Jump to content

Welcome to Card Game DB
Register now to gain access to all of our features. Once registered and logged in, you will be able to create topics, post replies to existing threads, give reputation to your fellow members, get your own private messenger, post status updates, manage your profile and so much more. If you already have an account, login here - otherwise create an account for free today!
Photo

The first results are in, a quick analysis of the first 8 tournaments (derived from the Annals)

- - - - - meta analysis

  • Please log in to reply
147 replies to this topic

#1
Barnie25

Barnie25

    Advanced Member

  • Small Council
  • 589 posts

So the first real batch of tournaments are in, ranging from 44 people to tournaments with 8 people. The sample size is small but it is already beginning to show certain trends.

 

 

The summery:

 

  

                      Represented in the field         Represented in the cut

Barathion                 21%                             50%

Greyjoy                    23%                             6%

Lannister                  10%                             13%

Martell                      7%                                -

Night's Watch            9%                                -

Stark                        7%                               6%

Tyrell                        2%                               -

Targaryen                 20%                             25%

 

 

 

So what do we see?

 

  • Well, Bara is doing really good and is making the cut more than two times more often than the relative attendance suggests (21% and 50%)
  • Lannister is doing slighty better with 10% and 13%
  • Targaryen is also sees an increase in slots in the cut compared to the number of attendees
  • Stark is slighty underrepresented in the cut
  • Martell / Tyrell / Night's Watch are totally absent from the cut

And now the biggest revelation:

  • Greyjoy is even though it was the most popular only making the cut in 6% of the times (23% and 6%)

 

These numbers are pure and raw and don't take into account which decks where present at the tournaments or which banners they used. These numbers are purely based on number of players per faction and number of players per faction that made the cut.

 

As you can see Bara is doing really good with only Lanni and Targ being able to put up positive records. These conclusions shock nobody most probably. But seeing how Greyjoy seems to do quite poorly is really interesting and I am wondering why this is the case. Stark is probably underrepresented as a whole and should see an increase when more players start realizing that it can do some damage.

 

So what are your thoughts about these numbers?

 

PS why can't I upload my table from a image hosting site?

Attached Files

  • Attached File  agot.png   25.97KB   3 downloads

  • PulseGlazer likes this

#2
ooo

ooo

    Advanced Member

  • Members
  • PipPipPip
  • 526 posts
So the question: how long until the fix for Bara arrives, given the R&D lead times?

Three years? It took about that long for the Criminal/Jinteki fixes to arrive in Netrunner.
  • Arkard and Barnie25 like this

#3
Top8Gamer

Top8Gamer

    Member

  • Members
  • PipPip
  • 17 posts

Shocked to see so few Tyrell. They have a really solid package. I guess they just make for a better Banner then a main house.  



#4
szczudel

szczudel

    Advanced Member

  • Banned
  • PipPipPip
  • 370 posts

I am not a native speaker so you must excuse my question.

What does it mean in the field or in the cut?

I think I can guess but I do not want to repeat "meta" misunderstanding.



#5
istaril

istaril

    Advanced Member

  • Contributor
  • PipPipPip
  • 1741 posts

Look, I love the Annals more than the next man... they're my pet project, and they're great. But it's *way* too early to make these sorts of statements. Things to keep in mind:

  • The data is of a very small sample size, over a very short period of time, and primarily from a single geographical area (most of it is from this weekend's Spanish meta tournaments)
  • Bara Fealty is known, and aside from being a solid deck, it's also the easiest to build. Any overrepresentation is likely magnified by that, especially in these early days.
  • There is a single instance of a Lanni N/A, a deck known to have a good Bara fealty matchup, and very few of some of the other constructions my testing partners have found to do well against it.
  • Of 8 Tournaments, only 2 had cuts (the rest were swiss champions), which changes 'overrepresentation' numbers significantly.
  • Of 8 Tournaments, 2 were won by the same player... World Champion 2013, Worlds Finalist 2014. Another was won by the European Champion (2011). 3 out of four tournaments with a European/National/World Champion in attendance were won by those players. 

In looking at the Annals, the only things I think are meaningful is that Fealty is quite popular, No Agenda is not. Lanni/Bara/Targ are popular as a main faction, Tyrell is not. 

Edit for szczudel: In the field means present at tournaments. In the cut means performing well enough to be chosen for the elimination rounds (if there are any).


