Last Update: Feburary 21st, 2015

A blog dedicated to Cardfight!! Vanguard

  • Moved to Github!!

check it out at


Site has been moved

Take advantage of:

  • Faster loading times. WordPress and other bloatware trimmed out to speed up blog.
  • No Ads. This goes hand in hand with faster loading times, but now you have no obstructions in reading solid articles.
  • IntenseDebate comment system. Has a ranking system and smart threading to make sure good discussions stay on top.
  • Interactive elements. Check out the Hypergeometric Calculator on the side, or the program below.
  • Open source. The blog is hosted on Github, so just head over to my repo to suggest changes or use ideas for your own blog.

I’ll stop posting articles on here so it’s time to move forward. You can check out my most recent article on rationality and optimization right on the new site.

When the Dust settles

The great resolution

Clunky resolution systems are those with a lot of repetitive and redundant nature to it, and require constant attention and memorization to both players. When combined with poor effect text, symbols and structure it can make any idea fall apart. Out of all the card games I have been playing (including Online CG that’s about a dozen) I admire Cardfight!! Vanguard’s brilliantly thought out rule set, and after 3000 different cards there has been no more then a couple of line changes to accommodate new mechanics1. All rulings are made by reading the handbook line-by-line. In comparison; MTG is the largest card game and houses an equally large 200 page rulebook. The runner-up, Yugioh, has dug itself into a giant hole by deliberately avoiding clarification on conflicting rulings and leave to the the floor judge to make the best of it. Vanguard is by no means the best, perfect model for resolutions. And the other resolution systems are great models as well with many benefits. Nonetheless, CFVG is a brilliant model to study and develop from. Of course, you can’t take my word for it. So I’m going to go through the different kinds of systems and explain the flaws and improvements they’ve made from previous iterations of resolution design.

To be indefinitely board and inaccurate I’m going to define a card effect structure in three parts:

  • Trigger (Red)
  • Condition (Yellow)
  • Result (Green)

If you are hoping to learn the details of actually playing these sample games, Ignore everything I write down and instead point to the references provided at the end of each section. Not that it’s wrong, but rulesharks exist to make you fall into infinitesimal loopholes and penalties. As I have the most experience with MTG, YGO and CFVG I’ll cover those primarily. You can thanks Blaise for indirectly aiding me to add one additional section.

The LIFO stack

I’ll take the 6th definition from wikitionary’s database to describe what a stack is
(computing) A linear data structure in which the last data item stored is the first retrieved; a LIFO queue.


(computing, accounting) Last In First Out (describing a stack data structure).

This is the most prominent standard in TCG’s. The core aspect relies on this:

  1. Play Card A. A goes into the “stack”.
  2. Play Card B. B goes into the “stack”.
  3. Play Card C. C goes into the “stack”.

To fill a stack and this:

  1. Take Card C out of the “stack”.
  2. Take Card B out of the “stack”.
  3. Take Card A out of the “stack”.

To empty the stack. The key thing to note is that you can not take Card A out of the stack before Card B is taken out. And Card B can’t be taken out before Card C. The Card on the bottom can’t be removed before taking out the ones on top.


FYI, spell := Card

To play a spell in MTG, you first must meet the Condition.


  • If you just played a spell, it is added to the stack, and you can either play another spell or let the opponent play a spell.
  • If your opponent just played a spell, it is added to the stack, and they can either play another spell or let you play a spell.
  • If neither player wants to play a spell (rather, both players consecutively ‘let’ the other play), then the stack empties.

However, the stack isn’t just dumped back out immediately like pudding. It goes one by one; a spell is taken out of the stack, the Result happens, and the active player2 can once again choose to play another spell (and hence goes back to [+]). Both players have to once again decline to play a spell before it the next spell in the stack Results.

Continuous effects are like “rules of the game” once in play. You would continue playing as if the game always had that rule. If you played it during a stack, once the Result of the card has past you treat the game as if the rule always existed.

Now we get to “counters”, which is really why WoTC3 decided upon a LIFO system. On the resolution of spells that “counter” another spell, the spell chosen to be targeted has all of its text removed, as if it is a completely blank card. When the stack reaches to that card, it ditches the card. Counters exist to stomp and roll all over an opponents plan. Say there was a card who’s Result would win the game. You don’t want that to happen so you play a counter spell to prevent that card from Resulting, therefore buying you time (possibly enough to win the game yourself). This miracle/comeback play is necessary to keep games dynamic and unexpected. If you think about it, it is possible to categorize effects into three areas: those that start a stack (the seed), those that are played in response to a stack (the counter), and those that do not start a stack or can be responded to (the passive/continuous).

For the finer details of stack and priority thank Yare for this well written article:

As the activate player, you can cast a spell every step of the stack resolution. Asking for each iteration being forced to announce every decision is incredibly tedious. A stack could be near the end and you can decide to cast something and build another long stack, making the constant unpacking and packing of cards is difficult to track. Long stacks are uncommon as it requires a large amount of mana and cards. But if you cast a spell that let’s you draw cards / add mana, then you grow the chain (because you get a chance to play every time a spell resolves). Have enough of this and you can get infinitely long chains. The MTG engine does not promote huge hands, extensive drawing capabilities or infinite mana, to combat the possibility of these large/long stacks. It does however, limit the growth of resources and therefore makes the game exceedingly long. A lot of this game is spent waiting to gain enough mana, cards, waiting to toolbox, draw non-land.

So the big problems MTG resolution system has are:

  • Large memory strain for ordered stack events
  • Redundancy in saying no, especially for long stacks
  • Constant peckering of both players
  • Resources are forced limited to reduce stack length4


FYI, chain:=stack

To activate a card in ygo, you first must meet the Condition.

  • If you just activated a card, it is added to the chain, and you can either activate another card or let the opponent activate a card.
  • If your opponent just activated a card, it is added to the chain, and they can either activate another card or let you activate a card.
  • If neither player wants to activate a card (rather, both players consecutively ‘let’ the other activate), then the chain empties.

I’d like to point out that Yugioh cards have a “Spell Speed” which acts like a miniature condition. You are not allowed to add to a chain if the Spell Speed of the Card you want to activate is lower than the Spell Speed at the top of the chain.

Just like before, the chain resolves one by one. Unlike stacks, chains do fall like pudding and all the Results happen one after the other with no interruptions. Here is where we get interesting. The Result or Condition of a Card could Trigger a different Card. It could Trigger multiple Cards. So when do we get a chance to apply its Condition or Result?

The simple answer is when the Result of the last Card in the chain is done. But then how do we determine the order of all these Cards? YGO has implemented SEGOC5 rules to resolve this issue. The order basically takes all the Cards that have Triggered and allocate them on a new chain based on a few rules.

Continuous effects also fall into the SEGOC rulings, so the Result or Rule the card would implement will not start until all the chain has finished (but before SEGOC clause happens). Note that this isn’t a real solid rule, as there are two rulings about these effects that conflict each other, but effects in this way make it more consistent to the practice established.

For more information about SEGOC try this article about two similar effects

Otherwise, you may explore the entire blog

Unlike MTG, YGO tried a new method by forcing chains to completely resolve before play resumes. That cuts out “I choose not to cast” by over 70%6. This also stops chains from being refilled constantly making each chain much shorter on average then stacks.

