Why we are disappointed in Ixalan's cards

FindingHeart8

Gilesclone said:

Is there actually any evidence to support this theory?

Beyond testimonies of other players and personal experience, no.  However, upon dissecting most other evidence presented in these threads, that is what it usually boils down to.

Of course, to be fair, that is why it is still a theory lol

Gunmix25

FindingHeart8 said:

HypnoticSpecter said:

Wow. Does @Volrak buy this?

some of us just have absolutely horrendous luck

About sums it up in a nutshell.

Volrak

FindingHeart8 said:

Gilesclone said:

Is there actually any evidence to support this theory?

Beyond testimonies of other players and personal experience, no.  However, upon dissecting most other evidence presented in these threads, that is what it usually boils down to.

Of course, to be fair, that is why it is still a theory lol

Forming a theory is obviously vastly easier than pursuing it.  It's not that pursuing it is greatly complicated (it's generally not).  It mostly just takes time and discipline.  In this case, all we need is to record a large enough set of unbiased card draw data over a number of games, and statistics will tell us the answer.

(It's also worth noting again that recall bias, a real phenomenon in all our brains, makes us tend to perceive the odds as being stacked against us when they're not.  That unfortunate fact makes most theories of limited value without some testing being done.)

FindingHeart8

Volrak said:

FindingHeart8 said:

Gilesclone said:

Is there actually any evidence to support this theory?

Beyond testimonies of other players and personal experience, no.  However, upon dissecting most other evidence presented in these threads, that is what it usually boils down to.

Of course, to be fair, that is why it is still a theory lol

Forming a theory is obviously vastly easier than pursuing it.  It's not that pursuing it is greatly complicated (it's generally not).  It mostly just takes time and discipline.  In this case, all we need is to record a large enough set of unbiased card draw data over a number of games, and statistics will tell us the answer.

(It's also worth noting again that recall bias, a real phenomenon in all our brains, makes us tend to perceive the odds as being stacked against us when they're not.  That unfortunate fact makes most theories of limited value without some testing being done.)

Thanks for the input Volrak.  I'm familiar with these concepts but a review is always nice.

And I agree with you that a large set of unbiased data would help in solidifying a theory.  I recall you presenting your database in past events, it was an impressive collection when I looked it over 6 months ago, and I'm sure it's grown greater since.  I'm interested in what steps you've taken with your database to insure accurate information has been inserted and to protect it from data tampering.  As I'm sure you know, rooting out any opportunity for malpractice is vital in pursuing accurate statistical results.

Skiglass6

I have always thought that hixus should disable every other turn. It could read disable all creatures when cast until end of opponent’s next turn. And when creatures do attack damage on Pw disable creature until end of next turn.

As for the destruction of overpowered supports, we need targeted support removal. And if not targeted, the support on the board longest should be targeted and should never include token supports. Just think if all of the creature removal were random and opponent has Olivia with token generator and every time you cast removal it hits the token. Even with this scenario Olivia can be delt with easier in other ways, disable, defender and berserk than the supports which the only other was is to match. 

babar3355

FindingHeart8 said:

Gilesclone said:

Is there actually any evidence to support this theory?

Beyond testimonies of other players and personal experience, no.  However, upon dissecting most other evidence presented in these threads, that is what it usually boils down to.

Of course, to be fair, that is why it is still a theory lol

I just don't really see how this could possibly be the case.  Whenever I have to cast 4 or more supports it feels like the cards are weighted to not draw supports.  When I need a support destruction spell I always get a kill spell and vice-versa.  If Hibernum was really that good at coding I would hope they could have fixed some of the masses of other bugs, errors, etc. 

No, I will stick with the most obvious and probable answer.  We are all a bunch of biased animals who think of ourselves as perfectly rationale beings.  The cards aren't rigged against us... Our brains are!

FindingHeart8

babar3355 said:

FindingHeart8 said:

Gilesclone said:

Is there actually any evidence to support this theory?

Beyond testimonies of other players and personal experience, no.  However, upon dissecting most other evidence presented in these threads, that is what it usually boils down to.

Of course, to be fair, that is why it is still a theory lol

I just don't really see how this could possibly be the case.  Whenever I have to cast 4 or more supports it feels like the cards are weighted to not draw supports.  When I need a support destruction spell I always get a kill spell and vice-versa.  If Hibernum was really that good at coding I would hope they could have fixed some of the masses of other bugs, errors, etc. 

