1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Statistical research on classes

Discussion in 'Guides' started by IzalithSky, Dec 16, 2015.

  1. IzalithSky

    IzalithSky User

    Likes Received:
    Hello there.
    The purpose of this research is to find an answer to question "What's the best class to play ?" from statistical point of view. In other word I am trying to calculate deficit/surplus of classes. This information is aimed for solo/duo players, who has no CP or clan with regular events. Research is focused on first 1-3 month when most of the characters are below level 40.
    If you don't want to read through all this stuff (and that's fine) you may jump straight to conclusions in the end.

    Input data:
    1) Class polls from this forum (Choose your class, [POLL] Choose Your Class, [POLL] What class will you play?);
    2) Poll on dualboxes from this forum (Will you dualbox and pay for two accounts?);
    3) Threads from recruitment hall on this forum and my personal experience of common group builds from rus. l2c server.
    Output: Deficit/surplus of classes represented by numbers.
    Instruments: Here is google sheet with calculations (it is pretty messy ! You have been warned !).

    Most of the data was taken from first poll (as it has most votes total), additional data (spoiler/crafter ratio among the dwarfs, pp/bishop ratio etc) was taken from polls 2 and 3 and dualbox poll. First poll has better accuracy (more votes) but suboptimal classes distribution among answers possible. The data from there was modified to take dualbox buffers into account (you see, 54% players will dualbox and their PP/SEs will come into play eventually).
    Having that data I calculated distribution of roles (for levels below 40 !). Roles are: p.dd, mag.dd, spoilers, tanks, supports (both buffers and healers are here because under lever 40 these are same guys (yeah I know shaman isn't good healer...)). To be precise:
    - m.dd: wizard, elven wizard, dark wizard;
    - p.dd: warrior, rogue, elven scout, assassin, raider, monk, artisan (after level 28);
    - spoilers: scavenger;
    - tanks: knight, elven knight, palus knight;
    - supports: cleric, elven oracle, shilien oracle, shaman;
    Aside from that I suggested most common types of groups formed from random players in the field (according to my experience from rus. l2c and recruitment hall). Group builds are:
    1) Mag. group: 3 mages (AoE), tank, 2 supports (heal + buff), spoil;
    2) P, group: 4 p.damage dealers, 2 supports, spoil.
    Also I calculated ratio of these groups (~1:2).
    With all that done I compared amounts of characters of each role with groups' needs. Formula for deficit/surplus is:
    ds = 1 - (gr * rn / ra ),​
    where ds - deficit (values below 0) or surplus (values above 0), gr - group ratio (mag. groups to p. groups), rn - roles needed, ra - roles available.

    Roles distribution (below level 40):
    1) Support 41.4%;
    2) P.dd 27.2%;
    3) Tanks 11.6%;
    4) M.dd 14.7%;
    5) Scavengers 5.1%.
    Here are result sorted from most to less wanted roles (below level 40):
    1) Spoil -0.78;
    2) P.dd -0.47;
    3) M.dd 0.12;
    4) Tank 0.63;
    5) Support 1.31.
    - Here values below 0 are deficit (you are wanted in groups), above 0 are surplus (you are not wanted);
    - Such surplus of supports is caused by dualboxes;
    - Artisans are included in P.dd (after level 28, before you got that teapot dude summon nobody loves you D: ).

    Here is results of calculations for levels above 40.
    Warning: These are much less accurate since group builds (for groups of random players) and ratio of those groups are hypothetical.
    Roles distribution:
    1) Buffers 36.6%;
    2) P.dd 22.7%;
    3) M.dd 11.4%;
    4) Archer 7.7%;
    5) Tanks 7.1%;
    6) Bishop/EE 5.8%;
    7) Bounty Hunters 5.1%;
    8) BD 3.0%;
    9) SwS 1.6%.
    Deficit/surplus of roles:
    1) SwS -4.12;
    2) BD -0.82;
    3) P.dd -0.06;
    4) Archer 0.10;
    5) M.dd 0.20;
    6) Tank 0.57;
    7) Bounty Hunter 0.79;
    8) Bishop/EE 1.03;
    9) Buffers 2.68.
    - Here values below 0 are deficit (you are wanted in groups), above 0 are surplus (you are not wanted);
    - Such surplus of supports is caused by dualboxes;
    - Warsmiths and summoners are included in P.dd;
    - Buffers are PP, WC, OL, SE;

