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

Research and Destroy

- - - - -

  • Please log in to reply
90 replies to this topic

#1
RegentOfTheWretched

RegentOfTheWretched

    Newbie

  • Members
  • Pip
  • 1 posts
Greetings motor sports fans.

Recently, the largest coc tournament in the LCG era took place in Liege Belgium. I managed to win this beast with a Miskatonic/Cthulhu deck. Somehow, in the last year or so, I feel like I’ve become the avatar of Miskatonic University. This year with Miskatonic/Cthulhu and last yr with my Miskatonic/Hastur deck, things just seem to be coming together for me and the University. At any rate - this is a comprehensive look at the deck and strategy that went into Euro 2013.


I can breathe a sigh of relief also. I had made the prediction (http://www.fantasyflightgames.com/edge_news.asp?eidn=4056) that Miskatonic would finally win one large LCG tournament.


Before continuing I would like to thank the rather brutal work of my test group. Roberto Carioli (as opposed to Bruno Carioli), who probably tested a million hours this year between several decks. Hastur/C which won almost all the regionals in Europe this year, and then Misk/C along with 3 other decks that didn't get played this time around. Also new comer Ray, who only showed up a few weeks ago but helped PT against Flux/Alien comboes. After last years rules "change/reversal" fiasco that left us with a deck in the garbage can and deckbuilding 10 minutes after the Euro tournament started, followed by me crafting a deck on the airplane on the way to worlds, this year we went a bit overboard on the test side. Bad experiences die hard, and we really put in quite a year this year.

This deck is a concatenation of several “Techs” that were discovered in 2011-2012, and have been bouncing around in my head for quite some time. It was one of five decks this year that were potential tournament decks, winning the tournament spot when the other 4 were hit by the late FAQ leaving only this one left.

I generally dislike the idea of posting lists, but since this has already has appeared in multiple tournaments, I will break the pattern this time.

"Research and Destroy" (R&D)

Miskatonic
3 Obsessive Insomniac
3 Matthew Alexander, Dark Companion
3 Archaeology Interns
3 Alternative Historian

3 Professor Morgan, Lean and Youngish
2 Lucas Tetlow, Eternal Curator <> Configuration Slot #1
1 Bruno Carioli, Hatching a Plan <> Configuration Slot #2
2 Coll Prospect <> Configuration Slot #2
3 Cafeteria Lady <> 3xAtlantis Configuration Slot #3

Cthulhu
3 Degenerate Serpent Cultist
3 Khopesh of the Abyss, Manifested Malice
2 Temple of R'lyeh, Frozen in Time
3 Dreamlands Fanatic
3 Uroborus, Fang of Yig
3 Naaginn

3 Deep One Assault
2 Flooded Vault


Neutral
3 Artifact of the Lost Cities
2 The Mage's Machinations


The Slots marked as Configuration are open elements for the side board. The deck went through 2 independent balancing sessions. Once for a Yithian environment and once for a rush. The slots are detailed below. The base list is the one that was played at Euro.

The deck has two primary “Tech” threads that encompasses its business plan.

1) Statistical Compression: This is encompassed mostly by its draw. The “Theory” here was that if you drew enough on average, you could reduce the raw number of cards delineated to each deck aspect. Normally in order to perform some function you are required to assign a minimum number of cards to be guaranteed to pull it off every game. I have a spreadsheet that delineates thresholds of cards drawn along with statistical “standard deviation” distributions which indicates the number of card slots I should dedicate to any specific function. For example, if I want to be able to remove 2 characters by T2 – I will need at least x removal in the deck. With compression, I can now shift the number of dedicated cards down because the number of cards I have available will be higher while maintaining the same level of performance. Since there was enough draw in the deck, I derived an expected number of cards per turn and reset the parameters to adjust for the new numbers. The deck sported a compression rate of 56% - meaning that it was conditionally drawing at a rate of 3 per 2 (rounds) vs a standard draw. For those of you familiar with some of the MTG spreadsheets that are available, this is the equivalent of dropping 2 rows vs a non “compressed” deck on the probability matrix. In real terms, taking my above example, normally I would need, say, 16 cards to guarantee my removal result. Now I may only need 12 or 13.



2) Stacking Triggers: In COC, card draw really isn’t that good. You are probably doing a double take here as I just told you that one of the primary deck functions is draw. The problem is that actions tend to be more valuable than resources in the competitive level. Most games are decided by turn 3 (even if they don’t necessarily end on that turn). And at a limit of 3 actions and 2 cards drawn, you can see that efficiency is normally more expedient that wasting actions to draw more resources. To give a concrete example, when I was building for Euro 2011, I had initially incorporated the “Artifact Of the Lost Cities” and “Khopesh of the Abyss” combination into the deck, and after a few derivations, I soon realized that the actions lost to artifact were more valuable than the cards drawn, and the Khopesh itself was really too slow for the metagame. The problem was that playing and losing 2 cards to remove 2 - 3 cards ended in a pyrrhic loss and a transfer of initiative to the opponent - except in very specific board configurations. I wrote extensively about this in an article (http://community.fantasyflightgames.com/index.php?/topic/59647-the-shivan-khopesh/) (*1). For that tournament I dropped the artifact completely, and was reduced to resourcing the Khopesh. Simply put, efficiency, is the most important quantity in COC, and cards that don’t work within the meta-game defined clock are always inferior. As a deckbuilder, one thing I take perverse delight in is when I can turn a detriment to a positive and warp the game. In this case, I had the idea that there were potentially enough cards that would trigger “for free” off the effects generated by the Khopesh that I could potentially override the games actions limitations (Turning card draw from “meh” to “wow” in the process as the card draw become synonymous with action generation) and subvert the Artifact/Khopesh into Action generators instead of action sink holes. If you peruse the list of cards you will see that 15 (and 2 free for a total of 17) of the cards can be triggered into play. The dialectic here being that if we consider the aggregate distribution of non costed chars, artifact should return a small net positive delta (2% in a vaccum, or up to 205% in a starting configuration in tandem with Khopesh, with the Khopesh also yielding a 200% ROI). Note that these numbers don’t take into account the compression rate above which now directly feeds the override, so you can see – even before playing a single hand – why the deck could be termed “explosive”. Originally I had “Forbidden Knowledge” in the deck – and in initial testing it proved to be highly unpredictable. In some games, I was able to drop 10+ chars on T2, and in other cases the deck tended to jam and crash. It also introduced some unexpected vulnerabilities to enfilade fire. Because of this susceptibility, I was forced to drop that card from the build.


