Duelyst Forums

Questions without answers (?)


The devs had said on a dev talk that you don’t have a deck but a pool of cards which the game randomly picks from.


Yes I am not questioning the fact that the cards are chosen randomly. But I am interested in learning when the game decides to “roll the dice”. Before the match or while you are playing?


If it’s the latter then it makes sense why i never open 2 drops and my opponent always opens the nuts.


i missed out 2 words so it completely changed the dynamic of the sentence.


yeah either way things never work out the way you want


I understood what you wanted to say.


If you replace a purgatos, you will never get a purgatos back. Even if you somehow had 22 in your deck
If you draw, you will probably get a purgatos if you somehow had 22


Is it relevant?

If I had to implement this myself, I would go for generating a fresh random number every time it is needed. Alternatively, one could set a random shuffling at the beginning of the game, but @epicflygon explained that there is no real deck, rather a pool of cards. With this in mind, the only plausible implementation choice is doing random samples on demand during the match.


That could work when drawing cards but when you replace its much more complicated. When replacing a card there are 2 main restrictions as I mentioned.


simple, it excludes the other cards in the pool.


I believe there is no easy way to reverse-engineer the replacing algorithm, so I can only guess what’s going on based on my development experience.

If I’m not missing anything about the mechanic, it seems to me that there is a straightforward implementation: remove all the copies of the replaced card before doing the random sampling from the pool and put them back afterwards. The only corner case is when you get an empty pool after doing the removal, in which case you are forced to get a duplicate.

I believe all these algorithms are trivial, there is not that much to conjecture about…


I’m really wondering if this is implemented or forgotten😅


I don’t know really, but it looks horribly hard to test without having access to the source code! Maybe the developers will appreciate the heads-up… :sweat_smile:


Just from my limited programming experience…

I’d definitely forget that, and since the situation is highly unlikely would never know about this bug :grin:


@ThanatosNoa which Dev do we mention to dispel this myth? Or can you do it?


What is this, the whole replace thing again?

I had some small talk with an engineer about a year ago, let me try to recap the conversation as I remember it.

I hope I can talk about this :fearful: Only because I want to dispel some myths


So the way replacing works (mind you, this is REPLACE, not mulligan, two completely different systems)

  • Replace card named “X”
  • Search ‘library’ for any card not named “X”
  • Add “not X” to your hand
  • EXCEPTION: If nothing but “X” shows up, add “X” back into your hand


Once again, completely different from replace, this only occurs when you start the match (where you pick your starting hand)

  • Replace card instance of “X1”

    Keep in mind you can have “X2” or “X3” of the same name in the deck, we’re only looking at this specific instance of that card

  • Find literally any other card that isn’t “X1” (this can include “X2” or “X3”)

  • Add that card to your hand

  • Repeat for the second mulligan (if any)

What does that mean? Unlike replace - you can totally get the same “named” card back.

Things I’m unsure about: Does it set aside both cards at the same time (both “instances”) if you do multiple mulligans or does it do it once at a time?

(In other words, if I mulligan “X1” and “Y1”, do both get pulled aside, or is it possible to get “X1” back into my hand when “Y1” gets mulligan’d)


:drum: Drum roll please…

  • Draw a(ny) card from your deck

:confetti_ball: Tada!

Ok on a more technical note: I believe I was told that the concept of a deck isn’t real (just figurative), the cards are in a “list” and when you draw, it just pulls randomly from the list. In theory, this works exactly like pulling from a stack of cards, because if you TRULY shuffled a card back into your deck (via “replace”) there is a chance that card could be at the top of the deck (unless your human urges instinctively tell you to shuffle that in the middle of your deck…)

Anyways, the point of the matter is that drawing from the deck is a completely disassociated system from REPLACING or MULLIGAN - there is no voodoo magic going behind the scenes to pull a card you just recently replaced (it really is just your own bias towards remembering the situation as opposed to the statistically relevant and more common occurrences of it not happening. It feels bad so you remember every time it happens).

The weirdest thing? Drawing into the card you replaced is mathematically not that rare to occur.

Your deck at most contains 39 cards (exclude the General) and could contain up to 3x copies of the card you just replaced. Upon your very first replace (exhuming the 5 card starting hand) you have a (1->3) / (34) chance of seeing the card you replaced that turn… drawn at the end of your turn. In other words, a 2->8% chance of having this occur.

The chance of this occurring only grows as you thin the deck of cards and retain copies of it.
Has it been 5 turns? Do you still have all 3 copies of Grandmaster Kraigon floating around? Well, then it’s a 3 / 30 [aka 10% chance] of finding that Kraigon you just replaced back in your hand after replacing… or once in every 10 games that last at least 5 turns with 3 copies of a card in your deck… this isn’t counting additional thining like card draw.

Want to know something spicy? This “math” also why it seems like everyone always maintains perfect answers into the late-game (you’ve thinned the deck every turn while poaching for your wincons) - every Magmar has Makantor on 6 because they’ve been looking for it every turn while drawing out of their deck!

To do that math you’d do 3/34 on the first replace, subtract a deck size for the end of turn draw, then do a 3/33 on the next turn, subtract a draw, that’s 4 searches in 2 turns for a rough percentage of 17% to run into one after 2 turns of replacing alone, and we’d only be at say 4 mana if we’re Player 1 so the chances of having one go higher and higher and if you take into account how many games you’d fight Magmar then…


Ok too much math. Too much rambling. I’m not a statistician.

TLDR: Game operates as game does because it do.

Draws are blindly chosen from the deck, much like if we were to truly shuffle a deck without any type of human error/bias (meaning that cards “shuffled” could actually end up at the top of deck).

Replace omits itself when finding a new card (of a completely different name) from the deck, mulligans just say “I don’t want THIS specific card back.” Card draw don’t care, it just goes “gimmie gimmie”

And there was some math.

No RNG in this game: LOOOL
Fan Card Design Hub: Discussion (+Card Design Contest Discussion)

can you add a line to that for me?

If Galaxydueler and turn 1, X becomes doom

thanks :kissing_heart:


regardless of actual deck and faction.


I would play Doom full time if I got that consistency boost

closed #21

This topic was automatically closed 5 days after the last reply. New replies are no longer allowed.