Bit on a rant intro, not necessary to understand the question. There is a heated debate between Arena players if (not) is the card shuffler (and matchmaker) rigged. I myself experienced very strange phenomenon. When I played on my PC, I was getting very poor draws with my Mono Red Deck and my win ratio was around 55%. However, when I swapped to playing on my laptop (first&new installation, this time not through Epic, but directly through WotC) I was getting incredibly lucky draws for roughly 50+ games easily getting 85% win ratio, most of the time being on the play and I would say getting favorable match-ups. I'm pretty sure about the win ratios and the quality of the hands (I used a tracker), the rest can really be my perspective. Anyways, I did not put too much thought into it back then.

Important part: However, today I came across this post about lands in the opening hand. Also this reddit post shows pretty thoroughly that the shuffler is not really random. I went and read some WotC forum posts and found people claiming that match maker altered their opponent picks according to their win ratio, providing (alleged) stats from trackers.

I'm okay with the shuffler and potentially matchmaker being 'rigged', to provide better game experience but only as long as the functionality is clear and public. Does anyone know of verified, up-to-date source which describes the shuffler and matchmaker rules, so one can play around it? So far I found nothing except small pieces of potentially outdated info or unfounded theories.

EDIT: I wonder if it was all my poor explanation or also people skim reading too much and ignoring comments but I intended this question purely as a search for information behind the shuffler/matchmaker. I think this is NOT opinion based and it feels to me that many people when see thread discussing shuffler/matchmaker automatically presume it is just salt mine thread, hence the vote for closing the question for being 'opinion based'. Anyways to sump up what I learned so far from comments and otherwise are:

  • Every time you shuffle in BO1, shuffle is done twice and the hand with the number of lands more representing the actual ratio of lands in your deck is chosen as your hand. This means it is extremely good to mulligan. This is very helpful to me because I avoided mulligan a lot in a fear of getting hand without lands or flooded by them, playing sub-optimal hands.
  • Fisher Yates is used for permutations. I also read Mersenne Twister is somehow involved, maybe for picking up the random numbers in Fisher Yates. Even though Mersenne Twister is not really very good at being random, I guess it is sufficient for this but I wonder if choosing a seed which would be unique for each id could somehow lead into what I experience. Anyway this is just a thought.
  • As for a matchmaker the only thing I noticed myself and also picked up from the threads I skimmed over and I'm confident enough to call without statistics is mirroring. You will face mirrors of your deck more often then it is represented in meta and people like CovertGoBlue who plays MTGA every day for hours claims to see this empirically too.
  • The following IS ONLY MY OBSERVATION but as this question is being voted to be closed as opinion based I think I will plug this in. I feel that MTGA deliberately 'tests' your deck. BTW I don't consider this to be a bad thing. I feel like it happens more than often that when picking new deck or running streak of wins or losses I run into very obscure cards which are either intended as a counters or a very favorable match-ups for me. I played Gruul yesterday running into mirrors, green, rogues, white aggro, black control, 0 Dimir's control. As soon as I switched to Cycling I get Ashiok's Erasure, which just shuts cycling with its little variability of cards which you actually play and who even plays this card anymore, and Dimir's control against. Like camaaan. :D I'm okay with it, but if this is random then what are the odds. There are many other theories and I don't want to be too much of a tin foil hatter but the only thing I'm trying to say here is: I do think that some meta table similar to this one but much more detailed, maybe even for specific cards, pair of cards or something is actually being built behind the scenes and it is somehow used to balance the gameplay further than just giving you mirrors more often. And again, I don't mind. Making the experience more balanced IS GOOD but I'm only trying to figure out its inner workings to understand if you should play more focused decks or more broad ones and so on.
  • 1
    This is how randomness works. Random does not mean "evenly distributed", until you have thousands or even millions of events. Humans are inherently bad at coping with the mercilessness of random, and believe we see patterns and intents everywhere. It is why we have prayed to weather gods for millennia, and it is most likely why you believe the shuffler to be rigged.
    – Arthur
    Commented Dec 29, 2020 at 15:23
  • @Arthur, your point is most apt. However, there are often bugs in software. There was a bug in the random number generator for Ultima Online that went undetected for YEARS, with many players discussing how to take advantage of it... It was very tricky to set up, but once you did, you could "farm" incredibly successfully. eXPRESS doesn't provide enough information to indicate if this is real, but 50 games at 85% winrate and 50 (??) games at 55% win rate is statistically significant.
    – John
    Commented Dec 29, 2020 at 15:54
  • 3
    @Arthur not really, they've actually outright said that their code doesn't intend true randomness, but does actually have code intended to smooth out card distribution between land and non-land. It still randomizes but lets call it random with an intentional bias to certain states.
    – Andrew
    Commented Dec 29, 2020 at 16:58
  • 2
    @Arthur behind the scenes it shuffles and draws twice then discards the less balanced hand of the two (based on the algorithm's idea of balance) they talked about it officially via forum announcement here: forums.mtgarena.com/forums/threads/26319?page=1
    – Andrew
    Commented Dec 29, 2020 at 18:13
  • 3
    You guys get all hung up about anecdotal desktop vs. laptop statistics, but you forget that each game, in its entirety (in particular shuffling), happens on the remote MTGA server, not on anyone's client. What insanity can lead anyone to believe that the server decides to favor players based on the platform they're playing on? I'm well aware MTGA has quite the history with bugs, but a correct shuffling algorithm is really trivial (google Fisher Yates), and would thus require extra effort to get wrong, especially in such a significant and easily detectable way. The whole premise is hogwash.
    – Hackworth
    Commented Dec 29, 2020 at 22:42