No, I will stick with the most obvious and probable answer.  We are all a bunch of biased animals who think of ourselves as perfectly rationale beings.  The cards aren't rigged against us... Our brains are!

I think it would really depend on how much Hibernum relied on coding template from previous Puzzle Quest games, and how easy it would be to adjust it.

It's not that difficult to imagine, considering we've seen different values of probability assigned to card draws, that you could have a similar method for the likelihood of what you draw each turn.

Again, I'm not claiming certainty, just my (and other players') suspicion.

Volrak

FindingHeart8 said:

I'm interested in what steps you've taken with your database to insure accurate information has been inserted and to protect it from data tampering.  As I'm sure you know, rooting out any opportunity for malpractice is vital in pursuing accurate statistical results.

The whole community owns the database, not me.  (I didn't even start it.)  As for protecting its integrity, since you asked, a few steps are taken.

Entries are inspected as they're made; if anything looks dodgy I can ask whoever entered it.  In the past this has tended to happen accidentally in good faith rather than due to tampering or vandalism.  If there are major bad edits (which has only happened once, and I again suspect due to an accidental change), we can restore data from backup.  Data can also be compared with backed up copies to find changes compared to old entries.

Finally, if a person had both infinite patience, and a deep desire to pervert the truth (both of which are pretty rare around here in isolation, let alone together, incidentally), and embarked on a program of systematically entering corrupt data instead of real data, then that would be detectable to the degree that the corrupt data differed statistically from other peoples' entries.  And here is a big loophole for you: avoiding detection altogether would be as simple as making up data that's statistically indistinguishable from the truth. :P

Ultimately though, the whole thing depends on care being taken when data is entered, and the level of care the community has taken when they've done this is one I think they can be proud of.

FindingHeart8 said:

It's not that difficult to imagine, considering we've seen different values of probability assigned to card draws, that you could have a similar method for the likelihood of what you draw each turn.

I think you mean different probability assigned to card drops (from packs).  The game can generally process each card drop in isolation, but card draws have to maintain a consistently ordered deck stack, as well as consistent totals within 40 cards, so it'd be somewhat more complicated to bias.

By the way, I hope this was already clear, but I don't claim to have proven or even believe that such a thing is absolutely impossible.  But any real effect must be very subtle, or everyone would notice it in every game; such a subtle perception has far likelier alternative explanations.  Indeed, if random variation didn't sometimes cause unfavourable outcomes (which, unless I missed something, is the sum total of what this theory is based on), it'd be strange indeed, and evidence of some non-randomness.

Rhasget

I once recorded every card drawn in every match of HoD: Samut (three nodes).

The sum of all matches put together matched the statistical output pretty close (in regard of how decks were built with the amount of creatures/spells/support in each).

But looking at each match there were really standout draws in many. The Samut node for example had three matches were I drew spells more than the others, in a deck with 4 creatures/2 spells/4 support. One match starting with Demolish x3.

But the sum of all matches were balanced out with several nodes drawing few or any spells at at all (as was more expected).

The trend though was that every match (all nodes included) was always weighted to two types of cards, regardless of deckbuild.
If starting hand was creature/support/support, fewer spells were drawn and so on.

And this is still my my feeling when playing. Some matches I just churn out creatures and support while Greg just fills card after card but not playing any since he probably draws no creatures and hold mostly buff spells.
And next time my deck is stalling when only drawing mana changers and buffs when you frantically try to force a draw of what you need more.
It's mostly one or the other, very rarely an even blend.

FindingHeart8

Volrak said:

FindingHeart8 said:

I'm interested in what steps you've taken with your database to insure accurate information has been inserted and to protect it from data tampering.  As I'm sure you know, rooting out any opportunity for malpractice is vital in pursuing accurate statistical results.

The whole community owns the database, not me.  (I didn't even start it.)  As for protecting its integrity, since you asked, a few steps are taken.

Entries are inspected as they're made; if anything looks dodgy I can ask whoever entered it.  In the past this has tended to happen accidentally in good faith rather than due to tampering or vandalism.  If there are major bad edits (which has only happened once, and I again suspect due to an accidental change), we can restore data from backup.  Data can also be compared with backed up copies to find changes compared to old entries.