But SEGOC and Triggers start the same deadly loop. At the end of a chain, a new chain starts with all the Triggered cards put into a new chain, which CAN be included in. So YGO just splits one massive chain into a series of smaller ones. Unfortunately, since the entire chain resolves, having copious amounts of cards that draw or add more cards into play will refresh stocks every time. In MTG by the time you get down to a particular spell the available the conditions and spells in the stack have changed. In YGO it starts all over so while you have a different scenario the abilities in play are all there. So you can tell YGO has to have the same restrictions as MTG to demote extensive chaining. I should also say that users barely have an option of how to order the different effects into the new chain.

So what problems did YGO solve or create?

  • (+)Memory strain is much less
  • (+)Redundancy is improving
  • (+)Less peckering
  • (-)Multiple chains that have no optional order
  • (-)Limited Resources to reduce continuous chains
  • (-)SEGOC Chains have limited user input

Although this last point does not concern the point I’m getting at, YGO has done a significant improvement in the understanding of card effects, now making them both consistent and logic-based.


So YGO and MTG each stack and chain off of each other ad infinitum, which forces their cards to cut significant amounts of resources. Buddyfight aims to shorten the unnecessary drive with a 1 counter clause. Therefore, using cards now works like this:

The turn player chooses to use one card or ability
The non-turn player may choose to use one Counter
If they do
	The turn player may choose to use one Counter
	If they do
		Resolve the turn player's Counter
		Resolve the non-turn player's Counter
		Resolve the turn player's card or ability
	If they don't
		Resolve the non-turn player's Counter
		Resolve the turn player's card or ability
If they don't
	Resolve the turn player's card or ability

Likewise, a MTG tower would be like:

The turn player may choose to play a spell
	If they do
		The turn player chooses to play a spell
		If they do
			The turn player chooses to play a spell
			If they do			
		If they don't
			The non-turn player may choose to play a spell
			If they do
				The non-turn player may choose to play a spell
				If they do
	If they don't
		The non-turn player may choose to play a spell
		If they do
			The non-turn player may choose to play a spell

It’s a simple and elegant solution. It doesn’t invite impressive plays, but it solves the problem of large complicated loops.

As it is so simple, my debriefing ends in exactly two sentences due to having nothing to talk about.

So what problems did FCBF solve or create?

  • (+)Almost no memory strain
  • (+)No redundancy
  • (+)Almost no peckering
  • (-)Stack has very little input
  • (-)Combos become impossible due to simplicity

The cache

So we have folly’d with this “ordered” resolution for a long time. After all, it works and everyone is content. We want combos and counter-play, and sacrificing that causes people to feel like the game is overtly simple and childish. That isn’t absolutely true, but it certainly reduces the depth of the effects that effects have.

So this is where I show what neat tricks Vanguard pulls to get the best of both worlds. We start with the definition of a cache
(computing) A fast temporary storage where recently or frequently used information is stored to avoid having to reload it from a slower storage medium.

Notice that in a cache, order is unimportant.


To activate an effect, it has to meet it’s conditions.

  1. Once you activate it, you Resolve it.
  2. Then for all effects that meet the Trigger, you add those effects into the cache.
  3. The activate player choose one effect that belongs to them and Resolves it.
  4. Then for all effects that meet the Trigger, you add those effects into the cache.

3 and 4 continues until you have no more effects left that is yours. It then becomes your opponent’s turn to resolve all of his effects, until he is done. It keeps going back and forth until the cache is empty.

Counters do not exist in the game, and continuous effects come into play as soon as they are put on the table (they do not go into the cache nor Trigger any effects)

WOW. That is really different from any other card game out there. Let’s see what problems it solves:

  • It reduces 100% of the “not choosing” plaguing card games. Redundancy is completely obliterated.
  • It reduces the need to constantly flip between two players
  • Card Condition and Result activate immediately. Exactly what you would naturally expect7
  • And unlike FCBF’s solution, it does not simplify the game or reduce combo potential.

The biggest or most obvious part about all this is the removal of counters. But actually, you can still create counter effects (When your unit is retired, call it to {R}) that are actually more interesting than normal counters. MTG/YGO counters are more like a stand off doing almost jack squat, or firing bullets at a Gundam shield. Vanguard is dynamic by letting you succeed, only to be caught by a surprise resurrection or counter attack. You can pull off your dream combo instead of it being violently stuffed by a single spell, and such counters do not merely neglect an action but turn the disadvantage into an advantage in another way (When your unit is retired, send a card from your deck to {Drop})

So what problems did CFVG solve or create?

  • (-) Some memory strain
  • (+) No redundancy
  • (+) No peckering
  • (+) Combos are completely free and malleable

Based on this small list it seems Vanguard has solved absolutely everything and there is little need to iterate further. However, there is a really really big hole in this.

Let’s have 2 card with these effects;

AUTO(R): During your End Phase, When you draw a card, you may pay the cost. If you do, this unit gets +1000.

AUTO(R): During your opponent End Phase, When one of their units gain power during the End Phase, your opponent draws one card.

Based on the way vanguard works, this can result in an infinite loop, provided there are enough cards in the deck to support the pair. While I documented a 2 card case, it could be viable for any number of units acting together. So that forces most cards to have a counterblast cost or some other cost to make sure if never repeats itself. (If have to say, that is a dumb effect above but it was only used to get the point across).

So if you want to get picky, here are some problems it suffers.

  • No counters to actually stop plays, therefore creating unstoppable wins8
  • All effects should have to prevent self-occurring loops
  • Logic base language does not work with deformed grammar9


So we had a short article here discussing the various ways to resolve multiple effects, optimally keeping it consistent, open to flexibility, reduces redundancy, and reduces load on memory for more important tasks. While there is no prefect system as so far, nor have I talked about any horrendous ones, the ones I have mentioned are all successful games. So while a well-defined rule set does not correlate to popularity, we know that people are willing to accept imperfections and devise short-cuts in order to enjoy the game keeping the rule sharks out.

Lastly, If you know a card game that has a completely unique but brilliant resolution system you can leave a link below.

  • I wrote this before the Stride and Deletor mechanics. Oh forbid.

  • The active player is the player who is performing his/her turn.

  • Wizards of the Coast, creators of Magic The Gathering.

  • Some may find the last point a positive since it gives you ‘more’ interaction, but in practice it’s very cumbersome and just plain messy. Luckily people have invented short-cuts to go through most of the talking but against a step-by-step AI…

  • Percentage to be taken with a grain of salt. If I play a card you counter I counter and we resolve, YGO has to say it EXACTLY 4 times while mtg is AT LEAST 9 times. But the longer the chain, the more it compiles.

  • I mean, when you throw a ball, would Time Stop for your opponent to reposition themselves?

  • Which, if you think about it, is why [HEAL] were created. Bushiroad’s developers really did think of everything.

  • To be covered in a future article?

Melodramatic Final Turns

It’s over 9000(0)!

