[Article] Playing in the Post-Goblin Era [Archive]

jwk

08-02-2006, 02:30 AM

This article was extremely well done :cool: ; I just wish they had a little (ok a whole bunch :smile: ) more details, such as:

Each of the match up analysis from both sides. ie If you are goblins and playing rifter down a game what are you looking for in your opening 7, what are good sideboard cards for this matchup, what should you considering having in your sideboard for his match up. And the same for the rifter side.

I'm extremely interested in how the Deadguy vs Threshold matchup went. You would think with all that discard and land hate and SO FEW lands in Threshold that some stratagy could be very effective aggaist them (whether to attack the lands or the hand or something).

Also something that was missing I felt was what some of them deck need for playskill. Goblins is the most played deck not only because It will beat up on the RANDOM decks but it is very easy to pilot. Loam-Confinement I've heard is well PhD level hard to play without errors. Did that affect testing and how so? It definitely affects what some people take to tourneys.

Since there were so many decks in the “random” category did that affect Goblins ranking or do all these deck perform the same against the field?

I was also disappointed that Iggy Pop was not done in this one but I do understand. God only knows how many hundreds of hours these guys both put into this article. But Iggy Pop seems to me a deck that people are considering playing so how does it far agaist these decks :confused:

Finally How do you beat Threshold :confused:

Over the course of time you guys can post (or maybe write articles for starcity) on some more in depth of these findings. Overall REALLY GOOD JOB
In another couple of Months maybe these two will come up with more analysis with decks like Dredgeatog, Iggy Pop, Affinity fair against these decks.

Also Anwar your article: Are you Playing a threat? Should be a MUST READ before anyone posts on this forum. IMO (definitely not humble though) I feel that is BEST legacy analysis article ever written as of this time. It is well presented and thought out with very good examples.

Mad Zur

08-02-2006, 03:01 AM

Why no Iggy Pop? Weren't there like 8 Iggy decks over the two days at the Duel for Duals? No, there were two on each day. However, this article was written before that happened, and the data they had to use included no Iggy Pop.

Sure, it didn't T8 or anything, but it showed up and is probably the 2nd best combo in the meta. I think the fact that you left it out skewed the data toward decks that get wrecked by combo (Survival and Loam). Any decks that showed up in multiples at both relevant events were included. It is not particularly useful to test against decks that show up as one-ofs or not at all, even if they are very good. Time is better spent confirming the data regarding the common matchups. By leaving out all unique decks (except those that can be viewed in a group like Rifter/Wombat, Sligh/Burn, and RGSA/RecSur), the data is indeed skewed in favor of decks that do poorly against decks in the "random" category. This includes metagame decks that are largely unprepared for portions of the field because they are designed around the popular decks, and may include decks that are unprepared for combo specifically. This was mentioned in the article. For reference, here is a list of what is in that "random" group for each event:

Duel for Duals 2 (linked because some names are vague; alphabetized because it was easier to get the links):

2 Landstill (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15743) (#2 (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15807))
1 3C Confinement Slide (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15783)
1 5/3 (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15784)
1 Affinity (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15669)
1 Belcher (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15805)
1 Elves (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15801)
1 Enchantress (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15655)
1 Life (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15776)
1 MBC (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15780)
1 Nausea (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15819)
1 Teen Titans (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15803)
1 Three Color Control (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15822)
1 UB Aggro (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15810)
1 White Stax (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15736)
1 WW (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15671)
1 Zoo (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15668)

Edit: Corrected error regarding Loam decks.

Total: 17 Random

Dual Land Draft 2 (unfortunately there are only T8 lists (http://mtgthesource.com/forums/showpost.php?p=67947&postcount=239)):

2 FlameVault (obviously irrelevant)
2 Salvager Game
2 Zilla Stompy
1 Affinity w/ Berserk
1 Train Wreck
1 Zombie Deck Wins
1 Scepter Chant
1 Angel Stax
1 Landstill
1 BR Aggro
1 Faerie Stompy
1 Garv.Dec
1 Curse of Dimir
1 Sensei Sensei
1 Death Cloud
1 GW Control

Total: 19 Random

So what decks does omitting the "random" section help? I have no idea, but there's the relevant information if you want to try to figure it out.

Machinus

08-02-2006, 04:45 AM

Thank you two for the article. Obviously I disagree with your Goblins numbers, but it was a good analysis.

Lukas Preuss

08-02-2006, 05:45 AM

One quick question, maybe I'm too stupid to see the answer right away, but:
If you look at the Breakout of Major Decks at the DLD, you listed 1 Solidarity, 1 Rifter/Wombat, 1 Deadguy/Suicide, 2 Angel Stompy, 2 Loam Control, 2 Gro, and 1 Random deck for the T8. That's a total of ten decks for the T8. That can't be right. Am I missing something?

Other than that, I can only say: thumbs up... nice article. When's the next one coming? :)

Zilla

08-02-2006, 07:19 AM

Great article guys. It's nice to see a solid breakdown of decktypes and how they relate to the rest of the field. I would have liked to see a sampling from a greater number of tournaments. Not necessarily with the question of "which deck is the best" in mind, but just to determine which decktypes are being played the most and what percentage of those played are making top 8 for all the biggest tournaments in the last, say, 6 months would be really really nice, and would provide us a tool for more accurately judging which decks belong in the LMF at the moment.

I have to say that I was taken aback by the data's apparent indication that Goblins is the 8th best decktype in the format. That comes as rather a large surprise.

I also don't understand why the Survival archetype appears to be the 4th best deck in the format according to the final chart, and yet it placed 0 out of 7 (decks played) in the Top 8. I'm not good at math, so maybe I'm missing something.

It's also kind of a surprise that the apparent second best deck in the format is an ill-defined entity known as Loam Control.

I'd also like more info on what defines "Loam Control", especially because Loam Confinement doesn't fit this category, if I'm understanding Zur's post correctly. What's Loam Control then? Something like Land Ho?

Bongo

08-02-2006, 07:29 AM

Very good and informative article, thank you!

1) What kind of build do you have in mind when you mention Loam Control?
I have played around with a few Loam builds with different colors, and the matchups vary a lot.

2) Could you elaborate a bit on the "randomness" factor?
I have found that Goblins and Deadguy, the two major archetypes that ranked poorly in the Calculated Strength section, do very well against randomness.

Also, Goblins and Deadguy have the "swinginess-factor" as I call it. Even against supposedly bad matchups you can score easy wins with good hands. Example: Multiple Sinkholes from Deadguy/first turn Lackey from Goblins.

3) While it is normal that people disagree with your matchup-assessments and there are different personal experiences, there is one archetype where my experience is so different from yours that it doesn't fall under the "personal taste" category. The archetype is Survival.
Both RGSA and RecSur have major problems against Goblins, UGR Gro and Angel Stompy in my testing.
I'd really like to hear how you got such positive matchups for Survival.