An interesting scenario came up when I was working with the Epic Loot crew. They were planning to build 4 of this deck but ran out of dreamlands fanatics for the 4th deck. When they asked if it would affect the deck, I did a quick calculation. If you re-plug in the numbers – you find that Artifact of the Lost Cities flips from above to below parity on its action generation! This means that even a shift of 3 free characters means that artifact no longer “works” and must be replaced in the deck. This then caused the Compression rate to fall, and that caused issues with removal distribution (most likely, other cards would need to be dropped and a 3rd temple would be added, along with Sacrificial Offerings).


Balancing the deck was an absolutely monumental task. As you can see from the above, one to two cards changed can have large impacts other aspects of the deck. Balancing the removal was a nightmare. DOA/Machinations are actually really really ugly "hacks" that appeared late in devolpment as a result of imbalances in the deck. I haven't used DOA in years in a Cthuhlu deck and I hated to use it this time, but it became imparative do to problems that arose with distribution and the Artifact, and also with the innate clumsiness of the Khopesh. Machinations is paralell to Temple, but it was to solve distribution problems between the Temple and the Compression tech.

About those Yithians, Euro and anti teching:


There is probably a lot of questions on the Yithians and the difference between Euro and US metas. I generally build decks with a sideboard, and initially this deck had 2 cards in the sideboard delineated for Yithians. The Yithian decks published so far follow a weak combo pattern (*2).

(1 of 3) & (1 of 3) - Distribution

From a technical perspective, the odds of this happening with Mulligan in any given game is strictly less than 34%, or 34% - the distribution offset. This means that in a 7 round tourney someone playing this deck can expect to hit the lotto, on average, two to three times at most. If you assume the opponent is slow, and change the imperative to “both cards in the top 12 cards” (eg by T2), it jumps up into the 40% range, but the expected net result against several fast decks is a fast exit from the tournament. Also I will note that I’ve “cheated” a bit here. The odds are slightly higher if a second minor loop is added….it’s not much better – I think in testing it was on the order of 5-10% increment. For simplicity I have limited this to the core.

There was another Yithian build that no one got (or at least no one played) that the % could be upped to

((1 of 3) & (1 of 3) + Dist) || ((1 of 3) & (1 of 5 where N = 18))

The odds of this Yithian deck is nearly double that of the first (Topping out in the high 60% range depending on cards chosen), but in my opinion I would not play either. I should point out that last year’s “correct” Logan Infinite discard build was

(((1 of 3) & (1 of 3) )|| (1 of 3)) & (1 of 6)) & (!(3 of 3) & !(3 of 3))

This deck is exactly twice as good as the Yithians deck, with even less opposing cards that interacted with it and a clock that was nearly half that of the Yithians (including 18% chance to win T1) – and in a tourney of 24 people, both got knocked out to bad distribution within 5 rounds.

I had done a lot of testing of “R&D” vs both iterations of the Yithians. In over 20 games, the “Vanilla” build (no sideboard), ran at slightly over an 80% win rate(*3). It had several ways to beat Yithians.

1) Misk C has ALOT of Arcane. 3 or more Arcane usually locks the Yithians down. If you choose that route, it can easily drop 3-4 A Icons in the first 2 turns.

2) Temple/Khopesh means no possible defense for story phase for the Yithians and free story runs. With most of your chars at 3 cost being equipped with khopesh, you can totally avoid Pushed from Beyond.

3) Bruno/Historian mean no A struggles resolve

4) Extreme Speed


The deck plays slightly differently against the Yithians than it does regular decks. For example Tetlow becomes paramount as he is your only C + A char. The C is important as you need to commit at least 1 combat to each committed story to avoid the myths Toughness Tax (this was his primary reason for inclusion). Ditto for Bruno Carioli, as he provides I and A for 1 (Bruno also at x1 balances out T1 drop probabilities – the + 1 char at 1 cost means that you can still play 2 chars on even a -2 Standard Deviation draw (7 secondary playable cards) – at 6x in this slot you can get clipped with a -1 deviation). Dropping Tetlow and Bruno on T1 means Studying The Void is nearly useless. In testing the sideboard, I had 2-3 copies of Atlantis. In 20+ games with sideboard, Misk C never lost vs either of the 2 builds mentioned above – nor was it close to losing with this card in the deck. However the decks performance was downgraded by about 6% vs “non” Yithians decks. It wasn't clear from above, but for the US Yithian environment the Prospect is dropped for more Carioli. In a more "generic" environ for this deck the Prospect is better. Hence I ran 3x Carioli for Worlds and 1x for Euro. The question for Euro became simple - take the 6% degradation for a guaranteed win on the Yithians, or max out vs. all other decks and have an 80% chance vs. the Yithians. There are 2 factors that influenced my decision.

1) I knew there was going to be quite a few players. This translates into more rounds played. More rounds played is really bad for a deck whose strategy is to flip 2 cards in the open. The odds of a Yithian deck making it through that many rounds into the final was really low.

2) The Euro players in general are more experienced than the US players (Most of the more organized metas there are from the CCG era), and also more diverse because of the number of countries involved. Many of the decks were fast – in the top 8 more than half the decks were running clocks at 3. The best Yithian clock possible was 3.4. Compare this with Gencon where most decks clocks are basically undefined (AFAICT). Also the Euro players are removal happy. Any deck that expects to win with a specific char in play isn’t going to do well in a tourney there. On the diversity side, there will always be several decks running any specific card. Power Drains…Flux Capacitor…Snow Graves…which means you will be guaranteed facing “Bane” cards at least twice in any tourney and maybe more. In my opinion, it is these factors that lead to the disparaging Euro vs US results. I mean no disrespect here – but I stopped reading US tourney results a long time ago. I just don’t understand them (Again not meant in the pejorative sense here – I just can’t analyze them and come up with anything I can quantify). My test results tend to mirror the Euro metas test results.

