[Article] March Type One Potpourri

[Dr. Sylvan]

Z said to put my articles in the Type One forum, and who am I to disagree?

Bram, it's another datafest for you. :-P

And the related percent of each set's card names that appeared.

March
19.9% - 57/287 Alpha/Beta/Unlimited
12.6% - 18/143 Urza's Legacy
10.9% - 18/165 Darksteel
9.0% - 7/78 Arabian Nights
8.7% - 25/286 Mirrodin
[...]

[jpmeyer]

YES.  POST ARTICLES HERE (lest we have a rehash of Hi-Val's last article.)

Odyssey
27 Standstill
24 Psychatog

:<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

[Dr. Sylvan]

Remember there's a Hulk list missing here. It's actually 27 Psychatog. No need to fret. :)

[jpmeyer]

Yes there is :<<<<<<<<<<<<<<<<<

[Smmenen]

Dr. Sylvan,

I think you should not count FoW as having a 5 cc.  The point is to see CCs, not what is printed on the card.  If you want a legitimate way of seeing this based on the text: how about viewing the UU3 as the alternate casting cost and 0 as the real casting cost?

Steve

[Dr. Sylvan]

[Openeth the floodgates of the data man.]

It might also be fair to conclude that Squee and Worldgorger Dragon are never cast for their cost (something untrue of Force of Will), and that they should be excluded like land. Unfortunately that requires a reporting bias.

For instance, let's say a billion games of playtesting Slavery vs. Hulk showed that it was pitch-casted 90% of the time by either deck (or overall). Or even, for the sake of argument, it could be the same for both of those decks. Should I then count it as a half-mana spell? Not necessarily. (Blue or colorless? Also not necessarily obvious.) What about Keeper? Maybe different decks have dramatically different pitch habits. Maybe a deck that features Misdirection will be more inclined to pitch for that whenever there's a counter war, preserving the hard counter for a time when it's cast normally. Maybe decks (or different variants of the same deck) with different numbers of blue cards are able to pitch more opportunistically, changing how often a zero would be accurate. Sweet merciful crap, it could take me a year to finally decide that FoW should be worth 0.76457674895 mana for March, and then when the metagame is a different distribution of deck types and matchups, the value is once again inaccurate to its contemporary situation.

And I definitely couldn't just pick Force of Will as the card to label by its ACC. Gush. Misdirection. Mogg freakin' Salvage. Pulverize. Exalted Angel. Decree of Justice. On and on. Each of these is sometimes played for an ACC... but sometimes not. Picking one or the other is arbitrary. Is their a proportion of the time that a card is played by its ACC that its cost goes from the corner to its 'real' mana cost? What specifically is that point? As I mentioned before, how do I determine what the proportion even is?

You can see how this is all less clear than it at first appears.

It might be possible to incorporate reasonable assumptions, but even so I think it's fair to leave it out. My logic here is that for all the alternate-casting-cost cards, there's a huge supply of zero-casting cost artifacts which are, realistically, treated in every way like speedy land, and are only in freak circumstances countered. I don't know what the precise ratio is, but I think the strictly accurate approach is reasonable.

One of the ideas I'm considering is that (since the issue you comment on bothers me to no end) I might try making them both unadjusted/raw like I've been doing, and adjusted/'realistic' like you recommend. Simply exclude cards that are treated like land or are never cast, and try to record the majority-use casting cost of each card. In general, I've been trying as much as possible to avoid doing anything other than representing actual data, which is why I'm so hesitant on such a seemingly obvious choice. Right now I'm in the middle of the Hulk dissection (or at least textifying it; I've had the numbers for a good week now), the Block critique with Tony, and some other trends. Plus school (haha, I are Cimmanon Bums Tan's excuse library), which is ramping up to finals. But I'll see what I can do on this one.

[Smmenen]

It basically boils down to what you are trying to accomplish.  If you are trying to show what the range of cc spells are that people generally play in Vintage, then you are making a serious mistake in not using the ACC.  

In my article last summer on the 10 principle of vintage, I wrote that no non blue or artifact spell that costs more than 3 is playable except Rector.

