 0 Def/DmgRes)
`-> observed 1500 base -> 750 taken (100% reduction, confirmed)
<III> TK Shield's Reinforcement adv adds ~16.64% DmgRes (not final) lingering effect:
(laser testing)
1500 base hit
@ 0% Def/DmgRes: 1286 taken (85.73%)
`-> 85.73% = 16.64% DmgRes added ( 1-0.8573 / 0.8573)
@ 29% Def/DmgRes: 1030 taken (68.67%)
`-> 29%+16.64% = 45.64% added
`-> 1 / (1+0.4564) = 68.66% (confirmed)
- Be safe and have fun, champs - for science!
Test char: No specs, no superstats, 33% def, 40% invuln, 2.3 pt mitigation. 1500 pt laser in powerhouse.
Full hit: (1/(1+0.33+0.4)*1500)-2.3 = 865 (actual = 866)
Ebon Void
Block, with no stacks = (1/(1+0.33+0.4)*1500)*(1/(1+2.5))-2.3 = 245 (246) (the EV linger procs after the hit connects)
Block stacks on separate layer from defense and defensive passives. We know this already, from the defensive passives thread.
However,
With 1 x EV linger = (1/(1+0.33+0.4)*1500)*(1/(1+0.1))-2.3 = 786 (786)
Note position of the 0.1.
Telekinetic Shield
Block = (1/(1+0.33+0.4)*1500)*(1/(1+2.5))-2.3 = 245 (246)
But!!!
TK linger = (1/(1+0.33+0.4+0.166)*1500)-2.3 = 789 (791)
This may seen trivial, but has MAJOR consequences for heavy-defense/invuln/defiance stacking characters. This is also not unique to invuln, because I tested it with AoRP also as well as no passive (just defense from equipment).
PS it seems that this was mentioned somewhere, buried in a thread from 2010. However, the calculations were not made clear, and no one apparently paid much attention to it.
- Be safe and have fun, champs - for science!
Yep, and yep. (That's Cryptic for you, indeed).
Epic Stronghold
Block timing explained
Force Sheath, No Gear, No Linger: 1439/1500 (+4.2% resistance; ego spec)
Force Sheath, No Gear, Linger: 1332/1500 (+12.6% resistance; +8.4%)
Force Sheath, Gear, No Linger: 969/1500 (+54.8% resistance)
Force Sheath, Gear, Linger: 920/1500 (+63.0% resistance; +8.2%)
If it was block layer, the second would be considerably larger
Phalanx Defense, no gear, no linger: 1500/1500 (0% resistance)
Phalanx Defense, no gear, linger: 1364/1500 (+10.0% resistance; 10% gain)
Phalanx Defense, gear, no linger: 1241/1500 (+20.9%)
Phalanx Defense, gear, linger: 1128/1500 (+33.0%; 12.1% gain)
That's consistent with block layer.
Epic Stronghold
Block timing explained
- Be safe and have fun, champs - for science!
Lots more from SS bonus, Pre bonus, role bonus, Compassion, and bonuses from specs...none of those diminish the bonus amount from gear. Even removing vs. adding all of those bonuses in the amount from sentinel brooches + gear was the same. If I had to guess, the bonus from gear doesn't appear to diminish because, like the bonus from Offense, it's in the severity layer. This is a pure guess though.
PARSER USAGE GUIDE: click here for an easy how-to ٩(๑・ิᴗ・ิ)۶٩(・ิᴗ・ิ๑)۶
Spectre beats Eido in an SG run: VIDEO HERE
- Be safe and have fun, champs - for science!
So hurray for healers, I guess.
- Be safe and have fun, champs - for science!
PARSER USAGE GUIDE: click here for an easy how-to ٩(๑・ิᴗ・ิ)۶٩(・ิᴗ・ิ๑)۶
Spectre beats Eido in an SG run: VIDEO HERE