To put it simply, and none too flatteringly, they probably don't care much. If Arena is making them money, and obviously it is, they're clearly not very interested in putting out a well-balanced product. Arena has a huge list of issues right now and not one of them has been addressed. The UI, for instance, is in massive need of an overhaul purely for clarity. The amount of things that do not work properly, either due to straight-up error (abilities not firing) or cruft on screen (try targeting multiple cards, or even one with keywords, where your card is blocking the right side of the screen, the arrows are in the way, and the reminder text on the abilities makes it impossible to see the card you're targeting). Instead of fixing that, they add the laughably overpriced Guild bundles that are so chintzy that they only give you some card styles, and not even any copies of the cards themselves. For $20, you get a few styles and sleeves, but no actual cards.
That alone shows they have no interest in creating an introductory product here, or a valid replacement for MTGO. They want only cash, and are charging pretty crazy prices for anything in-game. The game has no decent tutorial for new players. The UI is actively harmful to learning and playing the game. Building a decent card pool requires a lot of grind, which isn't itself a crime, or a huge expense, which when combined with the former is definitely a black mark.
I play Arena because it's fairly popular to stream and it's a way to keep sharp, practicing the game every day. I won't spend money because I won't support the insanely greedy practices WotC is showing here—and I used to think that Blizzard was bad with Hearthstone! Certainly I wouldn't recommend this game to new players, and Wizards needs to be told the many things they are doing wrong. Hopefully they can fix it and make Arena into something great. Right now, I definitely would not classify it as such.
Honestly I'm glad that there's no cards behind the styles. Cosmetic DLC is a way to offer people something they want that doesn't disadvantage players who don't want it or can't afford it - it's a strategy that's certainly worked for Fortnite and I expect there's a significant number of people who have clamored for it or something like it.
I'm going to be honest, I liked the tutorial even though I've been playing since 2006. It probably was one of the things that kept me interested in the early game. It taught you controls moreso than gameplay, but I've never thought the gameplay was the hardest thing to learn outside of the minutia. I also wonder what abilities you haven't gotten to work right, I haven't had any issues personally.
You're right that there's a grind involved, and that you need to pay with Time or Money does mean you have to really REALLY grind to get good. For some people that's a turnoff, but are you expecting them to just give you a playset of all the cards for free? Serious question here.
Private Mod Note
():
Rollback Post to RevisionRollBack
Top 16 - 2012 Indiana State Championships Currently Playing: GBStandard - Golgari Safari MidrangeBG RBWModern - Mardu PyromancerWBR RLegacy - Good Old Fashioned BurnR
Okay, my tinfoil hat looks like a dang medieval helmet with all the layers going on by now.
MAJOR UPDATE ON WIZARD'S NEFARIOUSNESS!
So it appears Wizards is VERY MUCH aware of the shuffler problem because with the newest patch, your list no longer exports as imported. This can only mean they KNOW that importing the cards in a single order fashion fixes their broke ass shuffler. While the exported file will not be in the same order as if you made it through Arena initially, they will group all the cards of the same type in one row now. THERE WAS NO UPDATE ABOUT THIS IN THE PATCH NOTES!!! https://forums.mtgarena.com/forums/threads/53949
This means:
They must know their shuffler is broken.
Instead of saying "Update to the shuffler coming soon." Or actually fixing the problem, they keep trying to hide it.
They are hiding changes they make when they update the patch notes section.
Well, if it looks like bull*****, smells like bull*****, and there is a bull standing right next to the *****, it's probably bull*****.
I'm sure we could come up with perfectly reasonable explanations as to why they would do this, with a ton of mental gymnastics.
Bottom line, THEY ARE HIDING SOMETHING, IT'S UNDENIABLE SINCE THEY DON'T MENTION THE ACTUAL CHANGES THEY MAKE. CHANGES THAT FIX THEIR BROKEN SHUFFLER. Either they are EXTREMELY INCOMPETENT and covering that up, or they have had designs on this the whole time and covering that up. I vote designs on this the whole time since they have denied their problem existed, then fix the one way you can work around their broken system. Also, the opponent decider is another way in which they have rigged it. Of course I "can't prove it." However, the evidence against them keeps stacking higher and higher.