Finally, if a person had both infinite patience, and a deep desire to pervert the truth (both of which are pretty rare around here in isolation, let alone together, incidentally), and embarked on a program of systematically entering corrupt data instead of real data, then that would be detectable to the degree that the corrupt data differed statistically from other peoples' entries.  And here is a big loophole for you: avoiding detection altogether would be as simple as making up data that's statistically indistinguishable from the truth. :P

Ultimately though, the whole thing depends on care being taken when data is entered, and the level of care the community has taken when they've done this is one I think they can be proud of.

FindingHeart8 said:

It's not that difficult to imagine, considering we've seen different values of probability assigned to card draws, that you could have a similar method for the likelihood of what you draw each turn.

I think you mean different probability assigned to card drops (from packs).  The game can generally process each card drop in isolation, but card draws have to maintain a consistently ordered deck stack, as well as consistent totals within 40 cards, so it'd be somewhat more complicated to bias.

By the way, I hope this was already clear, but I don't claim to have proven or even believe that such a thing is absolutely impossible.  But any real effect must be very subtle, or everyone would notice it in every game; such a subtle perception has far likelier alternative explanations.  Indeed, if random variation didn't sometimes cause unfavourable outcomes (which, unless I missed something, is the sum total of what this theory is based on), it'd be strange indeed, and evidence of some non-randomness.

Thanks for the explanation Volrak.  I'm glad to see you take the time and caution with the data you're protecting.

"And here is a big loophole for you: avoiding detection altogether would be a simple as making up data that's statically indistinguishable from the truth."

I'd take a moment to point out your circular reasoning here.  The 'truth' here is determined by the statistics of the sum data that is inputted into your database, not the other way around.  While you may be able to detect data input that is a deviation off, incorrect data input over long periods of time could warp your results.  Now I wouldn't expect that to be a common response by most players, but I wouldn't presume all players to be beyond error or honesty in their inputs.  And failure to account for this variable does leave room to question the solidity of your results.  A player submitting consistently but inaccurately would outweigh the occasional player submitting accurately, and you have no method (beyond extremely inaccurate input) to determine which is which.

Please don't be mistaken, in the overall sense I've always respected at the attempt to gather such a database to further help inform the players.  And while I would definitely agree the results of your database are at least worth serious consideration, the foundation of this data being based a player honor-system with no reliable method to detect inaccuracy does not make this database presentable as irrefutable evidence, which is what I've seen a lot of by some players in this forum.

"I think you mean different probability assigned to card drops (from packs).  The game can generally process each card drop in isolation, but card draws have to maintain a consistently ordered deck stack, as well as consistent totals within 40 cards, so it'd be somewhat more complicated to bias."

I've heard that the deck is coded to be 40 cards stacks, 4 of each card per stack, which resets upon reaching 40.  I'm not sure how the reset factors in there; many times I've drawn 5 of the same card before having drawn 20 cards total.

My theory is that if the deck is randomly sorted but maintains card order upon game start, that some of the more powerful cards (following a similar coding method as the likelihood of drawing higher tier rares/mythics from packs) are given a slightly higher probability of being beneath lower tier cards.  While this difference may not be enough to have a blatant detriment to each game, it may have some weight in long-term gameplay.  I can be quite confident in the high probability of my opening hand (or next 3 cards) being a Saving Grace, Deserts Hold, Renewed Faith, or War Oracle in my Espeth Deck (and have had multiple games where my first 3 cards in hand were 2-3 of one of those cards), but I've never had an opening hand of more than 1 Crested Sunmare or DtG.

That being said, in an attempt to be objective about this, I have always weighed this against the possibility that I just have really, really bad luck. 

Stormcrow

It is phenomenally unlikely that Hibernum would introduce code to ever-so-slightly weight card draws in matches.

In order to believe that such code exists, you have to believe that:

1.) Hibernum decided the negligible gameplay "benefits" (?) of having a draw order that is slightly weighted towards frustrating players by hiding their better cards would be worth the time and expense (time is money, people!) spent implementing this code. (Note also: this mechanism, if it exists as you describe, would punish players who paid money for powerful cards by not letting them benefit from those cards as often; this is backwards from usual F2P/P2W game design.)

2.) Hibernum then wrote this code (or imported it from some other PQ - I haven't played any others so I don't know if they even use "decks" like this one), and definitely got the code working correctly. Stop and think, here, for a moment. Think about the vast, unbiased-statistics-gathering-project you are contemplating in order to investigate this issue. Think about the fact that after Hibernum added this code to their game, they would have to do similar statistics-gathering to know if the code they wrote was even working. In other words: you have to believe that just this bit of code had more testing devoted to it than....well probably all the other playtesting that's been done with this game put together. Otherwise, there's a good chance that even if Hibernum tried to add code that did this, they failed!