Again, thanks for the awesome article! The article has some very promising loose ends (like the randomness factor, play difficulty, how different builds affect the matchups, etc.) that would be very interesting to expand upon, and I hope you can do that!

tivadar

08-02-2006, 08:20 AM

scrumdogg

08-02-2006, 10:48 AM

Interesting, I agree almost entirely with what Bongo was saying. Though I think lumping "Loam Control" makes some sense, we should specify a build (is it running solitary confinement especially).

I also don't know what crack you're on about Survival (can I have some?). I haven't tested its gobbo matchup, though I'd expect it to be bad. I know that it has a bad matchup against AS. At best pre-board it's 50/50 against AS, postboard, especially if they have pithings, it gets much worse.

Anyways, I do appreciate and agree with most of your match analysis. It's nice to have it all in one table. Keep up the good work.

EDIT: Oh, and how does Gro beat rifter? Is this really true?

Properly built modern SotF builds should be much better prepared for both Goblins & Angel Stompy specifically & Pithing Needle in general. It requires a heavier commitment main & side to artifact hate, but fortunately Tin-Street Hooligan is an extremely efficient critter (which also helps in the goblin matchup). Also, Survival needs to be running either Burning Wish (which allows you access to Pyroclasm & Hull Breach) or Enlightened Tutor - presuming you're not in a field full of Predict... (which gives you Seal of Cleansing & possibly Confinement & Crypt maindeck as options as well as a host of silver bullet goodness). Splashing white gives you STP, Hierarch, and Mystic Enforcer main with Chant/Abeyance & Rule of Law side and you always have Eternal Witness + Genesis. Modern SotF has zero excuse to not be prepared for aggro & aggro control or Needle, the downfall, as always, is the vulnerability to combo. Burning Wish Survival is pretty much an autoloss Game 1 and doesn't get much better 2 & 3. E-Tutor or white SotF have a few more options, but you have to have them packed and then be prepared for some aggressive mulliganning (and it still might not work....) especially since the 2 best combo decks in the format work in completely different ways.

AnwarA101

08-02-2006, 12:41 PM

I want to thank everyone for the compliments on the article. Brian did a great job on pulling all the data together as well writing a great article (I helped as well!). The VA group did a great job of helping us test. You guys know who you are.

On to the many questions -

Why no Iggy Pop? Weren't there like 8 Iggy decks over the two days at the Duel for Duals? Sure, it didn't T8 or anything, but it showed up and is probably the 2nd best combo in the meta. I think the fact that you left it out skewed the data toward decks that get wrecked by combo (Survival and Loam). Also, I would be curious what kind of Loam we're talking about.

This article was written before the Dual for Duels that happened on July 8 and July 9. The two tournaments that we used in our analysis did not have any Iggy Pop and that is why it is not mentioned.

Since there were so many decks in the “random” category did that affect Goblins ranking or do all these deck perform the same against the field?

The random category did not affect the rankings of any deck. It was simply excluded as part of the strength of deck calculation because the random category could not be given relative matchups because it would not be feasible to give these decks a win percentage against the rest of the field. There is a paragraph about this in the article itself.

One quick question, maybe I'm too stupid to see the answer right away, but:
If you look at the Breakout of Major Decks at the DLD, you listed 1 Solidarity, 1 Rifter/Wombat, 1 Deadguy/Suicide, 2 Angel Stompy, 2 Loam Control, 2 Gro, and 1 Random deck for the T8. That's a total of ten decks for the T8. That can't be right. Am I missing something?

This must have been a typo when the tables were being converted to HTML. The 2 for Loam Control is incorrect. It did not make Top8 at the Dual Land Draft. Hopefully I can get someone at StarCity to update this.

I have to say that I was taken aback by the data's apparent indication that Goblins is the 8th best decktype in the format. That comes as rather a large surprise.

I also don't understand why the Survival archetype appears to be the 4th best deck in the format according to the final chart, and yet it placed 0 out of 7 (decks played) in the Top 8. I'm not good at math, so maybe I'm missing something.

I believe confusion stems from what Brian and I are saying. Goblins is the 7th best deck if a tournament field looks like Table 3 (Expected Field at a Legacy Tournament). In a field like that Goblins is not a very good choice and in fact a fairly bad choice of deck. This doesn't make Goblins a poor deck it just means that the decks that are seeing play are highly ready to deal with Goblins and in that tournament Goblins will not do overwhelmingly well. If you need some confirmation of this look at Day 1 of Dual for Duels. 14 Goblin decks saw play but only 1 made top8. I haven't done any analysis of that day, but you can see how Goblins can easily not be the best deck to play at a given tournament. We were also somewhat suprised that Goblins was so far down on the list, but we went where the numbers took us.

Survival's 4th place spot is also suprising, but its lack of play explains its poor top8 showing at most tournaments. It also has glaring weakness against combo (mostly Solidarity) that hurt it in the metagame. Survival doesn't have 15-20 people playing it like Goblins in every tournament. If it did its top8 showing would probably be better than Goblins if the field looks similar to what we described in Table 3 (Expected Field at a Legacy Tournament).

I'd also like more info on what defines "Loam Control", especially because Loam Confinement doesn't fit this category, if I'm understanding Zur's post correctly. What's Loam Control then? Something like Land Ho?

Loam Control was a bit difficult to put together. Since there was no established list its matchups were an average of a few decklists that were similar. This is similar for what was done for Rifter/Wombat. The decklists were either Devastating Dreams based (Red removal) or Confiment with counterspells (Blue). These were the decks that saw play at the tournaments that we used in our analysis. You can take issue with how we grouped these and claim they are very different, but we found that their matchups were not in our testing. This may not be perfect, but Loam based decks are fairly new and unestablished and this is the best we thought we could do with what was out there.

While it is normal that people disagree with your matchup-assessments and there are different personal experiences, there is one archetype where my experience is so different from yours that it doesn't fall under the "personal taste" category. The archetype is Survival.
Both RGSA and RecSur have major problems against Goblins, UGR Gro and Angel Stompy in my testing.
I'd really like to hear how you got such positive matchups for Survival.

Survival decks by their nature have very different configurations, but by and large our testing was with RGSA. We found that against Goblins especially without Goblin King that RGSA was quite even with the deck. We have it at 55% currently. The Gro matchup was very even. I tested that myself and it was very close. Now alot of our testing was with UGW Gro because that has been more popular lately, but I don't see how Red could make the matchup much better? The burn can be used as a finisher so that might matter. Angel Stompy seemed pretty close as well. If you were able to equip a pro-red guy that could connect multiple times then the game was over, but an early active Survival usually spelled bad news for Angel Stompy especially if Baloths can provide lifegain while still having multiple creatures to beat in with. I wouldn't put it much beyond 50/50. But again like we stated card choices in either of these decks can simply swing the matchup (Goblin King is a great example of this).

Mad Zur

08-02-2006, 01:10 PM

Loam Control was a bit difficult to put together. Since there was no established list its matchups were an average of a few decklists that were similar. This is similar for what was done for Rifter/Wombat. The decklists were either Devastating Dreams based (Red removal) or Confiment with counterspells (Blue). I may have been mistaken, then. I see 3 decks (Diddly DAL (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15641), Loam Confinement (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15677), 3C Confinement Slide (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15783)) that can be described as Loam Control at the Duel for Duals, but the article says there were 2. Which one was omitted?