I'm curious what you mean, because when I exported my pseudo-randomized decklist to notepad, it was still in the same psuedo-random "cards added" order.
I also want to ask a dumb question: how does WotC benefit from having a clearly broken shuffler? The debacle was on the front page of Reddit, Chris Clay (the effective Boss of MTG Arena) was pinged multiple times in that thread. Wizards absolutely knows of the shuffler weirdness. And a lot of the players know that WotC knows. They can't ignore it at this point, and it's in their obvious self-interest to fix it.
You are asserting that Wizards is deliberately not fixing the shuffler they know is broken. This means that either A) It's in WotC's best interests to not fix the shuffler, or B) WotC is deliberately acting AGAINST its own best interests. So here's my question to you: Which do you think it is, and why?
Personally I would have very much appreciated at least a note that they're looking into things, and I'm a little concerned that there isn't one. Just because Artifact, once heralded as the game that would kill magic, has ceased to exist (their own numbers show less than 500 active players per day) doesn't mean Arena shouldn't keep being great.
Private Mod Note
():
Rollback Post to RevisionRollBack
Top 16 - 2012 Indiana State Championships Currently Playing: GBStandard - Golgari Safari MidrangeBG RBWModern - Mardu PyromancerWBR RLegacy - Good Old Fashioned BurnR
Open Beta went live on September 24, with the pre-releace events being September 29-30. Yesterday marked the 6 month mark for open beta.
Private Mod Note
():
Rollback Post to RevisionRollBack
Top 16 - 2012 Indiana State Championships Currently Playing: GBStandard - Golgari Safari MidrangeBG RBWModern - Mardu PyromancerWBR RLegacy - Good Old Fashioned BurnR
That's a good analysis Colt, and it seems like you have a lot more programming experience so I'll defer to you on that. I have also heard from a few people who are playing in the Mythic Invitational that the March 27th patch will also address the shuffling irregularities even though I can't see anything about it in the article. I'm not sure how much infrastructure it takes for Arena (The first 11 weeks had 250 Million games played, about 23 million a week - I expect it's closer to 27-28 million a week now) but hopefully they don't need too many more to apply this.
If it works, then I'll be rather pleasantly surprised at the quick turnaround. But we'll see fairly quickly if that's accurate.
Private Mod Note
():
Rollback Post to RevisionRollBack
Top 16 - 2012 Indiana State Championships Currently Playing: GBStandard - Golgari Safari MidrangeBG RBWModern - Mardu PyromancerWBR RLegacy - Good Old Fashioned BurnR
Okay y'all. "Randomizing" your deck in a text list, then importing it TOTALLY FIXES THE LAND/DUPLICATE PROBLEM! I know, I know, only one man, too small a sample size et.
I dare you to try it out for yourself. Instead of making it random, I actually put my list in a "semi-optimized order." As in, putting the cards in a list as I would hope to draw them.
This is all for bo1 ranked as my testing mode:
Well, I have been keeping tack of my win%, and also my land% drawn per game prior to this. My win ratio was at 50% and my % land drawn out of cards seen was at 50% or more, nearly every game, despite having only 23 lands in the 60 card deck. This is prior to the change.
I'm now 6 games in. My win:loss is 5:1 compared to 50%, and my land percentage drawn per game is an expected 30-40% of cards seen. (I know, too small a sample. Try it for yourself and see what you think.)
I'm not doubting you, I just want some more information, can you post the deck list (as exported) or explain where you've put the land to seemingly solve the problem? I'd like to do a bit of model testing myself, and knowing how you did things will give me some ideas on how I can test it further.
Private Mod Note
():
Rollback Post to RevisionRollBack
Top 16 - 2012 Indiana State Championships Currently Playing: GBStandard - Golgari Safari MidrangeBG RBWModern - Mardu PyromancerWBR RLegacy - Good Old Fashioned BurnR
EDIT!!!! Wait a second now!!! The deck builder automatically sorts everything right back into the usual order as soon as you click done, then reopen the list. That means it's IMPOSSIBLE to try and see if the order of the deck in the deck builder's list has an influence on the randomization factor of the shuffler.