In the end I chose to go without the sideboard. In the worst case scenario it would be an 80% chance in the finals which is still good. From conversations I had with the players – pretty much everyone at the tourney had come to the same conclusion. Finally, a year or so ago I read a really good MTG article that I think is applicable here (http://www.gatheringmagic.com/darwinkastle-082012-keep-it-simple-stupid/) , I almost never sideboard these days.

Conclusion for the tournament: Go with strength, and ignore the noise.


Epic Loot:

With this deck, I really wanted to do worlds this year, so it was a bit unfortunate when I found out that the 2 were less than 5 days apart. After 36 hours travel and a horrible case of jet lag, there was just no way I could turn around the next day and fly for 14 hours to Worlds. I actually planned to take a few months off from the game – when I got a call from Jim Black.

Jim was working a worlds deck, and said that several from his meta were going to worlds. I think initially he was just looking for 3rd party input to tweak what he had – but one thing led to another and I soon had become “Epic Loot: Tokyo Branch”. Several conference calls later, and 1-2 sit-ins on games and their 4 (which then became 3) players were training on this deck.

Interestingly the situation was a bit different for them, with regards to the sideboard. We knew there would be a few Yithian decks at the tournament, and they didn’t have the depth of schedule that I enjoyed at Euros. If you take a tournament like worlds you have the following:

12 players give or take
A minimum of 3-4 of those players will be casual or thematic
Most of the players will have less than 1-2 years experience / many will not have a full set of cards

Given this setup, and multiple Yithian decks, the odds at least one Yithian player will get a combination of easy schedule + “Flipping it” becomes a near certainty. When I showed the deck list, I gave it as is – the list and the sideboard swaps. I actually made 1 minor swap on the mainboard (Dropping Prospect for Carioli) for the expected Yog overload, but I let them choose individually to go with the high % or to onboard Atlantis. I think Jim Black ran with Atlantis, the other 2 decided to run without it.

I think I made an egregious error here. I had designed this deck in phases over the spring/summer (with the final revision post FAQ). Usually I complete it and then hand it off to Roberto with a list of test points. This year Roberto had decided to participate in several regionals all over Europe. (I must admit there was a bit of an agenda here, the deck I was aiming to plant at the regionals was designed to gather proof for the following spring FAQ). Roberto was so deep into testing/practicing with the deck for regionals that he didn’t look at Misk C until very late summer. When he started testing it, his initial results were wildly off from what I had produced over the summer. It was so disparate that I almost discarded the build. It took almost 3 weeks of testing and training before his results matched mine (and this disparancy was reproduced later with our new addition Ray). The real problem is the deck plays differently based on the opponent/draw. At Euro the three of us had different results. For example Ray lost vs MZI, but I think it was mostly because we didn’t have the time to show him how to change play vs that deck style. In my games, I knew how to change gears from the summers tests. In the end, both the Euro results and the level of time it took to get compatible test results that agreed, told me that the Epic Loot crew (who had about 2-3 days max to learn the deck) were going to have a rough time at the tourney. When I saw Brandon play in the finals(*4), I realized immediately that I should have main-boarded Atlantis before showing them the list and not given the choice to tech or not based solely on the complexity of the deck.


So ends another year of COC. Special thanks to Frédéric/Damien/Stavros for pulling off the largest yet COC tournament without a hitch. Indirectly, I almost got a full sweep this year. Derivations of my Misk/Hast deck won almost all the US regionals, and my Hastur/Cthulhu deck swept all the Euro regionals except for Spain where we didnt participate. Gen con and Worlds are the outliers, maybe next year....

G

Notes and addendum:

(*1) Unfortunately, a lot of people who really didn’t know what they were doing complained and the Khopesh got restricted. Sadly, I suspect that the flooded vault will suffer the same fate. In my testing it is simply too slow for most decks. I cut to x2, but the real reason it was in the deck was to plug holes where I would draw the Artifact and no Khopesh. Those hands were irrevocably dead hands and adding the +2 helped. Since it was implied that I had artifact in that situation, the net action loss for Vault was usually mitigated by the actions gained back from the trigger/draw (You get out of a bad situation with no momentum gain\loss but you clear the "Jam" out). I generally resourced it once the first Khopesh appeared and my hand dropped. In total, I think I used it a twice. I think that in normal decks it will be technically very hard to justify its use.

(*2) For those of you who may have taken classes in integrated circuits, you may recognize my shorthand as the abbreviated product of a mathmatical device known as a Karnough Map (for Advanced Calculus People "Veitch Diagram" ... and if I remember my collegiate years correctly, this is also similar to a cheat method for solving Fourier Transforms). I use these when I build or evaluate combo decks intra-tournament, as it makes evaluating them very fast – especially intra game when you need to figure out which part to remove to crash it. In COC, “1 of 3”&”1 of 3” is –in my vernacular – a “lottery ticket deck”. In less derisive language, it means that the deck should statistically fail enough times to not be able to win a tournament regardless of the opponents decks. Basically a person who plays this deck is no longer playing the game or strategizing, they are now merely “flipping” – against the odds – to win.

(*3) To the best of my knowledge, the deck went 3 for 4 (75% win rate) against Yithians at Worlds, were the loss could have been averted. Very close to my tests analysis. When I sat in on a Jeff/Jim test session, Yithians went first with 2 myths and a study and lost in 3 turns to Misk C. These results verified testing we had done earlier in the year.

(*4) Brandon actually had the game won off the draw, but made several critical mistakes between T1-T2. Years ago “Magic The Gathering” used to publish “puzzles”. You were given a board setup and a hand and you had to try and figure out how to win that turn. For those who like puzzles, you should watch the worlds video VERY closely and see if you can figure out how he could have locked the Yithians out on T2. There are at least 2 ways, and maybe a 3rd based on his resources (I couldn’t make out 2 of them).