*It’s so mainstream you’ll hear your great-grandchildren say it.

tl;dr – The data

When looking at a ‘final turn’ gambits people consider how likely your opponent can walk out unfazed, and whether the setup is just plain absurdity. So I’m going to go from very standard finishers to the 0.01% miracles and find out just when is it a plain overkill. We can use TehNACHO’s B.A.S.S values, but we are inclined to forget an extra stage sitting on every vanguard attack (for safe measures) and the 0.55 stages per attack due to triggers.

The Notation

Let’s consider an early Set 2 ‘Final Turn’, riding soul saver dragon. In essence this move dealt a 3/3v/4 or about 11.55 stages (A 3-stage rg, a 3-stage vg, and a 4-stage rg makes 10 stages; then add the +1 from the vg and +1 for triggers). Sitting with \60/ makes living through this turn no sweat. Though we are missing one type of card that would make this even sweeter; can you guess what it is?

Blau Dunkelheit

Yep. Perfect guards are able to shun out an entire 4 stages of Soul saver rage meaning you no longer have to worry about stocking up on a 6 triggers, making sure your guard stays nice and fluffy. This means SS[Final] is more around a 7.55 or \40/ shield. In fact, if you decide to guard the 4-stage RG instead, it drops to 7 or flat \35/.

Lastly, we also need to exercise extreme caution in using final turn gambits when the opponent is only at 4 damage. Since we know that everybody uses 12 Crit, we are inevitably forced to take a rg hit in order to stay in the game. At this point, we have a 1/3 chance to check a trigger and reduce all future stages by 1. If we put that into the already reduced stage count, we get a pleasurable 3.67 (PG the RG!) or \20/ and a PG to secure the chance of survival.

So the notation. I’ve thought about the different ways to go about it, but I think I’m going into one looking like this:

{# of Perfect guards} [ {stages for 5 damage} | {stages for 4 damage} ]

If you consider the entire SS:Final, the results look like:


0	[	11.55	|	7.88	]
1	[	7 	|	3.67	]
2	[	3	|	0	]

What are you looking at specifically though? It’ll be mainly the middle-right number. This is a frequent scenario in which your opponent is holding a PG captive while sitting at 4 damage, and you are considering on whether to make a break for it or not. If you are using SS:Final you need your opponent to have less than 4 cards, 4 cards with max 2\5/. As you can tell, SS:Final is bridging on the weakest final turn combo.

For quick reference, divide all the numbers here by two and add two for every PG to get the “Approximate number of cards in hand to live”. It isn’t accurate, but nonetheless you can see that you need to have the opponent to have < 6 cards to win at 5 damage, and < 4 cards to win at 4 damage.


0 {6|4}
1 {6|4}
2 {6|0}

Let’s add some fire. Introducing Palamedes and his good friend, Palamedes. So 4/4v/4 makes:


0	[	13.55	|	8.33	]
1	[	8.55	|	4.67	]
2	[	4	|	0	]

Which is just one stage more. Wonderful.

So from now on you’ll see this kind of notation pop up to judge the strength of final turns. And to judge them we would need a reference point or a ‘standard’ final turn. This is shown below for now, but we will derive this later on.


0	[	14.55	|	11.22	]
1	[	10.55	|	7.55	]
2	[	7	|	4.33	]

Kinds of Final turns

Except if you are going Glendios, a finishing blow come from getting enough hits in to seal the deal. This mainly falls into four categories:

  1. Power up vanguard
  2. Power up rearguard
  3. Attack with vanguard multiple times
  4. Attack with rearguard multiple times

Keep in mind the relative strengths; 5 1-stage attacks vs 1 5-stage attacks may favour the former, but 1 15-stage attack is slightly tougher to judge. So I’m just going to show off a bunch of different Final Turn situations and generalize the capabilities of each one.

Power up Melody

Let’s start with the easiest pair.

Powering up the Vanguard to nonsensical levels usually begs a Perfect guard to the face, but troublesome if they don’t have one. Then again, this is also beneficial to the rearguard situation too so this is probably the final turn ability you’ll want to avoid (despite the show drooling over these kinds of finishers).

I’m sure everyone is bored to death with the +10000 abilities, and to be completely honest that 39k Tyrannolegend or that 45k Ezel isn’t in the top-tier of vanguard power. I am also going to leave out DI for the next section.


Cosmolord had a hilarious combo way back in Set 3 that no one really tried. Combined with a BR, let’s assume a really typical scene of 5 rearguards and one being Anthrodroid. From reading these two effects, you can reason a combo of standing&resting Anthrodroid multiple times pumps Cosmolord by 15k. If you rest all the rearguards as well, you get another +15000.

Here is the coolest part. If you have Hungry Dumpty, Claydoll Mechanic, Muscle Hercules, or even Red Lightning in hand you get a free 6k (one for the rest, one for the unflip). If we also had Tough boy in hand the final output of Cosmolord is [58 + 6x + 3y] where x is the number of those extra unflip units and y is the nonflippers. You only need 7 cards in hand to reach…


*Death Army Bishop and Knight are actually more cost-to-rest efficient but even then this is just an absurd idea.


0	[	19	|	19	]
1	[	0	|	0	]

So next is rearguard power and oh boy do they like to do this one. So many decks imitate this but nothing comes close to the insanity of Dantarain+Abyss. It absolutely ridiculous to calculate the power because of how irregular the power tends to be. This is influenced on:

  • The amount of soul before the breakride
  • The number of Dimension Creepers to be blasted (Note extra soul charges can net another one)
  • The plausible use of Shirley and other soulchargers
  • As a fixed constant, Greedy hand will almost always bring in a Creeper

But let’s try a rough sketch of how powerful this can be. We can safely assume at least two but usually three creepers, a greedy, at least one other soulcharge unit, and 10 soul? So the math gets a bit wonky but trying to work it out

Just on the BR (w/o the CONT skill)

Generic		Abyss(10)	Generic
10		21		10
Doreen		Greedy		Doreen
6		5		6

After greedy and 3 creepers and stuff

Generic		Abyss(16)	Generic
10000		21000		10000
Doreen		Doreen		Doreen
30000		24000		30000

After Abyss x 2

Generic		Abyss(20)	Generic
10000		59000		10000
Doreen		Doreen		Doreen
42000		36000		42000

And finally the CONT

Generic		Abyss(20)	Generic
10000		59000		30000
Doreen		Doreen		Doreen
62000		36000		62000

Or for the 13/17v/17 stages:


0	[	48.55	|	34.88	]
1	[	30.55	|	17.22	]
2	[	13	|	0	]

Even at a portion of the power (2 less creepers, maybe 5 soul to start) you are still looking an incredible numbers here. Just for a check, compare this to “SS+P:Final” and you’ll find this to be close to 4 times better.


Reality Witch, RiasIronically, I won’t be discussing the pattern for Galaxy Beast, Zeal or it’s Gundam counterpart. Instead, we will stare into the eyes of these witches and their extremely powerful curses. At Legion’s best they can drop out anywhere from 25k to 45k from the opponent, which is the inverse of having all your columns gain the same amount. [STAND] will make this a nightmare but let’s just say that zealing is one of the greatest finishers in vanguard. I believe you can compute a matrix for a 11/13v/11 so I’ll leave this one aside for now.

The Emperors of Strength

Who can talk about multiple rearguard attacks without a mention at Aqua Force? So let’s bring up on of the best representatives of the AqF: Genovious. I know Maelstrom and other 3rd/4th types are also relevant but when it comes to multiple attacks this guy delivers.