It may sort things back into order in the deck building interface, but if you export the decklist to a text file it will be in "cards added" order.
Okay, I would love to "randomize" my deck order before it goes into play to see if that affects what the shuffler brings up. What I did was Export my current decklist, opened notepad, mixed up the lands and the cmc a bit, then imported my newly made "Random" list on Arena. All the cards were in the exact same order as they usually are. It seems impossible to make the deck builder deviate from the usual list of lowest cmc to highest cmc. Sure, you can save a list that's "Random" by clicking done, but it just puts them back in order when you do so.
Can you clarify how to arrange your own decklist order the way you want it? I would love to do so if possible. I anecdotally noticed that not only do certain cards "clump together" they also seem to be slightly organized by cmc as well. Like "Oh here's the 2 cmc section with a bit of 3s. Here's the 3 cmc section with a bit of 4s, Here the 4 cmc section with even more land.
I would love to be able to mix up my cmc and lands before it goes into the evil shuffler! Please explain to me how it's done. I'm sure other people here are curious as well.
I do not think the display will ever not sort them no matter what you do with that list. If you want to see if Arena preserves the list, then export the list for the deck you imported. If the card order is preserved then... well... we know that it's preserved. Unfortunately, if the re-exported list is sorted, it doesn't actually prove anything conclusive.
Okay, so I exported the "Random" list to notepad and they WERE still in the "Random" order. That does improve my confidence. You're saying, Arena will take the list order found under "Export" and not the usual display order under the deck builder feature? Do you know that conclusively? Or just guessing from a programming standpoint? Did Wizards explicitly state it?
Thanks for the tip!
I don't know this conclusively, but under my assumption of the source of the error it makes sense that the game reads your decklist in the same order it's shown in the export.
Private Mod Note
():
Rollback Post to RevisionRollBack
Top 16 - 2012 Indiana State Championships Currently Playing: GBStandard - Golgari Safari MidrangeBG RBWModern - Mardu PyromancerWBR RLegacy - Good Old Fashioned BurnR
DesolatorMagic's experiment is kinda meaningless, since he is only one "sample" in the population distribution. His experiment only proves that his experience is not the mean of the overall population (if indeed the true mean is 50/50), and not that the mean of the overall population is 50/50.
Edit: Basically DesolatorMagic's experiment does not test the true population mean in any way.
You're mostly correct but I'd like to note something. You're right that Dessy is only one data point. However, going first is a binary state - every game in Arena has to have one player go first, so getting a ton of data points from a ton of games will show a near perfect 50/50 of people going first, because half of everyone HAS to go first.
What Dessy's data set is showing (and 500 is probably a large enough number of trials to make a reasonable judgment) is that there is an aberration. It's notable, but it's also useless unless you can identify WHY there's an aberration and HOW that aberration is affecting the play/draw options. Again, hypothesis testing.
Private Mod Note
():
Rollback Post to RevisionRollBack
Top 16 - 2012 Indiana State Championships Currently Playing: GBStandard - Golgari Safari MidrangeBG RBWModern - Mardu PyromancerWBR RLegacy - Good Old Fashioned BurnR
EDIT!!!! Wait a second now!!! The deck builder automatically sorts everything right back into the usual order as soon as you click done, then reopen the list. That means it's IMPOSSIBLE to try and see if the order of the deck in the deck builder's list has an influence on the randomization factor of the shuffler.
It may sort things back into order in the deck building interface, but if you export the decklist to a text file it will be in "cards added" order.
RedditOP did not say he knew what methods WotC used but made some educated guesses because they are apparently near industry standard.
Regarding Dessy, the "who gets to play/draw first" thing is interesting but means nothing unless we have an explanation for why it's uneven, and can test that explanation.
Similarly, what RedditOP has discovered a very probable explanation for the phenomenon, but there's a fair degree of HARKing (Hypothesizing After Results are Known) which means his conclusion is circumstantial. Granted, it's enough circumstantial evidence for me to believe it as fact, but to statistically and scientifically PROVE it as fact requires hypothesis testing that as of yet hasn't been done. (Though I expect once it is done it will confirm the data we've discussed.)
