What is wrong with the CV system...

Discussion in 'General Archive' started by _Baragain_, Nov 21, 2015.

Dear forum reader,

if you’d like to actively participate on the forum by joining discussions or starting your own threads or topics, please log into the game first. If you do not have a game account, you will need to register for one. We look forward to your next visit! CLICK HERE
Thread Status:
Not open for further replies.
  1. _Baragain_

    _Baragain_ Living Forum Legend

    Well, BigPapa has his thread about what we don't like about CV, but now it is my turn to share with you what is wrong with the system... Be warned, it is a lot!

    I wrote this for support to pass to the powers that be, but I figured I'd share it with all of you.

    • The system does not truly allow direct comparison between characters, even those of the same class, due to the difference in builds. In the thread referenced above there is the idea of having separate scores for offence and defence and then as discussed in the other thread there could be an aggregate score for overall comparison. Doing this would allow two DPS focused builds to be compared or two tank builds to be compared. As a group lead and tank looking to invite a SM to my group, I would rather have a SM with an “offensive CV” that is off the chart than a SM who’s CV (in the existing system) looks really good, but it is because he is in a flag running build that would take forever to be able to kill even a basic monster. Additionally, if I were a SW trying to form a Balor group, I’d be looking for a DK with a “defensive CV” that indicates that he can take a beating and I wouldn’t care if he had virtually no damage. If the tool is supposed to allow us to compare relative strength, then let us compare apple to apples and oranges to oranges.

    • Armor, resistance, block rate, and block strength are all stats that offer diminishing returns as you gain more of them. However, resistance/armor have a flat rate that is independent from the total reduction. On the other hand, HP offers gains no matter how much you have. This is in direct to contrast to how CV is calculated. HP seems to be rated at 1 CV/HP where as armor and resistance seem to be rated at 2 CV/AP or CV/RP. This does absolutely nothing to account for the diminishing returns of those stats. Basically, if I had a DK with 15K armor, that would equate to 30K CV, but my damage reduction from armor would have capped out at 80% when I hit 9480 AP. Compare this to if I had 5k armor and 10K hp, I'd have a CV of 20K, but much better survivability with a damage reduction of 67.84%. I don't know about you, but as an experienced player, I would gladly swap 12.15% damage reduction for 10K HP (maybe even more once you pass it through the knowledge tree which would boost the CV a little at the same time), but the CV systems seems to indicate that this is a bad decision (30K CV-10K CV=20K CV).