  • Danigral, konx, Midian and 5 others like this

#2
HappyDD

HappyDD

    Advanced Member

  • Members
  • PipPipPip
  • 236 posts
Wow, thanks for the write up! I have a question about this statement:

I have a spreadsheet that delineates thresholds of cards drawn along with statistical “standard deviation” distributions which indicates the number of card slots I should dedicate to any specific function.


By "threshold" do you mean the minimum number of cards you have to draw to be able to do something (e.g., kill a target unit, kill a support, whatever). So your spreadsheet has X / 50 cards that do a certain thing, and you have a pre-established rule about when you want to be able to take that action (turn 2)? So you want to know what X should be?

Also, what is a standard deviation distribution? Do you mean you take into account the standard deviation when you are deciding what X should be? As in, you want to draw your unit kill (or whatever) card on turn 2, and if you have X kill cards that happens 60% of the time, but you want 80% and you increase X by some value?

Sounds like you've put a lot of effort into designing and testing this deck, congratulations on winning at Liege!

#3
Unnamable

Unnamable

    Advanced Member

  • Members
  • PipPipPip
  • 267 posts
I love the title you have given the deck! When Jim and I first tested the deck I wanted to name it 'The Wonderland Deck' because I felt like I had fallen down a rabbit hole! ;) We appreciated the opportunity to play the deck at Worlds and wish that we had played it better to give credit to all your work! I know it had to be excruciating to watch my first round match! The main reason I managed the W versus Tom was his horrendous draw! Another day or two to prepare would have done wonders but it couldn't be helped. Perhaps next year there can be better coordination with the two events being more spaced and a Euro contingent can make the trip?

#4
MagnusArcanis

MagnusArcanis

    Advanced Member

  • Members
  • PipPipPip
  • 331 posts
Hey Graham! Congrats on your win at Liege and a congrats to Rob, Ray and Mzi for their performance! Seriously way to go.

Solid deck and impressive anaylsis as alwasy Graham, but I can't help to feel that there are a few that need to be brought to light.

1) The reason why 'lotto ticket' is a negative term is that there is more to a deck that it's ability to draw a specific set of cards. Opponents and a player's ability to 'outsmart' said opponent to enact the deck's strategy is as, if not more, important than simply drawing the cards you need. By declaring a deck a "lotto ticket" makes it seem like the only reason a player ever won a game is because they happened to draw 'it.' Which... steps on the wrong nerves if you know what I mean.

2) While the Yithian deck is by no means perfect, but the fact that you (and others) think that it only wins by attempting to find a couple cards is just... I can't wrap my head around why theres such a gulf between how some people think Yithians work and how other people think Yithians work (especially the Shub and Miskatonic versions). It's one thing to disagree on a few aspects or something, but never before have I seen such a distinction between two seperate groups of thought for CoC.

Too, there are sooo many elements to playing a game beyond the math and they're one of the many reasons why i think your evalualtion of 80+% win chance against yithians is massively incorrect. Or to put it more succinctly... an irrelevant statistic in the world of card games. Not that analytical data isn't useful, I'm just saying its not as absolute as Graham is making it come across. Jeremy, nor I, didn't win just because we beat the odds of the deck and/or matchup. Well, we beat some odds, but... ah you know what I mean.

3) Hindsight being 20/20, especially when you have more knowledge about what the opposition has, I don't think it's right to throw Brandon under the bus. Granted, I haven't gone back and examined the match to see if there were any true mistakes, but there is a world of difference between analyzing a video and playing the game in person(even more so in front a with a crowd, not to mention other things possibly weighing on this mind). He played a good game, and played a good tournament. He gave himself a chance to win and thats all you can really ask for from a person. But just because it could've been possible for him to lock out his opponent based on what was in Jeremy's hand during that game... I think it's rather unfair to say Brandon was making mistakes, as he simply doesn't have the hand knowledge video watchers have and I certainly don't think playing against the Yithian deck is as robotic as you're making it out to be.

4) Are we really still operating under the thought that the only reason Khopesh was restricted was because a few people on the forums moaned about it? Really!? Didn't I and others dispel this way back when? Eh, in any case... if the Flooded Vault is hit with the nerf bat (something I very much believe won't be happening anytime soon)... it won't be because a few people moaned about it! Honestly people, give FFG more credit than that. They're not always flawless, but they're not ones to cave so easily to a few wheels just because they're squeaky.

Anyway, I digress. The die has been cast and the proof is in the pudding (ya.... I don't even know what I'm doing with those metaphors). Graham, you had a great year and again congrats. Kinda bummed we didn't get a chance to hangout this year, but hopefully we can remedy that in the future. ;)

#5
mzi

mzi

    Advanced Member

  • Members
  • PipPipPip
  • 30 posts
Thank you very much for the deck and the analysis.
That is a lot of food for thought, esp. the "Static Triggers" part.

@Tom:
1/ It seems that the community is divided about the Y-deck, but you have to admit that it has been tested by top players and that most of them decided NOT to play it (or any variation). This does not mean that the deck is bad, but it does not mean that the ones who have tested it cannot play it either.
2/ I do not think Graham "throws Brandon under the bus", but it would indeed be very interesting if he could explain how Brandon should have played, as sample material...
3/ At the time it went out, Khopesh combined every problem you could dream of in a card game: bad wording relying on classical mechanics, bad wording introducing a new mechanic, luck-based (draw) balance disruption and reinforcement of the already dominant meta.

PS: why not "Seek and Destroy"? :P

#6
konx

konx

    Advanced Member

  • Members
  • PipPipPip
  • 114 posts

1) The reason why 'lotto ticket' is a negative term is that there is more to a deck that it's ability to draw a specific set of cards. Opponents and a player's ability to 'outsmart' said opponent to enact the deck's strategy is as, if not more, important than simply drawing the cards you need. By declaring a deck a "lotto ticket" makes it seem like the only reason a player ever won a game is because they happened to draw 'it.' Which... steps on the wrong nerves if you know what I mean.


