Improve your deck with Trait Cards

Discussion in 'Deck Building' started by perrin144, Apr 11, 2016.

  1. perrin144

    perrin144 Kobold

    Hi everyone, I am new to Card Hunter but deeply interested in the game. I have recently put together a good Warrior/Warrior/Priest team and have been considering the merits of adding traits to the deck.

    It is common knowledge that trait cards improve draw chances, but how much does each additional trait card improve?

    This link (http://www.math.utah.edu/~thiriot/Math1040/homework/ch14review/ch14review.html) provides the formula as:

    1-( (total_cards-num_target_cards) / total_cards )^num_drawn

    For example, if you have a Cautious Mobility, and the two Elven Maneuvers are the only two trait cards in your deck.

    You have recently won a shinny new Bleneth's Skull with 3 Mass Frenzy cards, at the start of the game you would have a 24% chance (highlighted below) of drawing at least 1 copy of Mass Frenzy.

    Running the numbers through excel gives the following data:

    With 0 trait cards in the deck:

    1 card -- start round 08%, each turn 05%
    2 cards -- start round 17%, each turn 11%
    3 cards -- start round 24%, each turn 17%
    16 cards -- start round 83%, each turn 70%​

    With 2 trait cards in the deck:

    1 card -- start round 09%, each turn 06%
    2 cards -- start round 17%, each turn 11%
    3 cards -- start round 24%, each turn 17%
    16 cards -- start round 85%, each turn 72%​

    With 4 trait cards in the deck:

    1 card -- start round 09%, each turn 06%
    2 cards -- start round 18%, each turn 12%
    3 cards -- start round 26%, each turn 18%
    16 cards -- start round 88%, each turn 75%​

    With 6 trait cards in the deck:

    1 card -- start round 10%, each turn 07%
    2 cards -- start round 19%, each turn 13%
    3 cards -- start round 27%, each turn 19%
    16 cards -- start round 90, each turn 78%​

    Summary: Trait cards are great in terms of providing tempo/initiative, but consider adding more than a couple if you also want to improve the chances of drawing your deck's keystone cards.
     
    Last edited: Apr 12, 2016
  2. Omnomalisk

    Omnomalisk Kobold

    Hey! I really like this, I'm definitely gonna have to see what I can do to increase the consistency of my decks (using traits of course!)

    I've been working on my setup a lot lately and have hit the point where I liked two elf warrior setup but really need a priest to back them up with, do you have any suggestions?

    I'm guessing you have a lot of trait cards there too huh?
     
    perrin144 likes this.
  3. Killer74

    Killer74 Hydra

    Traits are pretty awesome, though if you run too many you run into the risk of exceeding your draw limit (10) in MP (more so if you run draw cards such as Unholy Energy / Elven Maneuvers. The max. # of traits is 20 for a priest, 18 for a wizard and 20 for a warrior (out of 36 cards).

    If you (I'm never sure whether to use we/one/you in whatever context, these are general yous) mush all three decks into one big one, with 108 cards, you can try and work out what might be a reasonable amount of traits so that you don't run the risk of too many overdraws. Ignore other drawing cards for the moment.

    Combinatorics ahead (hopefully correct), you've been warned.
    So, let T = # of trait cards. Probability you will overdraw = Prob. drawing >10 traits. C = choose, e.g. 5C2 = (5!/2!3!) = 10.
    You draw six cards per turn. Each trait card adds another draw.
    This is the same as saying you draw cards until you have drawn six non-trait cards (or else hit the draw limit).

    Out of the first sixteen cards, you'd want a maximum of 10 traits so as not to overstep the draw limit.

    Probability of overdrawing:
    P (# trait cards > 10 out of 16 cards) = P (#trait cards = 11) + P (#trait cards = 12) + ... + P (#trait cards = 16)
    P (#trait cards = 11) = (TC11)*(108-T)C(16-11) / 108C16
    .
    .
    .
    P(#trait cards > 10 out of 16 cards) = (TC11)*(108-T)C(16-11) / 108C16 + ...
    P(#trait cards > 10 out of 16 cards) = (TC11)*(108-T)C(16-11) + (TC12)*(108-T)C(16-12) + ... + (TC16)*(108-T)C(16-16)/108C16.

    Filling in various numbers for T... (Good old Excel, this is the sum of Hypergeometric distributions)

    If T = 10, P(#trait cards > 10 out of 16 cards) = 0
    If T = 20, P(#trait cards > 10 out of 16 cards) = 1.35215*10^-06
    If T = 30, P(#trait cards > 10 out of 16 cards) = 0.000252689
    If T = 40, P(#trait cards > 10 out of 16 cards) = 0.005757829
    If T = 50, P(#trait cards > 10 out of 16 cards) = 0.046228453
    If T = 55, P(#trait cards > 10 out of 16 cards) = 0.100687377
    If T = 60, P(#trait cards > 10 out of 16 cards) = 0.190714731 (Note that this is the max. # of traits you can have)
    If T = 71, P(#trait cards > 10 out of 16 cards) = 0.512879176
    If T = 78, P(#trait cards > 10 out of 16 cards) = 0.744113858

    This is for any given turn. There are considerations of the fact that past draws affect future draws, in a sense, but, I think (unsure) these can mostly be ignored, except for larger Ts.

