A perfect game of Kongai

Discussion in 'Kongai virtual card game' started by Nicol Bolas, Jun 10, 2008.

  1. Nicol Bolas

    Nicol Bolas New Member

    Definitions and rules, read carefuly!

    Version 1,8 is used !

    Deck A
    1st Ambrosia Thorn + necronomic tomes
    Higashi + Generals Insignia
    Helene + Valk charm

    Deck B
    1st Helene + valk charm
    Rumiko + scroll
    Tefari + death mask


    Rules:
    No SPAM, NO OFF TOPIC

    In your post: say which round you refer to, which move you refer to, and AS FORMAL AS POSSIBLE reasoning!
    (use logical operators and mathematical formalizations, its better then long text, don´t forget this thread will have many posts!

    I will assign your voice to that choice in current round, and then the move with most voices is selected
    A RNG will be used by Nicol Bolas, and Values posted, to determine the chosen moves!
    There is 48-72 hours for each round.

    You may choose for both sides!
    You must try hard not to know what opponent is doing, and what cards he has, don´t use that in your reasoning, I will not accept that!

    -----------------------------------------------------------------------
    VERY IMPORTANT
    Introduce numbers to your choices too. Then I will use them as weights (they must add up to 100 eg. Jash:"I would shield bash 30% of time, while intercept 70% of time,

    then I will sum it up :

    eg.
    bash, jash(30), jack(90) (30+90)/(30+90+70+10)= 0,60
    intercept, jash(70), jack(10) -II- = = 0,40

    and let the RNG select which move is actually selected. 60% for bash, 40% for intercept
    Otherwise we will be too biased. We should talk theoretically here, not really playing

    THOSE NAMES ARE JUST EXAMPLES.

    another example to avoid confusion:
    The
    A1 name(number)
    means that :
    name has decided to put value of number % to action A1
    Another Example:
    I have decided that in turn1-combat phase,
    i will give
    50% to switch
    50% to spectral choke
    sum must be 100%
    now on the summary list it will appear:
    A1-switch: supporters: nicolbolas(50) ,random player(40), jack(100) (jack would always switch)
    A2- choke: nicolbolas(50), random player(60) ,



    --------------------------------------------------------------------------------------------
    You may post out of current round, but it will not change the game.



    ----------------------------
    ----------------------------
    ----------------------------------------------------------------------------------------------------------------------
    ----------------------------------------------------------------------------------------------------------------------
    ----------------------------------------------------------------------------------------------------------------------
    Game summary, round by round

    Round 1 the range change phase:

    Player A:
    Ambrosia

    Far: nicolbolas(100)

    stay: PeteyC(100), jiaflu(100)

    close:



    Player B
    Helene:
    far:

    stay: garcia(100), EsieX3(100)

    close:



    Round 1 the consensus was Ambrosia stay Helene stay
    *RNG to simulate opponents thinking was not used, because this rule was created later
    ---------------------------
    ---------------------------

    Ok round1- combat phase
    We have:

    Full health Ambrosia (tome) with 100 energy

    at close distance vs

    Full health Helene (valk) with 100 energy


    A player Basic choices:
    A1) blood ritual
    A2) switch for Helene
    A3) Higashi
    A4) Intercept
    A5) Rest
    A6) Vampiric touch
    A7) spiked boots
    a8) spectral choke

    B player Basic choices
    B1) enchant
    B2) intercept
    B3) switch for Rumiko
    B4) Tefari
    B5) frenzied strikes
    B6) shield bash
    B7) Slash
    B8) (rest) (This would be useful against characters whose light defense is at least 3 higher then their dark defense)
  2. esiex3

    esiex3 Active Member

    Sure.

    B:

    Move close
    -Waste of energy. There's the possibility of Thorn not moving in, and even if you do, you can't frenzy because of Thorn's dark resist, so you can only sword slash, which MIGHT get you that 30 energy.