The data has absolutely no use whatsoever except as some esoteric sadistic meaning if you are going to count Gush and Force of Will as U4 and UU3 respectively.  Force of Will is the most played blue card in Type one and therefore all the blue averages are just pointless.  If you want the data to be useful, then I think common sense calculations are important.  

the tough issue is cards like Salvage becuase they are as often used at both costs.   But the fast majority of Type One spells are far more often used at one cost than another.  That's why Gush as an alternate casting cost of 4U and FoW as an alternate cc of 3UU.  If you are afraid of diluting the data with this sort of judgement, then make adjusted tables, and put the real tables in footnotes or something.  Squee and Dragon effects that have alternatve casting as spells.  I think that including them is also bad becuase if someone is trying to use your data for practical purposes for card design or deck construction, then they won't understand that 6 cc red spells are not playable and they might add jokulhops to a type one deck.


Steve

[Tristal]

In my article last summer on the 10 principle of vintage, I wrote that no non blue or artifact spell that costs more than 3 is playable except Rector.

This just came off as funny, because a good deal of the time Rector got Yawgmoth's Bargain.  I'll grant respite here because it's a rare occasion that Bargain ever gets 'cast', but I'd offer a few more examples:

Exalted Angel (Maybe not applicable last summer.)
Shattering Pulse/Allay (More applicable last summer than now.)
Tendrils of Agony
Nether Void
Mind Twist
The Abyss
Moat
Humility
Goblin Ringleader
Siege-Gang Commander


I don't think the ACC is really an issue.  If you look at the list of casting costs, the 4RR entry is not going to be so misleading given the number or R, 1R and 2R costs here.  It's blatantly obvious that all the curves are low; even with Blue's distortion at 3UU, the 1 and 2 casting cost spells far outweigh it.  I seriously doubt anyone reading Dr. Sylvan's articles is going to look seriously at the Blue rankings and think "Gee, I can build a mono-blue deck with a bunch of 5 casting cost stuff like Mawcor and Air Elemental!  Cause Dr. Sylvan said so!"

[Smmenen]

In my article last summer on the 10 principle of vintage, I wrote that no non blue or artifact spell that costs more than 3 is playable except Rector.

This just came off as funny, because a good deal of the time Rector got Yawgmoth's Bargain.  I'll grant respite here because it's a rare occasion that Bargain ever gets 'cast', but I'd offer a few more examples:

Exalted Angel (Maybe not applicable last summer.)
Shattering Pulse/Allay (More applicable last summer than now.)
Tendrils of Agony
Nether Void
Mind Twist
The Abyss
Moat
Humility
Goblin Ringleader
Siege-Gang Commander


I don't think the ACC is really an issue.  If you look at the list of casting costs, the 4RR entry is not going to be so misleading given the number or R, 1R and 2R costs here.  It's blatantly obvious that all the curves are low; even with Blue's distortion at 3UU, the 1 and 2 casting cost spells far outweigh it.  I seriously doubt anyone reading Dr. Sylvan's articles is going to look seriously at the Blue rankings and think "Gee, I can build a mono-blue deck with a bunch of 5 casting cost stuff like Mawcor and Air Elemental!  Cause Dr. Sylvan said so!"

I also said Bargain was an exception.  But Angel is generally morphed, Mind Twist uses Drain and is often cast for less then 4, SeigeGang comes into play with Lackey and one of the points of my article was that Moat and Abyss were no longer good.  It wasn't a hard and fast rule - it was a general rule of thumb.  Check it out becuase I don't remember exactly what I said:
http://www.starcitygames.com/php/news/expandnews.php?Article=5227

My actual rule was:
Principle #8: For the most part, Spells with a casting cost of four or more are unplayable unless they are Blue, Artifact, or have an Alternate Casting Cost.  I woudl dispute some of the other cards on your list as ever being "cast," or playable.

Steve

[Eastman]

In my article last summer on the 10 principle of vintage, I wrote that no non blue or artifact spell that costs more than 3 is playable except Rector.

This just came off as funny, because a good deal of the time Rector got Yawgmoth's Bargain.  I'll grant respite here because it's a rare occasion that Bargain ever gets 'cast', but I'd offer a few more examples:

Exalted Angel (Maybe not applicable last summer.)
Shattering Pulse/Allay (More applicable last summer than now.)
Tendrils of Agony
Nether Void
Mind Twist
The Abyss
Moat
Humility
Goblin Ringleader
Siege-Gang Commander


I don't think the ACC is really an issue.  If you look at the list of casting costs, the 4RR entry is not going to be so misleading given the number or R, 1R and 2R costs here.  It's blatantly obvious that all the curves are low; even with Blue's distortion at 3UU, the 1 and 2 casting cost spells far outweigh it.  I seriously doubt anyone reading Dr. Sylvan's articles is going to look seriously at the Blue rankings and think "Gee, I can build a mono-blue deck with a bunch of 5 casting cost stuff like Mawcor and Air Elemental!  Cause Dr. Sylvan said so!"