    • Direct comparison of items is ineffective at this time for three reasons.
      • First, as described above, the CV system already fails to account for the diminishing returns of some stats all together which makes comparing the actual usefulness of items with dissimilar stats impossible unless the system were adjusted to calculate based off of the % stats and then it compared your ∆CV based on if you were to equip this item in place of the current equipped item (while also ignoring the change of gems because that too would skeu things and make comparing the items wrong too).
      • Second, comparing two items thru the compare loot tool just looks at the stats of the items according to the rules above (1CV/HP and 2CV/AP) It doesn’t take into account multipliers like % armor (ex: from the Black Warlord or Defender's set) or %HP (ex: Ring of Life, the “Lost” set, or the Knowledge points) until you equip them.
        • A here is a perfect example: I have an Improved set of boots with 416 total Armor (833 CV) and a Magic set of boots with 597 HP on top of the base armor of 103 (804 CV). According to CV, the Improved Boots are better. Despite obviously not being better to anyone with even a week’s worth of experience playing the game (see the bullet point above again), it also fails to account for total HP after putting the boots on the character. In this case, if you have 75% HP already selected, that 596 HP comes out to 1043 HP. Add that to the boots base armor CV (for a total of 1249 CV) and the blue boots that are “worse” actually result in a higher CV over all.
      • Here is an excellent thread from earlier this year that delves into the complexity of balancing HP/armor/resistance and we see that it is vastly more complicated than the CV system makes it seem. In particular, I’ll start you at the post by nasabot.
      • Third reason why direct comparison fails at the moment: The treatment of % stat lines on the armor items and amulets (% HP, %CHV, %Armor, %Damage, %Attack speed, and %Run speed). It seems that there are two different variants of this issue.
        • First is something like attack speed. It is not factored into the CV displayed on the item at all. It is not factored in until the item equipped. From there, it is dependant on the base speed of the weapon in question and can range from 75 CV/%AS for a 1H sword with 1.18 attacks/second down to 50 CV/%AS for a 2H weapon with 0.83 attacks/second. This flat rate CV value has it’s own issues which can be inferred from the next topic (DPS), but for now, suffice it to say that it makes direct comparison of gloves (or other unique items) with attack speed impossible, even if you could adequately come up with a metric that could compare the value of offense and defence (again, apples and oranges)
        • The next category is % stats that do contribute to the item’s CV, but do not accurately reflect the actual change in CV that occurs when the item is equipped. Take for example, an Improved torso with 5.10% increased damage. It claims that it has a 1025 CV. However, when equipped to a character with no other gear, the gain is only 526 CV (off by 499). I then equipped a very nice 2h sword and the CV gained from equipping that item was 1273 (off by 248). I’m assuming that this is related to the % damage line assuming a “standard” average damage for the purpose of calculating the CV that the item should have displayed, but this once again highlights why the direct comparison tool should operate on an assumed swap and display the ∆CV, and not assign every stat a flat value.
    • By my calculation, crit is worth about 0.6 CV/CHV. Damage is worth 12 CV/Dam. Crit damage also is worth about 12CV/CD. Considering the max lines after R158 on jewelry items (not on weapons or off hand items) you can get 149 CV from a max crit line of 249, 420 CV from a max damage line of 35, and 204 CV for a max line of 17% crit damage. Right here, it is easy to see that damage has been vastly overestimated in its CV value, or the other two have been vastly undervalued.
    Even still, this fails to capture the complexity of the DPS problem because how useful a specific magnitude of a stat would be is dependant on the equation for DPS.​
    • DPS= An averaged DPS and our point of comparison.
      D1=Low end damage
      D2=High end damage
      S= Attack Speed
      R= Critical hit rate (in decimal format, not percentage. ex, 48% would be .48)
      C= Critical damage (Also in decimal format. ex, 237% would be 2.37)
      [​IMG]
    This is a true metric of the offensive capability of a character. Sure, you could modify it to use a floor rounding function to account for the nearest speed break, but this truly tells you the capability to dish out damage more than just by looking at individual stats on individual items. From this equation, you can see that the more damage you get, the more benefit you gain from stat line equivalent speed, crit rate, and crit damage line and the less benefit you get from the equivalent sort of damage line; which is again completely counter to what the CV system would indicate.​
    • There are several stats that are very useful, but do not have CV at all. Rage/Mana/Concentration/Steam all make it easier to attack and contributes to offense. It also contributes to defence by allowing use of crowd control skills, evasiveness skills, or HP regen skills. On the other hand, there is run speed which assists in defence by increasing evasiveness and the speed that one can escape from danger. It also helps offence by making it easier to chase down an opponent.
    • I have not examined Block Rate/Amount impact on CV due to not having enough shields to test this with, however, I’d look this one over with a more critical eye given the nature of the issues that exist in the other parts of the system.
    • Selected skills are supposed to apply CVs. At first, I thought that the developers had tried to guess which skills are the “best” and weight the points that way (which I would have had problems with, just for different reasons). Instead, upon testing, I see that they give a flat amount of points. How is this supposed to be a metric for differentiating between character quality? It only checks if you have used all of your available points. If something does not actually make a difference, then including it is a waste of time. It is analogous to a bias error in scientific measurement (which in an experiment you always seek to eliminate bias). At LVL 50, if all 50 points are used, you will always get 125 points from them (2.5 CV points per XP point spent). This is independent of gear, skills being in use on the skill bar etc. As said, this is a bias error and because of this, those 125 points have no practical meaning in comparing the relative strength of characters.
    • There are many other illogical and inexplicable artifacts that I have been unable to deduce the cause for yet, and I’m sure it has something to do with the equation that calculates this CV number. For example, with no skills selected, no knowledge points applied, no PvP talents active, on a character with no gear, I added various single pieces of gear with no gems or even gem slots. An improved pair of boots with an armor line on them and a total CV of 833 actually ads 474 to the CV. This is a difference of 41. Then a pair of boots with a resistance line and a total CV of 821, but it added 867. This is a difference of 46. Then I tried adding a pair of boots with an HP line and a total of 443 CV and it added 454 for a difference of 11. Using a pair of Magic boots with two lines of HP and a total score of 804 adds 814 for a difference of 10. Even allowing for rounding errors, this can’t be easily explained. The best I could come up with is that this hidden CV is related to how well rounded a piece of gear is but again Armor and Resist have been weighted in such a way that makes it look better than it is (as discussed several times above)
    • What can be done to fix this?
      • There are two primary attributes that matter in this game, Offence and Defence. For this reason, I and many others feel that there should be two separate CVs; one representing the offensive capability of the character and one representing the defencive capabilities. The two scores can then be added together for an aggregated score. This allows defensive builds to be compared to defensive builds and offensive builds to be compared of offensive builds. You could also consider things like extra Rage/Mana/Concentration/Steam or run speed or deceased cool down (the new gems) or skills selected or knowledge points spent as a miscellaneous category and assign them their own value that is displayed separately and added into the aggregate score.
      • There is only one stat that matters in Offence and one stat that matters in Defence.
        • In offence, we are concerned with DPS. This can follow the equation I referenced above when discussing CV of offensive stat lines. In this model, you only need to assign a CV value to damage and then all other offensive stats simply act as multipliers to the base damage CV. This makes the value of these others stats independent of opinion and based on hard facts. For ease of calculation, you could simply use attack speed as a number, but for a more indepth look you could look at the closest speed break point that they have achieved.
        • For defence, we are concerned with “Effective HP.” A character has their base HP and then every other defence oriented stat seeks to protect that HP pool and could be considered a “multiplier” to that HP. For example, if you have a shield with 50% chance to block and it blocks 50% of the damage taken, you effectively take 25% less damage over time when compared to a player without a shield. Alternatively, this could be looked at as 25% more “Effective HP.” The same is true of armor/resist and an excellent discussion of it’s effect and how to calculate it can be found here. In essence, it is E_HP=HP/(1-%Reduction). This could be calculated for armor and then resistance and then averaged together to get a value to be applied for the purpose of CV.
      • Comparing items: As described above, comparing items is currently useless due to the fact that it does not consider what will happen after the item is put on the character. I mentioned at that time that it should look at the ∆CV that you will receive when you put the item on (considering % enchantments, set bonuses made or broken, and knowledge points in particular) and give you an option to consider the gems that are already on the item as in all likelihood these gems will go onto the new item
     