    Stay
    -Lose no energy, but even if he goes far that gives you a chance to enchant

    Move Far
    -why the hell would you do this? I don't think I need to explain at all

    Player B passes
  3. Rufus

    Rufus Member

    In most situations the optimal strategy will be mixed, so there is no single best move.
  4. jiaflu

    jiaflu Member

    Why would you move far on Helen with Ambro? Choke = free 40. She can bash for some garbage dmg, frenzy doesn't do shit, and Slash is her best option at 27 dmg. She can always enchant blade but then SHE has to guess right twice. Will you move far, or pass? Will you keep choking or switch?

    -edit-
    Agreed, most of the times, it's definitely a mixed strategy, with no clear-cut dominant strategy or equilibrium.
  5. Nicol Bolas

    Nicol Bolas New Member

    Round 1
    Range change

    A:

    move close
    -50 energy, 100 % possibility of all helene attacks
    stay
    high possibility of helene attacks, small possibility of having 50 energy advantage
    move far
    -50 energy, possibility of helene not moving in, and getting free hit/intercpt
    or helene wasting 50 energy

    A optimal strategy is to move far at this stage

    Can anyone do player B, and then we can maybe see a nash or dominant equilibrium?


    Helene could pass-enchant, and if you passed-whatever with ambrosia, you will be in trouble next turn, and either guess twice, or finish the exchange and leave helene with 5HP

    Bash-bash-slash doesn´t seem too good, but in the best scenario, you took 34 damage and then guessed, in the worst scenario, you wasted energy and still got 34+ damage. And in nightmare scenario, Helene didn´t bash but got some free hits, and consused you eg. -enchant-bash-strikes

    This is why I prefer moving far, (given same skills opponent, and playing rationaly) Playing irationaly to confuse rational opponent is entirely different chapter



    ----
    On a sidenote, Do you guys think I should always copy the moves into edited first post, or just keep doing down? I would prefer editing the first post (giving credit to people of course), and discuss the moves downward the forum
  6. garcia1000

    garcia1000 World Champion Staff Member

    Theory fighter is fun!

    To be honest if I was player B, with that deck, I would always start with Rumiko.

    So I think we should switch to Rumiko.
  7. Nicol Bolas

    Nicol Bolas New Member


    That sounds reasonable, however EsIeX3, and jiaflu already did some work on helene, so I don´t wanna trash it now :(
    But I hope you will post some analysis too. Since you started the strategy guide, this might be sorta good for new people (and even experienced) to see the thinking behind the game of some of the veteran people, and improve their own game.


    Anyway, later today (in 6 hours from now, I ll officialy close the

    round1-rangephase-1

    So far the concensus seems Ambrosia - far (2far 1 stay) and Helene stay (


    Since these take a lot of time, I think we will play one round every 4 days.

    Even for the range phase 1, numbers could be assigned to declare what move would you choose realtively more often
  8. garcia1000

    garcia1000 World Champion Staff Member

    Well if you insist on sticking with Helene, there is not much to discuss. The only move is pass. I don't see how there is any argument here.


    This is totally off topic so I put it in a different font color

    So I have played the same deck against newbies a lot. I have Rumiko/Scroll and Tafari/Poison. They have Helene/Valk as starter.

    This is what has happened every single time:

    Rumiko vs Helene:
    Me: 1. Go far / rest Them: Pass / enchant blade
    2. Pass / switch to Tafari Go close / frenzied strikes

    Tafari vs 10 en Helene:
    3. Go far / poison dart Them: Pass / rest
    4. Pass / poison dart Them: Pass / rest
    5. Go far / rest Them: Go close / rest
    6. Pass / poison dart Them: Go close / sword slash
    7. Pass / poison dart Them: Leave game
  9. PeteyC

    PeteyC New Member

    I vote stay with Ambrosia, and cast blood ritual. If opponent shield bashes they are out enough power that they cannot move close and attack next round, so I can move back then. If opponent uses sword slash then we trade damage about equally, which is worthwhile since Helene is better than Ambrosia. Also, Helene has a miss chance in this case and I do not. If opponent uses Enchant Weapon, then I am in no worse a situation than I would have been had I gone far and used blood ritual, since a Helene who enchants weapon this round will almost certainly move close next round to make sure I don't take the easy way out of her attack, and if I move far and blood ritual I won't have the energy to maintain range.

