We've moved! Point your browsers here: http://v-mundi.com

Saturday, November 24, 2012

Building a Vanguard Deck

Preferably one that glows with blue fire
Deck building is a core essential part of any collectible or trading card game. The ability to customize and expand your experience is one of the basic fundamentals that draws people to card games. But building a deck on your own is rarely easy in any card game. It requires a certain skill set that is actually separate from the ability to play the game. In Vanguard, that skill set can be broken down into several different categories, and you'll need all of them to build the best deck for you. In order to fully understand this article, you will also need other skills sets, so there will be some prerequisite reading. If you have not read my articles on subgames, resources, pressure, and card advantage, you will be completely lost as I will be using key terms from those articles. It also helps to read several Deep Clans to understand how the analysis works. So let's jump in.

Vanguard is not 52 Card Pickup

Selecting a Clan

Vanguard is a very unique game in many of its aspects. Clans really are not one of them (though the execution does make them unique and well-done). Similarly to Magic: The Gathering, Duel Masters, Yu-Yu-Hakusho, Dragonball Z/GT and countless other games, there are base "colors" or "teams" in the game that largely filter the types of individual cards you will select to go into the deck. The difference is, in those other games, mixing colors or teams is very common and so color selection is not generally the first thing you do unless you're just a huge fan of mono-color. Vanguard is a largely mono-clan game as very few cards are "splashable" (able to be put into just about any deck and work equally well) and a lot of the better cards have specific clan requirements so that your card pool is limited. This is to keep an inherent balance in the game so that it doesn't become a runaway train of "What broken combos can you come up with?"

In the clan breakdown article and in the individual Deep Clans themselves, I try to give a good description of what each clan is capable of and their relative strengths and weaknesses. Let's make one thing clear up front: I'm not telling you how to be creative. Decks like Wharrrgarbl Bizarroblue, mixed clans, super hybridey things, ultra rushdowns and such are bizarre works of creativity—not the norm. The game mostly functions on your ability to do well within its inherent limitations. Being able to spot interesting and creative decks takes, well...creativity. That's not something that you can teach. So for our purposes, we're going to try and build decks to be optimal and efficient within the standard parameters of the game. And the first step in that parameter, with rare exception, is selecting a clan.

"Do you have a clear image of how you're going to win with this deck? If you can imagine that, you should be able to naturally understand which cards are necessary and which aren't."—Yuri Usui

Winning Image

The next step is to simply select a core concept of your deck that you want it to do. Provided that the clan can do it, of course. And actually, you can do this before clan selection if you already know what you want to do. You take this core idea—this image, and you maximize your ability to win with it—Winning Image. The way you do that is going to vary based on the idea. If you just love the hell out of drawing cards and believe that options lead to victory (they do), you could try some sort of deck that spams drawing cards like Nova Grappler with Hani and Kirara. Now this breaks down further into your ability to appraise and synergize. You have to identify what the costs or weaknesses are of the core of your deck and attempt to strengthen the benefits while mitigating the costs or risks. Just taking the same Nova Grappler draw engine example, you'd want to identify which of your two core draw units has the biggest risk or cost and find ways to make that cost disappear as much as possible. Hani's risk is poor defense for one turn if you ride it and requiring a full field. Well, it's just one turn and a full field is easy to swing in this game naturally. But Kirara is a solid 9000 card that costs two counterblast. Knowing that you could only activate this twice in one game without having a heal trigger activate, you would want to add someting that can unflip damage. Nova Grapplers can do that in many ways with cards like Clay Doll Mechanic, Hungry Dumpty, and Muscle Hercules. Clay Doll also happens to exactly boost her to Stage 2, which makes it more likely she will hit. With Kirara and Hani both in your deck, your opponent can't guard forever, so you're going to draw eventually.

This is one example. The hard part is identifying what a "core concept" can be, and even once you do it can be hard to appraise cards well enough to tell if they go into that concept. Take Shadow Paladins for example. What exactly do they do? Lots of things. Lots of their skills either kill units to draw or something generic like gain power. Then you have to think about what "kill units to draw" actually means. It's not just cycling and refreshing your unit selection, it's letting you get a bigger hand to guard if need be. But then they have generic power gaining cards like Dark Dictator and Garurubau. This is where you have to find a way that they tie together. Macha, Garurubau and the inclusion of a Rank 1 ride chain can allow you to have a ton of early field presence both from your Grade Ratio and from the skills of the lower-grade units. So it rushes early on. But it turtles in the later part of the game. And it's really consistent with the special ride chain, grade ratio, and hybrid cards like Garurubau on top of having the consistency of card cycling. So it doesn't appear to do anything particularly exclusively but it does do three things to a medium degree: Rush, Turtle, be Consistent. And so now you can select which cards fit this and that you'd like in your deck.