AnwarA101

08-02-2006, 02:44 PM

I may have been mistaken, then. I see 3 decks (Diddly DAL (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15641), Loam Confinement (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15677), 3C Confinement Slide (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15783)) that can be described as Loam Control at the Duel for Duals, but the article says there were 2. Which one was omitted?

I'm sure that we used Diddly Dal and Loam Confinement in our testing. I'm not positive in the 3C Confinement Slide. Brian was handling most of the testing for the Loam Control decks. He hasn't had a chance to get online, but maybe he'll to tell us if 3c Confinement Slide deck was tested as well. The article doesn't state what was contained in Loam Control - where do you see that?

Also Anwar your article: Are you Playing a threat? Should be a MUST READ before anyone posts on this forum. IMO (definitely not humble though) I feel that is BEST legacy analysis article ever written as of this time. It is well presented and thought out with very good examples.

Your compliments are appreciated. I'm glad you found it helpful.

jwk

08-03-2006, 02:48 AM

The random category did not affect the rankings of any deck. It was simply excluded as part of the strength of deck calculation because the random category could not be given relative matchups because it would not be feasible to give these decks a win percentage against the rest of the field. There is a paragraph about this in the article itself.

I guess I didn't ask the right question. I'll try and state it the right way.

Since at any given Legacy tourny if 25% of the decks are random and let's say theorically that the win percentages are as follows for:
Goblins 75%
Deadguy Ale 75%
Threshold 70%
Soldarity 55%
Rifter 55%
Burn 45%
Loam-Confinement 45%
RGSA 40%
Angel stompy 35%

against the overall "Random" decks.

Your article gives the projected outcomes if only those 9 major arcitypes show up and ONLY those nine (your article does a darn good job at that too). But Since 25% of the field or so won't use these decks and theoricatically Goblins and Deadguy are the best against those thats should be a major condsideration to what you take to a large tourney :confused:

However this article may be a great thing for Goblin players. As with Affinity in Extended (see Pro Tour: Columbus :eek: ) There are/were some many GREAT hate cards/decks and as more and testing was done people were like why play Affinity if it can be so hated out. No one prepares for it, the hate gets taken away and BOOM Affinity is again the best deck. I don't see that happening with Goblins in Legacy but you never know :cool:

Thanks again Anwar

Obfuscate Freely

08-03-2006, 09:55 AM

I guess I didn't ask the right question. I'll try and state it the right way.

Since at any given Legacy tourny if 25% of the decks are random and let's say theorically that the win percentages are as follows for:
Goblins 75%
Deadguy Ale 75%
Threshold 70%
Soldarity 55%
Rifter 55%
Burn 45%
Loam-Confinement 45%
RGSA 40%
Angel stompy 35%

against the overall "Random" decks.

Your article gives the projected outcomes if only those 9 major arcitypes show up and ONLY those nine (your article does a darn good job at that too). But Since 25% of the field or so won't use these decks and theoricatically Goblins and Deadguy are the best against those thats should be a major condsideration to what you take to a large tourney :confused:
There is too much variation in the "random" deck category to assume that any of the prevalent decks have better "random" matchups than the others.

Thanks to Zur, we can look at the random decks from these two tournaments:

Duel for Duals 2:

2 Landstill (#2)
1 3C Confinement Slide
1 5/3
1 Affinity
1 Belcher
1 Elves
1 Enchantress
1 Life
1 MBC
1 Nausea
1 Teen Titans
1 Three Color Control
1 UB Aggro
1 White Stax
1 WW
1 Zoo

Total: 17 Random

Dual Land Draft 2:

2 FlameVault (obviously irrelevant)
2 Salvager Game
2 Zilla Stompy
1 Affinity w/ Berserk
1 Train Wreck
1 Zombie Deck Wins
1 Scepter Chant
1 Angel Stax
1 Landstill
1 BR Aggro
1 Faerie Stompy
1 Garv.Dec
1 Curse of Dimir
1 Sensei Sensei
1 Death Cloud
1 GW Control

Total: 19 Random
I doubt very much that any one of the prevalant decks has a significant edge over the others against such a varied field, but it wouldn't be very meaningful if it did. Look at how different the two lists are. The bolded decks appear at both tournaments, but everything else is unique. There are 29 different archetypes listed there, between just two tournaments!

About 25% of any Legacy field will be comprised of an arbitrary selection of decks that may include, but will certainly not be limited to, anything in either of these lists, which means that a quarter of every Legacy environment is completely unpredictable.

The bottom line here is that it borders very closely on both impossibility and utter uselessness for anyone to try to prepare for this quarter of the field. It is therefore also silly to suggest that any deck has a better "matchup" against this unknown field than any other. That is why it was left out of the analysis, and that is why it should not affect your choice of deck going into a tournament.

AnwarA101

08-03-2006, 11:27 AM

cupajoe

08-03-2006, 11:48 AM

I'm doubting that Goblins has a 75 percent win chance against random

As a random player myself, I know the decks I design are designed to have a good chance against Goblins (not designed to specifically hate Goblins, mind you, but if your decks folds to goblins, it probably also folds to other aggro decks, which is a Bad Thing in the Legacy environment)

I would say Goblins has about a 95 percent win percentage against Bad Random, those players who show up completely unprepared, don't read up on the good decks, and play with some 8-year-old pet deck that they haven't tuned since 1999.....

I would say Bad Random has a slightly better chance to pull off a win against some of the non-Goblins decks, maybe a 25 percent chance to sneak out a win.....

But probably less than half of the Random players are Bad Random players, making it not very relevant.....

M.Maddox

08-03-2006, 11:54 AM

One has to question, with everyone reading that article, how useful it would actually be in winning.

Anusien

08-03-2006, 11:57 AM

Whit3 Ghost

08-03-2006, 12:00 PM

Because people aren't going OMG Lackey is t3h brokezors.
Goblins also isn't the undisputed best deck in the format anymore.
It's still a very good deck, but the format is finally catching up.

EDIT- Great article, very very nice.

dre4m

08-03-2006, 12:29 PM

One has to question, with everyone reading that article, how useful it would actually be in winning.
One also has to question the wisdom of 'everyone' being able to read this article or otherwise unable to obtain the information therein, and a great many people may simply not have looked at Star City that day, so I believe that the impact you refer to will be miniscule, for the compitive players who would take the time to read such an article are a hideous minority compared to the 80% of Legacy players who are just out to have a good time and couldn't give a bear's ass.

tivadar

08-03-2006, 01:23 PM

I imagine the Loam Control heading is only Eric Darland's RBG contraption. It was originally headed under "Loam Control" at the Duel For Duals, and the classifying of Confinement Slide under "Random" only reinforces this.

I really liked the article, and I encourage you all to keep up the excellent writing. My only question is, how can you call this the Post-Goblin Era?