    The only other moves I would consider would be move far and rest, stay close and switch my Helene in, and stay close and vamp kiss. The benefit of vamp kissing is that I come out ahead damage wise, rather than about even if my opponent uses sword slash. The disadvantage is that I lose 10 more energy, which is a big deal if she shield bashes and I lose the move anyway, or if she uses enchant weapon, in which case the healing doesn't matter. I prefer to blood ritual now, because I may move far next round, and if I move far the extra energy will allow two more blood rituals, rather than one more + resting. In practice, though, I would sometimes vamp kiss instead just to mix it up and make Helene afraid of slashing me instead of bashing me.

    There is no point in moving far and using blood ritual since my opponent will just enchant weapon and move close, which she could have done just as well if I stayed close.

    Edit: Note that Necronomic tomes make Vamp kiss more attractive than it would be, except that since they seem to be behaving erratically this patch, I hate to let them influence my decision.
  10. ChadMiller

    ChadMiller Well-Known Member

    I've noticed this lately too; people just leave if you kill their Helene. I wonder what they'll do post-patch...
  11. Nicol Bolas

    Nicol Bolas New Member

    Round 1 combat

    Round 1 range change is now closed (check OP) However, if you need to say something important concerning it, post it! but it won´t be part of the game anymore

    I ll close the round1-combat 48-72 hours from now. But that also depends yon you guys, so also post on how this case stdy game should be conducted, this is just my view, 24 hour intervals would seem like a second job to me.



    Important update: we will be using patch 1,8 , and I will actually use RNG to evaluate the moves we select !

    This is becouse we won´t finish before the 1,8 update anyway, so we would be in a choice to play with old rules or change them in midst of the game !
  12. esiex3

    esiex3 Active Member

    EDIT: what the hell, you changed it after i posted :(

    see my below post
  13. PeteyC

    PeteyC New Member

    As Ambrosia I use blood ritual for the reasons stated in my previous post.

    As Helene I suspect that my opponent will
    1) Switch Out
    2) Use Blood Ritual
    3) Use Vamp Kiss or
    4) Use Spectral Choke

    If 1, then I should Intercept
    If 2, then I should use Enchant Weapon to force a nasty choice
    If 3, then I should Shield Bash or use Enchant Weapon to force a nasty choice
    If 4, then I should Shield Bash

    If 2 or 4 then Sword Slash is a decent secondary possibility.
    If 1, 2, or 3 then Switch Out is a decent secondary possibility.
    If 1, 2, or 3 then Enchant Weapon is a decent secondary possibility.
    If 4 then Switch Out is bad, and if 1 then Shield Bash is bad, and if 2, 3, or 4 then Intercept is bad.

    Helene will Shield Bash because it is the safest move, and seeing what my opponent does will provide me with more information about their play style from which to make future judgments (note that in actual game play sometimes I would Shield Bash, and sometimes I would use Enchant Weapon. In 1.7 Rules I would Sword Slash, but I do not like paying the extra five power in this situation, so with 1.8 rules I generally would not sword slash).
  14. esiex3

    esiex3 Active Member

    blood ritual? I think you're trying to make ambro lose :p

    As ambro, I would switch to helene. The probability of the other helene picking shield bash is pretty high, so I would switch so that she wastes energy.

    As Helene, I would enchant blade. Shield bash is a waste of energy, because A) if she switches, you're down to 70 energy and B) what if she were to do a bleeding ritual or vamp kiss? Then you would be down some energy and C) Even if you do shield bash, how are you going to finish her off? Your sword slash is NOT going to be enough to finish her off. It's much more effective to enchant blade so that you can threaten her with frenzied strikes or sword slash.