Shadows are a rather undeveloped clan at the moment, so this may be descriptive of the whole clan for now, but lots of other clans don't have general statements like this. In those cases, the course of action to take is identifying what a clan can do and the attempting to execute those core concepts well. Being able to identify them is hard enough (Kagero has the ability to make an all-superior-ride deck that can unflip all its damage!) but determining how to best find the synergy for that concept can be a real challenge as well. It can also be hard to tell what a counts as a core concept. Here's a list of some to look out for:
  • High pressure. Having a field that can hit above 2/3/2 in formation of stages. Usually 3/3/3 is considered minimum grounds for a concept like Hell Marquis Amon or Lohengrin in Royal Paladins.
  • Tanking. Having the ability to guard a lot more, or guard a lot less in amount. Agravain is a good example of high defense that lets you guard for less. Decks that draw a lot or search guard like Tachikaze also count.
  • Advantage. This can be broken down further, but essentially it means either drawing a lot, killing a lot or superior calling a lot.
  • Damage. Can also be further broken down by specializations such as explosive "big plays" or early game rush, or abusing the critical mechanic.
  • Consistency. Can be broken down into areas of specialization such as ride consistency (not miss-riding is important), changing Grade Ratio by having special searchers, having Superior Rides, or skills that search a lot like Alfred.
  • Tech. Technique cores are typically decks that are either highly synergistic or have a lot of options like Granblue and Narukami.
  • Big Play. Things like megablasts, final turn gambits (Duke, Dudley Emperor, Stern etc). This usually comes as a sub-core of some other type like damage or high pressure but specifically does so through the actions of large (highly restrictive or costly) plays that try to turn the game around rather than lots of small plays that add up over a game.
Those are just some examples and lots of others will exist. It will be up to you to determine how the core of your deck functions. If you fail at this step, then every other step will be invalidated. As such, this can also be the hardest step depending on how new you are or how experienced at deck building you are.

Clearly the deck should be all Grade 3s

Grade Ratio

This is the framework for the deck. In choosing a grade ratio, you are setting the boundaries for the maximum number of copies of a certain property you will run, namely a specific grade. Grades in this game correspond to different solutions to different problems. Each one is balanced against each other, both by their numerical Grade and by the inherent skill and shield they each possess, along with their power and ability to have more devastating skills. The cards themselves are all balanced on a knife edge, and so it stands to reason that the ratio of different grades in your deck should be too. Too many of a grade means too little of another, and that means both grade-screw and grade-flood. In a general sense, grades fulfill two different roles: properly scaling the field for attacks or riding to achieve defense, skills and Twin Drive. Field scalability is important, but auxilliary to riding the proper vanguard. If you don't ride properly, your Early game means nothing and you'll quickly lose. In addition, riding determines what you can call, so riding is the primary function of a grade and calling is a secondary function. In order to determine the best Grade Ratio, a good starting point is determining the maximum possible copies of each grade to create the maximum total probability of not missing any of your rides.

Hopefully that makes sense to you. What we want is to know what ratio of grades will give us the largest total probability of never missing even one single ride. Since riding is very important. Now, we're obviously not counting special abilities like searching and drawing in this. Strictly from the standpoint of a vanilla set of cards, you can actually find that ratio. You should start with your Grade 0 lineup so you know how many of the 1-2-3 you will have to work with. Since every deck is required to have 16 triggers, and most decks run a 17th Grade 0 as a starting vanguard, that means most decks have 33 cards left to make a legal deck. If your deck has more Grade 0s, you'll have to read the rest of this and then use some logic to figure out your remaining ratios. If you simply split the 33 cards evenly among the three remaining Grade types, your ratio would be 11-11-11. Let's start with that because it would be logical in an independent probability system (like dice or coins) since you would want an equal chance of getting all three of them.

Now let's find what changes in vanguard from an independent probability system. Vanguard is a card game and not a dice game. That means that unlike dice, each time you draw upon a random element, the rest of the random elements change. That's called dependent probability. Each instance of chance is dependent on all instances that have occurred before it because the deck size shrinks or increases, or copies of a card enter and leave the deck. In order to find that ratio, you have to know what the probability of drawing a Grade 1  is before your Grade 1 ride, the probability of getting a Grade 2 before that ride, and the probability of getting a Grade 3 before that ride. Furthermore, vanguard has a mulligan system. Now obviously, we can't assume it will be mulligan 0, 1, or 2. The reason is because you draw 5 cards, this means that if you want to ride properly that game, the smart thing to do is keep one of each grade and mulligan away everything else. If you have one of each grade already, you keep at least three, meaning you mulligan 2, 1, or 0. Those mulligan amounts mean you already got all of your proper grades. Instead, we should start with the minimum (and actually the most likely on average) mulligan which is 3. But each turn you also draw a card and drive check a card which adds to the chances of getting what you need. That means that Grade 2 is easier to get than Grade 1 assuming you have 11-11-11 like we start with. And Grade 3 is therefore easier to get than 2 or 1 because it comes after both and you get more total draws (more total tries).