Since I tend to agree with Graham here, I can explain my point of view: while it is true that there is always the component "skill" in a game, it is also true that there is the component "statistics" (or luck, if you want, but I prefer the first term). And if you don't maximize the second, the first is just useless.

Now, one can have all the skill he wants, but if the deck has poor statistics (let's assume the 1 of 3, 1 of 3 example of Graham) it will happen that your statistics is all in the bottom part of the deck. At that point, the skill is irrelevant.

Example 1: Magic. If you don't draw lands/mana, all your skill is useless

Example 2 (from my experience in Liege): I have 11 cards in the deck doing the same thing, and I am in a board position waiting for 1 of those cards (doesn't matter which one). I drew 20 cards (almost 50% of the deck!!!) before seeing one of those cards. This to say that EVEN when the statistics is on your side (I mean, 11 out of 50 means that every 5 cards I should draw at least 1, right?) bad luck can influence the game.

Too, there are sooo many elements to playing a game beyond the math and they're one of the many reasons why i think your evalualtion of 80+% win chance against yithians is massively incorrect. Or to put it more succinctly... an irrelevant statistic in the world of card games. Not that analytical data isn't useful, I'm just saying its not as absolute as Graham is making it come across. Jeremy, nor I, didn't win just because we beat the odds of the deck and/or matchup. Well, we beat some odds, but... ah you know what I mean.


Well, I have a small trivia from the tournament regarding a similar concept. Ray was playing against a yog-stl bounce deck; the bounce deck won 2 stories and had tokens (3 or 4) in the last one. Ray wins the match. Discussion between Graham-Ray-Roberto after the match: "it played exactly like in playtest". On the long run, statistics in matchups is always coming out. Sure, in the single game everything can happen but that is a matter of the single game. You can beat the opponent, but statistics "in vacuum" is unbeatable.

I mean, there is a reason why, given the same decks and card availability, in Magic 90% of the people winning big tournaments is always the same. They maximize the statistics and then, they let the skill do the work. And while sometimes you see someone getting there by luck, I hardly recall someone with bad decks winning tournaments.

I think that what should be taken away from Graham article is the way of building and analyzing decks, which is the first basic step to make. The rest (skill, opponent skill, etc...) is a particular accident of the tournament.

Konx
  • mzi likes this

#7
Skelton

Skelton

    Advanced Member

  • Contributor
  • PipPipPip
  • 2081 posts
Great read, certainly a lot for the newer players among us to ponder.

#8
Wilbur

Wilbur

    Advanced Member

  • Members
  • PipPipPip
  • 630 posts
... Would the fact that, outside of some now mythical play testing, some version of the Yithian deck has never actually lost a tournament (so far as I know. It did NOT show up at Liege, is that correct?) count as statistically significant?

Given the insignificant sample size, I'd say simply no. Which is to say that to attempt to reduce a process both stochastic and skill-dependent to 'inevitable' statistical outcomes seems a bit... optimistic. When you factor in that every card of a given functional description (e.g. 'support removal') is asymmetrical with every other (Thunder in the East/Burrowing Beneath/DOA), things become far less predictable still. And this still doesn't factor in what your opponent actually does (the behavior of an actual opponent making his own evaluations and judgements is simply not identical to the pre-tournament projections of the metagame at any given tournament). I fully agree with Konx that "statistics 'in a vacuum' is unbeatable"; I would add that they are unplayable, as well. The Void can be Studied, but it cannot be made to play card games.

I understand how statistical optimization is supposed to work ('Yes! Another 80% of a victory! I can repeat this to infinity, until the margin of my victory becomes infinite, and I am truly a become a god!'), as well as how it works ('I've won far more games than I have lost. The deck really seems to be coming together nicely.'), and Graham certainly seems to have approached the matter thoroughly. However, a card game is meteorology, not celestial mechanics.

Congratulations to all participants, and especially winners, throughout the tournament season. Hopefully 2014 will see the game's popularity continue to grow.

#9
DarkSeph

DarkSeph

    Advanced Member

  • Members
  • PipPipPip
  • 55 posts

3) Hindsight being 20/20, especially when you have more knowledge about what the opposition has, I don't think it's right to throw Brandon under the bus. Granted, I haven't gone back and examined the match to see if there were any true mistakes, but there is a world of difference between analyzing a video and playing the game in person(even more so in front a with a crowd, not to mention other things possibly weighing on this mind). He played a good game, and played a good tournament. He gave himself a chance to win and thats all you can really ask for from a person. But just because it could've been possible for him to lock out his opponent based on what was in Jeremy's hand during that game... I think it's rather unfair to say Brandon was making mistakes, as he simply doesn't have the hand knowledge video watchers have and I certainly don't think playing against the Yithian deck is as robotic as you're making it out to be.
ut this year, but hopefully we can remedy that in the future. ;)


I knew I made the mistakes as it was the first thing that I said to graham after the match so its not really throwing me under bus when I know I made the mistakes from the start.
  • Midian likes this

#10
MagnusArcanis

MagnusArcanis

    Advanced Member

  • Members
  • PipPipPip
  • 331 posts

Thank you very much for the deck and the analysis.
That is a lot of food for thought, esp. the "Static Triggers" part.

@Tom:
1/ It seems that the community is divided about the Y-deck, but you have to admit that it has been tested by top players and that most of them decided NOT to play it (or any variation). This does not mean that the deck is bad, but it does not mean that the ones who have tested it cannot play it either.
2/ I do not think Graham "throws Brandon under the bus", but it would indeed be very interesting if he could explain how Brandon should have played, as sample material...
3/ At the time it went out, Khopesh combined every problem you could dream of in a card game: bad wording relying on classical mechanics, bad wording introducing a new mechanic, luck-based (draw) balance disruption and reinforcement of the already dominant meta.

