Legendary Rule Change 23rd May 2013

[Protoaddict]

I don't think you should be using the word objective then, as I don't think your definition is valid. There is no Objective situation in MTG, Period. Cards are always evaluated based on subjective critera. Objective evaluations are almost always useless unless the cards are "strictly superior" as in the case of shock vs bolt.

A change of a single card in a single deck in any tournament can change the relevance of any one card in your deck, even if you never face that deck itself. Very much Butterfly effect, chaos theory stuff here but it holds true. Top 8s also do not take into account player skill, Intentional draws, friends and teams colluding, etc, so  as a tool to evaluate something objectively it fails.

If you want to judge a card objectively you can only do so on its own merits (the card is objectively an artifact, it costs 0, etc), because as soon as you compare it to another card it becomes a subjective argument, based on the cards (the subjects) and all their possible applications, and then the relevance of those applications which is basically un-quantifiable.

These 2 cards are different enough that "strictly superior" is not something that applies as there will always be relevant situations where one is better than another. If anything I would flat out argue that the cards are 2 different card that have 2 different uses and are not comparable. I don't even think you can safely say that opal is strictly better or worse than even a true mox, because it is all contextual.

[vaughnbros]

Data is, by definition, objective.  The % of Top 8 decklists or the # of Top 8 decklists with a certain card is the paradigmatic example of objective data/facts.  

This is a false statement.  Data can easily be subjective.  Haven't you ever watched a tooth paste commercial?

Yes, when people ask poll questions that are slanted data can be manipulated.  

But this isn't a case of Morphling.de aggregating results in a biased way to help me prove my case.  This is more like a census than a brand performance test or a push poll.  

Of course data can be manipulated, but that obviously isn't the case with Morphling.de top 8 results.

The # of Mox Opals in a top 8 is not a subjective question.   Sorry.  It's an objective question as far as the word "objective" is used in the english language.  

However, this is partially the case.  There is nothing is preventing people from top 8ing with sub optimal deck lists.  As a result you can't say that the data is completely unbiased.  Magic deck lists, and top 8's, are inherently a very subjective matter.

Here are some reasons you may see lotus petal in more top 8's than mox opal after July regardless of which one is the better card
1.  Lotus petal is a card that is ingrained into many peoples minds as a great card since it is after all on the restricted list whereas mox opal still has to prove itself to some people.
2.  Lotus petal can be sub optimally thrown into any deck as it has no deck building restrictions whereas mox opal requires a decent artifact count to even be considered.
3.  Lotus petal is currently seeing more top 8's.  So due to people copying deck lists it will continue to see more top 8's
4.  Lotus petal is a more available card, about $3, versus mox opal, about $30.

[Stormanimagus]

I would follow up Proaddict and add that players in Vintage are highly prone to group-think. Often, as in the media, a narrative can take hold and remain in control even though it isn't the truth.

I personally think there are more mis-built decks than optimally built decks in Vintage. I also think that there are many mis-built decks out there that make top 8s. Does this mean they are all objectively better? If players never sufficiently test something in a variety of meta games and contexts I don't think there is grounds to say it is inferior or superior to something else. We simply don't know in that case. There are many tried and true strategies in Vintage. There are also many long-time "playable" cards and many long-time "staple" cards. Does this mean ALL of those cards should be regarded as mainstays forever? Absolutely not. There are also cards that see very little play and thus very few top 8s. Should these cards be dismissed forever? No. This is an imperfect game played by imperfect beings. While tournament results are verifiable metric of a card's worth they are not the only or even the TRUE metric.

-Storm

[hvndr3d y34r h3x]

Calling it now.  Within the next two blocks they are going to have Legendary Dual Lands.

I know this isn't the thread for this, but just imagine te boos domain zoo would get in modern. I don't see it happening.



Also if metal craft was such a non issue, people would play dispatch over stp and path.

[Smmenen]

I don't think you should be using the word objective then, as I don't think your definition is valid.

Look, this isn't rocket science.   The word "objective" has a standard, readily understandable dictionary definition in this context.   We are using the word "objective" to mean facts that are ascertainable and which exist independent of individual opinion or feelings.   

I'm not arguing that Lotus Petal is strictly superior to Mox Opal, as that would be very easy to disprove.   A single example in which that was untrue would falsify my claim.  However, claims of objectivity are possible based upon objective, underlying facts.   Such facts, frankly, exist.  They are called tournament data.

Vintage is nothing less than a set of tournament rules.   In the context of Vintage, we can resolve disputes about which cards are objectively better by reference to objective facts, namely tournament data. 

There is no Objective situation in MTG, Period. Cards are always evaluated based on subjective critera. Objective evaluations are almost always useless unless the cards are "strictly superior" as in the case of shock vs bolt.

For years people bickered and argued over card quality, utility and strength.   Those arguments, which are long rooted in the history of this game, have always ultimately been resolved not by debates over casting cost or card characteristics, but by reference to tournament performance.   Tournament performance is not simply the only objective fact, it is the ultimate one.  It distills all of the data into one, quantifiable objective measure.  Tournament results, no matter how much people may disagree with them, cannot be argued with or reasoned with.  They simply are. 

There may be no "objective situation" or point of view in Magic generally, but that's because of the variety of formats.  A card may be objectively better in one format and worse in another.   In the context of a particular format, however, that cannot be so.

And before the objection is raised that the format that may exist in some theoretical model is different than that which exists in actual metagames, let me address it: that does not make tournament results subjective.  Although tournament results may be amenable to metagame variation, which in turn can be influenced by player preferences, that does not make them subjective.  Moreover, the larger your data set, and the more metagames represented, the less influential any particular players set of preferences can be.

The data that constitutes tournament performance is always objective.   It is not susceptible to feeling, opinion or belief.  It simply is.  It exists independent of any subjective point of view.  Moreover, to the extent that tournament results differ from some theorized or abstract ideal of the format, I would argue that tournament results constitute the Vintage format as much as the results that govern them.  After all, this the basis for DCI policy.  DCI policy is based ONLY on tournament performance -- either as a result of dominant decks or how specific tactics or cards affect tournaments.  In the modern era, cards aren't restricted if they don't affect tournament performance.   The Vintage format does not exist outside of tournament results. 