Let’s also live in a pretend world where the opponent’s vanguard is 10k and we have a perfect field. Originally I had 4 Lysanders on the field, before realizing I can’t Counterblast to get the last attack so I replaced one with Diamantes. So on the breakride of Trans-core:

Diamantes Genovious Lysander

The resulting stage attacks goes something like 1/1/1/1/5v/1/1/1/1


0	[	14.55	|	10.88	]
1	[	8.55	|	5.22	]
2	[	7	|	4	]

One thing multiple rearguard attacks have over power columns is that the bottom-right value won’t be zero: The opponent will need additional guard even for the best case (4DMG and 2PG).

So this is where I would get into Nova grappler since they had that standing premise back in Set 1, but I’ve promoted them to a better position. So instead we’ll talk about the dirty Spike Brothers and one of the first examples of true final turns. The Emperor+Juggernaut combo has been notorious in delivering extremely powerful rushes. To keep this short, the following chart serves as a kind of ‘standard’ for a respectable final turn:

{Emperor+Juggernauts-BadEnd, and 4 Juggernaut attacks for 2/2/3v/3/3}

E+J-B:Final (or Standard:Final)

0	[	14.55	|	11.22	]
1	[	10.55	|	7.55	]
2	[	7	|	4.33	]


0 {7|6}
1 {7|6}
2 {8|6}

In numbers this is slightly superior to Genvious but in practice this combination is much more severe. Here I can explain the shortcoming of the simplified version. Looking at the 6’s down the side, and seeing that we have 4 attacks around 3 stages, the opponent really needs 8 or more cards to ‘ensure’ safety. I’ll be best to memorize what the opponent has in hand, but for those without solid short-term solutions then these keys here are easier to digest.

Bad End Dragger

With Set 10 we gained access to Bad End Dragger, one of the most insane skills in the game. While I would love to provide a chart for Bad End and Emperor, Set 15 threw gold at us with Bloody Ogre and Frozen Orge. Frozen extends an additional 3 attacks by and both vanguard skills give us +3 stages for any attacker. This is quite a force to be reckoned with and even hearing Frozen -> Frozen -> Frozen -> Silver sounds scary. The 3/3/3/5/6v/7/7 becomes


0	[	35.55	|	29.55	]
1	[	28	|	22.33	]
2	[	21	|	15.67	]

This is very very good. Interesting to note the clashing with Abyss:Final, who has a stronger top-half, but falters in the bottom half. It means that Abyss:Final is a strong gust that will force that last damage through, while Blood:Final is battering ram to strip all of your opponent’s remaining health. Bloody:Final is slightly better for those 5 to 3 damage match-ups as a comeback mechanism, but Abyss makes your death at 5 inevitable.

Re-standing 101

So now we are into the realm of re-standing vanguards (Or rather, multiple vanguard attacks). They have the benefit of always forcing an additional stage of their calculated values, gaining additional power from break-rides, fizzing the effectiveness of counter triggers, and multiple twin drives. Let’s start with the often called lamest one of the bunch — Dragonic Descendant and its 3v/4v/2/2 rage:


0	[	14.1	|	12.1	]
1	[	9.1	|	7.1	]
2	[	5.1	|	3.1	]

While it is about 1/2 a card worse than the standard, remember that the standard requires 2 Juggernauts, 2 counterblasts, becomes a self minus 3.45, and empties all attackers if the opponent manages to survive. Descendant however requires just 1 card, 1 counterblast, and is actually a self plus 0.1. Since the net effect is a plus, you can run the same final turn again. So I guess it isn’t a complete stomping final but the near perfect consistency makes Descendants game extremely difficult to face. If you have even one 3 stage column or pumped the vanguard by 5k, then you are looking at something that is strictly better than the standard.

The supreme successor to Descendant looks to be Omniscience of the Sacred Regalia, Minerva. In addition to her own ability we have Witch of Frogs and Witch of Birds that allow for rearguard attacks. Read this again: Multiple vanguard and rearguard attacks. This is the guide to death for most of today’s format. Nothing can come close to stopping so many insane attacks. But enough chit-chat, you want numbers. So assuming a Himiko BR granting an attack lineup of 3/4/4v/4v/1/2 creates:


0	[	21.1	|	16.43	]
1	[	16.1	|	11.77	]
2	[	11.1	|	7.1	]

And this sanity creates a +2 (on average. I know this isn’t Minerva’s full extent with the double soulblasting but it would not be much more then 1.5 cards)


Cat Butler

Cat Butler. Kat Bulter. KaxBuller. HaxBullet.
When the ruling was announced that Cat Butler can stand NG Legions, the world went crazy building Raizers. And justly so. After a bit of playing around, I found out that a pure (read: no Frozen Ogre fetchers) Nova deck has the best output with 3 Cat Butlers and two mates. With 3 Butlers, your vanguard will be standing three times which means 4 vanguard attacks. That’s like an octa-check or octative drive. 8 free cards and the chance of getting multiple triggers is immensely high. Then every time you attack with your vanguard your fellow rearguards also stand and gain a stage, so this snowballs into itself. The attack will looks something like this:


14 attacks! Is that even legal? Then you’ll love to see the table becomes:


0	[	45.2	|	39.87	]
1	[	39.2	|	34.2	]
2	[	34.2	|	29.53	]

And for reference


0	[	48.55	|	34.88	]
1	[	30.55	|	17.22	]
2	[	13	|	0	]

This just wounds Abyss in so many places. Multiple Raizer re-stands are terrifying and it justly serves Nova Grapplers as Tier 0, churning out quality competition from the very beginning. Make no mistake, Raizers are the embodiment of gruesome fatalities.

The king of vanguard overkill

So would it be surprising to know that this Nova Grappler deck ISN’T EVEN NG’S FINAL FORM!? It has ben personally coined in the Vmundi community as “Za Warudo”, and you’ll find out why in a moment.



Meet the Samurais. They each gets 3k for every Nova Grappler restood.



Meet Mond Blaukluger. He stands all rearguards, so that means all Samurais get (3000*5 = 15000) a 3 stage boost.

Asura Kaiser

Meet Asura Kaiser. He allows the Samurai condition to activate.


Now kneel before the king of vanguard, Immmortal Asura. He is a self stander and so he activates Samurai’s effect.


That activates this dudes skill


And triggers these guys again.