PS: why not "Seek and Destroy"? :P


@Mzi
1) Thats exactly why I can't wrap my head around it! Intelligent and experienced people playtested yithians all over the world. I don't get why theres such a major difference. I can get why people didn't play the deck though. I hated the fact that I took it to worlds for what I believe to be many of the same reason(s), but there are other reasons that I simply don't understand why the two groups are polar opposites on.
2) Perhaps it was the late hour, but the article read as if that if Brandon would've played against the hand that Jeremy had better, Graham's deck would've won. Thats throwing the guy under the bus to me, but I apologize if I misread it. I however, still think it's unfair to lay out a scenario that is set up against the situation Brandon was actually in. And based on the barrage of emails and texts I recieved on this point after this article went live... I'm aparently not alone in thinking that something was wrong.
3) Agreed, in fact there were more reasons than that! Thank you for bringing some of those to light.

Since I tend to agree with Graham here, I can explain my point of view: while it is true that there is always the component "skill" in a game, it is also true that there is the component "statistics" (or luck, if you want, but I prefer the first term). And if you don't maximize the second, the first is just useless.

Now, one can have all the skill he wants, but if the deck has poor statistics (let's assume the 1 of 3, 1 of 3 example of Graham) it will happen that your statistics is all in the bottom part of the deck. At that point, the skill is irrelevant.

Example 1: Magic. If you don't draw lands/mana, all your skill is useless

Example 2 (from my experience in Liege): I have 11 cards in the deck doing the same thing, and I am in a board position waiting for 1 of those cards (doesn't matter which one). I drew 20 cards (almost 50% of the deck!!!) before seeing one of those cards. This to say that EVEN when the statistics is on your side (I mean, 11 out of 50 means that every 5 cards I should draw at least 1, right?) bad luck can influence the game.


You're not wrong, you just missed my point. Of course probability plays a major role (alongside skill, composure, and game knowledge). I never, and I don't think anyone would, dispute that fact. I'm just pointing out why calling a deck a "lotto ticket" deck is insulting to the playerbase. I've asked Graham on several occasions to stop doing this and has continued to repeat his actions. I'm hoping that by explaining why people shouldn't do that would bring to light why I've asked him to stop.

Well, I have a small trivia from the tournament regarding a similar concept. Ray was playing against a yog-stl bounce deck; the bounce deck won 2 stories and had tokens (3 or 4) in the last one. Ray wins the match. Discussion between Graham-Ray-Roberto after the match: "it played exactly like in playtest". On the long run, statistics in matchups is always coming out. Sure, in the single game everything can happen but that is a matter of the single game. You can beat the opponent, but statistics "in vacuum" is unbeatable.

I mean, there is a reason why, given the same decks and card availability, in Magic 90% of the people winning big tournaments is always the same. They maximize the statistics and then, they let the skill do the work. And while sometimes you see someone getting there by luck, I hardly recall someone with bad decks winning tournaments.


My point is that there shouldn't be a vaccum. There are simply to many variables, tangible or otherwise, to consider. The odds between two set decklists played by the same players (whether they're real or not) can be calculated. Thats exactly what Grahams statistics portray. People, like yourself apparently, tend to take these as hard facts without realizing how flawed the math is because it doesn't take into account that rarely are two decks of the same archetype exactly the same, play styles between players is often different, etc.... Even Graham pointed out that changing just a few cards can radically alter the results...

Again, probability anaylsis is useful. No arguement there, but its' not absolute. That's my point here.

And to be fair. I won 2011 Worlds with a bad deck. It was solid enough to give me a chance, but it was by no means good at the time.


I think that what should be taken away from Graham article is the way of building and analyzing decks, which is the first basic step to make. The rest (skill, opponent skill, etc...) is a particular accident of the tournament.

Konx


I agree that what should be taken away from Graham's article is that having a strong statisitcal analysis behind the building of the deck is a great first step, but "the rest being a particular accident of the tournament"... wow. Thats... wow. I uhh... Konx, I know you're a good guy, but come on. I know you're not that narrow minded.


I knew I made the mistakes as it was the first thing that I said to graham after the match so its not really throwing me under bus when I know I made the mistakes from the start.


Then I think you're being unfair to yourself. Based on the scenario Graham laid out and from what you told me at the event. Not playing cards because you forgot to, or missing triggers, or making a move froggetting an on board element is one thing, but beating yourself up over in what you percieve to a be a miscaluations only after how the opponent played against you is unfair. Considering the butterfly effect each move has you can only ever make the best possible decision you can at the time based on the knowledge you can access. Sure, you could've made better guesses in some scenarios but even if you would've of 'made the correct move' still doesn't mean that you would've won the game. Obviously, it could affect the outcome, but if you moves change, so do your opponents which in turn change your later moves and before you know it, it becomes a completely different game which may or may not have the same result.

That being said, there is some good to beating yourself up for a percieved mistake. I know it has helped me over the years, but I honestly believe, that if one element changed... that early in the game... wouldn't necessarily result in a change of outcome match as the game would be completely different.

Anyway, I suppose people can agree with me in that there are other elements to winning that are just as important as draw/play probability and that you need the whole package if you are to be truely successful... or not. To each their own, but I feel that I am unequipped to say much more on the subject else I'll just end up repeating myself more than I already have.

PS. Can we stop using Magic as a baseline for compairson? CoC and MtG are on two different planets as far as most game philosphy goes. Resouce system, game play, player play styles, deckbuilding, community, and so much more are so different it's not even funny.
  • plasticXO likes this

#11
konx

konx

    Advanced Member

  • Members
  • PipPipPip
  • 114 posts

I agree that what should be taken away from Graham's article is that having a strong statisitcal analysis behind the building of the deck is a great first step, but "the rest being a particular accident of the tournament"... wow. Thats... wow. I uhh... Konx, I know you're a good guy, but come on. I know you're not that narrow minded.


I don't get just this part.

Step 1): build your deck to do something and do it in a reliable way, with statistics on your side.
Step 2): find the deck's weakness and decide if it is worth to consider them and build to fight them or not
Step 3): playtest against different strategies/decks/opponents until you get a clear grasp on how to play your deck at the best you can
Step 4): go to the tournament.
At this point:
- is the opponent an accident of the tournament (actually, of the single game. Opponent in this case reads: "skill of the opponent")? yes
- is a bad draw/good draw an accident of the tournament (actually, of the single game)? yes

= the only thing you can build on (in a reliable way) is your skill + statistics. I mean, I guess no one can really tech against "Graham" or "Tom" right? ^^

Maybe accident is the wrong word, I hope it is more clear this time.