As for the Pay to Play allegations, that's probably another argument entirely, though I'm going to lean on Hanlon's Razor: "Never attribute to malice that which is adequately explained by incompetence. (The first time)"
Private Mod Note
():
Rollback Post to RevisionRollBack
Top 16 - 2012 Indiana State Championships Currently Playing: GBStandard - Golgari Safari MidrangeBG RBWModern - Mardu PyromancerWBR RLegacy - Good Old Fashioned BurnR
Not sure what you're asking Malumdiabolus. Can you clarify?
Algernone25
Actually, the fact that this problem goes away after a mulligan could be proof that Wizards has nefarious intent. This was one confusing part of the study where he talks about "Maybe the mulligan fixes the problem because it 'shuffles' the deck again, thus making it more random." This doesn't make sense from a programming standpoint (I'm pretty sure at least, I'm not a programmer, Desolator gets into it in his video.) This is not a physical deck of cards that needs shuffling. They are using some kind of generator/algorithm to give you the opening hand and determine the card order sequence. A mulligan should in no way be influenced by how many hands you've seen in the digital world.
This is the best I can explain it. In paper Magic, you must shuffle your deck for a mulligan. On a computer, instead of having to shuffle that deck to get a new hand, the computer has another identical deck for you, already pre-shuffled. You just set aside the deck you mulliganed and pick up the fresh, pre-shuffled and randomized deck. Mulligan again? The computer has an infinite number of pre-shuffled decks containing all the same cards.
That's why taking a mulligan should have absolutely no impact on the randomness of the deck. If this is an error on Wizards part, it is a TREMENDOUS error. I say it looks more like a cover-up to me. Wizards has known that people complain about the shuffler and constantly tries to reassure them with placating words. Well, data is WAY more placating (or infuriating in this case) than words. Wizards could have run this study themselves with people in house and had a WAY larger sample size. Then they could say "Look everyone, here is your mathematical proof that the shuffler is correct." They didn't, probably because they knew they had something to hide.
I believe what Malumdiabolus is trying to ask is if there's a correlation between your player rank (bronze or gold or diamond or mythic) and how often/how much you get the skew from the expected values - that is, the game gives you worse mana issues if you're low rank and gives you better ones at high rank. Even if that's true, I find it likely that such data would be lost in the noise unless you had the extra effort to control for it, which I'm not sure his scraper is capable of.
I'll admit I'm not a programmer either, but I did a bit of reading on the randomization methods that the RedditOP talks about (Fisher-Yates Shuffle and Mersenne Twister) and I have what I think is a good guess of how the shuffler works:
-List cards in the deck from 1 to N, where N is the total number of cards in the deck.
-Use the Mersenne Twister to generate a random number between 1 and N.
-Find that card in the list and place it on the top (or bottom) of the deck.
-Repeat step 2 to generate a new random number, this time between 1 and N-1.
-Find that card in the iist, skipping over cards you've already placed in the list and place that card at the top (or bottom) of the deck.
-Repeat until all cards from the list are placed in a random order.
This looks all well and good from a technical standpoint. I suspect that the issue is caused by two factors - modulo bias and deckbuilding convention.
The Mersenne Twister dones't just pick a number from 1 to 60, it picks a number between 1 and 2^19937-1 which generates a gargantuan number. To make it fit, Arena probably takes the number and divides by your deck size (or deck size remaining to be shuffled) and uses whatever the remainder is as its random number. But since 60 or 40 doesn't evenly divide into that massive number, some remainders are going to be more common than others and that generates a bias, specifically towards the "top" cards in the list. Now consider the fact that when you start making a deck in Arena, it automatically loads lands into the list for you right as you start putting in cards. If you don't use the auto-land filler, lands are almost certainly the last cards you put into your deck.
This results in a state where the game's randomization table is more likely (not by a ton but by enough) to take cards from the top of your list first, and all the lands in your deck are either at the top or the bottom of that list. That's where I think the true issue lies. This might be testable by putting all your basic lands at the front of the decklist and all the non-basics at the back and see if you constantly see one or the other more frequently.