A change of a single card in a single deck in any tournament can change the relevance of any one card in your deck, even if you never face that deck itself. Very much Butterfly effect, chaos theory stuff here but it holds true. Top 8s also do not take into account player skill, Intentional draws, friends and teams colluding, etc, so  as a tool to evaluate something objectively it fails.

I was with you up until the last clause.  Top 8 data is objective.  The fact that many factors affect what makes top 8, some of then random or subjective, does not make the product: top 8 data, subjective.  Subjective preferences are baked into every objective cake.

Consider name trends.   Every year the Social Security administration reports the most popular baby names.  The names chosen are purely subjective: what parents want to name their children.  The aggregate reporting of those naming trends is objective: it's compiled by the government and reported as an objective fact, which it is.  The fact that subjective preferences are baked into objective facts does not render objective facts subjective.  http://www.ssa.gov/oact/babynames/

It is an objective fact that Jacob was the most popular baby name last year.   The name preference itself is subjective, yet it's popularity is an objective fact.  Same deal here.  While many subjective elements may be baked into the tournament cake, that is why we validate our data with larger sample sizes.   Over a long period of time, Lotus Petal has proven itself to be objectively better performer to Mox Opal, and I expect that this trend will continue despite the M14 rules changes. 

If that does not persuade you, let's return to something you said before:

Let's return to something you said before:

The reality though is that decks that would run opal and petal will now probably use opals first and then petal when they maxed out on opals,

This is a claim that is amenable to objective evaluation.   If you are right, then we will see decks that have Lotus Petal with no less than 4 Mox Opal.  If I am right, and you are wrong, people will run 1 Lotus Petal before the 4th Mox Opal. 

I think you are not only wrong, but the patent absurdity of your prediction underscores the extreme and gross flaws in your logic and understanding of the format.

These 2 cards are different enough that "strictly superior" is not something that applies as there will always be relevant situations where one is better than another. If anything I would flat out argue that the cards are 2 different card that have 2 different uses and are not comparable. I don't even think you can safely say that opal is strictly better or worse than even a true mox, because it is all contextual.

Did you miss the part where I said: "When I say that Lotus Petal is a better card that Mox Opal, I do not mean to say that it is "strictly superior."   Just wanted to be clear about that."

Calling something objectively better does not entail a claim that it is strictly superior.  I carefully distinguished between the two claims already. 

However, to say that the cards are so different that they are not comparable is laughable.  Of course they are: they are both 1) 0cc, 2) artifacts, which 3) can generate 1 mana of any color, and 4) function as artifact accelerants.   If they aren't comparable, then no two cards in Magic are susceptible of comparison. 

[Protoaddict]

Also if metal craft was such a non issue, people would play dispatch over stp and path.

This is certainly something that could happen in the future, but once again it's all situation. Mono white decks don't run many artifacts because they are devoted to hate, thusly it does not make sense in that deck. That being said you will see Path to exile over swords in some decks because it is situationally better when you have arbiters out or against decks with no basics.

Likewise Non mono white decks like bomberman tend to run a lot of sac artifacts, and since they are infinite combos or posion wins, swords is situationally better in those decks as the life gain is much less a draw back than the potential to not be able to cast it.

But if something like affinity every takes hold of in format, and mind you that deck just got much better with the change to opal and the tezzerets and that affinity as an ability is really good against sphere effects so prominent in the format, and that deck needs a swords effect, rest assured it would use dispatch.

[Smmenen]

Data is, by definition, objective.  The % of Top 8 decklists or the # of Top 8 decklists with a certain card is the paradigmatic example of objective data/facts.  

This is a false statement.  Data can easily be subjective.  Haven't you ever watched a tooth paste commercial?

Yes, when people ask poll questions that are slanted data can be manipulated.  

But this isn't a case of Morphling.de aggregating results in a biased way to help me prove my case.  This is more like a census than a brand performance test or a push poll.  

Of course data can be manipulated, but that obviously isn't the case with Morphling.de top 8 results.

The # of Mox Opals in a top 8 is not a subjective question.   Sorry.  It's an objective question as far as the word "objective" is used in the english language.  

However, this is partially the case.  There is nothing is preventing people from top 8ing with sub optimal deck lists.  As a result you can't say that the data is completely unbiased.  Magic deck lists, and top 8's, are inherently a very subjective matter.

Here are some reasons you may see lotus petal in more top 8's than mox opal after July regardless of which one is the better card
1.  Lotus petal is a card that is ingrained into many peoples minds as a great card since it is after all on the restricted list whereas mox opal still has to prove itself to some people.
2.  Lotus petal can be sub optimally thrown into any deck as it has no deck building restrictions whereas mox opal requires a decent artifact count to even be considered.
3.  Lotus petal is currently seeing more top 8's.  So due to people copying deck lists it will continue to see more top 8's
4.  Lotus petal is a more available card, about $3, versus mox opal, about $30.

I addressed your points in my response to protoaddict.  see my point about babynames and the SSA.  The fact that top 8 data may have baked in subjective preferences does not render it subjective.  Top 8 data is still objective data.

I would follow up Proaddict and add that players in Vintage are highly prone to group-think. Often, as in the media, a narrative can take hold and remain in control even though it isn't the truth.

I agree with that.  I'm constantly trying to argue against 'conventional wisdom' which is see as flawed.  But the reality is that group think cannot explain tournament results.  It only takes a single person who bucks conventional wisdom to perform well and change group opinion.  

Magic metagames operate like markets.   Markets are often driven by group think.  But when new entrepreneurs come into the market, and perform well, they change the entire metagame.  It only takes ONE person to actually change an entire metagame.  One person's tournament victory can constitute 12.5% of that top 8.  that's a huge number that can change the entire dynamic of Top 8 results.   Remember, most decks constitute much smaller % of Top 8s.   A single person can set of new trends which dramatically influence Top 8 results in any period of time. 

