Latest Diablo 3 News
DiabloWiki Updates
Support the site! Become a Diablo: IncGamers PAL - Remove ads and more!

[Upgrade MF items for defense] vs [less points into strength for more vitality]

Discussion in 'Theorycrafting and Statistics' started by Drunken_Tiger, Jun 28, 2005.

  1. Drunken_Tiger

    Drunken_Tiger IncGamers Member

    Joined:
    Jun 28, 2003
    Messages:
    648
    Likes Received:
    0
    Trophy Points:
    165
    [Upgrade MF items for defense] vs [less points into strength for more vitality]

    Current equipment for my magic find sorceress:
    Shako
    Skulders (upgrade?)
    Goldwrap (upgrade? once? twice?)
    Crystal Sword
    Monarch
    Magic Amulet
    2x Magic Rings
    Chanceguards (upgrade? once? twice?)
    War Travelers (upgrade?)

    Upgrading Goldwrap and Chanceguards aren't too taxing on the strength, but upgrading them twice? Upgrading Skulders? Upgrading War Travelers?

    How far should I take this upgrading fiasco? Is the huge strength spent for higher defense worth it? or is it better to have low strength, lower defense for more life?

    Oh also, would you guys recommend max block for a magic find sorceress?

    Thanks!
     
  2. helvete

    helvete IncGamers Member

    Joined:
    Jul 17, 2004
    Messages:
    6,807
    Likes Received:
    909
    Trophy Points:
    534
    I have a 800+ MF sorc, she doesn't have max block, although she would have had max if she had been using stormshield (which she used before she started mfing pindle only).

    Go with more life; you wouldn't get that much more defense anyway. And defense does nothing to avoid FE/CE/LE bosses and such.
     
  3. alexzed

    alexzed IncGamers Member

    Joined:
    Nov 10, 2003
    Messages:
    1,334
    Likes Received:
    0
    Trophy Points:
    255
    I think the other way of looking at this is to imagine what difference the 'upping' will make.

    Say you have a 1000def armour instead of 500. That makes your total def around 6000 when you have a mid-level shiver armour. Now look at the real defensive characters - 12-16000 def, and they still get hit. I say put it in life as well.

    But I like upgrading things too! Have a nice upgraded eth Que hegans and an Upped Eth Skin of the Flayed One (the regen mod is nice on that).

    Go with life...
     
  4. helvete

    helvete IncGamers Member

    Joined:
    Jul 17, 2004
    Messages:
    6,807
    Likes Received:
    909
    Trophy Points:
    534
    You are aware of the eth armor upgrade bug? You don't get the 50% base def bonus, and defense gets rerolled as well, so a couple of items actually get LESS defense when upgraded. (Hoz, Corpsemourne, and a good deal of others).
     
  5. Thrugg

    Thrugg IncGamers Member

    Joined:
    Jun 23, 2003
    Messages:
    1,527
    Likes Received:
    0
    Trophy Points:
    105
    I generally find that when it comes to defence, you need an awful lot, or else it doesn't matter. "An awful lot" is a bit vague, maybe 15K? But the difference between 5K and 6K is hardly noticeable.

    So, since your MFer is probably not going to have an awful lot of def anyway, I wouldn't go for small increases. I upgrade Goldwrap to get an extra row of potions. That's about it.

    I also like to be able to give items to new chars coming up too so the level req is something to think about too. With some items you'll find plenty of others but a good Skullders can be hard to find. Upgrading it would mean that much longer a new MF char couldn't use it yet.
     
  6. YellowDyeNo5

    YellowDyeNo5 IncGamers Member

    Joined:
    Jul 22, 2003
    Messages:
    185
    Likes Received:
    0
    Trophy Points:
    41
    You're already wearing a monarch, and thus have 156 str (assuming its not eth). So upgrading some of these items twice will mean nothing with regards to str.
     
