Jump to content

  • Curse Sites
Help
* * * * * 3 vote(s)

Attack vs. Precision?


  • Please log in to reply
99 replies to this topic

#31 Scarlet_Blossom

Scarlet_Blossom

    Seraph Guardian

  • Members
  • 1411 posts
  • Location:Right here.

Posted 22 May 2012 - 09:54 AM

View PostMammoth, on 22 May 2012 - 09:06 AM, said:

After thinking about it a bit I'm not sure it matters. The line retains its shape anyway, it simply moves up the y axis, which means that at any given point, you're still looking at more returns from each point of power than you are from each point of precision (up to a certain point of crit damage).
Wouldn't every single point in power make another point in power less effective and thus make one point in precision more effective?

#32 Mammoth

Mammoth

    Vanguard Scout

  • Members
  • 458 posts

Posted 22 May 2012 - 10:01 AM

View PostScarlet_Blossom, on 22 May 2012 - 09:54 AM, said:

Wouldn't every single point in power make another point in power less effective and thus make one point in precision more effective?

Not the first part, but yes to the second. Each point in precision makes each point in power more effective too. The graphs account for that, but my comment was probably a bit misleading sorry.

It would probably be a bit clearer if the y axis was 'damage gain from 100 pow/pre' rather than 'average damage', so you could drop the base weapon and power and replace 1403 with 100 in the equation, reducing the scale of the whole thing. You can actually replace 1403 with any number you like. The ability coefficient applies throughout so that's already been dropped. It looks correct to me, but I'm quickly finding out that I've lost all mathematical ability.

Edited by Mammoth, 22 May 2012 - 10:16 AM.


#33 Scarlet_Blossom

Scarlet_Blossom

    Seraph Guardian

  • Members
  • 1411 posts
  • Location:Right here.

Posted 22 May 2012 - 10:16 AM

View PostMammoth, on 22 May 2012 - 10:01 AM, said:

Not the first part, but yes to the second. Each point in precision makes each point in power more effective too. The graphs account for that, but my comment was probably a bit misleading sorry.

It would probably be a bit clearer if the y axis was 'damage gain from 100 pow/pre' rather than 'average damage', so you could drop the base weapon and power and replace 1403 with 100 in the equation.
My point is that if you have higher 'base' power (since amulets can give both power and precision) aftwerwards the increase diminishes faster than it would at a lower 'base' power.
I know that the graphs account for that but I think it might make a difference at which critical multiplier the effectivity of a single point in precision is higher than a single point in strength at a certain point.
I just don't think that anything below 1.8x crit multiplier does not really benefit from Precision at all compared to Power. I may be wrong though.

#34 Mammoth

Mammoth

    Vanguard Scout

  • Members
  • 458 posts

Posted 22 May 2012 - 10:25 AM

View PostScarlet_Blossom, on 22 May 2012 - 10:16 AM, said:

My point is that if you have higher 'base' power (since amulets can give both power and precision) aftwerwards the increase diminishes faster than it would at a lower 'base' power.
I know that the graphs account for that but I think it might make a difference at which critical multiplier the effectivity of a single point in precision is higher than a single point in strength at a certain point.
I just don't think that anything below 1.8x crit multiplier does not really benefit from Precision at all compared to Power. I may be wrong though.

I was thinking the same thing before, but discounted it due to the line retaining it's shape and only moving up the y axis. I probably missed something though.

Yep, you're correct, changing the attr value does change the shape of the line. Also, the ability coefficient drastically changes the shape of the line, so that shouldn't be dropped either. In fact, with a coefficient of .053 (such as stalker's strike), the positive section of the line is nearly flat, i.e, precision~power at any value we will see in game. This makes more sense to me, I only wish I could figure out why it's the case :P

I'm not good enough at math to figure out what the equation should be, but I think we have a problem.

Edited by Mammoth, 22 May 2012 - 11:26 AM.


#35 FrogReaver

FrogReaver

    Vanguard Scout

  • Members
  • 172 posts

Posted 22 May 2012 - 02:09 PM

View PostScarlet_Blossom, on 22 May 2012 - 10:16 AM, said:

My point is that if you have higher 'base' power (since amulets can give both power and precision) aftwerwards the increase diminishes faster than it would at a lower 'base' power.
I know that the graphs account for that but I think it might make a difference at which critical multiplier the effectivity of a single point in precision is higher than a single point in strength at a certain point.
I just don't think that anything below 1.8x crit multiplier does not really benefit from Precision at all compared to Power. I may be wrong though.

First of all I think there are a few misinterpretations about the graph.  The graph doesn't say that given a certain crit multiplier that you only need x amount of precision and then add the rest to power (the new graphs I requested should give us some insight in that direction).  Also, my base attack value is a little low for actual ingame numbers.  I think that the 1500 should be replaced by 1800 (I predict that this will shift the maximums a little more toward the center and if they do this means a more even split of power and precision is needed for the current power-precision pool).