If people then test that person's deck, and replicate the results, the meme takes off.   Similarly, if people are resistant, then that one person can continue to perform until people believe in that deck.   See my performance with Burning Tendrils.  At first, people didn't believe it could win tournaments.  But after I started winning a few, more and more people started playing it.  Same with Josh P and Landstill, etc.  

I personally think there are more mis-built decks than optimally built decks in Vintage. I also think that there are many mis-built decks out there that make top 8s.

Your opinion is subjective.  Tournament results, are by definition, objective.   You may be right, but we have no way of knowing.  It's an epistemological limitation.   The reality is that the only way we can know something objective about the format is tournament data.   That's why we trust tournament data over individual opinions.  

Moreover, Magic decks are 'properly built' only to the extent that they are designed optimally to win tournaments.   The problem is that optimality is deeply contextual: every local metagame condition changes, which can never be perfectly known in advance.  

Optimality only exists in relation to ability to win tournaments.  Tournament conditions are constantly changing.   That's why we don't rely on people's opinions, but the interaction of people's choices and other people's choices.   We look at tournament results, just as we look at market prices to determine where markets clear and the value of goods.   Market equilibirums are imperfect, but they are better than anything else we have.

Firms use market data to make supply chain decisions.   So Magic players use tournament results to make design decisions.  Sure, both are imperfect, but the results of their efforts tell us something about the value of their inputs, product, and thinking.   If a firm makes a bet or marketing decision based upon data, and performs well, then we can interpret that as correctly understanding the market.  Similarly, if a player makes a calculation based upon a metagame prediction, and performs well, then we must conclude that at least some of their design decisions contributed to their success.   In the aggregate, we can observe patterns and trends which give us even greater confidence in these results.  

If we observe, over time, a consistent relationship between a certain set of cards and tournament results, we are justified in making a conclusion that these cards contribute to the tournament results.  The larger the dataset, the more reliable our conclusion.  There are statistical methods for this.   This is objective, and not a subjective, methodology. 

Does this mean they are all objectively better? If players never sufficiently test something in a variety of meta games and contexts I don't think there is grounds to say it is inferior or superior to something else. We simply don't know in that case. There are many tried and true strategies in Vintage. There are also many long-time "playable" cards and many long-time "staple" cards. Does this mean ALL of those cards should be regarded as mainstays forever? Absolutely not. There are also cards that see very little play and thus very few top 8s. Should these cards be dismissed forever? No. This is an imperfect game played by imperfect beings. While tournament results are verifiable metric of a card's worth they are not the only or even the TRUE metric.

-Storm

I don't know what you mean by "true," but Top 8 data is certainly objective.  It is not something someone made up or a product of individual opinion or feeling.  

[psyburat]

.  I'm sorry that my words are drowning in a pool of my favorite posters' walls of text.

Considering my last reply was 3 lines, isn't that a bit overblown:

Just because a card can be situationally better or worse doesn't mean there is no objective winner. 

As I said,LED can be situationally better than Black Lotus.   Lion's Eye Diamond can be better than Black Lotus when a Root Maze is in play.   I've actually used an LED in that way at a Waterbury. 

The objective winner is the card that will see more play.  Mox Opal is no Lotus Petal. 

Not you, the other people in the thread.  Thank you for the clarification.

[Protoaddict]

I was with you up until the last clause.  Top 8 data is objective.  The fact that many factors affect what makes top 8, some of then random or subjective, does not make the product: top 8 data, subjective.  Subjective preferences are baked into every objective cake.

Consider name trends.   Every year the Social Security administration reports the most popular baby names.  The names chosen are purely subjective: what parents want to name their children.  The aggregate reporting of those naming trends is objective: it's compiled by the government and reported as an objective fact, which it is.  The fact that subjective preferences are baked into objective facts does not render objective facts subjective.  http://www.ssa.gov/oact/babynames/

It is an objective fact that Jacob was the most popular baby name last year.   The name preference itself is subjective, yet it's popularity is an objective fact.  Same deal here.  While many subjective elements may be baked into the tournament cake, that is why we validate our data with larger sample sizes.   Over a long period of time, Lotus Petal has proven itself to be objectively better performer to Mox Opal, and I expect that this trend will continue despite the M14 rules changes. 

In your baby names example, I do not see the parallel at all. In that situations you are measuring quantity based on quantity. You are not concluding that one name is better than another. Results are also mutually exclusive, while the magic top 8 results are not as one player can play both cards. Lets say in the next 6 month you have the following results:

100 people named Brandon
200 people names Ralph
200 people ran Lotus Petal
100 People ran Mox Opal

Objectively, you can gather from this quantity. However you cannot say based on these results that people named Ralph are better than people named Brandon. So I don't see how you can say that Petal is better than opal based on that.

Also you are making the assumption that the top 8, when used to gauge card quality, means that the players performed better based on the strength of their cards. and while this is mostly correct, we all know that in a great number of events, there are hand shakes. There are factors that cause people who should get to the top 8 to not make it that have nothing to do with card quality. People who performed well in Swiss and have a better win loss record than there opponent can still lose in the single elim bracket of a top 16 and not make it to top 8.

Hell what about the Twaunserious Invitational that just happened? would you count thoes top 8 results in your objective calculations fully knowing that most of the players were drinking and not playing optimally. Also knowing that some people brought decks they considered "lazy" and not optimal because they knew they would potentially get drunk?

These 2 cards are different enough that "strictly superior" is not something that applies as there will always be relevant situations where one is better than another. If anything I would flat out argue that the cards are 2 different card that have 2 different uses and are not comparable. I don't even think you can safely say that opal is strictly better or worse than even a true mox, because it is all contextual.

Did you miss the part where I said: "When I say that Lotus Petal is a better card that Mox Opal, I do not mean to say that it is "strictly superior."   Just wanted to be clear about that."

Calling something objectively better does not entail a claim that it is strictly superior.  I carefully distinguished between the two claims already. 