    It is up to you to decide what is an acceptable risk of losing a card or more. Obviously, with 9 draws (first turn), these numbers are higher. Further draw type cards will affect these numbers, some at the start of the round (e.g. Unholy Energy), some during (e.g. Demonic Power), though ones during the round won't affect the initial draws. You could consider such cards as traits (or double/more traits in some cases) if you want (at worst, but you do get to choose when to play them). Also, to get such large amounts of traits in your deck requires selecting from a much narrower pool of items, as well as the fact that you constantly knock off attachments (which can be good or bad).
     
    Last edited: Apr 12, 2016
    Omnomalisk, perrin144 and Scarponi like this.
  4. perrin144

    perrin144 Kobold

    Great write up, I would give reddit gold if it were possible. I am not sure, but the divisor of 108C11 might all be 108C16 since it calculates a total of 16 cards?
     
    Last edited: Apr 12, 2016
    Killer74 likes this.
  5. Killer74

    Killer74 Hydra

    Sorry, yeah. Corrected, thanks.

    Edit: Addendum to above post:

    The above is if the traits are divided evenly between the characters. If the traits are divided unevenly, the numbers skews quite dramatically upwards, and also you may have to consider that one character is going to reshuffle more often (Don't think this bit matters too much though).
    For example, if there are 20 traits on one character and 0 on the other two, then
    P(#trait cards > 10 out of 16 cards for all characters) = P(#trait cards > 10 out of 16 cards for character with traits) = (20C11)*(36-20)C(16-11) + (20C12)*(36-20)C(16-12) + ... + (20C16)*(36-20)C(16-16)/36C16 = 0.138375

    Similarly, if you have two characters with 20 traits each, and one with none, then
    P(#trait cards > 10 out of 16 cards for all characters) = P(#traits cards > 10 out of 16 cards for the two characters with traits) = (40C11)*(72-40)C(16-11) + (40C12)*(72-40)C(16-12) + ... + (40C16)*(72-40)C(16-16)/72C16 = 0.179528

    To be honest, I'm a little less sure about the maths (or rather, what exactly I'm trying to calculate) now after thinking about it a bit. It makes sense to me, but I may have gone wrong somewhere.

    Perhaps it's time for some experimenting :D.
     
    Last edited: Apr 13, 2016
  6. Deepweed

    Deepweed Thaumaturge

    The numbers actually don't look that good for traits. In practice traits are better than what these numbers depict, though. Quality is another variable that should be taken into account. Most MP-usable traits actually range from E (black) to C (bronze) in quality. Since the traits are on the lower end of card quality, the cards you end up with should have significantly better qualities if you're using E (black) and D (paper) quality handicaps.

    I've tried some sort of "min-maxing" approach to card quality with wizards before. The idea was to get as many E (black) quality traits as possible so that the rest of the deck had cards that were way better on the average. Of course, I had to avoid running traits that could trigger other traits, and find a way to nullify the effect of the traits in the first place (a purging / therapy priest was the solution). In this case, getting a purging priest will "dilute" the deck's strength because the great draws on the two other characters are offset by the draws of the priest, which are mostly used to break even.

    The key is to find a way to make all the traits you use synergistic with the rest of the deck so that you don't have to compensate for running traits. That's why builds which spam Squeamish and Blind Rage are so strong. If you feel you have to compensate too much for using a particular trait, I don't recommend running it. Only do if you're running some sort of "min-maxing" approach like aforementioned, and possibly getting advantages when compensating (Gene Therapy and Purge are crucial for this to work).
     
    ParodyKnaveBob and Krunail like this.
  7. visak13

    visak13 Ogre

    I would like to add what deepweed said.

    The percentage of drawing good cards in a single round shouldn't be your concern. The main idea behind good builds is to understand the role of each character. For example, I have a party wherein I have one able killer in the party who can solo kill 1-2 dwarf warriors. The next character's job is to finish off a weakened character or two and he does it quite efficiently. The third character is the support which takes care of the party when I don't have the upper hand so that I can make a comeback. This is how I play in general. And to add to it the defense of both the killers is quite strong.

    Now coming back to the topic, the trait cycling is of no use if the traits can't be balanced. For example, if you are running a fumble then you should have officer's harness in your hand. Sometimes parry draws fumbles and that is game over if you happen to be in enemy's range. Blind range should be accompanied with purge. And so on and so forth.

    Plus they reduce the draw limit for a round.

    Tl;Dr: Traits work good if the deck has a way to counter the negative effects. On top of that, knowing the strengths and weaknesses of your deck is equally important.
     
    ParodyKnaveBob likes this.

Share This Page