Anyways, the current graph does say that for a certain pool that you split between power and precision that their is a maximum.  What I hope happens is that if you decrease or increase the pool of available power and precision points that the maximums on those graphs still occur at near the same percentage of that pool being used for power and for precision.  This is yet to be verified though.  If this happens it would mean that for a given crit multiplier that you are most effective splitting your stats according to a given percentage for a specified crit multiplier.

Also, pvp ammys focus on 1 major stat and 2 minor stats.  They give a chunk of 800 to the major and 580 to the minors other than critdmg% and condition dmg.  (31% crit dmg - minor only and 901 condition dmg major and 643 condition dmg minor).

Runes typically can add up to 165 to a major stat and 50 to a minor stat (or you can trade the minor stat for something like extra boon duration).

Jewels add 25 major and 15 to two minor stats other than %crit dmg and condition dmg again).  This lets you get 150 major stat and 90 in 2 minor stats.

I don't want to talk about traits because I think the general consensus is it's better to get useful traits for your build and not worry so much about what stats they are adding too.

This means that from gear: 800 + 165 + 50 = 1115 for a major stat and 580 + 90 + 50= 720 for a minor (I don't think that you can get precision as a major and power as a minor).

Another interesting note:
Pvp armor values (I have the exact values written down somewhere.  I'll find them later and update this):
Heavy: around 1200
Medium is around 1050
Light is around 900

#36 Mammoth

Mammoth

    Vanguard Scout

  • Members
  • 458 posts

Posted 22 May 2012 - 02:22 PM

View PostFrogReaver, on 22 May 2012 - 02:09 PM, said:

Anyways, the current graph does say that for a certain pool that you split between power and precision that their is a maximum.  What I hope happens is that if you decrease or increase the pool of available power and precision points that the maximums on those graphs still occur at near the same percentage of that pool being used for power and for precision.  This is yet to be verified though.  If this happens it would mean that for a given crit multiplier that you are most effective splitting your stats according to a given percentage for a specified crit multiplier.

The vertices do vary along both the x and y axes when you vary the size of the pool, which throws a spanner in the works. They're also influenced by varying base power+weapon (moving left as that number gets higher, making crit more effective). Unsure if they change with the ability coefficients, but the flattening effect those coefficients have certainly makes traits like bleed on crit a lot more significant.

Edited by Mammoth, 22 May 2012 - 02:36 PM.


#37 Scarlet_Blossom

Scarlet_Blossom

    Seraph Guardian

  • Members
  • 1411 posts
  • Location:Right here.

Posted 22 May 2012 - 02:54 PM

View PostFrogReaver, on 22 May 2012 - 02:09 PM, said:

Also, pvp ammys focus on 1 major stat and 2 minor stats.  They give a chunk of 800 to the major and 580 to the minors other than critdmg% and condition dmg.  (31% crit dmg - minor only and 901 condition dmg major and 643 condition dmg minor).
+830 power and +580 precision don't seem that minor to me on the crit damage amulet?

#38 FrogReaver

FrogReaver

    Vanguard Scout

  • Members
  • 172 posts

Posted 22 May 2012 - 02:54 PM

View PostMammoth, on 22 May 2012 - 02:22 PM, said:

The vertices do vary along both the x and y axes when you vary the size of the pool, which throws a spanner in the works. They're also influenced by varying base power+weapon (moving left as that number gets higher, making crit more effective). Unsure if they change with the ability coefficients, but the flattening effect those coefficients have certainly makes traits like bleed on crit a lot more significant.

How much do the vertices vary along the x and y with changing the size of the pool?  And I totally agree the %chacne to bleed on crit may be a pretty significant source of dmg even it it only happens once every 4 or 5 attacks with a high crit chance (66%).

View PostScarlet_Blossom, on 22 May 2012 - 02:54 PM, said:

+830 power and +580 precision don't seem that minor to me on the crit damage amulet?

LOL.  I only called it minor because it didnt add as much as the one I called major ;)

#39 Mammoth

Mammoth

    Vanguard Scout

  • Members
  • 458 posts

Posted 22 May 2012 - 03:10 PM

View PostFrogReaver, on 22 May 2012 - 02:54 PM, said:

How much do the vertices vary along the x and y with changing the size of the pool?  And I totally agree the %chacne to bleed on crit may be a pretty significant source of dmg even it it only happens once every 4 or 5 attacks with a high crit chance (66%).

I have no idea :P

I'm just plugging values into the equation and graphing them with an online calc, not doing any math. Hmm, looking at it, the vertex moves to the right when the size of the pool is increased, meaning precision becomes less effective. I'm ready to give up on this, very frustrating to be blundering around failing at basic math. I can't figure out an equation that allows the attr value to be variable, you did a 3d graph earlier, any chance you could figure it out?

