Metagame analysis

[Nekrataal]

I am providing regular metagame analysis for a german website ( http://www.zkforum.de/showthread.php...193#post567193 ) on basis of DC and manual gathered data from german tourney websites. As you know the DC data is quite broad but lacks some good representation of american tourneys (data spread problem). The approach of results generation is powered by a database engine (MySQL) and quite flexible. It covers the MTGTheSource approach of computing DTB/DTW. Despite the flaws in raw data spread I think it is worth to post the result here of nearly 60 tourneys worldwide from May and June. The result tables are imported to Excel to generate some grafics and make it look nicer. The figures on each slice are the actual MTGTheSource scores each deck has gained.

Here it comes:



The computation does not take into account any limits on tourney size. So if you limit it to tourneys > 33 players it results in the following donut:



List of tourneys considered:

08.05.2010 DIMA Game 1K
02.05.2010 StarCityGames.com $5,000 Legacy Open
09.05.2010 Goblin Bologna Legacy
08.05.2010 Magickeller Hannover Legacy-Liga 05/2010
08.05.2010 ASL Berlin Legacy 08.05.2010
08.05.2010 Legacy Kiel 08/05
16.05.2010 Liga Valenciana de Legacy II - Torneo 8
15.05.2010 1º Regional Legacy Vitória
08.05.2010 Lliga Catalana de Legacy 5 - Maig - La MÃ*quina del Temps
08.05.2010 Legacy Nürnberg
03.05.2010 Beijing KaDou Weekly Legacy
09.05.2010 F2F Legacy May serie
23.05.2010 LwL #4
08.05.2010 Czech Vintage and Legacy Series 2010 - III
23.05.2010 Liga de Legacy de Cantabria I - Torneo 3
15.05.2010 Nenndorfer Magic Treff - Legacy Turnier 05/2010
15.05.2010 Bazaar of Moxen IV - Legacy Event
08.05.2010 5nto Torneo LLES4
15.05.2010 ASL Berlin 15.05.2010
29.05.2010 Legacy Mannheim 05/10
16.05.2010 Naoniscon 2010
08.05.2010 II Albacete's Legacy League - May
08.05.2010 Nürnberg Legacy Mai/10
22.05.2010 The Great 2nd TFF LEGACY
05.06.2010 ASL Berlin Legacy 05.06.2010
15.05.2010 8º Torneo LEL
06.06.2010 Summer Legacy at Goblin Bologna
08.06.2010 Magic-League Legacy Trial 08/06/10
15.05.2010 LubuLe Ludwigsburger Lagacy 05/10
06.06.2010 II Albacete's Legacy League - June
05.06.2010 FUNtainment Nürnberg Legacy
30.05.2010 Dragons League Verona
13.05.2010 Iserlohn Magic Treff 05
15.06.2010 Magic-League Legacy Trial 15/06/10
12.06.2010 Nürnberg Legacy 12.06.2010
29.05.2010 GPT Generic 2010 - San Juan / Puerto Rico
10.06.2010 Grand Prix Trial - Columbus, White Cap Comics, Grand Rapids, MI
19.06.2010 Kieler Legacy 06/10
19.06.2010 ASL Berlin Legacy 19.06.2010
19.06.2010 19.06.2010 Nürnberg wöchentliches Legacy
29.05.2010 29.05.2010 Nürnberg wöchentliches Legacy
12.06.2010 Magickeller Hannover Legacy-Liga 06/2010
12.06.2010 Großes Legcay im Spieleland Hamburg
12.06.2010 LotacyJune
27.06.2010 Lotacy 27/06/10
26.06.2010 Legacy Bäckerei Müller
22.05.2010 22.5.2010 Nürnberg wöchentliches Legacy
20.05.2010 3º LSL - Elche
23.06.2010 Magic-League Legacy Trial 23/06/10
30.06.2010 Magic-League Legacy Trial 30/06/10
27.06.2010 Liga de Legacy de Cantabria I - Torneo 4
26.06.2010 Legacy Mannheim 06/10
27.06.2010 Cachan Legacy V
19.06.2010 Lliga Catalana de Legacy 3 - Juny - Draw or Play
13.06.2010 Iserlohn Magic Treff 06
30.05.2010 Trader Liga Dülmen - Legacy Mai/10
20.06.2010 Trader Liga Dülmen - Legacy Jun/10
27.06.2010 20x Duals
26.06.2010 Legacy Spellkeeper


Have Fun and Discuss!

[SpikeyMikey]

So my BB doesn't render that graph very well. Can someone tell me what the top couple of decks are??

[xTrainx]

I can't tell the different between some of the colors. For the larger pieces, it might be helpful to label them according to their deck type...

But, from what I can tell, it seems that you put quite a bit of work into it. Nice job.

[DrJones]

The difference between colors is irrelevant, they are ordered clockwise from most succesful to less succesful.

The rogue archtype is about the creature type rogue?

[rockout]

Is there a deck list for rogues?

[MMogg]

Not to be a negative Nancy, but don't you think it's kind of a waste doing analysis on a format just as it has hit a major shake-up with the Mystical banning? With two of the top 8 decks neutered, surely it will have a ripple effect and shake up that meta.

[menace13]

Sorry I am apparently Color Blind and could you please post numbers and words as my spectrum of shades are entirely off. Good work and thank you for actually doing something so extensive, very informative.

[Nekrataal]

The difference between colors is irrelevant, they are ordered clockwise from most succesful to less succesful.

The rogue archtype is about the creature type rogue?

Exactly. They are sorted this way so counting does the trick. Or would you appreciate a text list? Rogue is a category where every deck that does not fit to any of the usual categories (Goblin, Zoo etc.) falls into (see deckcheck).