I understand people calling it the post-goblin area. The fact is that goblins dominated the format for a long time, with no real competition. Sure, other decks made a bit of headway (solidarity, landstill), but there was no question which was on top. Now that has definetly changed (along with the meta in my area at least). Gobbo and Thresh are clearly very close in terms of the "best" deck, and both these do very different things. Solidarity is picking up steam again, and there are a lot of up-and-coming new and potentially good decks thanks to things like dredge, loam, and lions eye diamond.

With the meta in such a flux, decks like rifter start to look significantly worse, as these decks sought to capitalize on the dominance of goblins. So now, we can no longer look at "how do we beat goblins, and, oh yeah, we may want to do good in other matchups". Now it's more a "how do we do good across the board, and, oh yeah, how do we beat goblins".

M.Maddox

08-03-2006, 02:36 PM

One also has to question the wisdom of 'everyone' being able to read this article or otherwise unable to obtain the information therein, and a great many people may simply not have looked at Star City that day, so I believe that the impact you refer to will be miniscule, for the compitive players who would take the time to read such an article are a hideous minority compared to the 80% of Legacy players who are just out to have a good time and couldn't give a bear's ass.

Good point.

URABAHN

08-03-2006, 04:29 PM

EDIT: Oh, and how does Gro beat rifter? Is this really true?

I don't think so and told my colleagues as such, but then again

Obviously these percentages aren't written in stone, and we have found that they can vary wildly depending on playskill. If our tested results disagree with your results it could be the result of minor card choices, a wider difference in play skill between the pilots of the two decks, or a simple case of having too small a test sample. We have tried to get the most accurate numbers possible, but we realize that card choices, tactics, and boarding strategies are constantly changing and the Matchup Matrix can be easily adjusted to follow these trends.

Volt

08-03-2006, 04:43 PM

Gro most certainly does not have a positive matchup vs. Rifter. Ask Bardo. I really have to question some of those matchup results.

SuckerPunch

08-03-2006, 05:54 PM

Thank you for that truly excellent article.

My only question about it is the deadguy matchup. You grouped it with Sui Black when it is a strategically far more aggressive and less disruptive deck. So I'm curious what build was used, a standard pikula list, or some weird hybrid.

Also, as a fellow person that plays Deadguy Ale I'm curious how many Deadguy players agree with the calculation that Deadguy is the seventh worst of the most commonly played decks, and how many think this is clearly untrue and the deck is still very strong against most of the popular archeatypes?

Obfuscate Freely

08-03-2006, 06:37 PM

Gro most certainly does not have a positive matchup vs. Rifter. Ask Bardo. I really have to question some of those matchup results.
Wait, they posted something having to do with Gro without asking Bardo if he agreed with it first? For shame! :rolleyes:

Seriously, what sort of matchup do you think Rifter has against Gro? The data Brian and Anwar used has an awful lot of testing behind it for you to simply dismiss it.

What other matchup results do you have to question?

Volt

08-03-2006, 06:53 PM

Perhaps I was overly harsh. It's a great article, despite whatever quibbles I (and others) may have with some of the published figures.

Regarding Gro vs. Rifter, I'm not the first person in this thread to question the published results regarding that particular matchup. Personally, I have only playtested the matchup a little bit. Bardo did tell me just a few weeks ago that he has been finding the matchup to be a poor one for Gro. We playtested the matchup a couple games, and it went very badly for Gro both times. Admittedly, that's a very small sample. I'm relying primarily on Bardo's experience here, as he does have a ton of experience with Gro.

Regarding other matchup results, I know that any decent Sligh deck will have a positive matchup against goblins. Perhaps the problem there is that they are lumping Sligh and Burn together. I would have thought that Burn has a good matchup vs. goblins as well, but I cannot speak from experience there.

SuckerPunch

08-03-2006, 07:44 PM

9 land Stompy always has had a good matchup against burn and sligh. And I've always seen Goblins to be akin to 9 land stompy in a lot of ways. So i'm not sure if burn does fare well against goblins or any fast weenie rush decks.

Keep in mind though that much of my experience with burn came before Flamebreak. So perhaps that turned things around.

Zilla

08-03-2006, 07:49 PM

Wait, they posted something having to do with Gro without asking Bardo if he agreed with it first? For shame! :rolleyes:
All due respect to Ewok and Anwar, but I'd be willing to bet cash money that Bardo has put in more hours of playtesting with UGw Thresh against the field than the two of them combined. It doesn't seem far-fetched to use people with firm knowledge on a particular subject as a point of reference.

PTBNL

08-03-2006, 08:02 PM

burn has a positive matchup vs goblins, about 70/30. any decent draw with flamebreak is an auto win.

Machinus

08-03-2006, 08:41 PM

All due respect to Ewok and Anwar, but I'd be willing to bet cash money that Bardo has put in more hours of playtesting with UGw Thresh against the field than the two of them combined. It doesn't seem far-fetched to use people with firm knowledge on a particular subject as a point of reference.

I think thats a reasonable assumption. What's the problem with giving credit to someone who published way more articles on the deck than anyone else?

Obfuscate Freely

08-03-2006, 09:46 PM

I think thats a reasonable assumption. What's the problem with giving credit to someone who published way more articles on the deck than anyone else?
That's pretty terrible logic. Getting articles published on the internet is hardly proof of expertise.

All due respect to Ewok and Anwar, but I'd be willing to bet cash money that Bardo has put in more hours of playtesting with UGw Thresh against the field than the two of them combined. It doesn't seem far-fetched to use people with firm knowledge on a particular subject as a point of reference.
It was mentioned in the article that the testing data used in the analysis comes from testing that we do as a group here in NoVa. The Gro results came about largely from testing with Mad Zur, overlord95, and myself. All three of us have actually been playing Gro in Legacy since before CavernNinja brought it to the world's attention, so I think we deserve a little more credit for our work.

Machinus

08-03-2006, 09:51 PM

That's pretty terrible logic. Getting articles published on the internet is hardly proof of expertise.

Bardo's expertise is self-evident. If you had read his articles, you wouldn't be so inclined to deny him the credit that he is due.

Volt

08-03-2006, 10:03 PM

That's pretty terrible logic. Getting articles published on the internet is hardly proof of expertise.

I hope you see the irony in that statement.

AnwarA101

08-03-2006, 10:29 PM

Bardo's expertise is self-evident. If you had read his articles, you wouldn't be so inclined to deny him the credit that he is due.

Bardo's expertise with deck doesn't preclude the expertise of other players such as Mad Zur, ObFreely, or Overlord95. Our testing results come largely from these players, but I'm not willing to concede that my playing skills are so bad that any testing I may have done with deck are immediatedly invalid. I believe Brian is a skilled player who can play Gro with a great deal of skill as well.

I'm not denying Bardo's expertise, but you can't easily dismiss the skill of others.

As a side note, I have yet to see Bardo post any complaints about our numbers.

bigbear102

08-03-2006, 10:32 PM

I hope you see the irony in that statement.

Quoted for.... just cuz i wanted to, and it's true.

Anyway, I do believe that Bardo is probably one of the best people to look to when talking about gro, and I would compare him to the Virginians who have played the deck almost as much. I do believe that it was the Virginians plan to have this article published so near to the DLD and GenCon that everyone would play Gro, and noone would play Deadguy or Goblins or solidarity, because they have conspired with IBA to overwhelm the tourney with some sort of amalgamation of black cards.