  7. sorcyone

    sorcyone Banned

    Joined:
    Mar 1, 2004
    Messages:
    501
    Likes Received:
    0
    Trophy Points:
    0
    I often find it useful to compare what charms would be required to get the same attributes. For a sorc 20 life = 10 strength = 50 def. But def is less important for a sorc so I generally modify this to 10 strength = 100 def. That is if you have to pump 10 strength to get 100 more def then it is a close call.
    If melee sorc then stick to 10 strength = 50 def.

    Max block ? You only said Monarch. If you mean spirit runeword then no, if stormshield then yes.
     
  8. slappy sam

    slappy sam IncGamers Member

    Joined:
    Nov 18, 2004
    Messages:
    261
    Likes Received:
    0
    Trophy Points:
    57
    If your highest str requiring item is monarch, then up everything that will (when upped) stay below this highest requirement, since you have to get that anyways. Like my sorc's highest is a monarch so I upped my magefists to exceptional, and my goldwrap. Since I'm on NL I can't go any farther.
     
  9. sangfagel

    sangfagel IncGamers Member

    Joined:
    Jun 27, 2003
    Messages:
    1,307
    Likes Received:
    0
    Trophy Points:
    120
    Let´s take it as a pure math:
    Chance to Hit (be hit):

    100 * AR / (AR + DR) * 2 * alvl / (alvl + dlvl)

    In case of high clvls the second part of this formula can be ignored since
    2 * alvl / (alvl + dlvl) ~ 1

    We have still this very simple

    100 * AR / (AR + DR)

    Now if we consider a chance to be hit by the same monster but changing our armour during the fight, AR= constant and DR is the ony variable.

    Chance to be hit = 100*con/(con + DR)

    What we got is... diminishing returns.

    Now I don´t really know what attack rating have monsters in Hell difficulty. I guess that it is something between 3k to 6k for regulars. If we take as example a value of 5000, we have:

    100% = 95% chance to be hit with 0 armour
    50% chance to be hit with 5k armour
    33% chance to be hit with 10k armour
    25% chance to be hit with 15k armour
    12.5% chance to be hit with 35k armour... etc

    So, a char needs 5k def to reduce nuber of hits by half and 10k more for next half and 20k more for next halving

    Am I wrong again? :scratch:
     
  10. Drunken_Tiger

    Drunken_Tiger IncGamers Member

    Joined:
    Jun 28, 2003
    Messages:
    648
    Likes Received:
    0
    Trophy Points:
    165
    Oh as clarification, the monarch is a white open socketed monarch (4 ists) Hmm... still deciding... in the mean time, leveling my sorc!

    Thanks for the replies so far!
     
  11. thegiantturtle

    thegiantturtle IncGamers Member

    Joined:
    Nov 18, 2004
    Messages:
    4,642
    Likes Received:
    0
    Trophy Points:
    120
    EDIT: i'm an idiot. nevermind.
     
  12. Thrugg

    Thrugg IncGamers Member

    Joined:
    Jun 23, 2003
    Messages:
    1,527
    Likes Received:
    0
    Trophy Points:
    105
    sang, yeah, I'm quite aware of how the to-hit formula diminishing returns works. The reason behind the numbers I quote is that most monsters in Hell have ARs up around the 15K-25K mark. Generally the curve of to-hit vs DR is sigmoidal (actually reverse sigmoidal) - flat up high and flat down low but steeply dropping near a critical point. And the critical point is right at the AR of the monster hitting you. 5K would be way up the early high flat part for such monsters, raising it to 6K won't make much difference.
     
  13. sangfagel

    sangfagel IncGamers Member

    Joined:
    Jun 27, 2003
    Messages:
    1,307
    Likes Received:
    0
    Trophy Points:
    120
    Thrugg, I am surprised of these AR numbers - I was never expecting they are so high as 15 - 25k!

    Anyway - yesterday I mad a simple check. In A2 Hell my amazon has 74% chance to be hit by regular Crusher. Her def is ~1k, so Crusher´s AR should be ~3k. This check was made in Canyon of the Magi, area lvl79 for clvl 83, ladder, one player in game. In this case 2 * alvl / (alvl + dlvl) = 0.98 and can be ignored.