3.) Then you have to believe that none of the updates to the game broke this code. That this code doesn't interfere with anything. That it didn't introduce bugs with draw spells or fetch spells, or card-drawing PW abilities, or events, or the AI's card drawing (including PvE AI decks that didn't exist when this code was theoretically written) or anything like that. In short, not only did Hibernum get it working, but it kept working ever since then and this code hasn't broken anything else in all the development since then.

4.) Lastly, you need to believe that it's strong enough effect to be noticeable to you....but subtle enough that nobody else ever noticed it before you came along!

Or, maybe you're just unlucky and/or subject to confirmation bias.

Guess which side of the razor I come down on?

Seriously, conspiracy theories are a helluva drug, but kids: just say no.

miss_zoey

https://kotaku.com/activision-patents-matchmaking-that-encourages-players-1819630937

Volrak

FindingHeart8 said:

"And here is a big loophole for you: avoiding detection altogether would be a simple as making up data that's statically indistinguishable from the truth."

I'd take a moment to point out your circular reasoning here.  The 'truth' here is determined by the statistics of the sum data that is inputted into your database, not the other way around.  While you may be able to detect data input that is a deviation off, incorrect data input over long periods of time could warp your results.  Now I wouldn't expect that to be a common response by most players, but I wouldn't presume all players to be beyond error or honesty in their inputs.  And failure to account for this variable does leave room to question the solidity of your results.  A player submitting consistently but inaccurately would outweigh the occasional player submitting accurately, and you have no method (beyond extremely inaccurate input) to determine which is which.

In an attempt at brevity I've used confusing language.

To unpack the words I used (data that's statically indistinguishable from the truth) into a more precise but correspondingly more boringly verbose form: data that's statistically indistinguishable from the rest of the dataset, under the assumption that the majority of data has been entered correctly.  This isn't circular, and the assumption here (that most data is correct) is one which holds unless, amongst the mtgpq community, there's a covert army of spreadsheet bandits who secretly conspire to insert flaws in statistical analyses, putting as much or more effort into their endeavor as the entire community who donates their data in a quest for truth, and have such impeccable subtlety that not one of their number has yet tipped their hand.  I hope this is not something your mind needs putting at ease about!

FindingHeart8 said:

...database presentable as irrefutable evidence, which is what I've seen a lot of by some players in this forum.

"Irrefutable" is a very strong word and I wouldn't agree with that if I saw it, although I can't say I've noticed such a term being used a lot.  I generally give confidence levels for any results I present, and these should be kept in mind.  For example, a 95% confidence level on a result would mean that in 100 parallel universes which are identical barring random choices, in 5 of those universes, the result may be false but appear to be true (but in such universes the result is still very likely to be close to true).  Confidence (both statistical and literal) increases with the amount of data we have.

FindingHeart8 said:

My theory is that if the deck is randomly sorted but maintains card order upon game start, that some of the more powerful cards (following a similar coding method as the likelihood of drawing higher tier rares/mythics from packs) are given a slightly higher probability of being beneath lower tier cards.  While this difference may not be enough to have a blatant detriment to each game, it may have some weight in long-term gameplay.  I can be quite confident in the high probability of my opening hand (or next 3 cards) being a Saving Grace, Deserts Hold, Renewed Faith, or War Oracle in my Espeth Deck (and have had multiple games where my first 3 cards in hand were 2-3 of one of those cards), but I've never had an opening hand of more than 1 Crested Sunmare or DtG.

That being said, in an attempt to be objective about this, I have always weighed this against the possibility that I just have really, really bad luck. 

Well, anything is possible, and if you ever do want to collect an unbiased set of card draw sequences then I'd be happy to decode your observations statistically.

FindingHeart8

Stormcrow said:

It is phenomenally unlikely that Hibernum would introduce code to ever-so-slightly weight card draws in matches.

In order to believe that such code exists, you have to believe that:

1.) Hibernum decided the negligible gameplay "benefits" (?) of having a draw order that is slightly weighted towards frustrating players by hiding their better cards would be worth the time and expense (time is money, people!) spent implementing this code. (Note also: this mechanism, if it exists as you describe, would punish players who paid money for powerful cards by not letting them benefit from those cards as often; this is backwards from usual F2P/P2W game design.)