However, to say that the cards are so different that they are not comparable is laughable.  Of course they are: they are both 1) 0cc, 2) artifacts, which 3) can generate 1 mana of any color, and 4) function as artifact accelerants.   If they aren't comparable, then no two cards in Magic are susceptible of comparison. 

It is not the similarities that make then comparable, it is lack of differences. Shock and Lightning bolt have 1 difference (2 counting name). That is it. So I think the strictly superior argument holds water

Mox opal and Lotus petal have multiple differences:

• One can be sacrificed without another card
• One has to be sacrificed to be used
• One requires metalcraft
• One is legendary and affected by cards or potential cards that affect legendary cards
• One can be in play in multiple on one side at a time
• One can come back in multiple from effects like open the vaults.

So while I think you can compare the 2 cards (certainly you can compare any 2 cards) and you can draw similarities and create situations where either one is better than the other, I don't think you can conclusively say one of these cards is "better" than the other. Better is a loaded word. Better in situations X or Y is fine, but better is objective and cards are never used objectively, ever. They are always subjectively strong or weak based on the sum total of the decks they are being used in.

[hvndr3d y34r h3x]

Also if metal craft was such a non issue, people would play dispatch over stp and path.

This is certainly something that could happen in the future, but once again it's all situation. Mono white decks don't run many artifacts because they are devoted to hate, thusly it does not make sense in that deck. That being said you will see Path to exile over swords in some decks because it is situationally better when you have arbiters out or against decks with no basics.

Likewise Non mono white decks like bomberman tend to run a lot of sac artifacts, and since they are infinite combos or posion wins, swords is situationally better in those decks as the life gain is much less a draw back than the potential to not be able to cast it.

But if something like affinity every takes hold of in format, and mind you that deck just got much better with the change to opal and the tezzerets and that affinity as an ability is really good against sphere effects so prominent in the format, and that deck needs a swords effect, rest assured it would use dispatch.

I'd say take mono white out because its awful but it illustrates my point. It would play dispatch already if metalcraft was at all reliable, seeing as how giving out a free land or 12 life is handing out free timewalks for them. They still dont play dispatch.

u decks run stp effects all the time. they never run dispatch even though metal craft is apparently not even a draw back. I guarantee you bomber man plays stp because metal craft is unreliable, not because it has a combo, the deck wins a lot of games beating with its duders. It would play dispatch 100% of the time if metal craft was actually as easy to obtain as people in this thread are stating.

I highly doubt affinity well see play outside of lodestone.dec, It would just get to blown out my snapcaster targeting claim or ancient grudge. the only good matchup it has, shop, just brings spash hate on it. Some awful deck that's a coin flip at best, and isn't making appearances anywhere doesn't really prove a point. That's like the earlier post about how 4x expedition map is playable because of mono u beltcher. Its just not, even if some deck that doesn't win anything with a degree of consistency plays it.

the point is no competitive existing decks outside of burning long, where its considered to be rather mediocre/cutable, plays metal craft cards. also, no imagined deck purposed so far can play it with out immoderately obvious glaring flaws.

[Smmenen]

Also you are making the assumption that the top 8, when used to gauge card quality, means that the players performed better based on the strength of their cards. and while this is mostly correct, we all know that in a great number of events, there are hand shakes. There are factors that cause people who should get to the top 8 to not make it that have nothing to do with card quality.

As I said, there are statistically valid methods for drawing inferences from objective data that would justify the conclusion that a card is 'better' than another.  These methods do not look at a single datapoint, but look for large patterns that minimize the influence of the kinds of variables you describe.   That's the benefit of Morphling.de   We can look at literally hundreds of tournament results all over the globe.   Over many years now, we are able to see how many Lotus Petals make Top 8s versus how many Mox Opals do. 

In any case, your point was that the data itself is not objective.  You said "there is nothing objective in Magic."  Tournament results are objective.   That's why they are the basis, in statistically valid data sets, for B&R list policy.

These 2 cards are different enough that "strictly superior" is not something that applies as there will always be relevant situations where one is better than another. If anything I would flat out argue that the cards are 2 different card that have 2 different uses and are not comparable. I don't even think you can safely say that opal is strictly better or worse than even a true mox, because it is all contextual.

Did you miss the part where I said: "When I say that Lotus Petal is a better card that Mox Opal, I do not mean to say that it is "strictly superior."   Just wanted to be clear about that."

Calling something objectively better does not entail a claim that it is strictly superior.  I carefully distinguished between the two claims already. 

However, to say that the cards are so different that they are not comparable is laughable.  Of course they are: they are both 1) 0cc, 2) artifacts, which 3) can generate 1 mana of any color, and 4) function as artifact accelerants.   If they aren't comparable, then no two cards in Magic are susceptible of comparison. 

It is not the similarities that make then comparable, it is lack of differences. Shock and Lightning bolt have 1 difference (2 counting name). That is it. So I think the strictly superior argument holds water

Except that no one -- and i mean no one -- is advancing a "strictly superior" argument in this context.   No one is claming that Mox Opal or Lotus Petal is strictly superior or inferior to another.    To the extent you are advancing that argument, it is a total straw man directed at no one.

So while I think you can compare the 2 cards (certainly you can compare any 2 cards) and you can draw similarities and create situations where either one is better than the other, I don't think you can conclusively say one of these cards is "better" than the other.

Well, it depends on what you mean by "conclusively."  Conclusive for all time?  No.  or with 100% confidence?  No.  But in a particular metagame at a particular period of time, you certainly can to a very high degree of statistical confidence, provided we have large enough data sets. 

Better is a loaded word. Better in situations X or Y is fine, but better is objective and cards are never used objectively, ever. They are always subjectively strong or weak based on the sum total of the decks they are being used in.

I think you are confusing the word "subjective" and "contextual."   Cards vary in synergistic strength based upon context, but from the point of view of the metagame system, we can certainly make judgments with a very high degree of confidence that a card is better than another. 