    I don't trust you with the RNG, you changed the rules AND the move after I made the post.
  15. Nicol Bolas

    Nicol Bolas New Member

    For A
    I pick switch for Ambrosia. (I wanted to move far in rangephase1, becouse I didnt want to face the shield bash, but i was overvoted)
    i would weight this chouce by 50

    spectral choke would be other 50
    the rest zero

    For B
    shield bash 100, other options are just too risky.

    Try to introduce numbers to your choices too. then I might do some maths with these, and let the RNG select which move is actually selected.

    Otherwise we will be too biased. we should talk theoretically here, not really playing
  16. Nicol Bolas

    Nicol Bolas New Member

    Important Update - The Last One

    -----------------------------------------------------------------------
    VERY IMPORTANT

    Assign percentages to your choices: example is in 1st post


    I will then evaluate
    eg.
    bash, garcia(30), nicol(90) (30+90)/(30+90+50+20)= 0,63
    intercept, sirlin (50), ilk(20) -II- = = 0,36

    Then:

    and let the RNG select which move is actually selected.
    Otherwise we will be too biased. We should talk theoretically here, not really playing

    --------------------------------------------------------------------------

    This will tremndously improve the simulation we play here, and really think of the best solutions in terms of hame theory, since the uncertainity of oppoent will be simulated here by a rational-by-weights-choices-RNG
  17. esiex3

    esiex3 Active Member

    wait, so how are you deciding our values? and why is garcia less than sirlin, he has a higher win/loss ratio than sirlin does, around 3:1
  18. Nicol Bolas

    Nicol Bolas New Member

    Hell no, I must have explained it so badly

    The

    A1 name(number)

    means that :

    name has decided to put value of number % to action A1

    Another Example:

    I have decided that in turn1-combatphase,

    i will give

    50% to switch
    50% to spectral choke

    sum must be 100%

    now on the summary list ot will appear:

    A1-swicth: supporters: nicolbolas(50) ,randomnplayer(40), jack(100) (jack would always swicth)
    A2- choke: nicolbolas(50), random player(60) ,
  19. PeteyC

    PeteyC New Member

    The problem with this example is that most strategies involve thinking 2-4 moves into the future, which you can't really do if there's a chance that move three in your 4 move combo will be switched into something else by a committee vote that goes against you. I suspect that all this will really prove is that decisions made by a committee are not a particularly good way to play a strategy game.

    That being said, here's my explanation for why I'd use blood ritual, since some people seem confused:

    Helene will probably either Shield Bash or Enchant Blade. If she Shield Bashes I lose very little energy, if she enchants her blade then I get a free hit before being put in the fight or flight position. And since I have enough energy to go far and blood ritual again, or go far, be countered, and vampire kiss, Helene will still have to pick move close after enchanting if she wants to guarantee that I will have the 50/50 damage split that she wants to stick me with. The important thing to note here is that whether Helene Shield Bashes or Enchants, Blood Ritual is the second best move. If she enchants, the best move is to spectral choke, and if she bashes, the best move is to switch out. However, given that she might do either one Blood Ritual is reasonably safe. Also note that two blood rituals and and a spectral choke guarantee her eventual death since she has no herbal remedy. Alternately, two spectral chokes and anything else (other than spiked boots) will kill her. The most energy efficient way to kill her is two bloods and a spectral, though, so the blood is not a wasted hit if it goes through.
  20. Nicol Bolas

    Nicol Bolas New Member


    The only way to solve it is by using the % system I just posted. The idea here is to run into and game theory equilibrium each round. But that impossible without formal expression.

    It will also be easier for others to follow you/ support you if you make explain your decisions shortly and with use of numbers.
  21. PeteyC

    PeteyC New Member

    In fact, this does not solve the problem, it exacerbates it.

    Allow me to explain:
    Move A may only be a good idea if it is followed by move B. If I cannot guarantee that I will be able to do move B after move A, then I will not do move A, I will do move C instead. However, doing move A and then move B together is better than doing move B.