Either way it was an awesome article....and I'm still palying Wombat Saturday.

Machinus

08-04-2006, 12:04 AM

Bardo's expertise with deck doesn't preclude the expertise of other players such as Mad Zur, ObFreely, or Overlord95. Our testing results come largely from these players, but I'm not willing to concede that my playing skills are so bad that any testing I may have done with deck are immediatedly invalid. I believe Brian is a skilled player who can play Gro with a great deal of skill as well.

I'm not denying Bardo's expertise, but you can't easily dismiss the skill of others.

As a side note, I have yet to see Bardo post any complaints about our numbers.

Obviously a deck can have many competent pilots, and certainly that is true about Threshold. However, good players don't necessarily know anything about design. When there are disagreements about the deck, the only way to resolve them is through experience.

In my experience, Bardo plays the deck very well, but he has also put a huge amount of time into developing it. It is appropriate to take advantage of his knowledge. His ideas about deck design are strongly supported by tournament data, even those about which there is disagreement.

I have no contentions about the UGW vs. WR matchup, by the way. If anything, I am skeptical of the low Goblins percentages.

AnwarA101

08-04-2006, 12:11 AM

Obviously a deck can have many competent pilots, and certainly that is true about Threshold. However, good players don't necessarily know anything about design. When there are disagreements about the deck, the only way to resolve them is through experience.

In my experience, Bardo plays the deck very well, but he has also put a huge amount of time into developing it. It is appropriate to take advantage of his knowledge. His ideas about deck design are strongly supported by tournament data, even those about which there is disagreement.

I have no contentions about the UGW vs. WR matchup, by the way. If anything, I am skeptical of the low Goblins percentages.

I am willing to say with confidence that Mad Zur, ObFreely, and Overlord95 have a had a great deal with the development of Gro in modern Legacy. I'm not saying that I would be against using the knowledge Bardo has, but I was relying on people I feel have been instrumental in designing the modern version of the deck. Bardo has alot of knowledge about the deck, but so do these players. In a perfect world I would have checked with Bardo as well as all the other players that have had a hand in designing the modern version.

Zilla

08-04-2006, 12:23 AM

It was mentioned in the article that the testing data used in the analysis comes from testing that we do as a group here in NoVa. The Gro results came about largely from testing with Mad Zur, overlord95, and myself. All three of us have actually been playing Gro in Legacy since before CavernNinja brought it to the world's attention, so I think we deserve a little more credit for our work.
Actually, Bardo's been working on Gro in Legacy since before CN popularized it as well... he originally discussed it over at TMD primarily as opposed to here. In any case, I wasn't suggesting that he was the end-all be-all authority on the archetype. I just don't think it's correct to dimiss the notion of asking his opinion on its matchups as somehow preposterous.

Given that the entire foundation of this article is based upon assumptions about how the various decks interact with one another, it would seem to behoove the authors to make every effort to reach as accurate a consensus regarding matchup percentages as possible. Failing to do so runs the risk of invalidating the majority of the conclusions reached therein.

Mad Zur

08-04-2006, 04:00 AM

Actually, Bardo's been working on Gro in Legacy since before CN popularized it as well... he originally discussed it over at TMD primarily as opposed to here. Nothing said so far contradicts this.

In any case, I wasn't suggesting that he was the end-all be-all authority on the archetype. I just don't think it's correct to dimiss the notion of asking his opinion on its matchups as somehow preposterous.Actually, it was the notion of trusting his opinion over their own testing, which is indeed preposterous.

Given that the entire foundation of this article is based upon assumptions about how the various decks interact with one another, it would seem to behoove the authors to make every effort to reach as accurate a consensus regarding matchup percentages as possible. Failing to do so runs the risk of invalidating the majority of the conclusions reached therein.Consensus does not affect data and should never get in the way of presenting it.

Machinus

08-04-2006, 04:43 AM

Bane of the Living

08-04-2006, 07:33 AM

Any thoughts on a follow up article guys? You certainly did a good job with this one but it would be nice to see one with newer data, updated tournament standings. Your goblin data does seem a bit low, especially for how well the archetype has been able to fight through all the hate aimed at it.

tivadar

08-04-2006, 08:36 AM

I don't think anyone is trying to say that your matchup analysis is horrible, one way or the other, only that certain matchups may be incorrect. Keep in mind this can happen very easily if you only test 10-20 games/matchups, as luck can play a big role in things. Also, you've been talking about how the people playing Gro were competent Gro players, but have said nothing about who was playing the Rifter deck or how competent they were playing it. Since Rifter seems to be losing, that's probably the more important question...

quicksilver

08-04-2006, 09:17 AM

SpatulaOfTheAges

08-04-2006, 09:42 AM

The person playing the rifter deck was competant, in fact all people involved in the testing alalysis of these matches were very competant magic players and almost certainly above the caliber of the average player you would meet in a tournement.

Which, note, is probably more practical for predicting tournement data than playtesting between two people who have been playing the same deck and fine-tuning it for years.

URABAHN

08-04-2006, 12:43 PM

Actually, Bardo's been working on Gro in Legacy since before CN popularized it as well... he originally discussed it over at TMD primarily as opposed to here. In any case, I wasn't suggesting that he was the end-all be-all authority on the archetype. I just don't think it's correct to dimiss the notion of asking his opinion on its matchups as somehow preposterous.

It was mentioned in the article that the testing data used in the analysis comes from testing that we do as a group here in NoVa. The Gro results came about largely from testing with Mad Zur, overlord95, and myself. All three of us have actually been playing Gro in Legacy since before CavernNinja brought it to the world's attention, so I think we deserve a little more credit for our work.

WEST COAST! Inglewood! Er, OREGON!
EAST COAST! V to the izz A!

Didn't the East Coast vs. West Coast thing die out in the mid-90s?

AnwarA101

08-07-2006, 02:17 AM

Consensus is the only meaningful way of interpreting data, however. It is not significant on its own.

@Anwar and Brian: How do you account for the success of Goblins if it has such a low ranking in your system?

I can only say that we have to disagree about the success of Goblins. Look at the two tournaments we looked at in the article. Duel for Duals had 19.40% of the field was Goblins that day. It had 2 people in Top8. Solidarity had 3 players and 1 was in Top8. Rifter and Angel Stompy only had 4 players each and each had 1 player in Top8. That is what is to be expected for a good deck, but hardly dominating. If Goblins was the deck of Legacy why isn't it dominating the Top8s if it is seeing so much play? Our testing and analysis lead to the conclusion that perhaps Goblins just isn't the strongest deck in the field.

Dual Land Draft 10.30% of the field was Goblins and none of them made Top8. By comparison Angel Stompy only had 3 players, but 2 of them made Top8. Gro was 12.10% and it also put 2 players into the Top8. Why didn't Goblins do better?

Machinus

08-07-2006, 03:40 AM

If you want to use tournment data to analyze the format, your conclusions will only be useful if they are drawn from statistically significant patterns.