Edited by Mammoth, 22 May 2012 - 03:16 PM.


#40 FrogReaver

FrogReaver

    Vanguard Scout

  • Members
  • 172 posts

Posted 22 May 2012 - 03:25 PM

View PostMammoth, on 22 May 2012 - 03:10 PM, said:

I have no idea :P

I'm just plugging values into the equation and graphing them with an online calc, not doing any math. Hmm, looking at it, the vertex moves to the right when the size of the pool is increased, meaning precision becomes less effective. I'm ready to give up on this, very frustrating to be blundering around failing at basic math. I can't figure out an equation that allows the attr value to be variable, you did a 3d graph earlier, any chance you could figure it out?

I'll do some more analysis.

#41 HatoraDe

HatoraDe

    Asuran Acolyte

  • Community Contributors
  • 80 posts
  • Location:California
  • Guild Tag:[JAM]

Posted 22 May 2012 - 04:32 PM

Definitely keeping an eye on this thread, good work so far guys!

#42 FrogReaver

FrogReaver

    Vanguard Scout

  • Members
  • 172 posts

Posted 22 May 2012 - 08:09 PM

On a side note: I've concluded that it's impossible to compare the effects +condition damage to power and precision without first having some idea of what base skill damages are.  If anyone has tested that or does test that please post the result here.

I'm having some issues pasting a graph from open office to here.  Any suggestions?

file:///C:/DOCUME%7E1/Killer/LOCALS%7E1/Temp/moz-screenshot-6.jpg

I checked things again and I have made one major mistake.  Increasing attack by 1% doesn't increase damage by 1%.  Increasing power by 1% does.  This means I shouldn't take the extra 900 from weapon for attack into account.

Basically (after correcting to fix the formula not to take into account attack, which may be a somewhat flawd assumption?) my current analysis is showing that for large possible to achieve ingame power-precision pools it's always best to max power and take whatever extra precision you can get.  For lower power-precision stat pools it's still better to go more power but it's not necessarily best to go all power.  For lower stat pools there are achievable maximums where you will want to skimp a little on power and take a little more precision depending on your crit multiplier.  Though, these maximums don't begin to show up till near the times 2 crit muliplier.

I'll be happy to show graphs if someone can help with my issue of getting them to show up here.

View PostMammoth, on 22 May 2012 - 01:43 AM, said:

Never mind, found some more testing that shows a multiplicative relationship, but it does throw up another problem:



This may not be much of a problem for most abilities though, as I imagine the coefficient directly relates to the cooldown/initiative cost (animation speed for 1skills) of the attack with exceptions for extra effects like leaps, aoe, evades, cripples, etc. It does mean there appears to be quite a few variables though, time, damage, coefficient, attack, defense, crit chance, crit bonus. I only add defense because I've also seen differing interpretations on which variables it affects, with one interpretation suggesting it may be difficult to remove from the equation.

If you're interested in doing some comprehensive testing during the next BWE shoot me a PM.

Even more interesting with your numbers is that slashes multiplier is exactly 3 times higher than the other skills.

Edited by FrogReaver, 22 May 2012 - 07:48 PM.


#43 EatThisShoe

EatThisShoe

    Vanguard Scout

  • Members
  • 369 posts

Posted 22 May 2012 - 09:39 PM

Are you counting prowess in your stat pool? Because you give up stats to get that crit multiplier in most cases.

#44 FrogReaver

FrogReaver

    Vanguard Scout

  • Members
  • 172 posts

Posted 22 May 2012 - 10:00 PM

View PostEatThisShoe, on 22 May 2012 - 09:39 PM, said:

Are you counting prowess in your stat pool? Because you give up stats to get that crit multiplier in most cases.

Nope.  It's another variable but I have taken it into account by looking at bigger/smaller stat pools and adjustable crit modifers.  While my analysis will not reveal the best power - precision - and crit multiplier you can get to max everything all at once, it does tell you that going crit over power without some in +%critDmg is generally not the best idea for raw damage output.

Another note: the only 2 pvp ammys  add power and precision.  1 of those adds 31% critDmg.  Theres a free 31% just from that ammy if you choose to take it.  The other adds condition damage which is interesting but not really in the scope of this thread to or really our ability at the time to compare condition damage to regular damage.

As far as stats from the trait tree... it's stupid IMO to build a guys traits around the stats you can get from the trait tree.  Sure if you are undecided about a trait in precision tree or a trait in toughness tree then use the stat as a tiebreaker but other than that ya really should build ur guy around traits and not so much the stats you get from the trait trees.  Stats should be a gear decision.

The 300 precision you can get from that tree makes a whopping 14% crit chance difference.  The 300 power is only going to be 10-20% of your total power (and less of attack if that's important at all).  300 in life is only going to be 3000 life, a 30% difference for an elementalist and maybe a 17% difference for a warrior.  The extra prowess at most is just a 20% increase in crit multiplier and probably closer to 15% increase for most characters.  Yes these small percentages are nice but the traits you can get seem for more important than 20% increases 2 or 3 of your stats.  And besides you are probably going to want 1 or 2 major traits from a given tree anyways so you are at least getting 1/3 of the values listed above.