    1) Class poll 1: https://4gameforum.com/threads/25730/
    2) Class poll 2: https://4gameforum.com/threads/26254/
    3) Class poll 3: https://4gameforum.com/threads/532398/
    4) Dualbox poll: https://4gameforum.com/threads/25779/
    5) Google Sheet with calculations: https://docs.google.com/spreadsheets/d/1tUy-Fz7e4G-iFJWMk1NL1lmSf73jb-qnL7YW-HKUu1M/edit?usp=sharing

    If you want to be needed in groups as solo player (no CP/clan) in early game (below level 40) you may want to be Scavenger. Buffers (cleric, elven oracle, shilien oracle, shaman) are not recommended for solo players. All other roles are allright.
    As for late game (above level 40) SwSs and BDs will be wanted BADLY, buffers (PP, SE, WC, OL) are not wanted, healers are not too popular as well. All other classes are allright.

    Thanks for attention.​
    Last edited: Dec 16, 2015
  2. AgNike

    AgNike User

    Likes Received:
    Good work! Very interesting :Thumbsup:
  3. ThreeNill

    ThreeNill User

    Likes Received:
    Interesting info (and great hard work!) but considering the expected population and the low amount of votes the calculations are not very accurate (as you have stated) just because the sample pool is small. The calculations would be very true if we consider the player base that had voted as the general tendencies for all players (which can only happen in a perfect world).

    I think a judgement based on past experiences would be more suitable as to knowing what will be needed and what will be abundant.

    But nevertheless it provides a good insight of what to expect in general. Maybe a good idea would be to isolate the boxed buffers(&healers) from the calculation because I highly doubt these will be a surplus.
    AgNike likes this.
  4. IzalithSky

    IzalithSky User

    Likes Received:
    Thanks for your feedback, guys.
    It would be nice to have larger amount of votes indeed.
    I'm kind of surprised by surplus of healers too. Not so much with buffers. Boxed supports are in fact separated (somehow. it could be done in the other ways, yes). To be precise:
    healers are separated from all other classes and calculated as
    h = (1-pp)ppbs + ee,
    where h - amount of healers, ppbs - total amount of PPs and Bishops, ee - total amount of EEs, (1 - pp) - percentage of Bishops among PPs/BSs (about 8%. that is where isolation of healers from boxed folk happens).
    Boxed buffers, on the other hand, are not isolated because (in the field where parties of random people are formed) dualboxers usually take their buffers in party with them (party is leveling both of your characters at the rate of one. isn't it nice).
    I agree my decision on how to calculate isolation may be wrong.
    From my experience of being elven oracle on rus. l2c: there was indeed surplus of healers, but it didn't felt like it was 1.03 (103%). I'd say it was about 50% (no calculations were involved back there).
    Thanks for feedback again :)
    ThreeNill likes this.
  5. DarkSerp

    DarkSerp User

    Likes Received:
    Gran Kain statistics

    (Shillien Elder)601
    (Bounty Hunter)335
    (Elven Elder)334
    (Sword Singer)256
    (Silver Ranger)133
    (Treasure Hunter)107
    (Phantom Ranger)104
    (Dark Avenger)74
    (Elemental Summoner)64
    (Shillien Knight)55
    (Phantom Summoner)41
    (Abyss Walker)37
    (Temple Knight)30
    (Plains Walker)24
    PsychoFreaq, Leeway and IzalithSky like this.
  6. Leeway

    Leeway User

    Likes Received:
    Geez...all these statistics are giving me a hypothetical boner.
    fodemhouse and AgNike like this.
  7. IzalithSky

    IzalithSky User

    Likes Received:
    Whoa !
    What an interesting piece of data. Thanks alot, DarkSerp.
    It's tempting to make some calculation on in right now.
    Though just by the look of it I can say of surplus of buffers and healers (which is still surprising for me after all I've seen here). We can also say mages were whining because of absence of tanks :)
  8. PartyPooper

    PartyPooper User

    Likes Received:
    Its not suprising at all - dualboxes. SE's are the most universal, and any class that can't benefit from se can benefit from pp, and there you have it.
  9. IzalithSky

    IzalithSky User

    Likes Received:
    Sure. What I wanted to say I was surprised by number of EE and Bishops. I wasn't clear on that, sorry.
  10. Lanochka94