2.) Hibernum then wrote this code (or imported it from some other PQ - I haven't played any others so I don't know if they even use "decks" like this one), and definitely got the code working correctly. Stop and think, here, for a moment. Think about the vast, unbiased-statistics-gathering-project you are contemplating in order to investigate this issue. Think about the fact that after Hibernum added this code to their game, they would have to do similar statistics-gathering to know if the code they wrote was even working. In other words: you have to believe that just this bit of code had more testing devoted to it than....well probably all the other playtesting that's been done with this game put together. Otherwise, there's a good chance that even if Hibernum tried to add code that did this, they failed!

3.) Then you have to believe that none of the updates to the game broke this code. That this code doesn't interfere with anything. That it didn't introduce bugs with draw spells or fetch spells, or card-drawing PW abilities, or events, or the AI's card drawing (including PvE AI decks that didn't exist when this code was theoretically written) or anything like that. In short, not only did Hibernum get it working, but it kept working ever since then and this code hasn't broken anything else in all the development since then.

4.) Lastly, you need to believe that it's strong enough effect to be noticeable to you....but subtle enough that nobody else ever noticed it before you came along!

Or, maybe you're just unlucky and/or subject to confirmation bias.

Guess which side of the razor I come down on?

Seriously, conspiracy theories are a helluva drug, but kids: just say no.

1) The purpose here would not be to frustrate players (which if the was a huge difference between card weight, I could see that being a perspective), but more player behavioral reconditioning through delayed reinforcement (which there is a lot of published research about the success rate of that out there in promoting addictive behaviors).  So yeah, there would be significant motive for a business (especially an app) to do just this, not a conspiracy theory.

2) I don't have the information on whether Hibernum developed the coding for this game or outsourced the project to a different company (and I doubt either do you, as neither of us were an employee of Hibernum or Octagon).  Given the other aspects of this game that were required to make this game functional (merging a unique version of MtG synced with bejeweled), it could easily be concluded that your claim of "just this bit of code had more testing devoted to it than...well probably all the other playtesting that's been done with this game put together," is inaccurate.

3) Appeal to ignorance fallacy in your argument here.  Just because an aspect of the game hasn't caused issues to the game is not an argument for it not existing.  My game has never crashed upon purchasing a planeswalker or playing story mode, but we all know these options exist in the game.  Additionally, you don't know what factors weighed in to mtgpq glitches.

4) Ad hominem fallacy here.  I never said I was smarter than other players, nor that I was the only one who noticed this.  I stated earlier in this thread that there have been testimonies from other players.

Again, let me reiterate that this is a theory.  I'm not trying to convince you that it is for 100% certainty.

What I'm saying is that I've seen weighted systems in other app games, I know from research that it can be an effective business strategy, and I've experienced (and heard similar experiences from others) what appears to be symptoms of a minor weighted system existing in this game.

Gilesclone

So,  not only is there a perverse conspiracy to give you bad cards (something you have provided no evidence at all to support) but now I’m part of the cover up because I provide card data to the spreadsheet?
Seems a bit X-Files to me.

Stormcrow

FindingHeart8 said:

it could easily be concluded that your claim of "just this bit of code had more testing devoted to it than...well probably all the other playtesting that's been done with this game put together," is inaccurate.

I will admit I was being a little bit hyperbolic there to make a point - a point which apparently escaped you anyways. Have you looked at the bug report section of the forum? (I mean seriously, look at this list! Prickleboar's still bugged!) Much as I enjoy this game, "bug free and thoroughly playtested" has never been an accurate description of this game - any part of this game. If we assume your weighted code exists, but is just as buggy as the rest of the game, then how well does it really work? The more so for being virtually invisible and much harder to test than say, low-level planeswalker abilities or coalition name changes.

What I'm saying is that I've seen weighted systems in other app games, I know from research that it can be an effective business strategy, and I've experienced (and heard similar experiences from others) what appears to be symptoms of a minor weighted system existing in this game.

What I'm saying is that your theory bears all the hallmarks of a classic conspiracy theory - assigns deep significance to what are far more likely to be random events; is effectively impossible to disprove; requires a substantial investment in time and effort (and secrecy) by a number of people in order to be true; but is psychologically reassuring to the person believing in the conspiracy, who typically sees themselves as a "victim" of the conspiracy in some way.