    The difference between 3k and 15 - 25k is rather big. What´s explanation?
    1.Laing character screen was laying again?
    2.Monsters AR depends on number of players?
    3.You was talking about act bosses or unique monsters with surely much higher AR?

    I am going to make some obseervations later in game and against Meph/Diablo/Baal. I think this question is important because if monsters´ AR is really so high as you said, there´s no reason to bother about def of light armoured chars. Nevertheless if a considerable number of monsters has AR about 3-6k every extra k can make a difference.
     
  14. sangfagel

    sangfagel IncGamers Member

    Joined:
    Jun 27, 2003
    Messages:
    1,307
    Likes Received:
    0
    Trophy Points:
    120
    I have made some more calculations using "chance to hit formula" and numbers from character screen. I used different characters and matched them to different monsters in A2 Hell. Calculated monsters AR was still about 3k.

    Something is wrong though.
    A char clvl 91 with armour 1756 has 66% chance to be hit by Duriel (mlvl **).
    The same char has 55% chance to be hit with armour 3073.

    When I calculated Duriel´s AR, I´ve got:
    In the first case 3621
    In the second case 3931
    So... Why the same Duriel would have two different AR???... Did he switched his weapon??? I don´t think so - he was already dead :p
    I used entire formula taking under consideration even this little discrepancy betweed lvls...
    This time I am not so completely sure that lying character screen is lying because given probabilities seems to work - my chars are hit with expected frequency - as far as I can see it. Anyway (lying) character screen does not use the formula from "Facts & Formulae Archive"
     
  15. Thrugg

    Thrugg IncGamers Member

    Joined:
    Jun 23, 2003
    Messages:
    1,527
    Likes Received:
    0
    Trophy Points:
    105
    The faults of the LCS are many and varied.

    For a start, it gets a monster's base mlvl and uses that in its calculations. Since monsters now use their area's level, this base level is wrong for most monsters (not Duriel though). The mlvl affects the to-hit equation twice, once because it is used directly, but also because monster AR is scaled by mlvl.

    The LCS also doesn't factor in any of the following in the displayed to-hit number:

    AR vs undead/demons
    -% enemy defence
    ITD

    There's probably more I don't remember. Not sure if it counts "-defence per hit" correctly. For monsters hitting you though, these aren't factors.

    Minor discrepancies working backwards from displayed numbers are not a surprise - remember the actual number is being rounded to the % you see. So in fact if it says 66% it actually means anything between 66.0% and 66.9999%. You could work out the range of possible AR values that satisfy that and you'll probably see an overlap that contains the actual value.
    In fact, according to the AS monster pages, Duriel is listed as 3874 in Hell with his main attack (he has two, but the LCS uses the first).

    However, I do have to apologise and I am glad you brought this up because it made me check something which I appeared to have remembered wrong.
    In every previous version of D2, listed monster AR was used multiplied by 4 due to either a bug or a compensating hack on Blizzard's part. They never bothered fixing it before because monsters were hitting at about the right rate and removing a big factor like that forces a rebalancing of all the numbers for every monster in the game, which is a big job. However, I had forgotten they fixed this in 1.10. So, I have been assuming you had to quadruple the AS numbers when apparently you shouldn't need to.

    That said, I am pretty sure I get hit more often than the AS AR values would suggest :p But, it does mean that the critical point for DR should now be much lower than it was in all earlier patches. Hmm.
     
  16. sangfagel

    sangfagel IncGamers Member

    Joined:
    Jun 27, 2003
    Messages:
    1,307
    Likes Received:
    0
    Trophy Points:
    120
    Thanks a lot for your commens! :)
    LCS is LCS and we have no choice but love it as it is...
    I just suspected something like this. In v .09 def didn´t matter for real chance to be hit (in pvm and with low/moderate def rate) and I never bothered about it. After v.10 I noticed that def started to work - it was just impression but anyway.