So logically, we know that our ratio will have more Grade 1s than 2 or 3, more Grade 2s than 3, and less Grade 3s than any other. You can't have any ratio be exactly equal to any other because that would violate cumulative draws per turn. You can't have a higher grade contain more copies because that's the inverse of the proper ride chances. So we definitely know the "shape" of the curve. It starts tall on Grade 1, and shrinks down as it approaches Grade 3. As it so happens, TehNACHO and I have conducted calculations for you to determine what the ratios are. TehNACHO focused on showing the large bulk of what each kind of ratio's total probability of riding would be (link above), and my focus was finding the natural optimal ratio as well as the optimal ratio for certain special decks. All things calculated, the proper ratio for a normal non-ride-assisted deck is:

tldr; 13-11-9

Oh golly gee gosh! That ratio is not one that hardly anyone uses ever (including me)! Just look at how gross it is with its 9 Grade 3s. But really, it's not far off from the one that most of the collective English player base has silently come to agree upon: 14-11-8. And we mostly figured that out from both collective experience and the fact that card copies are limited to 4 each, so typically you just pick two different Grade 3s and max them both out. Except in the case of the rare Grade 3 checker deck like Goku or Seifried. No, as it turns out, the proper ratio change from 11-11-11 (independent probability) is +2/+0/-2 or 13-11-9 (dependent probability) for Vanguard. Being a Grade 3 means that you are 2 card copies worth more likely to ride it than Grade 2, and 4 card copies more likely to ride it than Grade 1.

So we fix that by distributing its card copies to the Grade 1 (and Grade 2 didn't have to change). Remember, this is just the proper ratio when considering only the probability to ride. The maximum probability to ride an unassisted deck is almost 83% which is for the 13-11-9 ratio. Just shy of that is 14-11-8 with around 81%. So you can see that you don't really lose much consistency by just changing the one card over from being a Grade 3 to a Grade 1. The reason you would possibly want to do this is because you want to have a good viable booster-to-attacker ratio. Grades 2 and 3 are both attackers. Grade 1 (and 0, but almost never used) are boosters. So with 13-11-9 you have a 13:20 booster:attacker ratio. With 14-11-8, you have a 14:19 ratio which is slightly better. It's also even numbers which fit more into the game's card copy restriction.

Moving on from that, let's quickly go over some of the ratios for special decks. So far, I've only done the ones for ride chain ranks. If you are reading this article, then you damn sure better have read my ride chain article because you need to know why I rank them based on their probability of getting their rides. Some of the chains have special skills that increase the chance of other grades being ridden. When you do this, it can sometimes allow you to reduce that particularly higher chance grade by a few and increase any other grade you want. Typically, you prioritize getting the most consistent rides. Then you secondarily prioritize getting the best booster to attacker ratio, with the usual 14:19 being about a 74% amount and 1:1 being exactly equal amounts.
  1. Rank 1 chains - Ratio is 16-8-9. Really bizarre ratio because the Grade 1 directly searches the Grade 2 when ridden. Given the probability to ride is is very high, that means all excess copies of the Grade 1 can count as copies of the Grade 2 for deck construction purposes (in other words, getting the Grade 1 on the opening turn is the same as getting the Grade 2, so you can have less total Grade 2s). This can be easily figured with a simple logical fomula. Simply take the default 11 Grade 2 count, assume 4 of the Rank 1 chain Grade 2, then take away a percent of the remaining Grade 2s. That percent should be the probability that you ride the Grade 1 ride chain. That probability is 56%, therefore 11-(0.56 * 7).floor = 11-3 = 8. 9 is also acceptable so long as you subtract from the Grade 1 count and not the Grade 3 count. But this is the best possible ride ratio and booster-to-attacker ratio combined for this chain; at 16:17 which is nearly 1:1. If you care less about booster and attacker ratio. You could also play many more Grade 3s. A common criticism is you will have less shielding, but that's incorrect. You would have -5k total shielding in the deck per Grade 3 extra. You get the same reduction for each Draw Trigger used instead of a Critical or Stand. If you're using a 12 crit/stand deck, extra Grade 3s won't reduce your shielding at all.
  2. Rank 2 chains - Ratio is still 13-11-9. May seem odd at first, but consider the fact that they equally increase the probability of riding the next grade by the same amount. You always search the top 5 for the next grade during each ride step. So it's the same ratio. But it is an increased total probability over the natural probability. Therefore, Rank 2s are great for ride-consistency decks.
  3. Rank 3 chains - Ratio is 15-10-8. The grade 1 ride allows you to search, which increases the probability of getting a Grade 2 or 3 of the chain equally. Therefore, you can reduce each of them equally, but only by -2 total. However, since it's equally reduced, they must be reduced by the same amount each: thus +2/-1/-1 over the normal.
  4. Grade 3 searcher starting vanguard - Ratio is 14-12-7. Decided to include this for people who like using the Grade 3 searchers. They're the 4000 power starting vanguards that cost Counterblast 1, and move to soul to search the top 5 for any Grade 3. This does increase the chances of riding Grade 3 and not at the cost of any other grade ratio (just at the cost of -1 should you fail). If you mulligan normally, you can use this ratio. You can also go with 15-11-7 if you prefer it, but 14-12-7 has the same Grade 3 chance to ride as normal, with higher Grade 1 and 2 evenly. 15-11-7 only has a higher Grade 1, but normal of the other two. Depending on if you highly prioritize booster ratios or ride chances, take your pick. (Though I still do not recommend using Grade 3 searchers, due to poor boosting power).
If any of the logic of this was confusing to you, just know that you can check the bolded numbers for whatever your priorities are to get the proper grade ratio for that priority. Remember that each of these assumes 17 Grade 0s. With less Grade 0s, you can add 1 to any grade count you want. Like a "freebie". With more, I suggest having 3 extra Grade 0s, that way you know you distribute those -3 evenly and make 12-10-8 or 13-10-7. With the former being slightly better. If you only add 1 or 2 extra Grade 0s, you kind of have to choose what rides or other things are more important to you. Generally, riding Grade 1 is the most important since failing to means you lose. So almost never apply the uneven -1 to Grade 1. Typically, Grade 3 is the one you can wait just a little bit longer on.