Not to be a negative Nancy, but don't you think it's kind of a waste doing analysis on a format just as it has hit a major shake-up with the Mystical banning? With two of the top 8 decks neutered, surely it will have a ripple effect and shake up that meta.

No it is just the opposite. Especially because the meta will shift it is interersting to have an analysis BEFORE nad then AFTER ;) This analysis is the BEFORE (May/June). It is not too much work for me since everything is in a database. Well, of course it took some programming skill at first but that part is done.

[MMogg]

No it is just the opposite. Especially because the meta will shift it is interersting to have an analysis BEFORE nad then AFTER ;) This analysis is the BEFORE (May/June). It is not too much work for me since everything is in a database. Well, of course it took some programming skill at first but that part is done.

I see. It's a bit of an academic exercise though. Think about it: is anyone really interested in meta analysis from 2007?

Edit: in case that doesn't make sense, what I mean is outdated meta analysis is just that, outdated. Information has a short shelf life in MTG.

Anyway, I'm not trying to be so negative. Really cool work and I look forward to the post-Columbus era stats.

[Nekrataal]

But why is it outdated? The data goes up to end of July. Just because of the Mystical Tutor ban? The impact on the metagame could already be immanent because ppl adapt to the ban much earlier than actually at or right after the ban. So Reanimator is already at the end of DTW sections and will continue to fall. Iam sure about that. And ANT will still prosper because it can compensate the loss. So I do not see too much impact on the overall metagame on a large scale except maybe a slight aggro boost. And how will you be able to evaluate any changes if you do not have a baseline to compare to? I understand your argument that there will be change after the ban but I still believe that having any up to date information about the here and now is valuable. Anyway I can post the development in one month for June/July and we will see how much the change is in numbers ;)

[EssKay]

Rogue is a category where every deck that does not fit to any of the usual categories (Goblin, Zoo etc.) falls into (see deckcheck).

How is that different from Misc?

[keys]

How is that different from Misc?

That's what I was wondering.

[Srovex]

Perhaps Misc is bunch of established deck types that don't fit in any other category (like Dream Halls Combo, Angel Stompy, Madness etc.) and Rogue is real rogue decks?

[overpowered]

I'm going to assume that "Misc" differs from "[DTW] Rogue" since the "Misc" could mean anything from "b/g/u merfolk thopter painter" to "GW Survival" or other things not on the "established list of [DTW]"

Just a guess.

[Nekrataal]

Yeah sorry I didn't explain. When you do that statistics very often it becomes just too natural.

The further you follow the donut clockwise the weaker the appearance (points accumulated) of a certain deck archtype gets. E.g, Goblins could accumulate 207 points (3rd slice) whereas e.g. New Horizon could only manage to get 105 points. If the amount of points falls under a certain threshold the slices get to small to be either relevant or displayable. So everything below that threshold just gets subsumed under "Misc" meaning "All other not further distinguished". These are decks like e.g. Enchantress or White Stax which are categorized in principal but where their weak appearance doesn't make them relevant enough. They are all subsumed under "Misc" = miscellaneous.

It has nothing to do with the "Rogue" category which subsumes so called "rogue decks" that are not related to any known deck archetype but individual deck assemblies. You can check the Rogue deck section over at DC. Here is an exmaple of such a rogue deck: http://www.deckcheck.net/deck.php?id=36894 . Actually it is a Rebel deck. However since Rebels neither is a Tier1, 2 or 3 legacy valid strategy and usually isn't seen anywhere in the Top8 this deck is classified as Rogue deck. So if you bring a Ravnica pre-constructed deck to a tourney and happen to win with it the deck would be rendered "rogue" deck.
Side note: Lately you can find a significant amount of decks in the rogue category that try to get Emrakul into play with Polymorph, Eureka or Sneak Attack action. Please note that I check the rogue decks manually after adding them to my database because from mpov the categorization done over at DC sometimes is not that precise that means sometimes decks are called "rogue decks" which actually are variants of standard deck archetypes. But this is just details ....
Side Note 2: From mpov it is not a coincidence that the number of rogue decks is significant high when a shift in the metagame can be expected (Mystical Ban).

So if you have more questions or requests regarding data sets don't hesitate to ask.

[Gandalf_The_White]

Please note that I check the rogue decks manually after adding them to my database because from mpov the categorization done over at DC sometimes is not that precise that means sometimes decks are called "rogue decks" which actually are variants of standard deck archetypes.

For algorithmic deck classification, you could check out the following thread:
http://www.themanadrain.com/index.php?topic=40691.0
Might save you some time looking at lists manually.

[Nekrataal]

Hey this was a fun read. Nice to see that other guys are into it as myself :D However I am not sure it completely solves the problem of rogue decks. Rogue decks are made up by several sub categories:

a) New decks like Aeon Bridge that have not been categorized and therefore are classified "rogue"
b) Decks that could be categorized but are not because they are not Legacy staples (Rebel decks)
c) Random, fun or test deck from Johnny Casual (like Emrakul / Polymorph or Sneak Attack decks) => this is what rogue decks are
d) Decks that could be categorized into a standard category but have some odd choices so they were put into Rogue category
e) Decks that have been categorized by mistake into Rogue and actually belong to a standard category

I think the algorithm helps with e) and eventually could help with a) (when a path for the new deck is configured) but I am not so sure about d). Anyway it triggered an idea to implement a check into my program to see if the category for a given deck is correct. That is much easier to implement because you already have a category and you just have to check if this is correct. In case of a NO of course manual intervention is needed again. Currently this is not too much work for me to correct wrong categorizations (maybe 10 Minutes per week).