So what do you get, when you have a field of Samurais who gain 15000 when you attack with the vanguard, and gain 15000 again when you swing again? Oh, something like…

(1) 16000
(2) 16000
(3) 45000
(4) 52000
(5) 57000
(6) 73000
(7) 87000
(8) 92000

Have some fun with 2/2/8v/9/9/13v/15/15 and you get

Za Warudo:Final

0	[	76.1	|	71.77	]
1	[	60.55	|	56.55	]
2	[	45.55	|	41.88	]

Let’s do a massive reference right here to show you that nothing here even comes close:


0	[	11.55	|	7.88	]
1	[	7	|	3.67	]
2	[	3	|	0	]


0	[	13.55	|	8.33	]
1	[	8.55	|	4.67	]
2	[	4	|	0	]


0	[	14.1	|	12.1	]
1	[	9.1	|	7.1	]
2	[	5.1	|	3.1	]


0	[	14.55	|	10.88	]
1	[	8.55	|	5.22	]
2	[	7	|	4	]


0	[	14.55	|	11.22	]
1	[	10.55	|	7.55	]
2	[	7	|	4.33	]


0	[	21.1	|	16.43	]
1	[	16.1	|	11.77	]
2	[	11.1	|	7.1	]


0	[	35.55	|	29.55	]
1	[	28	|	22.33	]
2	[	21	|	15.67	]


0	[	45.2	|	39.87	]
1	[	39.2	|	34.2	]
2	[	34.2	|	29.53	]


0	[	48.55	|	34.88	]
1	[	30.55	|	17.22	]
2	[	13	|	0	]


0	[	19	|	19	]
1	[	0	|	0	]

Conclusion, The TRUE king of vanguard overkill

Thought this just an oversight? That’s because it probably is. But you know Bushiroad, they can’t let their pet clans get overshadowed by Mega Drill equipped Gurrens. So they snuck a small combo in Gold Paladin that is able to loop into itself infinitely. Basically it requires a Liberator Margaux, Blaster Blade Liberator, BBL’s unflipper, and a plusser every time you superior call. The details have just been released but you’ll have to wait for my optimization of this build (and the others I’ve talked about in this article). But for now:

c-c-c-c-combo breaker!!!



Ever had a non-trigger card in your hand and wished “Man. I hoped I could have gotten this card one turn earlier” or its converse “Why couldn’t I get this card a little bit later, I need to guard right now”? What you have experienced is usually known as ‘dead’ cards. With all the moving pieces in any sort of deck events like these are bound to happen. I’d like to present analysis on this property [Which has been coined ‘card entropy‘]* and provide a quantitative way to measure how much ‘card entropy’. As an overview, I will be:

  • Defining ‘regular’ entropy and ‘card’ entropy (which I would later just refer as entropy)
  • Explaining unavoidable and negligible entropy
  • Listing steps to measure card entropy
  • Dividing entropy into sub-games
  • Testing card entropy
  • Using a few decklists as examples

*I do want to note there is a difference between this definition of entropy and mine.

Defining Entropy

Thermodynamic entropy is a measure of the disorder in mass and energy. Thermodynamic entropy increases over time because the universe spontaneously disorders itself. Spontaneous means that there is a natural tendency for something to become disordered (contrary to the believe that it happens “instantaneously”, which is why spontaneous combustion is redundant and stupid). When you shuffle a deck it becomes more random, and hence more disordered. We perform tests to determine entropy to be able to make it more efficient (aka, reducing entropy). It is helpful to note that there is no solid way to calculate entropy alone*, but entropy changes are possible to measure. Lastly, I might use (del)S as notation for the change in entropy if I feel lazy.

*There is a such concept as absolute molar entropies and stuff but we aren’t delving into that today.

In general card entropy is loss of value in a card. This loss can be reflected in its inability to generate a plus, having a lower plus then normal, or just being unable to use the effect. With the concept of sub-stages, we can reasonably assume that the contributions of a card is split into these stages. Three possible situations exist:

  1. Activating a card effect in any sub-game returns the same value compared to the other two.
  2. Activating a card effect in one sub-game returns a lower value compared to the other two.
  3. Activating a card effect in one sub-game returns a higher value compared to the other two.

For (1), the entropy changes for these cards is zero. Perfect Guards are decent examples for this phenomena, since it is equally useful in all stages of the game. Number (2) is rare but is found mainly for units that superior call like High Dog Breeder Akane. By far the largest section is (3) which is just about every single clumped limit breaker / break ride card in existence.

To continue this train of thought we’ll let every three-four copies of a card be represent just one piece. For example, if we had 4 Akane (a type-2 card) Then in the specific sub-game she does not excel at (the endgame) the (del)S will increase.

The hardest for me is trying to pinpoint how entropic would a card be*. I went about it like this: Given a set of 3 or 4, if the card would not be used in that sub-game I’ll give it a solid 1. If it did have a use, but the condition to use it may be decently improbable, I’ll give it a 0.5. Point fives have also been given to on-hit rg skills whose effect diminishes and having a hard time to land. If you have a set of 1 or 2, you halve the entropies. Due to the nature of entropy, it is best to have a separate value of entropy score for each sub-game.

After we calculate (del)S for each sub-game, we can organize the information in a chart, and then provide terms to describe the build. Some terms here are provided:

Early entropic Options unavailable in the early game
Mid entropic Options unavailable in the mid-game
Late entropic Options unavailable in the endgame
High entropy Inconsistent, sub-game specific plays
Low entropy Consistent, sub-game non-specific plays

*Please note that I’m still trying to work out my thought process in deciding how to measure it. It ain’t an exact science.


Now before we get to deeply into this I’ll mention the very common entropies that is expected in all decks. Since I only want to get numeric values of units that separates it from another deck, we need to exclude some forms of entropy. Firstly it is reasonable* to expect that once your back row is filled in there is no need to replace the cards with others. This means that by the endgame all grade 1’s should have +1 (deL)S each. Obviously we can’t have this erroneous data floating about so we’ll cut that out. Same case applies to grade 3’s in the early game (Don’t you just want to draw a grade 3 whenever you need it and never extra?). So when you apply entropy evaluations you only take the set of cards which exists in that area.

Another form of entropy arises from triggers. In the endgame criticals don’t perform as well as they would during the rest of the game. Stands then take precedence in this area and actually outperform criticals. Since most decks should have the same line-up to avoid gracious flip-flops we can ignore them too. In general we just disregard the entire trigger setup.

*I know the retirement decks are quite relevant in the meta, but since there is no solid counter to this I don’t really have much to say.

The Test Run

These are close to the optimal builds for both decks, in case you didn’t know:

4 Dudley Emperor 4 Eradicator, Dragonic Descendent
4 Bad End Dragger 4 Eradicator, Vowing Sword Dragon
2 Juggernaut Maximum
4 High Speed Brakki 4 Supreme Army Eradicator, Zuitan
4 Dudley Mason 3 Fiendish Sword Eradicator, Cho-ou
2 Charging Bill Collector 4 Eradicator, Spark Rain Dragon
4 Cheer Girl Marilyn 4 Eradicator Wyvern Guard, Guld
4 Reckless Express 4 Sword Dance Eradicator, Hisen
4 Wonder Boy 4 Eradicator, Demolition Dragon
1 Medical Manager 2 Ceremonial Bonfire Eradicator, Castor
1 Mecha Trainer 1 Eradicator, Strike Dagger Dragon
8C/4S/4H 12C/4H

Early Game

So at this point, we do not consider Grade 3’s.