    Lanochka94 User

    Likes Received:
    thank you very much for the information!!
  11. IzalithSky

    IzalithSky User

    Likes Received:
    So that's what we have according to the data from rus. l2c gk server.
    Roles distribution:
    1) SE/PP/WC 21.1%;
    2) P.dd 19.4%;
    3) M.dd 19.9%;
    4) Archer 11.2%;
    5) Bishop/EE 10.1%;
    6) BH 5.9%;
    7) BD 4.8%;
    8) SwS 4.5%;
    9) Tank 3.1%.
    OLs are included in P.dd. That's kind of wierd, yea, but they are not buffers for a random party nor they are mages.
    Hypothetical deficit/surplus:
    1) Archer -1.65;
    2) P.dd 0.12;
    3) M.dd 0.34;
    4) SwS 0.45;
    5) BD 0.60;
    6) Tank 0.61;
    7) Spoil 1.05;
    8) Healer 1.87;
    9) Buffer 2.46.
    From that we can say archers parties are not too popular among random parties.
    Surplus on each role (besided archer) just means there are enough people to form mag. and melee parties. Guys with larger value of surplus are struggling to join a group more.
    So in conclusion I'd say I don't recommend to pick buffer or healer unless you got CP.
  12. noisy43

    noisy43 User

    Likes Received:
    Well, active buffer is always better than boxed.
    SweetDreamZ and Lilyputy like this.
  13. Dubiouss

    Dubiouss User

    Likes Received:
    Better with to many healers/buffers than to few
    a party can do fine with more than 1-2 supports, but not always without either healer or buffer

    Back when L2 came out I played dorf, but cause of the lack of healers I switched to Bishop and got parties before I could ask for one, up until healers got obsolete cause of VR and buffpets
    SweetDreamZ and IzalithSky like this.
  14. Yonnan

    Yonnan User

    Likes Received:
    Althoug statistics may provide you with some evidence your conclusions are wrong. First of all you need a buffer (or buffers) and a healer in every party. Furthermore you have to consider specific buffs exclusive for each buffer (Berserk and blesses - PP, EMP - SE, Seals - OL). A "proper" party should thus consist of 2 buffers (e.g. mages - OL+SE (EMP), meele WC+SE (Stigma, Mental)).
    Overrepresentation of SE is not at all a proof of this class being not wanted. You have 1134 mages of various classes (SPS SS NEC SORC) in Gran Kain statistics. There is only one buffer in this game that can give them EMP. Guess now, why SE is most popular class on mage heavy server?
  15. IzalithSky

    IzalithSky User

    Likes Received:
    Thanks for your replies. One who reads this guide should take them in consideration.
    My conclusions are just my take on these numbers. Readers are welcome to make their own.

    Well yes, now I see my statement about "not wanted" is way too drastic. Sorry for that.
    What I wanted to say of course was not "buffers will not be wanted at all". What I actually meant to say: "Being a buffer as main you will have troubles joining a group because the buffer's place(s) (more likely) will be taken already (more likely by someone's dualboxed buffer)".
    Yes, party would want to has two buffers (mystic buffers, with bd/sws it's 4), though 2 is just about 22% of the group. Thus, for example, server as whole needs less SE than there are mages (one SE can buff more than one guy), but what we can see: there are about the same amount of SEs and mages. So for optimal (for SEs) distribution each mage should go in the field with one SE.
    As for PP/WCs it's not so bad. Their numbers are not as overwhelming.
    Healers can count on being more popular than buffers. One can not effectively use healer for whole group while dualboxing.
    In the end I want to say I still do not recommend to pick buffer (mystic) as main, but that's just my opinion.

    Thank you, guys, for feedback again :)
    /Jedd and ThreeNill like this.
  16. wewdome

    wewdome User

    Likes Received:
    I get that you are trying to do a statistical breakdown but, as many ppl before me have pointed out, you are not taking into account the sheer amount of buff boxes ppl will use. I believe well over 60% of nukers and spoilers will have boxed se, many parties will have either bd or sws boxed, sometimes both, and probably at least 30% of archers will have pp. This is the situation until about 52lvl. After that i suspect full parties with active support will be more common and that means that most of the buff boxes wont be used anymore.

    Bottom line is that for the random solo guy it will still be hard to find active support, unless he joins cp or well organized clan. Another safe bet is that sws and bd will always be wanted in any party and there will never be enough of them
  17. IzalithSky

    IzalithSky User

    Likes Received:
    Hi, wewdome.
    I am sorry, but it's either I don't get what do you mean or we are saying the same thing.
    I actually do take boxed buffers into account (to be precise their amount will be 0.54*(spoilers+mdd+pdd) according to my guess).
    Also, please note I am talking about mystic supports here (not BDs and SwSs). As for BDs and SwSs that's exactly what I said: they will be wanted badly.
  18. SweetDreamZ

    SweetDreamZ User

    Likes Received:
    I love live support class players and i respect them for being one. Why pay extra for 2nd accound ? I Play mmo to party hard with yall and not just with myself <3