    Last edited: Nov 25, 2015
  2. BigPapa

    BigPapa Forum Overlooker

    Well stated, Baragain. Facts with support. I still like the idea of this tool, but it may be too complex for the team to tackle andmake it completely useful and applicable. At this point I would be happy with it being not misleading. This is why I have the same answer to many questions I get during gameplay these days. What's your damage, what's your speed, what's you CV, etc. all get the reply "alot" with a :) tacked on. Damage comparisons between classes are impossible post R155, why would something like CV that needs to be much more complex be a better measure until it shows what we need it too.
     
  3. Poor_man_song

    Poor_man_song Forum Greenhorn

    This is all of it


    Bonus / score

    HP / 1
    Armor / 2
    Resistance / 0.4
    Block rating / 0.5
    Block damage reduction / 3
    Min damage / 6
    Max damage / 6
    ATK Speed / 1.2
    Critical rating / 0.6
    Critical damage / 120
    Item Armor / 2
    Item block damage reduction / 120
    Item block rating / 0.5
    Item critical rating / 0.6
    Item HP / 1
    Item Min damage / 6
    Item max damage / 6
    Item resistance / 2
    Item ATK speed / 1.2
    Movement speed / 0.5 (arena only)

    well except a little bonus per character class & level which is added last in the count (and skill point), I have it all in my blog, forum rules forbids me from showing it, but even rocks knows

    besides I'm kind enough to share with youngsters, even though some thinks their something's bigger than other's :D
    yet experience forbids me from giving all at once, 'cause the copyright stealers are always up close :D with their "depending on the date, that may have been copied from me, for sure" (quoting Shakespeare)