Observing that Goblins underperformed in a few tournaments may be interesting, but it takes a large number of tournaments for this to provide significant information. Dedicated Legacy players are tempted to treat it as if it were a real format that developed on a regular basis, and in which predictions could be made. This is unfortunately not the case. Legacy is very slow to evolve, and if anything it takes more data than other formats to come to meaningful conclusions due to metagame fragmentation and much less rigorous design.

A hope intimated in your article (and one that I expressed in my last Stax article, back in January) is that the format is moving beyond Goblins. I want to believe this as much as anyone else, but I don't think it is true yet.

Goblins can often comprise a larger fraction of the field than any other deck, but fail to match this presence in the top8. I question the significance of this statistic. Goblins has been so successful in the past year that it would take another year of no top8s at all for there to be equal amounts of data. Or, to suggest a different argument - if you do believe that Legacy has evolved, etc., how do you account for day 2 of the D4D, where Goblins comprised 50% of the top8?

It is important to maintain objectivity in format analysis. I want Legacy to be diverse and healthy as much as anyone else. If it's even possible for us to effect that change, we have to be honest about what is really happening in the format.

EDIT: I just want to add that whether or not Goblins is the "best" deck in the format really depends on what factors one considers important. My answer to that obvious question depends completely on the decks it is compared to, and the type of environments they comprise. The complicated truth is that Goblins has created or influenced heavily the other decks in the format, so attempting to answer this question seriously should only make one realize that it is impossible to define "Legacy," and that ones perspective on Goblins is derived largely from local experience.

frogboy

08-07-2006, 04:15 AM

Sideboard games were considered when possible when determining the win percentage

Sideboarded games are more important than game ones when determining matchup percentage.

Can we see the raw data (win/loss record) and brief summary of games that were used to create the matrix?

Edit:

If Goblins was the deck of Legacy why isn't it dominating the Top8s if it is seeing so much play?

This reasoning would lead one to conclude that Affinity was not the dominant deck of Mirrodin Block Constructed season until the very end of the season.

Edit two: Can we see the lists used for testing?

jwk

08-07-2006, 11:43 AM

This reasoning would lead one to conclude that Affinity was not the dominant deck of Mirrodin Block Constructed season until the very end of the season.

Very good point.
Some of the decks that were tested are designed to beat up on A) red decks in general (Angel Stompy plays main deck Sliver Knight over the "Pump/Jump Knight") B) Creature Decks (ie Rifter)

if Goblins wasn't so Good these decks wouldn't be as good in the matrix. Goblins dominance is what makes decks like Angel Stompy and Rifter very compatitive. Goblin numbers go down then well Angel Stompy's win % goes down as more Thres and Combo players show up. Then Deadguy Ale's Goodmatch ups accour more often.

ps
Also what a said before about a theoriactical Goblins "75% win percatage" against "random" decks is not true (well I don't think it is). I was just suggesting that since Goblin's game plan hardly changes with each matchup that it can effectivily play each match up after turn one or two after its oppentent plays their first land. Where as Decks like Angel Stompy and Rifter might have completely different plans for each match up (and/or colors used). just saying. :smile:

Also I am no way shape or form stating that the "Random" decks can be tested/predicted/prepared for. They are Random! :eek: That's part of the fun of legacy is that at a given tourny 1/4 of the field is playing something completely Rouge and every once in a while it does well :cool:

Anusien

08-07-2006, 03:50 PM

The main problem is that I think the new Duel For Duals invalidates a lot of conclusions made about Goblins being down for the count.

dre4m

08-07-2006, 04:09 PM

The main problem is that I think the new Duel For Duals invalidates a lot of conclusions made about Goblins being down for the count.
Day two definately shot that down, yes, but day one either spoke for the success of the ever-present "Random" category, or the resiliance that certain rogue decks (consistant ones) can have against a predetermined metagame. That and excellent matchups a la Affinity.

AnwarA101

08-07-2006, 04:43 PM

If you want to use tournment data to analyze the format, your conclusions will only be useful if they are drawn from statistically significant patterns.

We stated that our data was somewhat limited in the article. Legacy is a format with limited tournaments and we addressed that in the article. That does not mean that limited data should prevent any analysis.

Observing that Goblins underperformed in a few tournaments may be interesting, but it takes a large number of tournaments for this to provide significant information. Dedicated Legacy players are tempted to treat it as if it was a real format that developed on a regular basis, and in which predictions can be made. This is unfortunately not the case. Legacy is very slow to evolve, and if anything it takes more data than other formats to come to meaningful conclusions due to metagame fragmentation and much less rigorous design.

I would hate to quote the article but the opening paragraph talks about how determining the best deck in a format like Legacy is difficult because of the limited tournament data.

A hope intimated in your article (and one that I expressed in my last Stax article, back in January) is that the format is moving beyond Goblins. I want to believe this as much as anyone else, but I don't think it is true yet.

Goblins can often comprise a larger fraction of the field than any other deck, but fail to match this presence in the top8. I question the significance of this statistic. Goblins has been so successful in the past year that it would take another year of no top8s at all for there to be equal amounts of data. Or, to suggest a different argument - if you do believe that Legacy has evolved, etc., how do you account for day 2 of the D4D, where Goblins comprised 50% of the top8?

I want to be clear. Our analysis did not say that Goblins was the 7th best deck in Legacy. It said that in the current metagame goblins was only 7th best. These are two completely different statements. If the metagame changed then Goblins position in that metagame would change. Its about its relative strength in the metagame not is objective power. I'll even agree that Goblins is a strong deck and can easily make Top8, but in the metagame similar to the 2 tournaments we used this won't be the case.

As far as I know no one has listed a breakdown of Day 2 of the Duel for Duals.
From my brief scanning of the decklists I believe Solidarity was the most played deck. If that is true then the metagame of Day 2 did not look like the metagame at the 2 tournaments we analyzed and thus Goblins relative strength could have shifted and probably did. Goblins won't always be the 7th strongest deck in a metagame. It could shift all the way to 1st depending what is played in any given tournament. But there is good reason to believe that metagames don't change overnight and so some analysis is most likely useful.

In no way did we want to imply that Goblins is an outdated deck. The name of the article to some degree represents the suprise that our analysis yielded.

It is important to maintain objectivity in format analysis. I want Legacy to be diverse and healthy as much as anyone else. If it's even possible for us to effect that change, we have to be honest about what is really happening in the format.

Where were we not objective?

Bardo

08-07-2006, 05:50 PM

As a side note, I have yet to see Bardo post any complaints about our numbers.

I thought the article was fine, but I kind of gloss over tables and charts whenever I'm reading a Magic article. And since it's obvious that SCG articles aren't some peer-reviewed piece that you might find in the New England Journal of Medicine, I try not to be too critical.

If I were to be a little more critical, I would focus on the methodology of the article, and not dwell on any particular set of data.

Like frogboy mentions:

How many games were pre- and post-board? (I'd say you need at least ten games pre-board, and twenty games post-board to have enough data to reach meaningful conclusions.)

What lists were you using? (U/G/r Threshold, vs. U/G/w has, a much better rating against goblins, for instance.)