  • scantrell24, kizerman86, PulseGlazer and 3 others like this

#6
Barnie25

Barnie25

    Advanced Member

  • Small Council
  • 589 posts

In my defense I did put a disclaimer up saying that it was early and that the sample size was small. But it does paint an interesting picture, it is interesting to keep track of how these charts will fluctuate and how the meta will shape up. Sure Bara is the 'easy to build' deck but it is not as if there are really strong decks that haven't been discovered yet, I think that most regulars on this forum would agree on this fact but if the people who participated in these events were running the types of decks that can be succesful in the current meta? The answer is probably not all of them which in turn would skew the results as you said, that a strong player wins multiple events by having an information advantage on the rest as well as a skill advantage.

 

Again the data is very raw and shouldn't be used to make definite conclusions but to comeback to your analysis, Greyjoy was the most played deck throughout the first 8 tournaments, a deck that was also 'easy to build' or so it would seem but yet it isn't making the cut as much as other decks. This could indicate that the 'easy to build' version of Greyjoy isn't the correct one, maybe?



#7
ooo

ooo

    Advanced Member

  • Members
  • PipPipPip
  • 526 posts
Greyjoy got left behind a month or so ago. It's a way off top tier.

#8
Scottie

Scottie

    Advanced Member

  • Members
  • PipPipPip
  • 1968 posts
Yeah, you made note of it being a small sample size. But you also didn't make note of the fact that the "Cut" numbers on represent 2 of the 8 events or that they had an overlap in attendance/general locations.

So not only does it look as if you a drawing conclusions from an overly small sample size but your representation of the data is pretty questionable.

The Cut numbers are honestly useless for analysis, because they only represent two events. 1 of those events had a Top Cut of 2 people.

#9
Barnie25

Barnie25

    Advanced Member

  • Small Council
  • 589 posts

Yeah, you made note of it being a small sample size. But you also didn't make note of the fact that the "Cut" numbers on represent 2 of the 8 events or that they had an overlap in attendance/general locations.

So not only does it look as if you a drawing conclusions from an overly small sample size but your representation of the data is pretty questionable.

The Cut numbers are honestly useless for analysis, because they only represent two events.

 

Ah my bad, that is indeed a problem.. :P also explains somethings but yeah it does make most of this quite pointless  :rolleyes:



#10
istaril

istaril

    Advanced Member

  • Contributor
  • PipPipPip
  • 1741 posts

Note that in a tournament with just straight-swiss rounds, the program treats it as a top "cut" of 1. So each tournament contributes at least one datapoint to the cut data, but obviously a 20 person swiss tournament (like one of the Padis ones) isn't contribution as much data as you'd expect (it would normally have a top 4) based on total number of attendees.

So yes, our "Cut" data is a total of 16 decks, 8 from Barcelona, 2 from Tulsa, and 1 from each of the remaining tournaments. 

 

That's why I'm pretty dismissive of the overrepresentation/performance numbers, and really only looking at things like popularity (for which we have at 137 datapoints).


  • Barnie25 likes this

#11
DrewDog

DrewDog

    Advanced Member

  • Banned
  • PipPipPip
  • 189 posts

A)  There are now "regulars" on this forum for a game that most players are just starting to get their hands on and play and their posts are gospel now?.....k

B)  After 1 or 2 low level/turnout events we're supposed to extrapolate the results as a great indicator of what is good/bad Houses/cards?

 

Yeah, I don't think so.  I'd pump the breaks a little bit before starting to say which cards and factions are the best.



#12
michaelius

michaelius

    Advanced Member

  • Members
  • PipPipPip
  • 214 posts

I'll be seriously surprised if later data change anything in top 3 houses.


  • ooo likes this

#13
fozzyfosbourne

fozzyfosbourne

    Advanced Member

  • Members
  • PipPipPip
  • 169 posts

It's weird to me how popular Banner of the Rose is in that data



#14
Barnie25

Barnie25

    Advanced Member

  • Small Council
  • 589 posts

A)  There are now "regulars" on this forum for a game that most players are just starting to get their hands on and play and their posts are gospel now?.....k
B)  After 1 or 2 low level/turnout events we're supposed to extrapolate the results as a great indicator of what is good/bad Houses/cards?
 
Yeah, I don't think so.  I'd pump the breaks a little bit before starting to say which cards and factions are the best.


I didn't wanted to sound smart or rude but a lot of the people here have been playing 2.0 for about 3 months now and there has been quite a lot of testing done but no actual tournaments. And playtesters have played with the core for even longer than that. The meta isn't all figured out, especially the rock paper scissors aspect of the deck building process is becoming more important as that is how the Lannister NA deck came to be.