Edited by FrogReaver, 22 May 2012 - 10:18 PM.


#45 misterdevious

misterdevious

    Sylvari Specialist

  • Members
  • 699 posts

Posted 22 May 2012 - 10:41 PM

Everyone can get the same stats from gear, but the stats you get from traits will vary by profession.  That makes some professions better than others when you try to get specific combinations of stats, especially when you are getting stats that aren't available on amulets or jewels.


Rangers can get higher Power/Precision/Prowess than Warriors, because they only need to invest in two traitlines (Marksmanship, Skirmishing) and the Warrior would have to invest in three (Strenght, Arms, Tactics). You can't maximize three traitlines which means the Warrior would have less of one of the three "big crit" stats.  He may have the same Power and the same Precision, but the ranger could have +20% Prowess.

Warriors though can have better on-crit bleeds than Rangers, because they can get Precision, Malice, and Expertise in only two lines (along with Power), but the Ranger would require three.

#46 FrogReaver

FrogReaver

    Vanguard Scout

  • Members
  • 172 posts

Posted 22 May 2012 - 11:23 PM

View Postmisterdevious, on 22 May 2012 - 10:41 PM, said:

Everyone can get the same stats from gear, but the stats you get from traits will vary by profession.  That makes some professions better than others when you try to get specific combinations of stats, especially when you are getting stats that aren't available on amulets or jewels.


Rangers can get higher Power/Precision/Prowess than Warriors, because they only need to invest in two traitlines (Marksmanship, Skirmishing) and the Warrior would have to invest in three (Strenght, Arms, Tactics). You can't maximize three traitlines which means the Warrior would have less of one of the three "big crit" stats.  He may have the same Power and the same Precision, but the ranger could have +20% Prowess.

Warriors though can have better on-crit bleeds than Rangers, because they can get Precision, Malice, and Expertise in only two lines (along with Power), but the Ranger would require three.

I think we already have enough information to answer that question.  It would just take the creation of two or 3 of the graphs like the one guy posted earlier.  A warrior would require 2 or 3 such graphs but it would be very doable.  A ranger can max both power, precision and prowess in traits.  So his analysis just takes one similar graph to what the earlier poster posted.

#47 Milkhead

Milkhead

    Asuran Acolyte

  • Curse Premium
  • Curse Premium
  • 53 posts
  • Guild Tag:[THAI]
  • Server:Blackgate

Posted 23 May 2012 - 03:38 AM

View PostFrogReaver, on 22 May 2012 - 08:09 PM, said:


I checked things again and I have made one major mistake.  Increasing attack by 1% doesn't increase damage by 1%.  Increasing power by 1% does.  This means I shouldn't take the extra 900 from weapon for attack into account.

Basically (after correcting to fix the formula not to take into account attack, which may be a somewhat flawd assumption?) my current analysis is showing that for large possible to achieve ingame power-precision pools it's always best to max power and take whatever extra precision you can get.  For lower power-precision stat pools it's still better to go more power but it's not necessarily best to go all power.  For lower stat pools there are achievable maximums where you will want to skimp a little on power and take a little more precision depending on your crit multiplier.  Though, these maximums don't begin to show up till near the times 2 crit muliplier.


Clearly it gets more complex when other variables need taken into account (effect of trait, condition duration, condition dmg etc.), but can we simply conclude the following:


when Crit Multiplier <1.8 we should just foucs on Power

when Crit Multiplier >1.8 we should aim for Extra Power apporx. 800 Extra Precision approx. 500

extra Precision: 500 is around 24%

can we simply conclude that when Crit Multiplier is >1.8 the amount of extra crit chance we need is 24%

instead of obtaining more extra crit chance, it's more effective to either obtaining extra power or extra crit multiplier?

and which has more value then, extra power or extra crit multiplier?





is the shape of the dmg vs power/precision pool graph consistent with varying size of total stat pool (in other word, is it valid to apply this model for the whole level range 1-80)?
Siberz [Thai]
Blackgate

#48 FrogReaver

FrogReaver

    Vanguard Scout

  • Members
  • 172 posts

Posted 23 May 2012 - 05:06 AM

Here's my graphs.  I modified the equation and took attack out of it since we observed a 1% power increase giving a 1% damage increase (instead of a 1% attack increase giving a 1% damage increase).  So the new equation is:

AvgDmg = (900+x)*(0.96-(totalStatPool-x)/2100)+(900+x)*2.1*(0.04+(totalStatPool-x)/2100)

*900 is for starting power
*Also please note that I considered baseStat+1403 to be the most obtainable value in one stat.  This should be pretty close to the actual ingame value.  Therefore, I cut all my graphs off at 1403 power even when they had larger statPools.