I also said Bargain was an exception.  But Angel is generally morphed, Mind Twist uses Drain and is often cast for less then 4, SeigeGang comes into play with Lackey and one of the points of my article was that Moat and Abyss were no longer good.  It wasn't a hard and fast rule - it was a general rule of thumb.  Check it out becuase I don't remember exactly what I said:
http://www.starcitygames.com/php/news/expandnews.php?Article=5227

My actual rule was:
Principle #8: For the most part, Spells with a casting cost of four or more are unplayable unless they are Blue, Artifact, or have an Alternate Casting Cost.  I woudl dispute some of the other cards on your list as ever being "cast," or playable.

Steve

Steve this isn't directed just at you, I realize you're  reacting to someone elses comments and much of that article was certainly VERY useful to a lot of people.

but..

how is any of this discussion relevant? What is the relevance of how we describe the casting cost of FoW? What is the relevance of how many non-blue/non-artifact 4+ casting cost non-x spells (what a requirement lol) were played between Labor Day and Easter? Why should this site's THOUSANDS of readers care?

We need to keep in mind whether what we are writing can then be used by a player to build or play a deck.

[jpmeyer]

That sort of stuff is more for R&D

[Smmenen]

It's very relevant.  People need to know what's good and what's not.  Drawing upon established knowledge is what magic writing and sites like this are all about.

A key part of that is seeing what the average blue spell cc is becuase it reflects on the viability of unused cards that someone may have discovered and the evaluation of new cards.  The CC of Force of Will distorts that average and decreases the utility of the data - data I think is very practically important.

In fact, the work Dr. Sylvan is doing is probably the most important Type One analysis on the web - which is why I even card.  I'm not criticizing Dr. Sylvan - he's doing a great job - I just want his data to be most practically useful.

Steve

[Eastman]

It's very relevant.  People need to know what's good and what's not.

A key part of that is seeing what the average blue spell cc is becuase it reflects on the viability of unused cards that someone may have discovered and the evaluation of new cards.

I disagree. Numerical knowledge of how many X were played is useless without appropriate context. You need to know the meta to interpret these kinds of figures to any benefit. Anyone who knows the meta and is capable of that analysis has an intrinsic knowledge of how many cards of each type were played.

And again, even your argument, should your supposition be correct (and again, I disagree on this point), does not suggest any relevance for knowing how many cards from each set were played.

[Dr. Sylvan]

Much of my statistical analysis is, like JP says, most beneficial for R&D. My stuff is generally not going to help someone build a deck except in a general way (though maybe my Keeper article could guide someone making card choices). The casting cost averages might specifically let us keep an eye on Mana Drain's influence. You never know when we're gonna have solid proof that Raph Caron is right, after all. :)

The set analysis is partly to disspell old myths about the format's staleness, and mostly to inform R&D. If they know which sets had a lot of cards appear (and they simultaneously can see which cards those were to what extent), they know what kind of cards will be helpful to us in the future. (With greater precision even then JP's survey article a few months back.)

So no, I'm not the man to come to for tech, though I appreciate Steve's praise. My data has an inherent delay anyway, and doesn't even reflect the kind of idealized metagame that Steve and JP can talk about thanks to their Meandeck testing. What I most importantly do is focus peoples' thoughts to emerging trends, which will hopefully make it easier to reduce the randomness at the next tournament any of us goes to.

[Smmenen]

It's very relevant.  People need to know what's good and what's not.

A key part of that is seeing what the average blue spell cc is becuase it reflects on the viability of unused cards that someone may have discovered and the evaluation of new cards.

I disagree. Numerical knowledge of how many X were played is useless without appropriate context. You need to know the meta to interpret these kinds of figures to any benefit. Anyone who knows the meta and is capable of that analysis has an intrinsic knowledge of how many cards of each type were played.

And again, even your argument, should your supposition be correct (and again, I disagree on this point), does not suggest any relevance for knowing how many cards from each set were played.

I don't agree.  Your sentence after "I disagree" assumes that cards are only measurable in a context. (hence why you say that it is useless).
  This is not true.  Some cards are objectively powerful and some are not.  These lines need to be drawn and are best determined through this sort of analysis.  Keep in mind we aren't counting how many were played - we are counting only cards that were in decks that did well.  

I have no idea how the idea of how many cards for each set was involved in my comment - that doesn't concern me. The point is that Force of Will should be counted as a 0 cc spell becuase 96%+ of the time, that is how it is cast.  