For example, we can conclude:

* Black Lotus is objectively better than Lion's Eye Diamond, although no strictly superior.

* Ancestral Recall is objectively better than Deep Analysis, although not strictly superior.

* Tarmogoyf is objectively better than Grizzley Bears, although not strictly superior.

That doesn't mean that these determinations hold for all time.  There could be a card that allows you to win the game if you can resolve Grizzley Bears, but not Tarmogofy.   Nor does that mean these conclusions are made with 100% certainty.  You don't have 100% certainty in anything in life.  But they are made with a very high degree of confidence (approaching 99% in some cases) that are given by the methods of statistics. 

[vaughnbros]

I addressed your points in my response to protoaddict.  see my point about babynames and the SSA.  The fact that top 8 data may have baked in subjective preferences does not render it subjective.  Top 8 data is still objective data.

Exactly I can't say that a name is better than another just because it was used more often.  In fact someone else might say a name is better if its used less often.  

There has to be an objective criterion for evaluating something in order to objectively call it better than something else.

Also you are making the assumption that the top 8, when used to gauge card quality, means that the players performed better based on the strength of their cards. and while this is mostly correct, we all know that in a great number of events, there are hand shakes. There are factors that cause people who should get to the top 8 to not make it that have nothing to do with card quality.

As I said, there are statistically valid methods for drawing inferences from objective data that would justify the conclusion that a card is 'better' than another.  These methods do not look at a single datapoint, but look for large patterns that minimize the influence of the kinds of variables you describe.   That's the benefit of Morphling.de   We can look at literally hundreds of tournament results all over the globe.   Over many years now, we are able to see how many Lotus Petals make Top 8s versus how many Mox Opals do.  

That doesn't mean that these determinations hold for all time.  There could be a card that allows you to win the game if you can resolve Grizzley Bears, but not Tarmogofy.   Nor does that mean these conclusions are made with 100% certainty.  You don't have 100% certainty in anything in life.  But they are made with a very high degree of confidence (approaching 99% in some cases) that are given by the methods of statistics.  

Yes there are statistically valid methods for drawing inferences from objective data.  However, some assumptions need to be met in order to do so.  When there are very few assumptions to work with there you can't perform a proper statistical analysis.

If we look at the data you are presenting, frequency of top 8's, it doesn't lend itself well to a statistical analysis.  There are a litny of problems with the data, so I'll just highlight some of the biggest problems I see with it:
1. The number of people playing in each tournament generally isn't the same
2. The number of each card being played in a tournament varies as well as the cards its competing against
3. The skill of each player in every tournament is different
4. There is variance in each individual match of magic since we only play a best of 3

[Phoenix888]

It would play dispatch 100% of the time if metal craft was actually as easy to obtain as people in this thread are stating.

While I'd agree that Lotus Petal not needing metalcraft is a huge plus and arguably makes it a better card, don't forget the obvious (I'm sure you haven't, but I'll state it anyways).  Dispatch requires 3 other artifact cards to function.  Mox Opal only needs 2 other artifact cards to function.  That aspect is non-negligible.

[hvndr3d y34r h3x]

It would play dispatch 100% of the time if metal craft was actually as easy to obtain as people in this thread are stating.

While I'd agree that Lotus Petal not needing metalcraft is a huge plus and arguably makes it a better card, don't forget the obvious (I'm sure you haven't, but I'll state it anyways).  Dispatch requires 3 other artifact cards to function.  Mox Opal only needs 2 other artifact cards to function.  That aspect is non-negligible.

Running 4 also means you've got three artifacts that will never count towards crafting it.
Not to mention I've seen opals do nothingness preventing it from turning its self on many a time.

I've played the card a lot and its still not a given it is turned on. If crafting it wasn't an issue more people would be playing the solo opal already for the freebie. Don't you think?

[Smmenen]

I addressed your points in my response to protoaddict.  see my point about babynames and the SSA.  The fact that top 8 data may have baked in subjective preferences does not render it subjective.  Top 8 data is still objective data.

Exactly I can't say that a name is better than another just because it was used more often.  In fact someone else might say a name is better if its used less often.  

There has to be an objective criterion for evaluating something in order to objectively call it better than something else.

And there is: tournament performance as measured either by % of Top 8 decklists or # in Top 8s is an objective criterion depending on whether a card is restricted or not.

After all, that is EXACTLY the same criteria that Wizards uses to Restrict or Ban cards in Vintage and other formats.  That's because it is an objective criterion.   

Also you are making the assumption that the top 8, when used to gauge card quality, means that the players performed better based on the strength of their cards. and while this is mostly correct, we all know that in a great number of events, there are hand shakes. There are factors that cause people who should get to the top 8 to not make it that have nothing to do with card quality.

As I said, there are statistically valid methods for drawing inferences from objective data that would justify the conclusion that a card is 'better' than another.  These methods do not look at a single datapoint, but look for large patterns that minimize the influence of the kinds of variables you describe.   That's the benefit of Morphling.de   We can look at literally hundreds of tournament results all over the globe.   Over many years now, we are able to see how many Lotus Petals make Top 8s versus how many Mox Opals do.  

That doesn't mean that these determinations hold for all time.  There could be a card that allows you to win the game if you can resolve Grizzley Bears, but not Tarmogofy.   Nor does that mean these conclusions are made with 100% certainty.  You don't have 100% certainty in anything in life.  But they are made with a very high degree of confidence (approaching 99% in some cases) that are given by the methods of statistics.  

Yes there are statistically valid methods for drawing inferences from objective data.  However, some assumptions need to be met in order to do so.  When there are very few assumptions to work with there you can't perform a proper statistical analysis.

If we look at the data you are presenting, frequency of top 8's, it doesn't lend itself well to a statistical analysis.  There are a litny of problems with the data, so I'll just highlight some of the biggest problems I see with it:
1. The number of people playing in each tournament generally isn't the same
2. The number of each card being played in a tournament varies as well as the cards its competing against
3. The skill of each player in every tournament is different
4. There is variance in each individual match of magic since we only play a best of 3

That's why you use large sample sizes, which increases their statistical validity.  The factors you cite do not render performance metrics invalid.  They just mean that the amount of data changes the confidence level in your conclusions.  You can never get 100% certainty, but you can get a high enough level of confidence in the data to justify calling a card "better." 