Note that in a tournament with just straight-swiss rounds, the program treats it as a top "cut" of 1. So each tournament contributes at least one datapoint to the cut data, but obviously a 20 person swiss tournament (like one of the Padis ones) isn't contribution as much data as you'd expect (it would normally have a top 4) based on total number of attendees.
So yes, our "Cut" data is a total of 16 decks, 8 from Barcelona, 2 from Tulsa, and 1 from each of the remaining tournaments. 
 
That's why I'm pretty dismissive of the overrepresentation/performance numbers, and really only looking at things like popularity (for which we have at 137 datapoints).


Would it be possible to change the annals to have each tournament count towards the cut, so each top 4 for each swiss tournament. It would mean that small tournaments skew the results somewhat but in overall it would probably be more representative. But thanks for the rundown.

#15
fozzyfosbourne

fozzyfosbourne

    Advanced Member

  • Members
  • PipPipPip
  • 169 posts

It looks like a 137 participant tourney in Zaragoza was entered today? 

 

e: istaril points out that I'm just bad at reading spreadsheet



#16
istaril

istaril

    Advanced Member

  • Contributor
  • PipPipPip
  • 1741 posts

Would it be possible to change the annals to have each tournament count towards the cut, so each top 4 for each swiss tournament. It would mean that small tournaments skew the results somewhat but in overall it would probably be more representative. But thanks for the rundown.

 

Possible - and even easy, but I'm not sure it would be wise. First of all, that's a loss of a lot of data for a large tournament, but if you just threshold it to a minimum of 4, you get the problems you mention with 4 person tournaments... finally, there's the aspect of people 'playing to reach the cut' where people might play different knowing there's a cut than if there isn't, especially after their first loss

(or once they've made the cut, if it's large/too large). 

 

Overall, I think using the actual cut is more meaningful (and representative) - and allows the annals to be used to track cut matches, when they actually happen. I could pull out a separate 'pseudo-cut' data of the top 15% of every field, or something - but that'd be a little bit of work, and while it might be valuable, I'm not sure if it's valuable enough to warrant the additional clutter. I could potentially be convinced, if someone can make a strong case for it.



#17
istaril

istaril

    Advanced Member

  • Contributor
  • PipPipPip
  • 1741 posts

It looks like a 137 participant tourney in Zaragoza was entered today? 

 

No, you're looking at calculation only sheet (Flattened), which runs totals from all the sheets after it when I run a script. The textual values are taken from the most recently added sheet, while the numerical ones are sums. The actual tab marked for Zaragoza is here, and was only 16 players. To simplify things, I've hidden that sheet... in the hopes that doesn't mess up the way the script treats it. We'll find out next time I run it :P.



#18
fozzyfosbourne

fozzyfosbourne

    Advanced Member

  • Members
  • PipPipPip
  • 169 posts

Oh ok thanks for clearing that up



#19
DrewDog

DrewDog

    Advanced Member

  • Banned
  • PipPipPip
  • 189 posts

whole 3 months... that doesn't mean much.  I'd give it a bit more time to really flush out the meta considering how small a window that is for determining what's the best at this current time.  As for playtesters... I don't really put too much faith into them when combined with FFG.  SWLCG has taught me that lesson.



#20
Barnie25

Barnie25

    Advanced Member

  • Small Council
  • 589 posts

Possible - and even easy, but I'm not sure it would be wise. First of all, that's a loss of a lot of data for a large tournament, but if you just threshold it to a minimum of 4, you get the problems you mention with 4 person tournaments... finally, there's the aspect of people 'playing to reach the cut' where people might play different knowing there's a cut than if there isn't, especially after their first loss
(or once they've made the cut, if it's large/too large).

Overall, I think using the actual cut is more meaningful (and representative) - and allows the annals to be used to track cut matches, when they actually happen. I could pull out a separate 'pseudo-cut' data of the top 15% of every field, or something - but that'd be a little bit of work, and while it might be valuable, I'm not sure if it's valuable enough to warrant the additional clutter. I could potentially be convinced, if someone can make a strong case for it.


In MTG they track top 8 results for every tournament but the sheer number of tournaments give more significant results, in Thrones there are not that many tournaments which makes it harder for see action patterns that's why k would imagine all the data matters. Maybe there should a standard for how many players make the 'cut' for swiss tournaments to not skew the data such as top 2 decks for 8 to 15 man tournaments and top 4 for everything above or something like it, shouldn't be to hard.