1 Mecha Trainer 1 Eradicator, Strike Dagger Dragon
Skill can be used early to quickly fetch any grade 1 to facilitate the field {+0} Used for superior riding Descendant,If you don’t have him in hand {+0}
1 Medical Manager 2 Ceremonial Bonfire Eradicator, Castor
As a 1 copy, it doesn’t really matter what is here, it is almost always +0. {+0} When you have two copies, you halve the entropies. Here Castor’s wash is sightly 0.5, but halving it makes it negligible +0. {+0}
4 Wonder Boy 4 Eradicator, Demolition Dragon
Great for rushing and general support {+0} Great for rushing and general support {+0}
4 Reckless Express 4 Sword Dance Eradicator, Hisen
As this deck is inclined to the BR, Reckless’ effect wouldn’t be used at this point. It technically can work, but in general it is sitting there. {+1} In the early game this facilitates the superior ride,So this is useful. We aren’t going to get it every game,but certainly when we don’t have EDD in hand. {+0.5}
4 Cheer Girl Marilyn 4 Eradicator Wyvern Guard, Guld
Never a dull moment {+0} Never a dull moment {+0}
2 Charging Bill Collector 4 Eradicator, Spark Rain Dragon
Bill Collector is a solid beat-stick {+0} Another solid beat-stick to throw early {+0}
4 Dudley Mason 3 Fiendish Sword Eradicator, Cho-ou
Puts on the pressure since turn two {+0} I’m kinda mixed on this since you won’t always use it but this is an option you would sometimes activate {+0}
4 High Speed Brakki 4 Supreme Army Eradicator, Zuitan
Like Reckless, I’ll increase entropy here {+1} This can serve two uses, one for the ride chain, and another for its pointless on-hit {+0}

Middle Game

Now we add in Grade 3’s. (I’m omitting the cards that are {+0} to save space)

1 Mecha Trainer 1 Eradicator, Strike Dagger Dragon
Skill is still viable {+0} At this point he does nothing but as one copy, he isn’t hurting anything either {+0}
4 Reckless Express 4 Sword Dance Eradicator, Hisen
Not yet. Just wait {+1} Nice descendent booster for any kind of push, but this requests you have 2 of them. {+0.5}
4 Dudley Mason 3 Fiendish Sword Eradicator, Cho-ou
The opponent has enough to stop this threat and likely you’ve already built a field too {+0.5} Fair game here {+0}
4 High Speed Brakki 4 Supreme Army Eradicator, Zuitan
Like Reckless, I’ll increase entropy here {+1} What are you counterblasting at this point? {+1}
2 Juggernaut Maximum
You would usually tag these so you can recall them later {+0}
4 Bad End Dragger 4 Eradicator, Vowing Sword Dragon
All the Limit Break cards are locked out {+1} {+1}
4 Dudley Emperor 4 Eradicator, Dragonic Descendent
{+1} {+1}

End Game

By now, the row is filled and our grade 1’s will leave the set.

4 Dudley Mason 3 Fiendish Sword Eradicator, Cho-ou
Still okay if you miss Brakki/Juggernaut {+} Still okay here {+0}
4 High Speed Brakki 4 Supreme Army Eradicator, Zuitan
It’s time to show your stuff {+0} Will it even hit? {+0.5}
2 Juggernaut Maximum
4 Bad End Dragger 4 Eradicator, Vowing Sword Dragon
All the Limit Break cards are locked out {+0} {+0}
4 Dudley Emperor 4 Eradicator, Dragonic Descendent
{+0} {+0}


To put the numbers together in a nice chart:

(del)S in its respective sub-games

Subgame BED EDD
Early +2 +0.5
Mid +4.5 +3.5
End +0 +0.5
All +6 +4

We can kinda guess BED has a higher entropy then EDD, and it should be obvious. After all, they are less consistent and suffer from a more clumped prototype. Everything is used to set up the one-turn explosive crush rush and that leaves no room for any other gambits. But let’s go into each sub-game:

  • Early :: BED does have Brakki and Reckless, but unfortunately it is unwise to use their effects at this point. (For BED-specific decks at least) Hence the large entropic gap as your options are limited. EDD however has a chance to perform a superior ride, cutting back the inability to do things. In fact, any deck with ride assistance reduces the entropy of the early game.
  • Mid :: The main problem with Limit Breaks is that they don’t do anything mid-game. Break rides further facilitate this fact and makes the mid game very very boring. All Set 5 and lower have low entropies for mid-game since their skills aren’t restricted. [Legion] is attempting to bring back the mid game, and I just we’ll have to see if it does.
    BED is atrocious at this stage as Juggernaut is the only thing they can do, and to add insult you don’t want it to eat too much soul! EDD however has a personal booster, a retirement option and an option to swap. Unfortunately, it is nothing like mid-game centred decks.
  • End :: The great thing about BED is that once you BR, everything works. Even triggers become deadly and you have tons of targets and options here. This is partly why BED can haul itself back form a poor mid-game with this extensive toolbox. EDD doesn’t fair any worse since the entire gambit relies on just one card (which again, makes EDD so consistent).
  • All :: As I’ve said, combining the scores doesn’t give much information. The only thing you can really tell is that BED is less consistent on average. Think about this; 0[0[3 and 1[1[1 both give a final score of 3, but are completely different in execution.

Okay, so hopefully after going through my example you kinda get how entropy works. Here I’ll provide a couple more decklists and cover some extremities:

SDR Daikaiser AquaForce Rush
4 SDR Daikaiser 4 Tear Knight, Cyprus
4 SDR Goeagle 4 Tear Knight Theo
4 SDR Daiyusha 4 Storm Rider, Eugen
1 Enigman Storm 1 Random 8k
4 Dimensional Robo Kaizard 4 Splash Assault
4 Enigman Wave 4 Random 8k
4 Diamond Ace 4 Emerald Shield, Paschall
4 Enigman Ripple 4 Random 8k
4 Dimesional Robo 4 Shortstop Brave Shooter
1 Enigman Flow 1 Advance Party Brave Shooter
12C/4H 16C

(del)S in its respective sub-games

Subgame SDR AqF
Early +0 +0
Mid +7 +1
End +2 N/A
All +9 +1

Most of the entropy stems from the fact that ride chain pieces and grade 3’s are useless after turn 3.

So from here it looks like this tier 1 deck has some serious entropy issues, while the tier 2 deck list looks spot on perfect. This just goes to show you that entropy does not effect the tiering system. What’s more important then entropy is consistency, and the SDR deck is sufficiently consistent. There is a strong secure ride and options to force the vanguard as Daikaiser. (In fact, future sets make this even more consistent). I’d like to relate to you that if the ride chain was removed, the resulting entropy in the early game would increase drastically. And this is true in general:

decrease in consistency \Rightarrow Increase in entropy

However, its converse is not true. In other words:

Increase in entropy \not \Rightarrow  decrease in consistency


And this point needs to be stressed. Entropy is not the same as consistency, but it is a measurement of what kinds of plays you can make at any given sub-game. Entropy is not advantage, but does tell you when you are going to gain it. Entropy is not something that is easily calculated nor has a set amount, but outlines the areas where options are limited. Entropy is not something that entails a decks success, but entropy should be minimized once all other qualities are set.

I wonder if you asked yourself why you should care about entropy. Well, an important part of card games is the ability to make choices, and decks are more interesting to play with the more options you have. But even if you try to fill your deck with lots of options, entropy can still increase and limit your available scope. But as I’ve tried to show, entropy doesn’t have a significant impact. And this is one of the reasons I find vanguard well built: every card is flexible you can’t truly* be screwed over by a ‘bad hand’. Hence why we don’t need multiple mulligans, mandatory searching cards nor a side deck

Well grade locking is a thing and it sucks, but the chance of getting grade locked is far less then getting a bad ygo/mtg hand. And that is really the only fault of vanguards brilliant strategy of mitigating entropy.