Album: power vs. precision graphs
7 images
1 comments

View PostMilkhead, on 23 May 2012 - 03:38 AM, said:

Clearly it gets more complex when other variables need taken into account (effect of trait, condition duration, condition dmg etc.), but can we simply conclude the following:


when Crit Multiplier &--#60;1.8 we should just foucs on Power From my new graphs it appears Crit Multiplier needs to be closer to 2.0 or higher and this only happens when our power-precision pool is between 703 and 2003.  Anything more or less than that is better maxing power and then taking on what precision if any that they can.

when Crit Multiplier &--#62;1.8 we should aim for Extra Power apporx. 800 Extra Precision approx. 500 extra Precision:  It's more accurate to say that the most power we should ever trade for precision is 350ish from max and this only happens around the 1403 pool.

500 is around 24%

can we simply conclude that when Crit Multiplier is &--#62;1.8 the amount of extra crit chance we need is 24%
This is not true and not what the graphs actually show.  We never get to so much crit% and then stop.  We always want to add more.  It's simply we don't want to have to trade very much precision for power in order to obtain it.  

instead of obtaining more extra crit chance, it's more effective to either obtaining extra power or extra crit multiplier?
First of all from a quick glace crit multiplier doesn't appear to keep up between graphs.  Even obtaining a +30% crit multiplier (going up 3 lines) didn't ever appear to be as good as 300 extra in the stat pool to spread around.  (look at the max average damage on the next graph on the same crit multiplier and compare it to the max on the same graph 3 lines higher).

2ndly the graphs aren't saying stop adding to precision they are saying don't trade out much power in order to obtain it.  

and which has more value then, extra power or extra crit multiplier?
Crit multiplier is important but I'm beginning to be more convinced its not as good as 300 more stats in the power-precision stat pool.  




is the shape of the dmg vs power/precision pool graph consistent with varying size of total stat pool (in other word, is it valid to apply this model for the whole level range 1-80)?

I'm not sure how level effects the results.  We know the starting power is different which does have an effect and the crit formula may also be different depending on level (My guess would be that it is).  

Edited by FrogReaver, 23 May 2012 - 05:13 AM.


#49 Milkhead

Milkhead

    Asuran Acolyte

  • Curse Premium
  • Curse Premium
  • 53 posts
  • Guild Tag:[THAI]
  • Server:Blackgate

Posted 23 May 2012 - 07:32 AM

View PostFrogReaver, on 23 May 2012 - 05:06 AM, said:

Here's my graphs.  I modified the equation and took attack out of it since we observed a 1% power increase giving a 1% damage increase (instead of a 1% attack increase giving a 1% damage increase).  So the new equation is:

AvgDmg = (900+x)*(0.96-(totalStatPool-x)/2100)+(900+x)*2.1*(0.04+(totalStatPool-x)/2100)

*900 is for starting power
*Also please note that I considered baseStat+1403 to be the most obtainable value in one stat.  This should be pretty close to the actual ingame value.  Therefore, I cut all my graphs off at 1403 power even when they had larger statPools.








I'm not sure how level effects the results.  We know the starting power is different which does have an effect and the crit formula may also be different depending on level (My guess would be that it is).  


sry i should communicate better.

trade off between precision and power was what i meant.

We never get so much crit and then stop. I simply meant, with all the analysis undergoing, should we raise crit to a certain number and after that we should try to raise Power / Crit Modifier instead as the tradeoff for crit is not as good as that point.

which u later commented you believe Power is more valuable than Crit modifier.

---------


and as things get more n more complex we dig deeper into it.

is there any useful information that can be obtained at this point???

my feeling is yes but im not yet sure how to arrive at a conclusion.

Edited by Milkhead, 23 May 2012 - 07:34 AM.

Siberz [Thai]
Blackgate

#50 FrogReaver

FrogReaver

    Vanguard Scout

  • Members
  • 172 posts

Posted 23 May 2012 - 02:00 PM

View PostMilkhead, on 23 May 2012 - 07:32 AM, said:

sry i should communicate better.

trade off between precision and power was what i meant.

We never get so much crit and then stop. I simply meant, with all the analysis undergoing, should we raise crit to a certain number and after that we should try to raise Power / Crit Modifier instead as the tradeoff for crit is not as good as that point.

which u later commented you believe Power is more valuable than Crit modifier.

---------


and as things get more n more complex we dig deeper into it.

is there any useful information that can be obtained at this point???

my feeling is yes but im not yet sure how to arrive at a conclusion.

The difference isn't very extreme unless you go all crit compared to the maximum on the graphs.  It seems to me that on almost all the graphs you can stay anywhere from 5/14 to 6/14 of max power and lose no more than 10% damage (max power = 1403 for statpools > 1403 and the stat pool size for ones less than 1403).  So basically after putting a moderate amount from your stat pool towards power I would feel more than free to go either precision or power from there.