    Now go forth and farm, farmers of Duria :D
     
    Last edited: Nov 21, 2015
  4. _Baragain_

    _Baragain_ Living Forum Legend

    Interesting list you have there, but right off the bat, you may need to double check on some of them. You say "Item ATK speed / 1.2" but this is a little bit ambiguous. Is this per 1.00% speed? Is it per 0.10% speed? Is it per 0.01% speed. Additionally, it ignores that the amount a % speed line or item adds to your CV is base on your base speed and will give you more CV if you have a fast 1H weapon vs if you are using a slow 2H weapon. Also, you are missing a couple of stat lines like the relative enchantments that can be found on amulets (%HP), Helmets (% CHV), Shoulders (% Armor), Torsos (% Damage), and you only briefly touched on gloves (% Attack Speed) and Boots (% Run Speed).
    Agreed, this tool has a lot of potential. My "What can be done to fix this" will draw from some of the ideas you already put forward, but with some of the finer points flushed out.
    This was my biggest goal when I started this project. As an experienced player, my gut told me that something was wrong. At first I thought "who is BigPoint to tell me what is best for my character?" After all, I know my goals and if DPS matters to me or if I'm going for tanky. But now, after my research, my question has become "Who is BigPoint to tell anyone how to build their characters? The don't even understand what makes a character good, yet they presume to tell everyone that the new item that they got is junk when it is actually a godlike item." My favorite example is Attack Speed gloves... Because it does not get reflected on an item's base CV and is only factored in once it is put on the character, a Legendary pair of gloves with 35% attack speed would have a displayed CV of 206 (on a DK due to the base armor of 103). Compare that to a purple set of gloves with 1000 total armor that would have a displayed CV of 2000. I'd hope that a noob would see this and use their brain, but if they trust the CV system, they will never even put the gloves on and see that they would give a huge CV boost. With a DK's 1h Sword, that 35% would give you 2625 CV on top of the 206 base CV. Plus, there are many players who would trade plenty of defensive stats for a set of speed gloves that amazing, total character CV be damned.
     
  5. Poor_man_song

    Poor_man_song Forum Greenhorn

    what's really missing, as pointed out, are some pieces of information on how to use the given values and their units

    that's cause at beginning I thought about giving it away in one shot, then changed my mind :D

    for today, this is one more giveaway :cool:
    [​IMG]

    some more pieces shall follow as long as the discussion deserves it
     
  6. _Baragain_

    _Baragain_ Living Forum Legend

    I'm sorry, but your model needs some work.

    Here is a white hammer and a white sword.
    [​IMG]
    First take the average damage of both of these. The average damage for the hammer is 303. The average damage for the sword is 220. Multiply both of these by 12 (which is the same as damage on armor or jewelry) and you get 3636 and 2640 which are very close to the displayed value. The fact that the one with the lower speed has a +1 difference and the one with the higher speed has a -3 difference shows that speed is unlikely to be relevant in the calculation of the displayed CV. Now take my 2H for example.
    [​IMG]
    You could calculate this a couple of way, but I'll go through the easiest way.
    First, the base damage is 625-832. Average damage is 728.5. Multiplied by 12 you get 8742. Then add in the 570*(0.6)= 342 and you get 9084 which is really close to the displayed 9089 (-5) and could be rounding errors or the artifacts that I discussed in my last bullet point.

    The biggest reason that I'm confident in my model as opposed to yours is that because, so far, my model is self consistent. That is to say that damage or crit has the same value on jewelry and armor as it does in a weapon and that speed is not accounted for until you put the item on.

    Over all, good start, but I'd do a bit more research before you say you have the answer as to how CV is calculated. I spent about 8+ hours playing with this to get the detailed understanding I have and I'd still only say that I have it 95% figured out.
     
    Last edited: Nov 21, 2015
    thebearreturns likes this.
  7. Rhysingstar

    Rhysingstar Forum Ambassador

    Your math is certainly compelling. I went searching to see if I could find variables in the calculations made by this CV value and the way it adds up isn't obvious as there seems to be some random factors in their math that effect the outcome.

    I grabbed my 2H axe and found another for comparisons using your formula which does in fact get you close every single time.