I am totally not a secret Dimension Police spy...

Perfect Guards

Coming from the Japanese 完全防禦 (romaji: kanzen bougyo) meaning "perfect guard". Two kanji for wholeness or perfection, and two kanji for defense. I have seen some people using "Null Guard" but that's not really a good term since perfect defense is the official term. The term "Null" should not refer to the numerical 0 on the shield's value since the number zero and the concept of a null set are different. The term "Null" should not be short for "Nullify guard" since it's an adjective describing 'guard' and would therefore mean you have a useless defense. Since it's neither of those things—Perfect Guard. So why is this section here? It's just one type of card. Yes, and it's the card you know will be in every incarnation of a normal deck. So this is the first unit you should be putting into your deck regardless of its Winning Image. And it needs to be put in at maximum copies.

Unfortunately, many deck builders have been a bit confused about Perfect Guards. Their purpose during play as well as the copies that should go into the deck. It seems that 4 copies, while very common, is not unilaterally accepted yet. Some people still favor 3 and 2 copies. There also seems to be some misconceptions about them as well—especially among new players. Since the skill of a perfect guard is used by discarding a unit of the same clan while it is on the (G) circle, it is seen by some to be a severe loss of advantage. It's actually true. Guarding with it and discarding one is a net -2 overall to your advantage. Doing something like this in the Mid or Early game can certainly spell doom for you. However, this goes for guarding any attack. At all. Ever. Every single attack guard is at least -1 net, if not more. But you can't eat damage all day. In fact, since pressure is just going to keep going forever, you have to guard, and it's best to guard as little as possible as early as possible. Guarding small 1 and 2 stage attacks early means you can take big hits later. It's the same 1 damage either way in vanilla circumstances. But non-vanilla circumstances occur in Late game. This is when you can't afford to take a critical hit. This is when you can't afford to take that 6th damage from a Blond Ezel that just hit 4 stages of power, or a Dark Lord of the Abyss that hit a ridiculus 8 stages of power. How can you possibly guard that? Are you really going to throw down anywhere from 5-9 cards just to not take the damage? Can you even do that?

This is where Perfect Guards come in. They are the perfect solution to this problem. Any time an attack would require -3 or more, you can take it or perfect guard it. Taking it is not always an option, or the best option. So simply from a survival point of view; running maximum Perfect Guards is optimal for maxing out the chances that you survive these kinds of hits. And they do exist. Boy do they ever. If your opponent has no rear-guards, you can perfect their vanguard to completely stop all damage and waste their drive checks that turn. If they're dumb enough to attack with the vanguard last—same deal. The drive checks are now wasted. The existence of Perfect Guards means that all decks that don't run stands now have to attack with the vanguard first or second. Think about it—their mere existence changes the order of attacks. For both players.

For most cards, any question like "how many copies of X do I use?" should be entirely dependent on the structure of the deck. Your core idea is going to shape that for the run of the mill card. But when you have something this powerful that's available to exactly every single clan, the answer isn't the same. You run maximum. As many as legally allowed. In Vanguard, that's 4. Not just 4 of one card, but 4 of any perfect guard. You can't have more than 4 total of that card archetype in your deck. Bushiroad themselves understand how powerful this is and if we ever get hybridization in our clans, 8 perfect guards would be nonsense. Now, obviously if there were ever some sort of broken defensive card in this game that gets 12000 or more defense without a cost, you can run less Perfect Guards in that deck. Until, of course, hitting 3+ stages against your new special snowflake defender becomes the norm. Then you're back to 4 again. Every other normal 10000 and 11000 defensible deck will need 4, because this game has some seriously heavy hitters.

Limit Breaks are a huge majority of those heavy hitters. The Power Breakers that nearly every clan will eventually have will get +1 stage of power during Late game. Other Limit Break skills usually grant some amount of power which easily translates to another stage. Now, this is just vanguards, of course. But there are other vanguards that cause this power surge. Demon World Marquis Amon and Dark Lord of the Abyss hit absolutely criminal levels of stages. Most non-11k vanguards have some innate ability to make 3 stages with an 8000 vanilla booster behind them. Using a 10000 defender means you're susceptible to 20000 columns, which aren't hard to make. Some decks like Soul Paladins and Amon Dark Irregulars practically live on huge pressure in the rear-guard too. So there's no escaping them if criticals start showing up. Guarding this plethora of power rape in the Late game is nearly impossible (and definitely impractical) without a Perfect Guard. And some On-Hit vanguards absolutely must be stopped like Stern Blaukruger. Or you will lose the game. Another reason to run 4. 