What was your position on theoretical 'perfect' play? (i.e. did you allow take-backs)

And once I saw that you'd lumped all flavors of "Gro" into a single item, I really didn't focus on your results. But now that I look at them more closely, I think your percentages for Goblins, Deadguy, and Rifter are overly optimistic.

But, really, whatever; it was a fine article and I'm happy to see any decent Legacy content on a big site like SCG.

Bardo's expertise is self-evident. If you had read his articles, you wouldn't be so inclined to deny him the credit that he is due.

Motherfucker, please. I'm just some guy who plays the deck a lot. But thanks. :)

SuckerPunch

08-07-2006, 06:26 PM

Actually, more significant that grouping the various Thres variants together, was the grouping of Deadguy with Red Death and SuiBlack...

Like I said...

The latter two are a strategically far more aggressive and less disruptive deck. So I'm curious what build was used, a standard pikula list, or some weird hybrid.

It's possible to bring Pikula closer to traditional sui by running 4 Negators, but the strategy and the matchups change significantly just by adding that one card.

So at the very least, it should made clear which listed/variant was used for the testing purposes as your results for Deadguy are far less positive than the experiences of the people who play the deck a significant amount.

Other than that though, the article was truly excellent. Thak you for taking the time to put it together.

P.S. Grouping together Sligh with Burn was very significant too, as burn in my experience is definatly positive versus Thres variants not running Worship/Counterbalance.

Bardo

08-08-2006, 02:16 PM

Week
of
August 4th

$50: Brian Diefendorf and Anwar Ahmad's Playing in the Post-Goblin Era - Analyzing the Legacy Metagame

This excellent analysis of the Legacy metagame is impressive in both scope and execution. As it's co-authored by two writers, each wins $25!

Congratulations, fellas!

Awesome. Congrats guys!

Man, I wish he'd run my article. :(

SuckerPunch

08-09-2006, 01:54 PM

can you guys please post your test builds. your results aren't at all meaningful considering you grouped together so many very different decks int the same category (like grouping sligh with burn, red death with pikula, thres with u/g/r thres, ugw thres, ugb thres and ugw thres with worship).

But if you posted the decklists used, the results you laid out would be very meaningful, for people running those specific builds that you tested with.

Not dissing the article, it was good, but it left off the most important part, what builds you actually used in your testing.

The 4 Negators that deadguy doesn't run but that red death and sui do change every aspect of the deck, from it's speed and goals, to it's vulnerability to all sorts of things.

Same for burn and sligh. The two decks are nothing alike, the latter depends on creatures, which means it has a far far harder time against other creature decks and suddenly has unfavorable matchups against thres and the like.

The article clearly had a lot of work put into it and was well written and appreciate what you're trying to do. But without posting your builds, your conclusions and tables lose all significance.

Ewokslayer

08-09-2006, 03:50 PM

So Anwar told me about this thread, I was completely unaware of it until today at 3 thanks to me moving and no internet connection.
A few things, in no particular order and without quotes to the relevant post above.
The article was written and the numbers for the Matrix put in about a week and a half before the Back to Back Duel for Duals (Duel for Duals 3 and 4).
Since that time additionally testing have changed some numbers.
Off the top of my head:
Survival is down to 40% vs Goblins (Pretty much if Goblins continues to run Port/Waste as the land base this won't be favorable for survival, Goblin King also sucks)
Solidarity is now 35% vs Gro thanks to a better boarding plan.
With the introduction of Red Death, Suicide has been grouped with it and been separated from Deadguy.
Iggy Pop had been added previous but wasn't included in the article because no one played it before then.
As for the Goblin numbers, Yes Goblins is strong and capable of doing well. But it always under-performs for the percentage of the field in comprises with the exception of Duel for Duals 4. Which percentages are off for Goblins by a significant amount. Can you really say that it is good against Rifter or Angel Stompy?
The Loam decks used were indeed DAL, Confinement Slide, Loam Confinement, and a little of quicksilver's build from the Dual Land Draft 2.
Matchups were based on a 10 game min pre board and most of the more likely matchups have many more games played.
The Gro vs Rifter issue is most likely a sideboard thing. All colors of Gro can board in a powerful Sorcery (Echoes, Flashfires, Geddon) that trumps Rifter. Rifter is then required to board against that (Sacred Ground, Gilded Light, etc) weakening its board control elements that allow it to stop Gros creatures. Plus 55% isn't that much of a win percentage. If generally means a game 3 which in the case of Rifter v Gro doesn't actually occur in tournament play due to time.
Oh, Burn sucks. A lot. It almost never does good at big tournaments. Stop playing it. Though it does beat Red Death.
I will post some deck lists when I can get to them.

Bardo

08-09-2006, 04:18 PM

Oh, Burn sucks. A lot. It almost never does good at big tournaments. Stop playing it. Though it does beat Red Death.

Heh! This is the same point I made in my currently-unpublished-though-craig-said-by-email-twice-it-would-be-up-weeks-ago article too.

Just to underscore your point, yeah, stop playing Burn folks. Reuseable damage es muy bueno.

Ewokslayer

08-09-2006, 06:21 PM

Decklists Used for the Article

U/G/W Gro (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=16903)

Angel Stompy (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=16906)

Solidarity (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=17989)

Loam Control 1 (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15641)
Loam Control 2 (http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=15783)

Deadguy
4 Sinkhole
4 Vindicate
4 Confidant
4 Hyppie
4 Hymn
3 Shade
2 Verdict
2 Swords
2 Scroll
2 plague
4 ritual
3 Duress
4 Scrubland
4 Delta
1 Tainted Field
1 Tomb of Urami
4 Wasteland
8 Swamps
Sideboard
4 Wretch
2 Swords
2 Verdict
2 Plague
3 Edict
2 Perish

Rifter ("http://sales.starcitygames.com/deckdatabase/displaydeck.php?DeckID=16910)
As well as
4 Swords
3 Dragon
2 Pyroclasm
2 Starstorm
2 Wrath
2 Rune
3 Decree
3 Renewed Faith
2 Vengence
2 Abeyance
2 Orim's Thunder
3 Humility
3 Slice and Dice
3 Rift
9 Plains
5 Mountains
4 Plateau
3 Steppe
3 Cave
Board
2 Boil
2 Pulse of the Fields
3 Rule of Law
2 Abeyance
2 REB
2 Pyroblast
2 Pyroclasm

Machinus

08-09-2006, 07:42 PM

Heh! This is the same point I made in my currently-published-though-craig-said-by-email-twice-it-would-be-up-weeks-ago article too.

I submitted my report over 4 weeks ago. I have been told regularly that it would be going up "tomorrow" or something to that effect. I have articles lined up, but this is riduculous.

SuckerPunch

08-09-2006, 07:45 PM

Thank you, that's a big help.

Atleast we can rest assure that the suiblack/pikula list you used didn't run negators.

Interesting choice in cutting a duress, and not running any jotun grunts/withered wretches but running 2 verdicts by the way, but the list looks very solid otherwise.

Machinus

08-09-2006, 08:22 PM

But it always under-performs for the percentage of the field in comprises...