See my Black Lotus / LED comparison.  We have enough data to support that conclusion.  Ditto Tarmogoyf and Grizzly Bear. 

[heiner]

I agree with Smenenen that tournament results is the only objective means of measuring card quality. However, just counting occurences in T8 is difficult as:
- a restricted card is likely to appear less often than a four-of (when counting raw numbers and not decks containing a specific card)
- cards (and their quality) interact with each other

For example every deck that plays Opals will also play a single petal, hence, by definition the number of decks playing petal will be equal or greater than decks playing Opal. Opal basically makes Petal a better card. It is also interesting to note that while the disadvantage of Petal (sacrifice) is pretty much self contained (Yawgwill would be an exception), the negative impact of Metalcraft highly depends on the rest of the deck. Assume we had artifact-dual lands, I m pretty sure this would heave Opal over the top. Actually, RIGHT NOW Metalcraft is perfectly balanced in Vintage as you can just barely play Opal without loading your deck with crappy artifacts. This makes Petal right now the better card.

My predictions on the impact of the new legends rule:
- burning long decks might now up their Opals from 2 to 3 but not to four
- Tezzeret that plays one Opal might try 2
- The combined amount of Opals and Petal occurring in T8s will be roughly the same (some playing 1 petal 0 opal, some playing 1 petal 2 opal)

[Smmenen]

I agree with Smenenen that tournament results is the only objective means of measuring card quality. However, just counting occurences in T8 is difficult as:
- a restricted card is likely to appear less often than a four-of (when counting raw numbers and not decks containing a specific card)
- cards (and their quality) interact with each other

For example every deck that plays Opals will also play a single petal, hence, by definition the number of decks playing petal will be equal or greater than decks playing Opal. Opal basically makes Petal a better card. It is also interesting to note that while the disadvantage of Petal (sacrifice) is pretty much self contained (Yawgwill would be an exception), the negative impact of Metalcraft highly depends on the rest of the deck. Assume we had artifact-dual lands, I m pretty sure this would heave Opal over the top. Actually, RIGHT NOW Metalcraft is perfectly balanced in Vintage as you can just barely play Opal without loading your deck with crappy artifacts. This makes Petal right now the better card.

My predictions on the impact of the new legends rule:
- burning long decks might now up their Opals from 2 to 3 but not to four
- Tezzeret that plays one Opal might try 2
- The combined amount of Opals and Petal occurring in T8s will be roughly the same (some playing 1 petal 0 opal, some playing 1 petal 2 opal)

Heiner, I already addressed your main point: you count top 8 decks with each card, not the # of each card.  A deck with 4 Mox Opal is counted once.

[heiner]

I agree with Smenenen that tournament results is the only objective means of measuring card quality. However, just counting occurences in T8 is difficult as:
- a restricted card is likely to appear less often than a four-of (when counting raw numbers and not decks containing a specific card)
- cards (and their quality) interact with each other

For example every deck that plays Opals will also play a single petal, hence, by definition the number of decks playing petal will be equal or greater than decks playing Opal. Opal basically makes Petal a better card. It is also interesting to note that while the disadvantage of Petal (sacrifice) is pretty much self contained (Yawgwill would be an exception), the negative impact of Metalcraft highly depends on the rest of the deck. Assume we had artifact-dual lands, I m pretty sure this would heave Opal over the top. Actually, RIGHT NOW Metalcraft is perfectly balanced in Vintage as you can just barely play Opal without loading your deck with crappy artifacts. This makes Petal right now the better card.

My predictions on the impact of the new legends rule:
- burning long decks might now up their Opals from 2 to 3 but not to four
- Tezzeret that plays one Opal might try 2
- The combined amount of Opals and Petal occurring in T8s will be roughly the same (some playing 1 petal 0 opal, some playing 1 petal 2 opal)

Heiner, I already addressed your main point: you count top 8 decks with each card, not the # of each card.  A deck with 4 Mox Opal is counted once.

Thats why I am saying this counting technique is flawed. Opal basically requires you to play Petal, so the number of decks playing Petal will always be greater/equal than the number of decks playing Petal. Does this automatically make Petal the better card? No!
Assume the following card:
A: 0, this card can't be countered, you win the game if you play card B in your deck. B: 0, you loose the game. B is obviously worse than A although T8 numbers would be identical.

If cards were independent of each other, than your counting technique was fair, in the other case it is not. The interdependence of Opal and Petal is particularly important as (as I said before) Opal is currently just on the borderline of beeing playable. Would we all play only artifact lands, than it would be feasible to play Opals without Petals and in this case Opal actually might be better than Petal.

[Protoaddict]

And there is: tournament performance as measured either by % of Top 8 decklists or # in Top 8s is an objective criterion depending on whether a card is restricted or not.

After all, that is EXACTLY the same criteria that Wizards uses to Restrict or Ban cards in Vintage and other formats.  That's because it is an objective criterion.   

Wotc looks at a lot more than just top 8s when determining a ban. Top 8s may weigh a little more, but they look at overall attendance, diversity of decks, overall numbers, etc.

Best example is when Jtms was in standard (before cawblade). He was strong. He was not unbeatable. There were still a number of jund lists and bushwhacker boros lists putting up amazing results. But tourneys were dominated by Jace in attendance and people were vocal about him because a walker like that was still unprecedented.

So while your top 8 at a given even may have been 4 decks with Jace and 4 without, still a solid showing, overall tourney attendance may have some 75% decks with Jace. Statistically speaking, the top 8 of every tourney will have a more centerist metric of attendance of the tourney overall, as the more any one deck shows up the more likely they are to have to face off against one another and knock each other out of contention, proportionality reducing their presence in the top 8 a lot more on average, and reducing that decks actual presence at the event.