Why a mulligan fixes things, I assume it takes your already shuffled deck as the seed list instead of reverting to the original decklist, and since the lands aren't all at the top or bottom you now get the expected distribution, or one that's within error.
As for wizards not having done this math, I can think of a couple reasons but the biggest one is this: Wizards putting out a claim that the shuffler has no bias just as proof raises a lot of questions behind their logic - similar to putting out a new cereal that's "100% certified asbestos-free". Is it factual, sure. But you've just raised a hell of a lot more concerns than just breakfast cereal.
Private Mod Note
():
Rollback Post to RevisionRollBack
Top 16 - 2012 Indiana State Championships Currently Playing: GBStandard - Golgari Safari MidrangeBG RBWModern - Mardu PyromancerWBR RLegacy - Good Old Fashioned BurnR
I have always been fervent in stating that I wanted to see proof of any rigging in shuffling before admitting it's there, that I'm not going to accept just anecdotal data and the rants of a salty YT crapposter.
This is a sufficient sample size and the data analyzed in a very well-thought out and consistent manner, and the data is presented in a very clear format that makes the point rather clear. I was wrong, there is a problem with the shuffler.
I'm convinced that there is an issue that's going on, and Wizards needs to address this rather quickly because it's a VERY damning issue for them. Until then, about the only thing you can do is arm yourself with the knowledge. For those not wanting to pour through all the technical details, here's a good tl;dr:
-- Limited Decks will constantly be mana-starved, no matter how many you play. Fill up on stuff low on the curve.
-- 22-23 lands is the sweet spot for not getting issues - Play too many more and you get flooded more often, play too many fewer and you get screwed more often compared to expected value.
-- 3 land opening hands are great. 2-land or fewer hands get starved more often, 4-land or more hands get flooded more often compared to expected value.
-- Taking any mulligan seems to put the land/spell ratio back to where it should be.
--By all accounts this appears to be a common mistake in implementing the shuffler algorithm in which the deck is randomized, but not randomized enough. This is why taking a mulligan fixes it. This is an act of incompetence moreso than malfeasance.
-- There is no data to suggest that Arena gives you more copies of specific cards more often than expected.
I am hopeful that the people in charge of this at WotC catch wind of this and are able to affect a fix of some kind. It shouldn't be that hard, from what the report claims.
Top 16 - 2012 Indiana State Championships Currently Playing: GBStandard - Golgari Safari MidrangeBG RBWModern - Mardu PyromancerWBR RLegacy - Good Old Fashioned BurnR
Honestly I'm glad that there's no cards behind the styles. Cosmetic DLC is a way to offer people something they want that doesn't disadvantage players who don't want it or can't afford it - it's a strategy that's certainly worked for Fortnite and I expect there's a significant number of people who have clamored for it or something like it.
I'm going to be honest, I liked the tutorial even though I've been playing since 2006. It probably was one of the things that kept me interested in the early game. It taught you controls moreso than gameplay, but I've never thought the gameplay was the hardest thing to learn outside of the minutia. I also wonder what abilities you haven't gotten to work right, I haven't had any issues personally.
You're right that there's a grind involved, and that you need to pay with Time or Money does mean you have to really REALLY grind to get good. For some people that's a turnoff, but are you expecting them to just give you a playset of all the cards for free? Serious question here.
Currently Playing:
GBStandard - Golgari Safari MidrangeBG
RBWModern - Mardu PyromancerWBR
RLegacy - Good Old Fashioned BurnR
Clan Contest 3 Mafia - Mafia Co-MVP
I'm curious what you mean, because when I exported my pseudo-randomized decklist to notepad, it was still in the same psuedo-random "cards added" order.
I also want to ask a dumb question: how does WotC benefit from having a clearly broken shuffler? The debacle was on the front page of Reddit, Chris Clay (the effective Boss of MTG Arena) was pinged multiple times in that thread. Wizards absolutely knows of the shuffler weirdness. And a lot of the players know that WotC knows. They can't ignore it at this point, and it's in their obvious self-interest to fix it.
You are asserting that Wizards is deliberately not fixing the shuffler they know is broken. This means that either A) It's in WotC's best interests to not fix the shuffler, or B) WotC is deliberately acting AGAINST its own best interests. So here's my question to you: Which do you think it is, and why?