Another point of consideration.  It seems that 1% crit damage is rarely if ever worth 10 more in the stat pool.  Therefore it seems getting more power or precision from traits will tend to help more than getting more prowess.

Edited by FrogReaver, 23 May 2012 - 02:14 PM.


#51 Yourgrandmother

Yourgrandmother

    Asuran Acolyte

  • Members
  • 137 posts
  • Guild Tag:[KGB]
  • Server:Fort Aspenwood

Posted 23 May 2012 - 02:32 PM

Anyone that has played multiple MMO's knows that real DPS/Burst comes from Crit Damage builds.

#52 FrogReaver

FrogReaver

    Vanguard Scout

  • Members
  • 172 posts

Posted 23 May 2012 - 03:04 PM

View PostYourgrandmother, on 23 May 2012 - 02:32 PM, said:

Anyone that has played multiple MMO's knows that real DPS/Burst comes from Crit Damage builds.

That's not what the numbers show...

Also.  +% crit gear/traits skews the graphs even more in favor of going max power and then tack on what precision you can get.  I'm sorry to inform you but in GW2 it seems power should be your primary focus for direct damage builds.  (I can't come up with a valid test between Condition Damage and Direct Damage yet without knowing base damages and how toughness/armor affects direct damage).  But anyways, for direct damage, precision and prowess both definetly seem to be secondary stat focuses for damage because only in a small handful of situations will it be worth trading any power out for more of them at all.

That being said as long as you can get some power then going for all precision from there can't really hurt your damage by more than 10%.  So in the end it's going to probably come down to determing whether a little better chance of procing effects is worth more than 10% more overall direct damage.

I also wanted to add, that you can have a better chance of high burst damage over 3-4 attacks (You could even have a decent chance of critting on every attack in that 3-4 attack (not skill) series, maybe a 10% or 20% chance for that to happen)  If you can ensure victory in those 3-4 attacks (not skills) then I would suggest going more crit just for the chance to kill an opponent in 3-4 attacks (not skills, but attacks).  However, if it's going to take you 10 attacks to kill a guy then you're probably better off looking at which build gets you the higher average damage.  I think guild wars to generally takes more than 3-4 hits to kill a guy?  However if you do manage to have a build that can kill others in 3 or 4 critical attacks (not skills) then it makes total sense to go for more crit for the higher chance of an extremly high 3-4 attacks.

Edited by FrogReaver, 23 May 2012 - 03:19 PM.


#53 Yourgrandmother

Yourgrandmother

    Asuran Acolyte

  • Members
  • 137 posts
  • Guild Tag:[KGB]
  • Server:Fort Aspenwood

Posted 23 May 2012 - 03:19 PM

Forget the numbers.

BURST DPS builds come from having ideal trait/sigil synergy to kill your opponent in seconds. If you are talking about long PvE fights against bosses sure consistant DPS by going with more Power may be helpful. In PvP ideal DPS is always from Burst (which comes from near guaranteed crits + crit damage multipliers). Why do you think people are bitching about Doom and Intelligence sigils?

Many professions have multiple traits when used with certain utility abilities and their main DPS ability (hundred blades, unload, pistol whip, blurred frenzy, frenzy, whirling axes etc etc etc.) will far out DPS non crit high power builds. You can play with your numbers all you want, after you've played several PvP MMO's you'll understand that they key to winning outnumbered zerg bombing fights or being relied on as a DPS in SPvP 5v5/8v8 versus actual human beings is from burst damage and not sustained damage against braindead AI.

#54 FrogReaver

FrogReaver

    Vanguard Scout

  • Members
  • 172 posts

Posted 23 May 2012 - 03:30 PM

View PostYourgrandmother, on 23 May 2012 - 03:19 PM, said:

Forget the numbers.

BURST DPS builds come from having ideal trait/sigil synergy to kill your opponent in seconds. If you are talking about long PvE fights against bosses sure consistant DPS by going with more Power may be helpful. In PvP ideal DPS is always from Burst (which comes from near guaranteed crits + crit damage multipliers). Why do you think people are bitching about Doom and Intelligence sigils?
People complain about alot of stupid stuff.  Usually its because x guy killed me he's OP he needs nerfed.  99% of the time the guy isn't actually OP and smarter play or a change in build could easily handle him.  Ideal PVP DPS depends on the number of attacks you have to make in order to kill your opponent.  If you can get a high chance for that to be a low number with crits then by all means up them a little more.  However, multi-hit skills don't help in this scenario.  It's better to go high average dmg with any multihit skill over going crit with them in this game.