    I guess you are right - my char was a shield user. Even if the chance to block wasn´t big (~55%) it took these "unexpected" hits giving me an illusion of balance. A blocked hit doesn´t sound every time, does it?

    To take a next turn of our maths: I think that progress of attack effectivity is rather like hyperbola.

    In the equation:

    Y = chance to hit
    B max = 2 * alvl / (alvl + dlvl)
    Kd = DR
    X = AR

    Chance for effective defense would be just 1-Y or (1-Y)*100%. (we must reverse the curve)

    If we now take DR as a variable and fix AR as a constant value I still can´t see any reason why progression would be reverse sigmoidal (what´s reason for platau? :cheesy: ) It is reversed hyperbola – nothing else. If we consider chance to avoid hit as a function of DR, we will again get a hyperbola: beginning at 0 witth no defence and progressing toward 1 (100% chance to don´t be hit) but never reaching this value.

    As far as I understand this issue there´s no sources easy enough to use and giving correct info about monsters´AR. It doesn´t make me very sad since I am not going to calculate optimal defense for every of bilions fights in the game :D

    Generally, according to AS, Hell monsters have AR in range 3 - 6k but for some extra guys as Meph, Diablo, Minions of Destruction and Baal it would be rather 6.5 - 10k. All these is an underestimation though.
    My WILD, WILD guess is that average hell monster has effective AR of 5-8k and â€specials†10-20k.

    What´s about practical conclussions than?...

    As it is for me, most effect is gained by increasing def up to monsters´ AR value. For light armoured chars it should be anything to 10K; for heavy armoured – 10 to 25k. In almost any case block is more important than DR. For most chars tactical skills/abilities and mobility works better than DR… See… â€faith is stronger than any armourâ€â€¦
     
  17. Thrugg

    Thrugg IncGamers Member

    Joined:
    Jun 23, 2003
    Messages:
    1,527
    Likes Received:
    0
    Trophy Points:
    105
    While you are right of course, the function is a hyperbola, but it is a matter of perception. I'd suggest that DR is perceived by the player on more of a logarithmic scale than a linear one. You don't think of going from 50 to 100 as the same as going from 1000 to 1050. You think of it more like going from 1000 to 2000 (ie, doubling your defence).

    As a result, try looking at your hyperbola on a log x scale, or alternatively look at a graph of log(AR/(AR + DR)). There's your sigmoid.

    Well, maybe that's just me :)

    It is possible to calculate actual AR for any monster in the game (provided you know which area it is in) but I have not seen an accurate lisitng online anywhere.
     
  18. helvete

    helvete IncGamers Member

    Joined:
    Jul 17, 2004
    Messages:
    6,807
    Likes Received:
    909
    Trophy Points:
    534
    Well, I have a paladin with >20k def, and monsters in tower cellar level 5 (hell, but I didn't need to say, did I?) have something like 6% chance to hit according to the LCS. His level is 86.

    However, they NEVER hit him! Because he never loses life. They bang on his holy shield at times, but they never ever hit. I would know, because I left him in the moddle of the countess mob without killing anything, thought I hit Esc to pause, went to the kitchen to make tea, returned in panic when tea was done, and find him... alive!!
     
  19. sangfagel

    sangfagel IncGamers Member

    Joined:
    Jun 27, 2003
    Messages:
    1,307
    Likes Received:
    0
    Trophy Points:
    120
    Of course you´re right.
    As I wrote before I can´t see a really need of such a listning. Nevertheless it would sometimes help to have a new factor value - similar to this "4" in .09. - Just to calculate still aproximate but more correct values from numbers in AS or LCS.
     
  20. RTB

    RTB IncGamers Member

    Joined:
    Jun 21, 2003
    Messages:
    5,108
    Likes Received:
    2
    Trophy Points:
    120
    Instead of updating the rest of the items here on dii.net, I decided to make a monstats.xls and offer it to Elly when I'm done. This will take a lot of time, so don't except to see it anytime soon.
     

Share This Page