Jace ultimately was banned because of the legend rule. More people were playing him as a means to deal by legend ruling him than anything else, because there was no great walker removal at the time. The meta was not ready for that card and turned into can't beat em, join em.

This obviously has nothing to do with deck, player, or card quality, and everythign to do with statistics and luck, which leads into why top 8 is not a proper analysis of a meta or deck strength.

As I said, there are statistically valid methods for drawing inferences from objective data that would justify the conclusion that a card is 'better' than another.  These methods do not look at a single datapoint, but look for large patterns that minimize the influence of the kinds of variables you describe.   That's the benefit of Morphling.de   We can look at literally hundreds of tournament results all over the globe.   Over many years now, we are able to see how many Lotus Petals make Top 8s versus how many Mox Opals do.  

Just because your sample size is huge does not make the question you are asking of it more accurate. A sample size of 50 people who are all asked what gender they are will return a very accurate report on gender. A group with a sample size of 500 will not return a good report if the question they are asked is gender and then from that you try to extrapolate what their favorite website is.

I would also point out that intentional draws and collusion happen at a substantial number of events, so much so that it would have to be a substantial part of your sample size. This isnt to say that these are malicious, but people handshaking and IDing to get into the top 8 is not a skill based factor yet it alters the data you can get from a top 8. If that event is not removed from the equation (it cannot be) you will always have a corrupted sample, because that act alone ensures that skill and card power levels are not the determining factors.

[hashswag]

I think the "Mox Opal vs Lotus Petal" discussion might be getting a bit out of hand.

It's often said that the effectiveness of seemingly similar cards varies contextually (as they don't exist in a vacuum), but in the case of Mox Opal vs Lotus Petal, this is even more true, as the effectiveness of Mox Opal is directly dependent on Lotus Petal. Consider that while opening with Lotus Petal increases your chances of activating metalcraft for a Mox Opal in your opener, drawing Mox Opal a turn or two later when you have potentially used your one Lotus Petal could be very bad, since you effectively need to have had 4 artifacts for metalcraft.

So no, Mox Opal is not a better Lotus Petal with the new rules, nor is it a worse Lotus Petal; it's not a Lotus Petal at all. It's a situationally good mana source that is both helped and hindered by the existence of Lotus Petal. With the new rules, it's the definition of a high risk - high reward card. Storm pilots get the most benefit from the potential extra moxen multiple Mox Opals provide, but they also get the highest risk, since being able to bring large amounts of coloured mana online in the first 1-2 turns is crucial to the viability of their game plan. Other decks get a smaller boost from being able to use Mox Opal to power out one or two particular permanents, but they also take a smaller risk, since their decks are generally able to take advantage of it in the midgame when metalcraft is more likely to be online.

As far as determining whether Mox Opal should be restricted, I think it would be better to look at the percentage of Top 8s with one Mox Opal vs the percentage of Top 8s with multiple Mox Opals and see if there's an overwhelming difference. I don't think it will get even close to restricted, though, simply because it effectively pressures its pilot into treating it as a restricted card due to the exponential increase in inconsistency multiples introduce.

[DubDub]

Calling it now.  Within the next two blocks they are going to have Legendary Dual Lands.

I know this isn't the thread for this, but just imagine te boos domain zoo would get in modern. I don't see it happening.

Like that would matter.  They would just ban cards in the Domain Zoo deck and keep laughing all the way to the bank.  And it would be a seriously large bank, so that it could store the massive amount of money they'd make off Legendary Duals.  Modern players are growing increasingly accustomed to repeated and frequent waves of bans in Modern.  They can do whatever the crap they want there because they hold a few PT Moderns a year, that's pretty clear.

[nedleeds]

Calling it now.  Within the next two blocks they are going to have Legendary Dual Lands.

I know this isn't the thread for this, but just imagine te boos domain zoo would get in modern. I don't see it happening.

Like that would matter.  They would just ban cards in the Domain Zoo deck and keep laughing all the way to the bank.  And it would be a seriously large bank, so that it could store the massive amount of money they'd make off Legendary Duals.  Modern players are growing increasingly accustomed to repeated and frequent waves of bans in Modern.  They can do whatever the crap they want there because they hold a few PT Moderns a year, that's pretty clear.

They've already banned 2 mana Shock and a 1 mana conditional 3/3 that dies to every removal spell played in the format (except the 2 mana shock they banned). What else would they ban in a zoo style deck? Tribal Flames? I guess I wouldn't put anything past them with respect to modern.

[hvndr3d y34r h3x]

Punishing fire was way too good... i miss my 3/3 though. The player base has been pretty happy with recent banning.
The major downside of domain zoo is that you basic start at 11-15 life. The deck only runs one of the selected shock land. Having a vintage mana base while everyone else takes random 2-3's doesn't seem like something wotc would miss given they want the format to work...
... Then again, tarmogoyf.

[John Jones]

Yes but think about what this does for combo based decks. Maybe it will balance out the format.

[Cruel Ultimatum]

No.  "So on" means possible game 4s or 5s in matches that have drawn games. 

Good. Thanks for clearing that up for me.

Also, not all matches are best of 3. Some are best of 5, like pro your top 8s and the scg invitational t8.

[serracollector]

A big part of the lack of opals in the current vintage scene was the fact they were legendary. If I played 4 and my opponent played 4 then its off to the races. Now it is not like that their is nothing stopping me from playing my 4 additional moxen. Metalcraft is not nearly as hard to get in "certain" decks that are built to achieve metalcraft or just happen to be artifact heavy anyways such as mud, neo academy, and even affinity. How hard is it to acieve metalcraft in a deck with artifact lands? For your example about a land that couldt work unless you had three lands, well wat if certain creatures and moxen counted as lands? Then it wouldnt nearly be as much an issue. We have artifacts, artifact lands and artifact creatures. Even somthing as "silly" as a deck running 4 phyrexian revokers, 4 porcelain legionnaires, 4 etherswron cannonist, 4 shardless agents, and 4 metamorphs could run 4 mox opals and benefit from it greatly more than lotus petal. Its a permanent that stays in play to tap under wire, sac to smokestack, and provide mana after its first use. These are all benefits. I dont know how many people I know who would rather use a mox pearl over petal in regards to mud alone. I am sorry about my quote earlier that opal is better than petal, what I meant to say isthat in decjs built around/for the abuse of opal that opal is better in those decks. In a deck running 4+ rituals, 4 oath, and broken spells sure having 1 mana extra of what you need turn 1 is better, but it is not better in all deck types or versus all deck types.