Many professions have multiple traits when used with certain utility abilities and their main DPS ability (hundred blades, unload, pistol whip, blurred frenzy, frenzy, whirling axes etc etc etc.) will far out DPS non crit high power builds. You can play with your numbers all you want, after you've played several PvP MMO's you'll understand that they key to winning outnumbered zerg bombing fights or being relied on as a DPS in SPvP 5v5/8v8 versus actual human beings is from burst damage and not sustained damage against braindead AI.

Hundred blades , whirling axes, and I don't know how many other of those skills you listed are multi hit skills.  Multihit skills hit multiple times which means you are already giving a huge headstart for a high average damage build to overtake them.  Not to mention, most high average damage builds are still going to have good crit rates.  So really you aren't OMG ganking that many more players with a high crit build than a high average damage build which automatically has a decent crit rate can.

If you really want to test this give me some power/precision and trait tree you think would be good.  I'll come up with a more power focused build and we will number crunch to see who is better.  Just tell me what weapon skills we are going to be using and such.    I guarantee you my build more power focused build will win more % of the time than your build.  Assuming you don't already take the gear I would take for a more power focused build.

Edited by FrogReaver, 23 May 2012 - 03:37 PM.


#55 Artaz

Artaz

    Vanguard Scout

  • Members
  • 151 posts

Posted 23 May 2012 - 03:39 PM

View PostMilkhead, on 23 May 2012 - 03:38 AM, said:

Clearly it gets more complex when other variables need taken into account (effect of trait, condition duration, condition dmg etc.), but can we simply conclude the following:


when Crit Multiplier <1.8 we should just foucs on Power

when Crit Multiplier >1.8 we should aim for Extra Power apporx. 800 Extra Precision approx. 500

extra Precision: 500 is around 24%

can we simply conclude that when Crit Multiplier is >1.8 the amount of extra crit chance we need is 24%

instead of obtaining more extra crit chance, it's more effective to either obtaining extra power or extra crit multiplier?

and which has more value then, extra power or extra crit multiplier?





is the shape of the dmg vs power/precision pool graph consistent with varying size of total stat pool (in other word, is it valid to apply this model for the whole level range 1-80)?


Probably correct.  Except that what is missed is the Prowess trait which is hypothetically increased as Precision is increased.  Just from the +300 Precision trait increase, +30 Prowess also increases critMultiplier (not sure if +1 Prowess = +1% critMultiplier).

This would mean that as gear increases Precision, Prowess is also increased.  This would mean > +180% critMod threshold appears to be relatively easy to achieve just through gear and traits.  Also, Rangers have +33% critMultiplier just from Honed Axes, Elementalists have +3% to +9% critMultiplier from Arcane Lightning, etc. weighing Precision further over Power.

TL;DR, the effect of Prowess (where Precision is greater than Power after 180% critMultiplier) needs to be added to the calculation.

#56 FrogReaver

FrogReaver

    Vanguard Scout

  • Members
  • 172 posts

Posted 23 May 2012 - 03:49 PM

View PostArtaz, on 23 May 2012 - 03:39 PM, said:

Probably correct.  Except that what is missed is the Prowess trait which is hypothetically increased as Precision is increased.  Just from the +300 Precision trait increase, +30 Prowess also increases critMultiplier (not sure if +1 Prowess = +1% critMultiplier).

This would mean that as gear increases Precision, Prowess is also increased.  This would mean &--#62; +180% critMod threshold appears to be relatively easy to achieve just through gear and traits.  Also, Rangers have +33% critMultiplier just from Honed Axes, Elementalists have +3% to +9% critMultiplier from Arcane Lightning, etc. weighing Precision further over Power.

TL;DR, the effect of Prowess (where Precision is greater than Power after 180% critMultiplier) needs to be added to the calculation.

You can already look at that.  Each graph has about 9 different crit multipliers (each point in prowess ups this by 1%)  So each line going up on the graph can represent 10 more prowess.  Or 3 lines represent 30 more prowess.  So whatever power value you want to look at just look at the average damage on one graph and then go to the next one at the same power value and move the line 3 up.  Then move up 100 power and down a line.  Move up 100 more power and down a line.  And do it one more time.  Whichever of those is higher tells you how you should add to traits for stats.  (You will need to do this for each graph and try to draw some general conclusion from the results you obtain... GOOD LUCK!)  I assume most people will pick useful traits instead of just going out in a tree for the stats therefore I didn't really spend much time answering the question of whether its better to add to a power trait line or a precision/prowess trait line.

Edited by FrogReaver, 23 May 2012 - 03:52 PM.


#57 Brilhasti

Brilhasti

    Asuran Acolyte

  • Members
  • 127 posts

Posted 23 May 2012 - 04:35 PM

View PostFrogReaver, on 21 May 2012 - 06:32 PM, said:

If two warriors are fighting with identical weapons, skills and traits which stat should I focus my gear on for the most average damage? Or is there a sweet spot like half power - half precision that I should aim for?