How is this relevant to anything?

I know this measurement is sometimes used to characterize performance. I fail to see how it applies to anything useful in Legacy.

If players want to construct useful theories about this format, they need to accept some basic premises about it. One of them is that Legacy is poorly defined and insufficiently developed. As I discussed in my TLE (http://www.themanadrain.com/index.php?topic=25217.0) thread, this means you cannot use tools that are meant for use in more structured formats. Many other factors play a more dominant role in determining tournament success than field fraction; in fact I challenge that it has any importance at all in Legacy. (In my opinion, there has been only one tournament where these tools have been applicable.)

Machinus

08-09-2006, 08:53 PM

We stated that our data was somewhat limited in the article. Legacy is a format with limited tournaments and we addressed that in the article. That does not mean that limited data should prevent any analysis.

I would hate to quote the article but the opening paragraph talks about how determining the best deck in a format like Legacy is difficult because of the limited tournament data.

Obviously I think we should attempt to analyze the format as well. What I disagree with you on is how much we can really determine from the very limited data that we have - and I think our data is severely limited in quality, not just quantity. Legacy is unlike any other format, despite it's apparent similarities to some. I agree that it's something we should pursue, but we have to be prepared to accept inconclusive evidence.

I want to be clear. Our analysis did not say that Goblins was the 7th best deck in Legacy. It said that in the current metagame goblins was only 7th best. These are two completely different statements. If the metagame changed then Goblins position in that metagame would change. Its about its relative strength in the metagame not is objective power. I'll even agree that Goblins is a strong deck and can easily make Top8, but in the metagame similar to the 2 tournaments we used this won't be the case...

If that is true then the metagame of Day 2 did not look like the metagame at the 2 tournaments we analyzed and thus Goblins relative strength could have shifted and probably did. Goblins won't always be the 7th strongest deck in a metagame. It could shift all the way to 1st depending what is played in any given tournament. But there is good reason to believe that metagames don't change overnight and so some analysis is most likely useful.

In no way did we want to imply that Goblins is an outdated deck. The name of the article to some degree represents the suprise that our analysis yielded.

I'm actually not concerned about how Goblins is characterized in Legacy. It is the obvious central force for deck design in the format, so there is really no way to accurately describe it's place in the format with a simple ranking. However, I am very concerned with how players talk about the format - about the format's voices being clear about what Legacy is, and is not. Enjoying this format makes it more difficult to acknowledge it's shortcomings, but they are significant and it can only be destructive for us to ignore them.

Where were we not objective?

Yes, as I said, I liked the article. However, I think we should be more careful about making bold claims about what the metagame is or is not just based on testing. Besides being limited from the perspective of tournament analysis, there is a large amount of uncertainty in the testing process - skill, experience, design, and randomness all contribute to this. Furthermore, I am skeptical about how deep a consistent understanding of Legacy can even be.

Please don't be misled by my criticisms. I would like to see more examination of the format from this perspective. I just believe that Legacy requires a new set of ideas to understand it correctly.

Ewokslayer

08-09-2006, 09:13 PM

How is this relevant to anything?
It isn't really except to defend the matchups for Goblins. People see Goblins making several Top 8s and assume that it is still the best deck in the format. Our analysis suggests that the only reason it is still making consistant Top 8 appearances is the large numbers it has at each tournament. While it isn't the best deck in the format it is still a format defining deck in that the Goblin matchup must be considered when playing in a tournament. That will probably change as its presence decreases slightly as Gro and Solidarity gain in popularity. Unfortunately, the percentage of decks in the "Random" category remains relatively unchanged.

Obviously I think we should attempt to analyze the format as well. What I disagree with you on is how much we can really determine from the very limited data that we have - and I think our data is severely limited in quality, not just quantity. Legacy is unlike any other format, despite it's apparent similarities to some. I agree that it's something we should pursue, but we have to be prepared to accept inconclusive evidence.
I agree that any analysis will have a lot of "wiggle room" as a result of metagame shifts, luck, card selection, and the large presence of just random crap at tournaments, which generally have at least one good matchup between Gro and Goblins and all seem to suck against Solidarity.

AnwarA101

08-09-2006, 09:56 PM

However, I am very concerned with how players talk about the format - about the format's voices being clear about what Legacy is, and is not. Enjoying this format makes it more difficult to acknowledge it's shortcomings, but they are significant and it can only be destructive for us to ignore them.

What exactly are your concerns? What shortcomings? You are going to have to be more specific to make it clear what it is you are talking about.

We merely took a couple tournament results and analyzed the major decks and then tested them against each other to see what came out ahead. Sure our testing could be flawed as well the dynamics could change based on card choices. We never intended our results to be something perfect and unchanging, but that hardly makes our results invalid. I don't see the harm in doing a reasonable analysis such as ours.

Yes, as I said, I liked the article. However, I think we should be more careful about making bold claims about what the metagame is or is not just based on testing. Besides being limited from the perspective of tournament analysis, there is a large amount of uncertainty in the testing process - skill, experience, design, and randomness all contribute to this. Furthermore, I am skeptical about how deep a consistent understanding of Legacy can even be.

Please don't be misled by my criticisms. I would like to see more examination of the format from this perspective. I just believe that Legacy requires a new set of ideas to understand it correctly.

Perhaps the name of the article is the source of your concern. In part the name of the article is a bit provocative, but it is also reflective of our results. There was no bold statement except to say that this analysis lead to this result. You are welcome to disagree with that result, but I don't see how you could necessarily think we made a bold statement. We never stated Goblins was bad only that it did not fair well in our final calculations.

Machinus

08-09-2006, 11:24 PM

What exactly are your concerns? What shortcomings? You are going to have to be more specific to make it clear what it is you are talking about.

My explanation of this turned out to be a lot more complicated than I'm willing to go into here. Don't misunderstand, I really like playing Legacy, but it is a poor format for analysis and theorization. Basically, development is very slow, tournaments are small and random, and the card pool necessitates large information samples. Together these factors make format analysis very difficult.

We merely took a couple tournament results and analyzed the major decks and then tested them against each other to see what came out ahead. Sure our testing could be flawed as well the dynamics could change based on card choices. We never intended our results to be something perfect and unchanging, but that hardly makes our results invalid. I don't see the harm in doing a reasonable analysis such as ours.

Actually I am sure your testing was fine; I don't doubt the abilities of the players involved. The anaysis was reasonable, also. I am merely suggesting that players consider the factors I mentioned above, and their implications - mainly that coming to meaningful conclusions about Legacy is more difficult than with other formats, yet we have less to work with.

Perhaps the name of the article is the source of your concern. In part the name of the article is a bit provocative, but it is also reflective of our results. There was no bold statement except to say that this analysis lead to this result. You are welcome to disagree with that result, but I don't see how you could necessarily think we made a bold statement. We never stated Goblins was bad only that it did not fair well in our final calculations.

Yes, I am a little worried about that. Directed at experienced Legacy players, it may just be a matter of opinion, but for the general magic player I do think it is misleading. As robust as your testing may be, it can't possibly support that result, even if it's a suggestion.