Buddyfight Strategy

Welcome to part 1 of 0* of this card game created by Bushiroad. We all know this company hailing the Cardfight!! Vanguard franchise, but they have also worked on other products like Weiss Schwarz, Victory Spark and Milky Homes. Their latest effort is Future card Buddyfight(BF). But why am I talking about it? Well, it relatively simple for me to quickly test and push a finding.

*Don’t expect anything else about BF by me, I’m pretty sure more meta-theory will spark somewhere else.

Pure Advantage

Cardfight vanguard is the only game I’ve played that throws 3 cards a turn at you. This cuts entropy* without turning the game into one-off gambits (like PKMN). It also means that a draw only increases tempo by 1/3 (yet another reason why draw triggers are bad). Most other games limit advantage gaining as it literally doubles the amount of options they have. Buddy fight falls somewhat in the middle of the spectrum with its ‘charge and draw’ (which gets terribly repetitive) though there is still a gap from +0.33 and +0.5. With some cards being tied to specific events this pretty much serves as a mitigation technique for its own entropy it created. I’ll put a little more depth later.

*Entropy in thermodynamics is the disorder of matter/energy. In this context it means the amount of dead/useless cards you have. Naturally, as the game progresses more cards will become dead.

Elite Attack Force

There are three renewable resources. They are: Deck, Gauge and Attacks.

  • The deck should never be of concern.
  • Players start with 2 gauge and +1 every turn. The statistical average implies optimally the average of all your monsters gauge should be 1. <=> The sum of all your monsters should be no more than 50 gauge.
  • The attacking game is core to a card game. In general the faster you the field fills up the quicker the game propels itself. Most card games have an inhibitor mechanic to slow the pace of the battle, not for dragging it out but to provide time to set up. MTG monsters has ‘sleeping sickness’, while YGO used to need ‘tributes’. PKMN has more of a skill based lag, and CFVG has about three-four* turns of leverage. Buddyfight’s inhibitor is non-existent. We already see a problem here. Any advantage you gain is likely to be washed away by brute force. You’ll see this become more apparent later.

The average damage a monster deals is approximately worth one card. In other words: 2.5D = 1C = +1 advantage.

*Though it isn’t as solid


Weapons as I’ll hope to prove are bad. They are inefficient and increase entropy.

Postulate 1:

 What is the spread of monsters-spells-weapons-impact? Logically, the easiest way to deal damage is with attacks, and to get attacks you’ll need monsters or weapons. However you may have up to three monsters while you can only have one weapon out at a time. It is easier to remove monsters then weapons. Hence you need to have more monsters then weapons. We can also state that monsters have a chance to last more then one turn while most spells are lost once cast. Hence #monsters > #spells or #weapons.

Postulate 2:

 Let’s consider a 1/1/1 field against a 2/W/1 field. The opponent has two options when you attack the centre: taking the hit or use a counter-spell. Both are -1. Now how many counter-spells do you expect to have at hand? [Hint: not enough].

2 W 1

1  1  1

The “Triple Offensive Maneuver” may look cool but you should avoid it at all costs. With three attacks for both sides, the opposing player will end up with a -3. The difference lies in how the minus is distributed. The top player will take three of some form of counter use and damage loss. For the bottom row the player loses 1 monster* and 2 of the aforementioned. If the opponent removes the centre monster (Dragonic destroy) to force the loss of 3 in anything but monsters, he traded a -1 for your -1. Also, since #monsters > #spell > #damage, it is natural to say that losing a monster is worth less then losing a spell. Therefore it is better to have a monster in the center.

Postulate 3:

 This monster to card trade applies to any field setup as long as one monster exists in the centre. Weapons only work when the centre is cleared.


 This implies weapons are strictly inferior to the optimal strategy (monster in center)   Certainly there are decent weapons, notably the magic world blaster. Monsters can also have ‘Move’, protecting you from arm by constantly moving back and forth. There are also support cards that only work if you are holding a weapon. In this case you need to ask yourself whether the advantage you gain is worth the extra 8 slots (4 for weapon, 4 for support), lower probabilities, increased entropy and wide open center. I say no*, just don’t use weapons and keep it simple. They are just too situational to be employed effectively.

*If you have ever played some game called Robotek, you will know exactly a unit is the only thing standing in the way of a serious laser show.

Two for one

Many cards are cleverly disguised washes, but there is always this one thing that leaves the door to advantage slightly ajar. Some can be obvious: If one card draws or destroys more than 1 card. Likely, those cards will have conditions like “if you have 5 of less life” or something equally limiting. But others may not be as visible:

  • Double attack
    • Extra attack will lead to another -1 the opponent has to deal with.
    • The pluses increases the longer it is out
    • Countered by the fact that the card can just be destroyed to prevent the plus from going off
  • Soulguard
    • To be fair, this needs to be combined with a charge to be actually a plus.
  • Opponent will need to waste an extra resource to remove the monster.
    • Cerberus and similar cards are not pluses. It only counts if the soul came from the deck.
  • Penetrate
    • Remove a monster AND deal damage.
    • Like double attack, it can be paused before the plus
    • Dependent on a monster able to be destroyed, and not to get countered by 50 different means.
  • 7000+ Defense
    • To explain briefly, it should require a link attack to be killed saving you from dealing with one more attack.
    • Spells can obscure this plus.
  • Counterattack
    • Normally the opponent forces a -1, but you get the +1 back and so you nice nice and tidy
    • A hidden gem is the combination with soulguard as you can kill even if you lose!
    • And as always, destructive cards takes this plus away

Do you see that soulguard is pretty much the only plus that is almost* impossible to remove? I’m pretty much convinced that soulguard is your go to and you should always attempt to stack your deck with these buggers.

*Magical Goodbye!

Essentially Extra

Lastly, I’ll wrap up some loose ends.

Impacts, as much as I dislike them, are quite useful. All will refund everything you use to cast them. The big giant sword of Armageddon for instance trades 1 gauge for 1 damage, and an empty centre practically screams punishment. (Checkmate!! is bad though). Impacts pack a heavy dose of advantage into one card (think: BR’s) so they are nice even if it does cause small consistency errors. The best analogy I have is that impacts are not so much as a win-more but more like a reduced life bar.

Most grade 3’s need to pay a hefty amount of gauge that could be better spent on Impacts, as they convert about 3 gauge into a +1 [the upped defense counts as +1 and then you add the other skill]. So I’m still questioning it’s utility. But they can troll real hard with counters (especially ones that give it counterattack), and those with soulguard AND 7000+ defense are terrifying. But many grade 2’s have good effects.

I really want to like these Set cards. They provide nice pluses over the course of the match, if the match exists long enough. They are like the inverse of impacts as they work better as soon as possible instead of near the end. But is it more likely to get an Impact by late game or a Set by early game?

Put as many counters as you possibly can. [Shields, 3+Counterattack, Destroy, etc.]They are vital in saving your bacon. If a non-counter isn’t giving you a lot of advantage (Dragonic Grim-ore) then it is in your best interest to leave it out.