Using 4 perfect guards also puts you at no disadvantage over those that run less. In Game Theory, this means that running 4 perfect guards dominates the strategy of running less. Since running less than 4 puts one player at a disadvantage, neither player wants to deviate from running 4 if both are being rational. This is called a Nash Equilibrium—where the best response of either player is not to deviate from said equilibrium strategy. What do I mean by it doesn't put you at a disadvantage? I've seen a lot of people make statements suggesting that Perfect Guards can "clog" the hand. However, that doesn't make any sense. They are Grade 1s, so they can be ridden to keep your rides going. They have the boost skill and 6000 power, which means they can go behind any normal Grade 3 and some Grade 2s with no problem. Every deck of which should have the requirements for that kind of column. The fact that you draw 1 and drive check 2 each turn means you have 3 cards in hand. Thus a Perfect Guard can only "clog" the hand if you don't plan your guards properly and end up with just the perfect as your last card. And even then, it may not be a clogging, but the more optimal guard strategy. 

Even having 2 of them in your hand means that you can simply discard one for the other. You shouldn't allow their wonderful skill to enslave your mind to what they can do. They still fill the role of any other Grade 1 except in the super specific and rare circumstance of being the only card you have in your hand, not having any interceptors, and the opponent is attacking for 1 stage (which probably isn't even an issue unless it will be game). That's too rare and circumstantial to even be on anyone's radar. Their non-guard uses aren't commonly required, and when they are, it's not debilitating. Using them as a guard, is never less efficient than simply using any random two 5k shields. If you discard a draw trigger, Grade 1, or Grade 2, you just net +0 shielding against a 2 stage attack. It's an inefficient guard, but so would be guarding for any 2 given 5k shields. Discarding a Grade 3 is like guarding with 15k. It vastly increases your total shielding capacity and makes uses of those extra Grade 3s floating around in your hand.

I also understand the desire to not have to to use 4 Perfect Guards; as opposed to the desire to specifically use less than 4. They can be tough on the wallet when most of them are $11 USD or above. Especially now that retail chains have figured out they're staples for every single deck and keep artificially jacking up their prices. But price of the card doesn't make it any less optimal for your deck. Running less of them means you're on a budget, but it doesn't have anything to do with efficient deck building. And even most budget builds have to run 4, they just cut corners elsewhere. I also get that some people have a mindset of "It doesn't leave much room for my other cards". But that's more like saying "I wish I didn't have to use so many of these super awesome cards so I could use these slightly less awesome cards". 4 Perfects is a cemented thing for your deck. The copies of the other cards can be altered slightly to fit around that just fine. If they can't, your deck idea is probably not going to work very well so you need to refine it.


Team Assemble!

Unit Selection

By now you have your framework. The core idea. The grade ratio of units you can put in. And you've drawn first blood with 4 Perfect Guards resting nicely in your Grade 1 count. Now you need to fill up the rest of the deck. But how? This is going to be one of those really shitty answers—it depends on the deck. Yeah, sorry, but it's true. The first thing you do is start adding in core cards. The number of copies per each should probably be 4 with rare exceptions. They are core after all. If your deck were Nova Grapplers 13-11-9 and focused on making redundancies with Asura Kaiser to spam stand skills as much as possible, then you'd obviously want 4 Asura Kaiser; 4 of another Grade 3 that works with him. One such Grade 3 could be Azure Dragon. Drive checking a copy of Azure will activate Asura. Later on, you can ride Azure if you run out of checkables and you have 2 or more copies in hand to use his Persona Blast. If you get stuck riding Azure first, you can simply wait out checking Azure's persona as well. To be further redundant, you can max out all the Death Army units that stand when a Grade 3 is checked. And round it off with Best Deity White Tiger as your starter since he can stand any Azures you might call. From there it's simple: you want to keep standing, so include units that are good stand targets. Like 11k Grade 2s and 10k Grade 2s. That's where Jack and King of Sword would come in. These are support and not part of the core so it's not necessary they are at max. But you should have an idea of how this would work for a simple core idea like that.