Konx

#12
MagnusArcanis

MagnusArcanis

    Advanced Member

  • Members
  • PipPipPip
  • 331 posts

I don't get just this part.

Step 1): build your deck to do something and do it in a reliable way, with statistics on your side.
Step 2): find the deck's weakness and decide if it is worth to consider them and build to fight them or not
Step 3): playtest against different strategies/decks/opponents until you get a clear grasp on how to play your deck at the best you can
Step 4): go to the tournament.
At this point:
- is the opponent an accident of the tournament (actually, of the single game. Opponent in this case reads: "skill of the opponent")? yes
- is a bad draw/good draw an accident of the tournament (actually, of the single game)? yes

= the only thing you can build on (in a reliable way) is your skill + statistics. I mean, I guess no one can really tech against "Graham" or "Tom" right? ^^

Maybe accident is the wrong word, I hope it is more clear this time.

Konx


I'm glad we're on more of an misunderstanding... cause I didn't want to be mad at you Konx. The way you said it before it made it seem like if an opponent wins against the supposed match up average.... it was only by luck and that the player didn't matter.

For the most part... We're in agreeance.

@ "- is the opponent an accident of the tournament (actually, of the single game. Opponent in this case reads: "skill of the opponent")? yes"
I couldn't disagree more. By definition (an event that happens by chance or that is without apparent or deliberate cause), having a good/bad draw is an 'accident', but opponents and their "skill" I don't think is an accident. You may not be able to account for everything (which is my point on most of this stuff), but I believe to be apart of the 3rd step you listed. Learning how you own deck plays is important, but learning how your 'guess of competition' plays their decks is almost as valuable.

Opponent's matter. By their deck build decisions and/or by gameplay skill, they do influence outcomes and largely invalidate most 'vaccum' statistics.

@ "= the only thing you can build on (in a reliable way) is your skill + statistics. I mean, I guess no one can really tech against "Graham" or "Tom" right? ^^"
Technically you can.... but I'm with you. The further you get outside your best friend and into larger tournaments, especially with those invovling those outside your normal circle of community... it becomes less and less of a good idea. ;)

I dunno, may you still disagree, but I've personally witnessed play skill beat "match advantage" on many occasions, accorss many different games (as in, not just CoC), and I see it way too often for it to be an 'accident'. In fact... I would say that my frist 3 championships were based FAR more on skill than anything else. Each year, the consensus was that 'runner up' had the advantage in their respective matchups. In fact, the top 4 matches for 2010 and 2011 weren't in my favor either.

So... I dunno. Perhaps we merely see the world from two different lenses. =/

#13
Danigral

Danigral

    Advanced Member

  • Small Council
  • 1358 posts
~If this is what's required to build truly excellent decks, then I fear I am stuck making subpar ones. Unless I want to take community college courses just so that I can understand what the **** Graham is talking about. My google fu was slightly helpful, but really, it's my liberal arts background that seems to be letting me down the most. Math for Music Majors doesn't go into probability distributions. :P
  • badash56, Yipe, HomerJ and 2 others like this

#14
mzi

mzi

    Advanced Member

  • Members
  • PipPipPip
  • 30 posts

~If this is what's required to build truly excellent decks, then I fear I am stuck making subpar ones.

I think this is what's required to technically explain how winning decks work.

I do not understand Tom's concern about Graham's phrase "lotto ticket". That simply means Graham does not want to play that deck because its main feat relies on a probability that is lower than the one he can actually accept. [Note: I have no problem with that. I even have no problem with anyone thinking that a deck I play is bad as long as the game remains fair.]

On the other hand, Graham almost never post a deck. If the discussion goes around the way he expresses his analysis rather than the deck itself, I can understand that he becomes really upset or disappointed.

I also do not understand why we should stop the MTG comparison. Any cardgame implies hypergeometric law, any card game implies board control evaluation, any card game implies a lot of things in common (drawing cards, resource management, discard pile manipulation, etc.). There have been tons of articles about those aspects of MTG on the internet and it is highly valuable that someone points out what articles are relevant to the analysis of the CoC mechanics. Additionally, Graham has explained very precisely the specific implications of drawing effect in CoC, not in MTG.

I'm glad we're on more of an misunderstanding... cause I didn't want to be mad at you Konx. The way you said it before it made it seem like if an opponent wins against the supposed match up average.... it was only by luck and that the player didn't matter.

1/ English is not our native language.
2/ It often happens that the player does not matter too much. As you have said before, sometimes, you just make the best choice at one specific moment and it turns out to be a wrong choice two turns later. When you play a good deck versus another good deck, against a real challenger, and none of you makes any real mistake (only bad decisions, which is not the same), it is really luck that drives the game. It has happened often around here during my playtests against B_P.
  • Jhaelen and Carioz like this

#15
konx

konx

    Advanced Member

  • Members
  • PipPipPip
  • 114 posts

@ "- is the opponent an accident of the tournament (actually, of the single game. Opponent in this case reads: "skill of the opponent")? yes"
I couldn't disagree more. By definition (an event that happens by chance or that is without apparent or deliberate cause), having a good/bad draw is an 'accident', but opponents and their "skill" I don't think is an accident. You may not be able to account for everything (which is my point on most of this stuff), but I believe to be apart of the 3rd step you listed. Learning how you own deck plays is important, but learning how your 'guess of competition' plays their decks is almost as valuable.

Opponent's matter. By their deck build decisions and/or by gameplay skill, they do influence outcomes and largely invalidate most 'vaccum' statistics.


Ok, I will be more clear: the opponent you face in a tournament is completely random. You don't know (for the most part) who is going to show up at the tournament and even if you do know, you don't know how you are going to end up paired, since the pairings are random.