    My equipped axe falls -2 (10460.4) and the exo adds + 2 (6988.2).

    I'm left to think that somehow the % increase or + dmg or maybe even the given dps are fudging the math in their calculations as I am not able to see how speed changes the + or - using your formula. There are always variations given the same speed.

    I for one am truly looking forward to seeing how you work this mess out, hopefully DSO will listen to you on this screwed up CV system.

    [​IMG]

    Compared to a hammer with identical speed, the hammer gets a +1 using your formula.

    [​IMG]
     
    Last edited: Nov 22, 2015
  8. Poor_man_song

    Poor_man_song Forum Greenhorn

    @Baragain
    I know that's your way of saying "thanks" because you can't spell it directly (yet it's easy "t-h-a-n-k-s")
    and your attitude, is your only flaw*
    cause it's not some sacred stones that makes you flawless, it's when you open heart & mind to people that you get flawless

    but I'm nice so here the internals
    [​IMG]

    footnotes:
    min+max times 12 is just using property of add- operation, it ain't anything special that you discovered or something :D
    nor a self consistent something whatever
    the game formula is 6 times min + 6 times max
    and there are simple reasons for it, but this is not a coders community
    which in turn by using the prop. of + can be expressed as 12 times x
    and many other ways to express it, depending on the properties you want to use

    hours: the time to collect all the data on the CV system is 6 minutes :oops: not hours
    cause there are three strong sources to get it quickly, including the game developers provided you have the right attitude

    the picture from before may have errors (drawn in seconds from a test char i never used before), you actually didn't prove it, you just stated that the negligible difference in points must come from anything but not the speed, which can be true, but that's not the way to prove neither a nice way to communicate with people,

    research: at post1 you had reckoned on very few parameters so your research were somewhat like 30% not 95% with my data you can now claim you can get to 95% and even to 100%
    you love diminishing returns so i can call you the accountant here :rolleyes:
    so if you wanna calculate the extact percentages i won't stop you
    I always hated maths :mad: even after getting an engineer's degree, that cannot be helped i was born this way

    but i already have the 100% data, and i was just nice enough to share with you
    and to the others, because that is the kind of person I am

    in the next chapter i'll post the data on the points given for skills and for the character class, data which is actually already published elsewhere :rolleyes: and i'm bringing to your community like an early xmas present

    the CV surely s11cks but that's a minor issue :D
    given the overall game situation

    ------------
    * same goes for the hp bug but which is actually graphical (well a little more complex, because some details of RakNet are involved and makes it salty) but you fought to hard against some users that wanted it fixed, no matter if it's gfx or not, basically you're always fighting on the wrong side of the wall :D
     
    Last edited: Nov 22, 2015
  9. Mindship

    Mindship Forum Inhabitant

    This is great stuff, guys (Baragain, you're like the Sheldon Cooper of this game). Thank you all for presenting something to read and ponder during my metacarpal recovery sessions.
     
  10. trakilaki

    trakilaki Living Forum Legend

    I can see him on my side of the wall ... which automatically means he is on the right side.
    Same here ... I hate math too that is why I stay away from the topics where math is included ... not because I can't do math but because I hate math ... even though just like you I have engineer's degree in CS.
    But that is not stopping me to use common sense and express how much I hate this CV system (or tool ... whatever) ... because not a single piece of junk code can tell me whether I should use some equipment or not. Because at the end we are coming to the ... who is stronger ... "Batman Vs. Spiderman" eternal debate.
    In fact I want an option in the settings to remove the CV from my screen.
     
  11. Dylannn

    Dylannn Board Analyst

    I think the CV system is useless and a very bad-calculating system, some values are way too overrated.

    Personally I think a player himself should decide and should know what piece of equipment he prefers, some prefer dmg over speed, some prefer speed, etc. It's a personal issue.

    Now really I hope they will never make PvP balancing based on this system, even tho we can trick that, God help us if they do so :D
     
  12. Aemon

    Aemon Advanced

    Would you mind giving me a hint how to find your blog? I'd love to read it :)
     
  13. _Baragain_

    _Baragain_ Living Forum Legend

    Added my recommendations on how to fix the system.
     
Thread Status:
Not open for further replies.