Others aren't necessarily as obvious, but it's not "is this unit good?" when you're deciding to put it in your deck, but "is this unit relevant?" If you have a crapton of Grade 3s and the vanilla booster, then you have a perfect opening for a Special Interceptor to make up for shielding. If you can search out a unit with Conroe and Mecha trainer, it basically becomes a tech-fest. Techs are those things you run less than 4 of, usually 1-2 copies. They make a slight difference during a game. Lots of them added up can make a big difference or blindside your opponent depending on the choice and its efficacy. You have to worry about cost as well. Individual unit cost and total cost. If you have a deck like Pale Moon that focuses on getting more attacks in and scaling the field dynamically, then you're likely to use Sarah and Alice. That means you have a lot of Grade 3s and you need your counterblasts open to be used with Alice. Including units that use counterblast will directly interfere with your core mechanic unless they also do the same thing equally well as her. If they don't, then it's not worth it. So your other skills should likely be costless or have a cost that doesn't include counterblast. On the flipside, some decks favor options. These are tech decks like Narukami that have maximum Deathscythe and Dragonic Kaiser Vermillion. Heavy costs to be sure. Yet it's up to the player to choose which ones to use since they both fulfill similar ends. Determining if your deck can have lots of redundant costs like Narukami or varied and different costs like Pale Moon is key. You don't want to run into an issue where you use a Megablast deck as the primary, but keep losing because you use up all your counterblast or soul.


So much of this step is based entirely on experience and logic. The thing you do not want to do is start adding crap in that's tangentially related to what you want to do, unrelated, or something you "feel" works. No matter how much you like them, the Tachikaze savages just don't work. Especially not with the current card pool or the pool of the foreseeable future. If you can think of a card that does a better job than one you're deliberating putting in, or does the job at a lower cost, then put it in. If there are some trade offs, consider the implications of those trade offs. For instance; one example is you might need to decide on your draw power in the deck. Between Kirara and Hani. Hani is weak and needs a full field. Kirara needs to spend counterblast. So instead, an option is Street Bouncer which has none of those requirements (it doesn't even need to hit). Instead, it sacrifices one column of pressure to draw 1. If it would hit for 2 stages, that's actually a viable trade off since +1 to you is the same as -1 to the opponent in terms of relative advantage. For a generic Novas deck, it actually doesn't work due to the other things you would be running, but as a rule, it's good to consider how those alternatives work relative to each other.

When considering how many copies of a card to include; beyond simply wanting to appeal to the core idea of your deck or make backup ideas and redundancies, you should consider when you want to draw it. Obviously the more copies of a card, the more likely you are to draw it at any given time. Since times are cumulative, that means that more copies of a card directly translates to drawing more copies in game and likely that means drawing some earlier too. If you want 4 copies of a card, it should be something you don't mind riding at some point. At 4 copies, that means it will very likely have to be ridden and somewhat often relative to anything at less copies. And you'll be almost sure to get it some time during the Early game. This is why a lot of Seifried Spike Brothers decks run Seifried and Dudley Dan at maximum. You want to draw Dudley Dan early so that you don't have to waste a counterblast that Dan would otherwise use on Mecha Trainer searching him out. You do get 2 Dan uses and 1 extra counterblast per game, but if you use the 1 first, that means you have to wait until you take more damage to use Dan. It isn't always about the amount of costs paid as sometimes the timing can be important as well. Similarly, you absolutely have to ride Seifried. So 4 copies is required. Dimension Police don't need their Laurel until some time in the Mid game since it's specifically for use with the Grade 3. Thus, they can run 3 copies of Laurel so they're not in danger of riding it. If you never want to ride a card, consider running 1 copy, or at most 2 copies. This goes for things like Nemain and Badb Catha. So 4 copies is early game material, 3 copies is mid game material, and 2-1 copies is late game material. You can safely run 2-3 copies of a Limit Break card if it isn't something you need to ride as soon as possible. Things like Luquier or Dragonic Lawkeeper can wait until Late game to be ridden. Cards like Leo-Pald and Dragonic Kaiser Vermillion really need to be active immediately, due to their other skills or defensive power.

Make sure you make field scaling selections. If your Grade 1 lineup is all 6000 boosters, and you don't have any 10000 Grade 2s, then you're screwed. You need to be able to make proper 16k columns in this game. Typically, a rule of thumb is have a 1:1 ratio of boosters to interceptors that become 16k. Since most Grade 3s can be boosted by any Grade 1 for good stages, and since interceptors will eventually get swapped out for them, boosters are the ones that remain. So it's fine to have some fudge room. If your Grade 2 lineup has 8000s and some 9000s, then your Grade 1 lineup needs 8000s and 7000s. This is mostly going to be common sense. If your core image is good enough, you should already have all your Grade 3s, half your Grade 1s, and half or all of your Grade 2s. Anything that remains is either tech, personal choice (sometimes), or picking things that don't make the deck strictly worse when you put them in. Try to think carefully about proper selection. You don't need Genocide Jack in every Nova Grappler deck. Just in ones that abuse stand mechanics since 11k can hit 11k vanguards, whether stood alone or stood with a trigger. You don't need vanilla interceptors in every deck either since not all decks have 6000 boosters beyond the Perfect Guards (which can scale with Grade 3s anyway and at most 1 would be called). Believe it or not, you don't need 8000 boosters or 12000 Grade 3s in every deck. They're nice to supplement or fill space, and they do go in decks that need them for proper columns, but required? Hardly. Exploiting 10000 vanguards isn't as strong anymore anyway what with the increasing number of 11000s for each clan and ride chain usage.