Personally I would have very much appreciated at least a note that they're looking into things, and I'm a little concerned that there isn't one. Just because Artifact, once heralded as the game that would kill magic, has ceased to exist (their own numbers show less than 500 active players per day) doesn't mean Arena shouldn't keep being great.
Currently Playing:
GBStandard - Golgari Safari MidrangeBG
RBWModern - Mardu PyromancerWBR
RLegacy - Good Old Fashioned BurnR
Clan Contest 3 Mafia - Mafia Co-MVP
Currently Playing:
GBStandard - Golgari Safari MidrangeBG
RBWModern - Mardu PyromancerWBR
RLegacy - Good Old Fashioned BurnR
Clan Contest 3 Mafia - Mafia Co-MVP
If it works, then I'll be rather pleasantly surprised at the quick turnaround. But we'll see fairly quickly if that's accurate.
Currently Playing:
GBStandard - Golgari Safari MidrangeBG
RBWModern - Mardu PyromancerWBR
RLegacy - Good Old Fashioned BurnR
Clan Contest 3 Mafia - Mafia Co-MVP
I'm not doubting you, I just want some more information, can you post the deck list (as exported) or explain where you've put the land to seemingly solve the problem? I'd like to do a bit of model testing myself, and knowing how you did things will give me some ideas on how I can test it further.
Currently Playing:
GBStandard - Golgari Safari MidrangeBG
RBWModern - Mardu PyromancerWBR
RLegacy - Good Old Fashioned BurnR
Clan Contest 3 Mafia - Mafia Co-MVP
I don't know this conclusively, but under my assumption of the source of the error it makes sense that the game reads your decklist in the same order it's shown in the export.
Currently Playing:
GBStandard - Golgari Safari MidrangeBG
RBWModern - Mardu PyromancerWBR
RLegacy - Good Old Fashioned BurnR
Clan Contest 3 Mafia - Mafia Co-MVP
You're mostly correct but I'd like to note something. You're right that Dessy is only one data point. However, going first is a binary state - every game in Arena has to have one player go first, so getting a ton of data points from a ton of games will show a near perfect 50/50 of people going first, because half of everyone HAS to go first.
What Dessy's data set is showing (and 500 is probably a large enough number of trials to make a reasonable judgment) is that there is an aberration. It's notable, but it's also useless unless you can identify WHY there's an aberration and HOW that aberration is affecting the play/draw options. Again, hypothesis testing.
Currently Playing:
GBStandard - Golgari Safari MidrangeBG
RBWModern - Mardu PyromancerWBR
RLegacy - Good Old Fashioned BurnR
Clan Contest 3 Mafia - Mafia Co-MVP
It may sort things back into order in the deck building interface, but if you export the decklist to a text file it will be in "cards added" order.
RedditOP did not say he knew what methods WotC used but made some educated guesses because they are apparently near industry standard.
Regarding Dessy, the "who gets to play/draw first" thing is interesting but means nothing unless we have an explanation for why it's uneven, and can test that explanation.
Similarly, what RedditOP has discovered a very probable explanation for the phenomenon, but there's a fair degree of HARKing (Hypothesizing After Results are Known) which means his conclusion is circumstantial. Granted, it's enough circumstantial evidence for me to believe it as fact, but to statistically and scientifically PROVE it as fact requires hypothesis testing that as of yet hasn't been done. (Though I expect once it is done it will confirm the data we've discussed.)
As for the Pay to Play allegations, that's probably another argument entirely, though I'm going to lean on Hanlon's Razor: "Never attribute to malice that which is adequately explained by incompetence. (The first time)"
Currently Playing:
GBStandard - Golgari Safari MidrangeBG
RBWModern - Mardu PyromancerWBR
RLegacy - Good Old Fashioned BurnR
Clan Contest 3 Mafia - Mafia Co-MVP
I believe what Malumdiabolus is trying to ask is if there's a correlation between your player rank (bronze or gold or diamond or mythic) and how often/how much you get the skew from the expected values - that is, the game gives you worse mana issues if you're low rank and gives you better ones at high rank. Even if that's true, I find it likely that such data would be lost in the noise unless you had the extra effort to control for it, which I'm not sure his scraper is capable of.