    In this situation, the fact that my doing move B is subject to a decision by committee followed by an RNG means that move A is made artificially less tempting than it would be in the context of a real match. The dynamic is changed, and the example is no longer an accurate reflection of in game strategies.

    If the decision is made by committee, then I can at least gauge my chance of convincing the committee to do what I want, then act accordingly. However, using RNG as you've described it, means that unless the vote is unanimous I cannot guarantee that I will be able to do move B.

    Sorry I couldn't think of a way to express this using mathematical formulas, I'll try harder next time.

    Given:
    If Blood+Blood+Choke then Result(Helene=Dead)
    If Choke+Choke+(Blood or Vamp) then Result(Helene=Dead)

    Cost(Blood+Blood+Choke) = 110 Energy + 3 Rounds
    Cost(Choke+Choke+(Blood or Vamp)) = 130 or 140 Energy + 3 Rounds
    CostRounds(Blood+Blood+Choke)=CostRounds(Choke+Choke+(Blood or Vamp))
    CostEnergy(Blood+Blood+Choke)<CostEnergy(Choke+Choke+(Blood or Vamp))

    Since Result(Blood+Blood+Choke)=Result(Choke+Choke+(Blood or Vamp)) AND CostRounds(Blood+Blood+Choke)=CostRounds(Choke+Choke+(Blood or Vamp)) AND CostEnergy(Blood+Blood+Choke)<CostEnergy(Choke+Choke+(Blood or Vamp)) then (Blood+Blood+Choke) > (Choke+Choke+(Blood or Vamp))

    /sarcasm

    Sorry for that folks, I couldn't resist. Now I feel oddly ashamed of myself, though.

    Seriously, not trying to rain on anyone's statistical parade but I sense a fatal flaw in the way game theory is being deployed here. I'll withdraw from making further comments on the matter, though.
  22. Xanthus

    Xanthus New Member

    Your problem is not that others can't see into your mind, it's that you don't trust them to take opportunities later.

    You can look at a game as a series of positions (that may repeat), what you're actually saying is that you evaluate a future position more accurately than you trust others to (because they might miss the sequence that makes the position more favourable than on first sight). Well, you might be right, but it's obviously a problem with the committee not being good enough at Kongai, not with game theory ;)

    Actually, I was thinking that theoretically if you could evaluate the position resulting from all 8*8 combinations you could put them in normal form and solve the matrix, get your probabilities, and with that an evaluation of the current position, from there. You could do that recursively (also add 3*3 for the range), and you have a recipe for a ridiculously inefficient Kongai solver! Might actually work for low energy end-games that also include a small amount of guessing, although you'd need to trim useless moves and handle cycles. Maybe I should write an article: "Solving Kongai, the impractical brute-force way"! (If anyone has some time to spare, might be a fun programming project ;) Otherwise I might have a go at it myself in about two months.)
  23. Nicol Bolas

    Nicol Bolas New Member

    Ok. maybe trash it altoghether, and one post will just = one phase ? Purely decided by the poster?
  24. garcia1000

    garcia1000 World Champion Staff Member

    Hey guys, this is way too complicated!
  25. PeteyC

    PeteyC New Member

    I think Xanthus is right and I'm wrong, at least in theory. In practical terms the fact that people have different priorities, for instance, some people may care more about being safe, while others may be more risk taking likely means that a decision by committee will result in a screwy play style and some apparently illogical maneuvers, but hey, it could work.
  26. x1372

    x1372 Active Member

    yeah. this whole thing has gotten way too confusing

    Honestly, IF we're going to have any set of "theory kongai" then it
    A) needs to be two separate threads for the two sides
    B) need both sides to discuss options
    C) have a balanced, pre-set method for deciding each move.

    As is, this thread just gets more convoluted and useless. 3 pages in and the first turn hasn't even been decided for sure.

Share This Page