You will not get your selection by "listening" to your cards.
And if they talk back to you, seek help immediately.

Trigger Selection

This step seems like it might be the first, but not really. It can be if you're already sure of how your deck should work based on the Winning Image you chose, or based on what the clan is forced to have due to availability. If all your clan has are Rainbow triggers (1 of each type), then just use that. You're forced to, so do it. If you just have some typical selection of rainbow plus one extra of just one type, well, most of the time you're going to be shooting for 8 Criticals, 4 Draws, 4 Heals. But if you have a clan with a true selection like Nova Grappler, then you can make some real decisions based on the Winning Image and available units.

If your Winning Image was, as before, to abuse stand mechanics, then it should be pretty logical and easy to determine your triggers: Stands. So why then, do so many people make the mistake of running Criticals in Azure/Asura? There's this notion that the only trigger that exists is Critical. They're amazing triggers, but they're not the only one in the library. You have to think of each type of trigger as a tool for a job. Heals are going to be staple at 4. Just like Perfect Guards, there's no good excuse for not running 4 Heals in every normal deck. Remember that special creative decks aren't being talked about here. So now you can decide between 12 of any Criticals, Draws, or Stands. Some of those triggers will have skills that you need. Most trigger skills are not worth using. Since most of them are -1 for an effect that does not always translate into +1, you typically net a minus along with losing a 10k shield. Sometimes only 5k like in the case of Margal. In other words, you mostly will not decide triggers based on the skills they possess. If the skills happen to be useful after their trigger type is selected, go for it.

So back to the example. Why aren't Criticals a good decision for Azure/Asura? Well, they actually are. But they aren't too. It can go either way. The deck will not be optimal with Criticals instead of Stands, since you'll be just constantly using Asura to stand things for 1 stage (or a Death Army column, which is 2 regardless of selection). It's essentially two ways to play the same deck. The difference is that with the Stand variant, the Genocide Jack and King of Sword copies aren't going to waste. And the Stand variant has a very strong Late game in addition to the strong mid game given naturally by the skills present. The Critical version is just a "safe" method to play the deck that really makes two of your entire Grade 2 lineups wasted slots and falls back on the tried and true method of dealing hard damage fast instead. But then the deck becomes a hybrid and you have to switch modes in your mind to play it differently. Both have drawbacks and benefits, though it cannot be said that they are always going to be equal.

So now the question of which offensive trigger is settled. But do you run 12 stands or some mixture of stands and draws? If draws, what mixture? These are tricky tricky questions. 12 of any type of trigger says two things: the first is that the deck is completely focused on that mechanic. It gets advantage some other way or doesn't need raw advantage from draws. This does actually imply that some amount of draws are default—and they are. The second thing is that you could make this decision to have a net increase in shielding over running any amount of Draw triggers. Since draws have half the normal shielding of any other trigger, this keeps people from running all draws and taking advantage of the game. It also means that since they default at least 4; running all offensive activation triggers actually gives you more total shielding. About +20k more to your deck in total. So instead of "do I run draws?" the question is "why shouldn't I run draws?". If you can't answer that, you run draws. And probably 4.

If you can say why you don't run draws, it's likely for the defense reason or the focus reason above. But that doesn't automatically mean you run 12 Critical or 12 Stand. You can mix Critical and Stand for a weird offensive hybrid. Some decks can actually do this to great effect like Great Nature. Their ability to stand high powered units as well as granting them more critical are equally good. Those same decks typically have their own advantage engines and don't need draws. Therefore, they usually mix Stand and Critical together with some middle-ground amount like 6 and 6 or 7 and 5. The selection of which is usually personal preference.

Running the draws can be part of a standard setup. A standard for almost every deck is 8/4/4. With 8 being the offensive trigger (Critical or Stand) and 4s being the remaining two types. This is a "safe" amount that you can usually go back to with any deck. It generates half offensive power, half defensive/advantage power. using more rogue lineups are typically dependent on the special circumstances of the deck (such as mixing Crits and Stands). One often overlooked possibility is 6 Draws and 6 of some other offensive such as Critical. Tsukuyomi Oracle Think Tank puts this to great use by having enough draw power to facilitate the core loop of the deck, without sacrificing the critical offensive capabilities of the deck loop once you actually reach it. Knowing how to trade these off is going to be part intuition, part logic, part testing.


Aichi getting his empiricism on!

SCIENCE! Test it

And that leads to the most important part of deck building—testing. If you've been following thus far, you might now notice that deck building vaguely follows the scientific method. Where you've gotten an idea (Winning Image), gathered up the necessary data for it (card selection), and designed an experiment (deck). Everyone knows that after Hypothesis - Gather Data - Design Experiment - the next step is Run the Experiment. The biggest issue seems to be that people fall into the traps of this step. Humans are lazy and egotistical. Unless you have an extreme innate desire for scientific validity and empiricism, you likely are too. We want to think our deck ideas are perfect or that we can just play any old games and a few wins or losses will tell us everything we need to know about the deck. Wrong.