I do not know if it will be worthy of restriction or not, time will tell, but I 100% do believe that now that there is no longer the worry of losing your mox opal to your opponent playing one, or the fear of not being able to use yours because your opponent has one out, and the fact it can be used ss a petal once metalcraft has been achieved, will get a lot more people to try and run 3-4 copies in possibly new decks.

[hvndr3d y34r h3x]

A big part of the lack of opals in the current vintage scene was the fact they were legendary. If I played 4 and my opponent played 4 then its off to the races. Now it is not like that their is nothing stopping me from playing my 4 additional moxen. Metalcraft is not nearly as hard to get in "certain" decks that are built to achieve metalcraft or just happen to be artifact heavy anyways such as mud, neo academy, and even affinity. How hard is it to acieve metalcraft in a deck with artifact lands? For your example about a land that couldt work unless you had three lands, well wat if certain creatures and moxen counted as lands? Then it wouldnt nearly be as much an issue. We have artifacts, artifact lands and artifact creatures. Even somthing as "silly" as a deck running 4 phyrexian revokers, 4 porcelain legionnaires, 4 etherswron cannonist, 4 shardless agents, and 4 metamorphs could run 4 mox opals and benefit from it greatly more than lotus petal. Its a permanent that stays in play to tap under wire, sac to smokestack, and provide mana after its first use. These are all benefits. I dont know how many people I know who would rather use a mox pearl over petal in regards to mud alone. I am sorry about my quote earlier that opal is better than petal, what I meant to say isthat in decjs built around/for the abuse of opal that opal is better in those decks. In a deck running 4+ rituals, 4 oath, and broken spells sure having 1 mana extra of what you need turn 1 is better, but it is not better in all deck types or versus all deck types.

I do not know if it will be worthy of restriction or not, time will tell, but I 100% do believe that now that there is no longer the worry of losing your mox opal to your opponent playing one, or the fear of not being able to use yours because your opponent has one out, and the fact it can be used ss a petal once metalcraft has been achieved, will get a lot more people to try and run 3-4 copies in possibly new decks.

I think you are wrong about the legendary thing, I'm pretty sure its the reliability of metal craft I've been talking about. The problem is that the "certain" decks built to achieve metal craft are just bad, at least all the examples I've seen (and anyone who knows me can tell you I'll test some pretty junks stuff on the off chance its good and I just don't see it). Going for metal craft is just not a good goal to shoot for, it doesn't achieve anything on its own accept for opening you up to commonly played hate spells, and the cards with "metal craft" written on them just aren't particularly high impact. At least not enough to play the bad cards, or bad combinations of cards, required to get craft beyond ring, crypt, moxen, lotus, top, sometimes vault/key, and the occasional mana vault. Anything beyond that and I would start taking a hard look at the list and asking myself what I'm giving up for what I'm getting. The number of artifacts people play currently may seem like a lot, but it just really isn't enough and they're typicality cards you don't want to see in high numbers in your opener.

[DubDub]

I'm pretty sure there's a whole new level of unintentional hilarity in that the rule is still named the Planeswalker Uniqueness Rule.

[Smmenen]

I agree with Smenenen that tournament results is the only objective means of measuring card quality. However, just counting occurences in T8 is difficult as:
- a restricted card is likely to appear less often than a four-of (when counting raw numbers and not decks containing a specific card)
- cards (and their quality) interact with each other

For example every deck that plays Opals will also play a single petal, hence, by definition the number of decks playing petal will be equal or greater than decks playing Opal. Opal basically makes Petal a better card. It is also interesting to note that while the disadvantage of Petal (sacrifice) is pretty much self contained (Yawgwill would be an exception), the negative impact of Metalcraft highly depends on the rest of the deck. Assume we had artifact-dual lands, I m pretty sure this would heave Opal over the top. Actually, RIGHT NOW Metalcraft is perfectly balanced in Vintage as you can just barely play Opal without loading your deck with crappy artifacts. This makes Petal right now the better card.

My predictions on the impact of the new legends rule:
- burning long decks might now up their Opals from 2 to 3 but not to four
- Tezzeret that plays one Opal might try 2
- The combined amount of Opals and Petal occurring in T8s will be roughly the same (some playing 1 petal 0 opal, some playing 1 petal 2 opal)

Heiner, I already addressed your main point: you count top 8 decks with each card, not the # of each card.  A deck with 4 Mox Opal is counted once.

Thats why I am saying this counting technique is flawed. Opal basically requires you to play Petal, so the number of decks playing Petal will always be greater/equal than the number of decks playing Petal. Does this automatically make Petal the better card? No!
Assume the following card:
A: 0, this card can't be countered, you win the game if you play card B in your deck. B: 0, you loose the game. B is obviously worse than A although T8 numbers would be identical.

If cards were independent of each other, than your counting technique was fair, in the other case it is not. The interdependence of Opal and Petal is particularly important as (as I said before) Opal is currently just on the borderline of beeing playable. Would we all play only artifact lands, than it would be feasible to play Opals without Petals and in this case Opal actually might be better than Petal.

the degree of interdependence is not as strong as you suggest.  Workshop decks, for example, could run Opal without any Petal at all. 

Even still, there are ways to mitigate any degree of interdependence in measurements that still use Top 8 frequencies by looking at marginal differences if not absolute #s.

[Thecheese]

Gaeas cradle is at $152 on ebay (regular not foil) abd still has a day left.......i did go and pock up a pset geist for around 65$. Lets see hat happens