I'll admit I'm not a programmer either, but I did a bit of reading on the randomization methods that the RedditOP talks about (Fisher-Yates Shuffle and Mersenne Twister) and I have what I think is a good guess of how the shuffler works:
-List cards in the deck from 1 to N, where N is the total number of cards in the deck.
-Use the Mersenne Twister to generate a random number between 1 and N.
-Find that card in the list and place it on the top (or bottom) of the deck.
-Repeat step 2 to generate a new random number, this time between 1 and N-1.
-Find that card in the iist, skipping over cards you've already placed in the list and place that card at the top (or bottom) of the deck.
-Repeat until all cards from the list are placed in a random order.
This looks all well and good from a technical standpoint. I suspect that the issue is caused by two factors - modulo bias and deckbuilding convention.
The Mersenne Twister dones't just pick a number from 1 to 60, it picks a number between 1 and 2^19937-1 which generates a gargantuan number. To make it fit, Arena probably takes the number and divides by your deck size (or deck size remaining to be shuffled) and uses whatever the remainder is as its random number. But since 60 or 40 doesn't evenly divide into that massive number, some remainders are going to be more common than others and that generates a bias, specifically towards the "top" cards in the list. Now consider the fact that when you start making a deck in Arena, it automatically loads lands into the list for you right as you start putting in cards. If you don't use the auto-land filler, lands are almost certainly the last cards you put into your deck.
This results in a state where the game's randomization table is more likely (not by a ton but by enough) to take cards from the top of your list first, and all the lands in your deck are either at the top or the bottom of that list. That's where I think the true issue lies. This might be testable by putting all your basic lands at the front of the decklist and all the non-basics at the back and see if you constantly see one or the other more frequently.
Why a mulligan fixes things, I assume it takes your already shuffled deck as the seed list instead of reverting to the original decklist, and since the lands aren't all at the top or bottom you now get the expected distribution, or one that's within error.
As for wizards not having done this math, I can think of a couple reasons but the biggest one is this: Wizards putting out a claim that the shuffler has no bias just as proof raises a lot of questions behind their logic - similar to putting out a new cereal that's "100% certified asbestos-free". Is it factual, sure. But you've just raised a hell of a lot more concerns than just breakfast cereal.
Currently Playing:
GBStandard - Golgari Safari MidrangeBG
RBWModern - Mardu PyromancerWBR
RLegacy - Good Old Fashioned BurnR
Clan Contest 3 Mafia - Mafia Co-MVP
This is a sufficient sample size and the data analyzed in a very well-thought out and consistent manner, and the data is presented in a very clear format that makes the point rather clear. I was wrong, there is a problem with the shuffler.
I'm convinced that there is an issue that's going on, and Wizards needs to address this rather quickly because it's a VERY damning issue for them. Until then, about the only thing you can do is arm yourself with the knowledge. For those not wanting to pour through all the technical details, here's a good tl;dr:
-- Limited Decks will constantly be mana-starved, no matter how many you play. Fill up on stuff low on the curve.
-- 22-23 lands is the sweet spot for not getting issues - Play too many more and you get flooded more often, play too many fewer and you get screwed more often compared to expected value.
-- 3 land opening hands are great. 2-land or fewer hands get starved more often, 4-land or more hands get flooded more often compared to expected value.
-- Taking any mulligan seems to put the land/spell ratio back to where it should be.
--By all accounts this appears to be a common mistake in implementing the shuffler algorithm in which the deck is randomized, but not randomized enough. This is why taking a mulligan fixes it. This is an act of incompetence moreso than malfeasance.
-- There is no data to suggest that Arena gives you more copies of specific cards more often than expected.
I am hopeful that the people in charge of this at WotC catch wind of this and are able to affect a fix of some kind. It shouldn't be that hard, from what the report claims.
Currently Playing:
GBStandard - Golgari Safari MidrangeBG
RBWModern - Mardu PyromancerWBR
RLegacy - Good Old Fashioned BurnR
Clan Contest 3 Mafia - Mafia Co-MVP