1 Answer 1


On the surface, 85% win rate over 50 games compared to 55% win rate over (??) games is statistically significant.

But, the danger of doing a single variate statistical analysis (which is rampant in modern discourse) is you aren't examining investigating other potential causes.

For Arena, specifically, I have found the following to have significant impact in my win rates.

  • Time of day. In my time zone, if I play between 10 am and 3 pm my win rates are significantly higher than if I play between 5 pm and 10 pm.
  • Event. You haven't indicated event. If I am on the "play" queue, I often win at much higher rates than if I'm on the "standard ranked" queue. I suspect other events have similarly higher variability.
  • Rank: My win rates in gold on the Best of 1 ladder are vastly higher than my win rates in Platinum.
  • Play Queue: My son likes brewing decks and playing on the play queue. His win rate is very low (he is a fan of jank) if he plays a lot, and I start playing on the play queue, I win all the time.
  • The Selection Effect: (Credit to Phillip Kendall) No one posts if they change computers and see no impact. In 2019 there were 3 million active players of Arena, the probably that at least one of them changes computers and saw variations of this size is non-trivial .

I'm sure I've missed others.

  • 3
    The biggest thing you've missed is the selection effect. Nobody posts when they charge machine and their win rate doesn't change, only if it does. Commented Dec 29, 2020 at 16:16
  • I don't play queue, only ranked, so the variability should be low. The stats are from Plat to Mythic. The laptop was only Diamond+ so that should also mean lower win ratio as people are generally better. I wonder how time of the day could make the difference, I did not see any effect on my win ratio but it is definitely interesting observation. To the Sel. Eff., not only that the person must change the computer but also observe this by playing long in a row and if this applies only to new players to boost their ego it might be unnoticeable. Also they have to care to mention and you notice it.
    – eXPRESS
    Commented Dec 29, 2020 at 16:40
  • Follow up, I don't make any accusations about WotC, I just think that noticing such a thing could really be tricky as we can be talking about very small fraction of a percent of players to be able to experience it let alone notice it, report it and the statement to reach the reader without getting lost. Also the posts are provided show it is not random per say, drawing 7 cards from a deck once, I'm okay with that, I really just want to know exactly how the mechanism works so I can better hedge my odds when mulligan, creating very specific deck or countering meta decks.
    – eXPRESS
    Commented Dec 29, 2020 at 16:42
  • @eXPRESS, there are two issues here. 1. Win percentage based on hardware. 2. How does the "land smoothing" algoritohm affect gameplay. My answer addresses the first issue. That issue is completely independent of the second, as it affects both game players equally. If you have a specific question on 2. ask it in a separate question.
    – John
    Commented Dec 29, 2020 at 18:43
  • @John Sure, but does it though. At least the things you mentioned does not seem to have effect on the stats I acquired. I think that my measurements are most likely just result of pseudo-random luck and my initial intention was not to claim otherwise. The reason why I mentioned it in a first place was only to provide the context why I started to dig into this. My question is actually about point 2 all along, but not only about land smoothing but generally about rules used in the shuffler/matchmaker. I see now that I maybe did not make myself entirely clear, so I'm sorry for that.
    – eXPRESS
    Commented Dec 29, 2020 at 19:18

Not the answer you're looking for? Browse other questions tagged .