That is not how proper testing works. Firstly, you have to get a large sample size. No, your local tournament is not a legitimately large sample size. Something like 9 rounds does absolutely nothing to tell you how optimal your deck is. Your testing needs to be done with an expected error margin. For 50 tests, you're looking at a standard deviation of 2% performance. This is the error margin. So for 100 tests, it's 1% and 200 tests it's 0.5% (see the pattern?). You have to decide on a margin of error you are willing to take. Let me repeat that, because it bears repeating: you have to decide how wrong you want to be. How wrong you want to be. Science is about being wrong, not about being right. It's about being wrong constantly and fixing those things that were wrong. In science, you try to disprove the hypothesis. The hypothesis is your deck; so what you need to do is show that your deck is total crap and garbage. And every time it fails to be crap or garbage, it's closer to being accepted as a good deck.

But a large sample size isn't everything. You'd be wasting your time if the entire sample size was against some weird non-normal deck. That doesn't tell you how that deck is going to perform in the average scenario. Nor does picking a deck that gives you a strictly favorable or unfavorable advantage. Try picking either a neutrally opposing deck (this takes intuition and logic) or try running the experiment less times each but against many more decks so that you at least have a wider selection. You don't have to pit it against everything in the game, just a good sample size. If there's a popular metagame (knowledge about the game) such as Gold Paladins, you could just exploit that knowledge by making a ratio of your tests the approximate ratio of Gold Paladin players. Data is out there for the 2012 Championship that showed clan popularity and you can use it.

Furthermore; vanguard relies on three major key elements for a deck's victory:
  • Skill
  • Chance
  • Deck strength
This is a bad thing. You actually have to control for these in a scientific experiment so that you know the data you're getting back is truly relevant. You don't want data skewed by chance, wonky skill, or imbalanced decks. In fact the last one is one you're trying to test. In order to control for chance, you need that large sample size we talked about before. Honestly, anything less than 50-60 games is just child's play and nonsense. If you're serious about a deck, you need to test at least that many times. To control for deck strength, you have to take what's been shown as a "standard candle" deck and go up against that. Like I discussed before, these are decks in your meta or the like. They're proven to be sturdy decks. Things like Goku Kagero are very stable and consistent decks that scale well with any metagame. And you also need to control for skill. There are two ways to go about this.

You could just fight yourself. Shadowboxing is a skill set honed by competitive athletes and players in many other card games. If you're particularly good at compartmentalizing information that a given player should have (and pretending to be both players) and you have decent skill, then you've controlled for the skill element. This really requires you to be honest with yourself and your deck at all times. If you can't do this or you're not sure about your own skill level, then it's best to use the second method: play people who don't suck at vanguard. Find someone who's really good, maybe your equal or better. Play them with your deck several times. Then switch off and have them use your deck. Record the differences. If there's a difference-in-skill, then get better at the game. Truly, your ability to build decks ultimately lies on your ability to play the game.

This part is important: record everything properly. When a game ends, indicate which deck won and by what method. For me, I make sure to record abrupt chance-based wins properly. If a deck is at 3 damage and it makes no sense to guard at that time, if it gets double critical checked for game, that's obviously an abrupt chance-based win that the deck is not at fault for. You don't fail to record it. You do it anyway and count the win properly, but make a note that you saw abnormalities in that game. It isn't something like a really cool play at the end or a single critical. Those are all normal parts of the game. Do not be ruled by Impact Bias. Humans are egotistical and will easily fall into the tricks of misleadingly vivid results. Just because a deck seems to check more critical that others does not mean it does. If two decks have the same ratio of the same card, they should get them equally. Not doing that means it's just a weird chance fluke.

All of this may seem like hard work, but that's what it takes. The less work you do to determine how good your deck is, the less accurate you will be in your findings. As you should know, the step after Experiment is to Analyze the results. Like I said before, this takes good notetaking and shying away from impact bias and other fallacies. The key component of science is Peer Review: anyone should be able to repeat your results by running the exact same experiments. When I put up my decklists on V-Mundi, those are me offering my absolute best and most optimal results for peer review. So you have to look back at those experiments and make sure there are no oddities. If you're seeing a lot of chance weirdness, you need to run more experiments. If you see a lot of deck strength imbalances, you need to get a better standard candle. If you see a lot of misplays, you need to simply get better at the game. And lastly, you need to repeat the results after refining your deck. If you're not getting the results you want, fix some of the cards and start testing all over again. Put up your list for others to see, comment on and test. 

Above all, remain logical about your deck. This was not an article about how to construct a deck "for your play style" or "with art you like" or something of that nature. This is explicitly a way to make competitive optimal decks in vanguard. If you want to be creative, try gimmicks, use cards for art, or some other preferential thing, you should do that instead. It's not for anyone else to tell you what your payoffs for playing this game should be. If you're on a budget, you might not be able to put the best stuff in (though actually you can still use this guide, just with more restrictions). But if your payoff is to win at vanguard, I hope this guide can help you develop the skill set needed to build decks.

The outcome of the game is in your hands.
Related Posts Plugin for WordPress, Blogger...