You are relatively new to type one - but you have to understand that the day of BD and even later, people would claim that almost any cards was good in Type One.  These rules are needed for newbs to build better decks on the first try.  Additionally, they provide princples to evaluate new cards that are printed - a rebuttable presumption, but at least a rule to work with.  

These sort of determinations and delineations are what interests me in Type One becuase it goes against the Type One grain - the idea that people can play whatever they want - becuase they cant.

Steve

[MixedkNuts]

The problem here is that you run into a situation where you start wiggling on all sorts of cards.  Magic is filled with conditionals, and stats aren't a big fan of conditionals when working with large, varied swathes of data.  I'm confortable with the idea that Force/Foil/Thwart/Gush/Misdirection all cost 0 for the most part because the conditions for those cards are almost always met, but when you start bringing things like Exalted Angel and Siege-Gang into the mix I get a bit uncomfortable.  

If the Angel costs "3" to cast, then she technically costs 7 to make useful, so is her casting cost 3 with an Echo of 2WW, or is it just 4WW?  Siege-Gang most certainly costs 3RR when there's no Goblin Lackey around, and that's a condition that is met less frequently.  Do all artifacts start costing 2 mana less in decks that run Workshop?  Did all artifacts in an unbanned Tinker environment technically cost 2U?

Obviously some of these examples are more out there than others, but every one of them is contextual and could legitimately be applied in certain situations.  I mention this because most of the time, leaving the numbers "clean" and getting into the dirty stuff in the analysis is the best way to go.

--Knut

[Eastman]

If the Angel costs "3" to cast, then she technically costs 7 to make useful, so is her casting cost 3 with an Echo of 2WW, or is it just 4WW? Siege-Gang most certainly costs 3RR when there's no Goblin Lackey around, and that's a condition that is met less frequently. Do all artifacts start costing 2 mana less in decks that run Workshop? Did all artifacts in an unbanned Tinker environment technically cost 2U?

This is what i'm talking about when I say we're discussing things that are not relevant to anyone playing/building a deck.

[MuzzonoAmi]

It's very relevant.  People need to know what's good and what's not.

A key part of that is seeing what the average blue spell cc is becuase it reflects on the viability of unused cards that someone may have discovered and the evaluation of new cards.

I disagree. Numerical knowledge of how many X were played is useless without appropriate context. You need to know the meta to interpret these kinds of figures to any benefit. Anyone who knows the meta and is capable of that analysis has an intrinsic knowledge of how many cards of each type were played.

And again, even your argument, should your supposition be correct (and again, I disagree on this point), does not suggest any relevance for knowing how many cards from each set were played.

I don't agree.  Your sentence after "I disagree" assumes that cards are only measurable in a context. (hence why you say that it is useless).
  This is not true.  Some cards are objectively powerful and some are not.  These lines need to be drawn and are best determined through this sort of analysis.  Keep in mind we aren't counting how many were played - we are counting only cards that were in decks that did well.  

I have no idea how the idea of how many cards for each set was involved in my comment - that doesn't concern me. The point is that Force of Will should be counted as a 0 cc spell becuase 96%+ of the time, that is how it is cast.  

You are relatively new to type one - but you have to understand that the day of BD and even later, people would claim that almost any cards was good in Type One.  These rules are needed for newbs to build better decks on the first try.  Additionally, they provide princples to evaluate new cards that are printed - a rebuttable presumption, but at least a rule to work with.  

These sort of determinations and delineations are what interests me in Type One becuase it goes against the Type One grain - the idea that people can play whatever they want - becuase they cant.

Steve

Unless you've actually sat down and calculated that 96%, you can't go around making those claims based soley on your experience. As Dr. Sylvan pointed out,decks vary in their pitch habits, and for many relatively blue-light decks, FoW is hardcast significantly more often than it would be in a heavly Blue deck like Hulk.  I agree that cards like Squee and Worldgorger Dragon shjould be excluded from the count because they are only cast in the most uncommon and desperate situations.

I think it would be interesting to see how FoW's casting cost changes as a function of game length. I'm not saying that Phil should do it, but it would help sort this problem out. You could apply Resource Point Theory in conjunction with examining game state.

[Zherbus]

Ugh, I cringed when I saw Smmenen throw out 96%. He really should have forseen someone get anal about it instead of see his point that it is 0cc most of the time (not 96% - I see that as an exageration).

[MuzzonoAmi]

I realize that it's pitched far more often than it is hardcast, but in my experience the desparity between the two ways of playing Force hasn't been anywhere near that great.