I've read on tons of report sentences like: "I was playing deck XY which has bad matchups against deck KZ, but I got lucky and I didn't face any of those during the tournament even though there were many". This is what I call "the opponent is an accident". This you cannot (and should not, IMO) take into account while building the deck. You can, in fact, say "I don't care about deck KZ and I hope either not to face it or be ready to lose against it".


I dunno, may you still disagree, but I've personally witnessed play skill beat "match advantage" on many occasions, accorss many different games (as in, not just CoC), and I see it way too often for it to be an 'accident'. In fact... I would say that my frist 3 championships were based FAR more on skill than anything else. Each year, the consensus was that 'runner up' had the advantage in their respective matchups. In fact, the top 4 matches for 2010 and 2011 weren't in my favor either.

So... I dunno. Perhaps we merely see the world from two different lenses. =/


I am not denying that skill can get you an advantage. I am denying that you can rely on that or making it the main strength of your deck. I am denying that you can use "the skill" as the factor to do an analysis on a deck.

Since we are in Graham's deck thread, let's take that as an example: first thing Graham did was maximizing the statistics. The second thing he did was learning how to play his deck the best he could. So, first you maximize the deck and then you maximize the skill.

Bottom line: skill, luck, matchups...everything matters. But the ONLY two things that you can rely on in a consistent and continuous way are your deck and your skill. The rest is given by "chaos theory".

Konx

#16
Carioz

Carioz

    Member

  • Members
  • PipPip
  • 29 posts

~If this is what's required to build truly excellent decks, then I fear I am stuck making subpar ones. Unless I want to take community college courses just so that I can understand what the **** Graham is talking about. My google fu was slightly helpful, but really, it's my liberal arts background that seems to be letting me down the most. Math for Music Majors doesn't go into probability distributions. :P


Do not despair, it took me more or less two years of almost daily dialogue to be able to get what Graham was saying in a snap (granted, language barrier didn't help). Sometimes I will post about the day when I "got it" and was able to contribute in a meaningfull way to the discussion. It was like day and night. For the record what I mostly do now is take the brunt of the real cards testing.

#17
badash56

badash56

    Advanced Member

  • Members
  • PipPipPip
  • 437 posts

~If this is what's required to build truly excellent decks, then I fear I am stuck making subpar ones. Unless I want to take community college courses just so that I can understand what the **** Graham is talking about. My google fu was slightly helpful, but really, it's my liberal arts background that seems to be letting me down the most. Math for Music Majors doesn't go into probability distributions. :P


I'm with you man, lol.
  • kamacausey likes this

#18
dboeren

dboeren

    Advanced Member

  • Contributor
  • PipPipPip
  • 2834 posts
As someone who does have a math background, I wouldn't worry about it. If you've got a 44% chance to draw a key card by turn 2, you've got the same 44% chance whether you understand how to calculate it or not. If you don't know what your exact chance is, if you're a good player you still are going to have a gut instinct that tells you that you get it roughly half the time, and being slightly off on your estimate does not affect the actual probability one bit.

Not to mention that by necessity these sort of analysis are (at best) only looking at a small part of the game. It doesn't take into account what your opponent is doing for example, or what stories are in play, or a number of other important factors. This is not just true of card games but pretty much all tournament style games other than maybe the abstracts. When I was mainly playing minis games we saw the same thing there - people made assumptions, predictions, and proclamations based on analyzing one aspect of the game but ignoring others. Only there the issue was known well enough that it had a name and everyone knew about it. "Theoryhammer" for the Warhammer players, "Theorymachine" for the Warmachine players, etc… But in a real game there were always complication factors that cropped up as well as opponents who didn't want to follow the plan you assumed they would and the nice simple odds you thought held true may or may not work anymore.

There is enough variability in the card draw and deviations from the plan prompted by the immediate needs of the moment that I believe piloting skill will significantly outweigh this level of deck optimization. That's not to say that you can't or shouldn't optimize your deck if you want to and you do need to have a decent deck that you understand how to play as a first criteria, but there is a point of diminishing returns and a lot of optimization is better done through play testing rather than computation.

#19
konx

konx

    Advanced Member

  • Members
  • PipPipPip
  • 114 posts
Anyway, after discussing some trivia facts about maths/playtesting/skills/luck whatever..it is interesting to notice that no one has discussed yet the main point of a deck like this presented with an article like this.

How would you beat it?

Because I have seen this deck win through a board were the opponent had alienist+flux+rays of dawn (which are kind of the first cards that comes to mind to me to stop the deck and the triggers). I have seen the deck winning through snow graves locks+destruction on table. And, as mentioned, I have seen the deck recovering from a "2 stories already won + a few tokens on the last story" situation.

What are your thoughts about it?

Konx

#20
Danigral

Danigral

    Advanced Member

  • Small Council
  • 1358 posts

On the other hand, Graham almost never post a deck. If the discussion goes around the way he expresses his analysis rather than the deck itself, I can understand that he becomes really upset or disappointed.


So yeah, let's discuss the deck. Statistics aside, it's easy to see that this deck is streamlined and efficient, just in what is included. My mantra has long been "free is good," and Graham expressed this well when talking about translating card advantage into free actions.

I put this together last night to goldfish with it, and I must say that it is brutal how fast it can flood. I can see that the idea is not to just draw cards ad infinitum, but to really just use Khopesh/Artifact to create a sort of jump-start. It's not even necessary if you already have 5-6+ characters and are able to wound 1-2 of your opponent's characters every round. The Uroborus loop is just flogging a dead horse.

Funnily enough, I was considering Morgan for another deck I was working on with an infinite draw combo, but I never considered adding in ALL the "free enters play" characters (well, not quite all, Stalking Hound is a notable omission).

I'm not trying to "improve" the deck, just try to gain more insight about why certain cards were not included. So just a couple questions about a couple of cards:
  • Was Stalking Hound ever considered in an iteration of this deck, even just one copy? It would seem that since this deck doesn't try to do anything to shut down cards like MotM, it would be fairly easy to trigger.
  • What about one copy of Laban Shrewsbury? It helps with potential long games if they happen to get out an invulnerable AO that can't be wounded with Khopesh, and still works with the draw tech.