I came up with this formula for average damage:
AvgDmg = BaseDmg*Attack*(1 - Critical%) + BaseDmg*Attack*(CriticalMultiplier)*(Criticial%)

Since BaseDmg is the same for two identical warriors then we can disregard it.  Therefore,
AvgDmg* = Attack*(1 - CriticalProbability) + Attack*(CriticalMultiplier)*(CriticalProbability)

From another guy's work he stated that for level 80:
CriticalProbability = .04 + (Precision - 916)/21000

Therefore,
AvgDmg* = Attack*(1 - .04 - (Precision - 916)/21000) + Attack*(CriticalMultiplier)*(Precision-916)/21000

Now I pick a critical multiplier for both characters.  Let's say the base multiplier which is 1.5 (you can try other multipliers but others don't appear to affect the look of this graph at all)

AvgDmg* = Attack*(1 - .04 - (Precision-916)/21000) + Attack*(1.5)*(Precision - 916)/21000

I then graph this 3 variable equation (I used http://www.livephysi...ion-grapher.php)

let x = attack and y = precision
x min = 1500 (916 + weaponAttack)
x max = 3700 (not positive what the actual max attack value is but this should be close)
ymin = 916
ymax = 3116 (916 + 2200 where 2200 is the difference between max and min power because you should be able to get as much extra precision as you can power)

From this graph it shows that when both precision and power are maxed we get the most damage, however we can't max both of them.  Therefore you need to draw a diagnol from (maxattack , minprecision) down to (minattack , maxprecision).  This diagnol represents the trade off of attack with precision.  From it we see that z or AvgDmg* is highest when attack is maxed and precision is left at base.

Therefore I conclude that maxing attack without worrying about precision is the way to do for the most AvgDmg*.  If there were any flaws in my analysis please let me know.  Thanks.

My god,your equations just solved the whole "travelling through wormholes" difficultys we had over the last few years.

Thank you,thank you.

Kindest regards,NASA.

#58 Yourgrandmother

Yourgrandmother

    Asuran Acolyte

  • Members
  • 137 posts
  • Guild Tag:[KGB]
  • Server:Fort Aspenwood

Posted 23 May 2012 - 04:48 PM

View PostFrogReaver, on 23 May 2012 - 03:30 PM, said:

Hundred blades , whirling axes, and I don't know how many other of those skills you listed are multi hit skills.  Multihit skills hit multiple times which means you are already giving a huge headstart for a high average damage build to overtake them.  Not to mention, most high average damage builds are still going to have good crit rates.  So really you aren't OMG ganking that many more players with a high crit build than a high average damage build which automatically has a decent crit rate can.

If you really want to test this give me some power/precision and trait tree you think would be good.  I'll come up with a more power focused build and we will number crunch to see who is better.  Just tell me what weapon skills we are going to be using and such. I guarantee you my build more power focused build will win more % of the time than your build.  Assuming you don't already take the gear I would take for a more power focused build.

Yes I know that Doom/Intelligence only work on the first hit of a weapon swap and not the combination. Those skills were examples of burst abilities that you can use even WITHOUT those sigls for burst damage with crit % stacking traits and stat allocation which can kill opponents in its duration.

I'm done on this subject, have known guys like you for 10 years and things never change. Burst damage kills, even more so in a twitch-like game like GW2 with dodging. DPS after 20 seconds is meaningless in PvP, but hey keep plugging away and stacking a bunch of power and 10% crit with no prowess.

#59 FrogReaver

FrogReaver

    Vanguard Scout

  • Members
  • 172 posts

Posted 23 May 2012 - 05:36 PM

View PostYourgrandmother, on 23 May 2012 - 04:48 PM, said:

Yes I know that Doom/Intelligence only work on the first hit of a weapon swap and not the combination. Those skills were examples of burst abilities that you can use even WITHOUT those sigls for burst damage with crit % stacking traits and stat allocation which can kill opponents in its duration.

I'm done on this subject, have known guys like you for 10 years and things never change. Burst damage kills, even more so in a twitch-like game like GW2 with dodging. DPS after 20 seconds is meaningless in PvP, but hey keep plugging away and stacking a bunch of power and 10% crit with no prowess.

You are assuming that I can't do high bursts with a near maximized average damage build.  You are also assuming that my crit% chance will be 10%.  That's laughable as the power-precision ammy gives +27.6% all on its own.  My crits with nothing else are going to be 31.6%.  Now if i go all power and precision gear my crits can get up to around 38% (Without traits).  Considering most of this gear also adds %critdmg I'm goina be floating at about a 2 crit multiplier.

I'll add more later.

Edited by FrogReaver, 23 May 2012 - 05:39 PM.


#60 Sephiroth

Sephiroth

    Vanguard Scout

  • Members
  • 386 posts

Posted 24 May 2012 - 12:37 AM

Can you make that Critmultiplier list along with the recommended Power-Precision for each stage?  There are too many graphs so I don't quite understand. thanks.




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users