And now you're attacking people who express doubts about the validity of your conspiracy, and suggesting that someone (for no apparent motive) might be deliberately tampering with any statistical results which do not support your theory. You've long since passed being on the wrong side of Occam's Razor and now you are deep into irrational, Kool-Aid-drinking territory here. (At this point, I think the person with the strongest motive to tamper with Volrak's data is pretty clearly you.)

FindingHeart8

And now you're attacking people who express doubts about the validity of your conspiracy, and suggesting that someone (for no apparent motive) might be deliberately tampering with any statistical results which do not support your theory. You've long since passed being on the wrong side of Occam's Razor and now you are deep into irrational, Kool-Aid-drinking territory here. (At this point, I think the person with the strongest motive to tamper with Volrak's data is pretty clearly you.)

I would like to start off by saying at no point have I attacked anyone in this thread.  Nor do I appreciate false claims of doing so, or baseless accusations of sabotage which cross the lines of civil discourse for these threads.  Let me make this clear:

• I have stated repeatedly that I respect and support Volrak's pursuit, and have stated repeatedly that my argument is not 100% certain.  Ironically there is a line saying this directly above the last thing I wrote that you quoted.

• Any unbiased experiment needs to have the capability of filtering out data-tampering, the inability to do so puts the results at risk.  Volrak has taken some precautions, but the small number of contributors to this experiment makes it a sample population, and these factors must be looked at if you're considering it as a representation of the entire mtg community.  This is not me having a conspiracy theory, nor threatening to sabotage the data, it's Statistics 101.

• I am not claiming to be a victim, nor start a revolution.  I also mentioned further back that if the weighted card-draw system hadn't been implemented, that it might be a good idea to consider to solve card-balance disputes.  Instead of nerfing powerful cards, why not make them just harder to draw?  Doesn't sound like a bad idea to me.
  

I will admit I was being a little bit hyperbolic there to make a point - a point which apparently escaped you anyways. Have you looked at the bug report section of the forum? (I mean seriously, look at this list! Prickleboar's still bugged!) Much as I enjoy this game, "bug free and thoroughly playtested" has never been an accurate description of this game - any part of this game. If we assume your weighted code exists, but is just as buggy as the rest of the game, then how well does it really work? The more so for being virtually invisible and much harder to test than say, low-level planeswalker abilities or coalition name changes.

Well we don't know it's as buggy as the rest of the game.  Some parts of this game have never bugged (in my experience at least.  Ex: story mode, purchasing a planeswalker).  However, updates they were making on card mechanics has prevented me (and other players I know) from logging into the game entirely, which makes me wonder how one bug in mtgpq effects other aspects.

FindingHeart8

Gilesclone said:

So,  not only is there a perverse conspiracy to give you bad cards (something you have provided no evidence at all to support) but now I’m part of the cover up because I provide card data to the spreadsheet?
Seems a bit X-Files to me.

1) this is a negative exaggeration of the points I was making.
2) not a conspiracy (see my post to Storm Crow above)
3) not accusing you of being "part of the cover up,"  Pointing out the vulnerability of an experiment to data tampering is not the same of accusing players of data tampering.
4) X-files was a great show

bk1234

[MOD MIC ON] If we cannot restore this thread to a reasonable degree of civility it will be closed [//MOD MIC]

Stormcrow

Oh there was an obnoxious bug for a little while in the early days (I wanna say around when Oath of the Gatewatch dropped?) where we were all (or a lot of us, I don't remember exactly) totally locked out of Story Mode for like a week. That was an odd one. Pretty sure there have been bugs with purchasing planeswalkers too, though understandably bugs with purchasing anything tend to be fixed very quickly. (Unlike Prickleboar....man I'm still mad about that, that bug totally ruined my first red deck ever...*shakes fist*) Anyways, the specifics aren't really the point: the point is, nothing about the game that we can observe suggests there are bug-free parts of the game we can't observe. It just adds to the extra-ordinariness of your extraordinary claims. And you know what they say about extraordinary claims.

I will add that I think making broken cards harder to draw is a good way to introduce massive bug and lots of player frustration. For one thing, there have been previous issues (and may be again!) with getting the deck's first 40 cards to actually be 4 of each card. For another thing, people who build decks around a single broken card usually aren't silly enough to wait to actually draw it. At least not without enough card-draw in the deck to blow right past any weighting mechanism. I, for one, learned long ago with paper cards that MtG would not be kind to me if I made decks that assumed I wouldn't be unlucky.