World lines

So I’ll spend a few sections on the cards  currently* out:

Danger World

  • Size 0’s: Nice extra
  •  Armorknight Succubus : size 1 with soulguard!
  •  Armorknight Golem : god tier stats
  • Fighting Dragon, Demongodol: soulguard AND counterattack?
  • Lightning Strike Dragon Garjion: meat shield
  •  Emperor Dragon, Gael Karn: Amazing card to replace an impact with (as weapons sucks)

Dragon World

  •  Dragon knight, Iwamoto: Without any destruction spells this is literally a +1
  • Dragon knight, Rudel: Debatable
  • Drum Bunker (All of them): The entire lineup is powerful
  • Rising Flare: OMG sooooo many good size 2’s. This is probably why they have terrible and 1’s
  • Inferno Armor Dragon: Big ass defense, but grade 2’s are so crowded
  • Proud Song of Soul: If the early game favors you…
  • Day of the dragon: Yum

Magic World

  • Magic Knights: Hard conditional but worth it
  • Center of the World, Mary Sue: Toolbox
  • Demon Realm negotiator, Gusion: Great plus
  • Gunrod, Bechstein: The only good weapon, if you want to use it
  • Begone!!
  • Devil Advantage
  • Great Spell, My Grandfather Clock: (getting lazy)
  • Great Spell, Saturday Night Devil Fever
  • Nice one!: really, pot of greed

*Dated March 18th, 2014


If you ask me right now (At the time of this writing), I’m going to say Danger World possess the best build. With a pair of fluffy grade 0’s (So you can have a 2/0/1), relatively cheap healing (letting you take one more hit then most decks), soulguarding for all grades, a powerful field clearance like spell (Devastation starts at +2), draw engine, and abundant denial cards. But I’m certain Bushiroad will provide boatloads of strong support* to their pet world (Dragon World) until it outranks everything.

In terms of strategy; keeping a monster in the center at all times is optimal, no weapons, max out counters and soulguards and field wipers. Yep. Not too much to it.

*So with that said I’m done with this game. Have fun exploring the [tiny and dragon-dominated] world of BuddyFight!

Current Interest in VG

*This should not be used or substituted as any sort of scientific research or logical analysis. You have been warned.

Well, let’s look at how popular Cardfight Vanguard is after about, say two years now. So let’s just search it up:

Google: “popularity of cardfight vanguard”

  1. articles.
    • Not at all sure why this comes up first. Nothing instrumental or solid.
    • www
    • (1) MTG, (2) YGO, (3) PKMN, (4) CFVG
    • Interesting

> While I could go to page 2, that probably wouldn’t be very useful

Let me just scroll down and check behind that rock. Annnnd ... nope, page copyright year starts with '19'. Oh God, is this a WEBRING?

Google News: “popularity of cardfight vanguard”

    • Shortages of expansions mean good business, and bad palnning. Then again, the Bermuda booster are coveted being filled to the brim in pre-orders…
    • Hey there are references
    • “Top 10 Hobby Channel Collectible Games” – ICV2
      • CFVG places 5th (Marvel got 3rd!?!)
    • Another dead end

Time to get off bad search engines and find something better:

Google Trends: “vanguard, yugioh, pokemon, magic”



You’ll get this neat chart. Out of the data collected by google, these are some nice things to say:

  • YGO  gets interest from US and the Philliphines
  • MTG is almost soley US
  • CFVG is still only important in Japan and Brunei
  • PKMN has both US and Japan fairly evenly
  • YGO’s interest acutally comes from the Anime; the CCG is sitting at 20.
  • Typing magic gets you to mtg., but “vanguard” doesn’t (for fresh searches)
  • The forecat predicts MTG to stay the same
  • YGO is just falling down
  • PKMN will live in a perpetual state of weakness
  • CFVG is being predicted to drop

Eh, Let’s look at at a website stats:


  • Ranked: 1045
  • About 1 million in traffic


  • Ranked: 4037
  • About 278, 000


  • Ranked: 37388
  • About 20, 850


  • Ranked: 1526
  • About 814, 042
  • Of course, this is for and not the tcg site



And that’s me being lazy. And about that program I’m working on:

Making grade 3 searchers work

I lie. Since I write up a report after my research I can tell you not to run them. But let’s be blissfully ignorant and start off with declaring some numbers:

Probability_Calculator ChanceToGetMin1(0,6,4) = 0.67171555
Or about 67% to get a certain grade 3 on turn three assuming you run 4. That also means that (1-.67)*.67 = 22% of the time you’ll have to ride the ‘other’ grade 3. And the remaining 10% is for failing to get either (and that’s after you have the other ride pieces)

Chance to pull any grade 3 from the searcher 1-(41*40*39*38*37)/(49*48*47*46*45) = 0.60700
Chance to pull the specific grade 3 from the searcher 1-(45*44*43*42*41)/(49*48*47*46*45) = 0.35929

The generic deck

So we look at two possibilities:

The normal, non searcher deck

(1) 67% * [Advantage with main] + 22% * [Advantage with backup]

The searcher deck

(2) 67% * [Advantage with main – 0.40] + 12% * [Advantage with main] + 19% * [Advantage with backup – 1]

12% comes from multiplying 33% by 35.929%. Since we want (2) better then (1):

0.67(M-0.4) + 0.12M + 0.19(B – 1) – 0.02 > 0.67M + 0.33B
0.79M + 0.19B – 0.46 > 0.67M + 0.33B
0.13M – 0.13B > 0.46
0.29M – 0.29B > 1

~roughly speaking~

0.29M - 0.29B > 1

The overwhelming pluses

Well let’s say your backup gives you no additional advantage* (generic 12k beat stick) while the main gives you a +3 throughout the game. Is it worth it?
*Note: since both grade 3’s twin drive, we cancel their advantages out and are left with net 0.

0.29(3) – 0.29(0) ≯ 1
0.87 ≯ 1

So no, you are better suited to having no starter at all! VG must supply 3.5 units of advantage to cope with the shortcomings of the searcher.
But after 3.5 you actually do get more advantage then a non searcher! Very important find!

Oh, but what if the either of the following occurs:

  • A different starter can garner you a +0.5.
  • The other grade 3 has a on-ride or other ability that provide a plus 0.5?

In both cases your vanguard will work overtime, having to give you +4.5 to justify a searcher!

The inexcusable negatives

Looking at the equation you can also go in reverse, have the backup so bad at vanguard you actually minus for riding it (then if you just stuck with a grade 2). Not as easy to do, as your disadvantage must outweigh your vanguards advantage + additional!

So let’s have an OP grade 3 whose effect works in RG/hand/soul/drop/deck but has no effect and 3k as vanguard. That’s like a -6. Your vanguard only really provides a +2, but that’s fine because your backup is so good. Should you consider a searcher?

0.29(2) – 0.29(-6) > 1
1.16 > 1

Yay! I can say that there a reason to use this sonofafag. What’s that? There are no grade 3’s like that? Well you are right there.

The negligible selection

Some other things I will just skim over, provide as a list since lists are convenient

  • Break rides: Even if your breakride generates an insane amount of advantage, chances are your backup will also have some sort of advantage as well and not some generic throw-in.
  • Crossrides: Calculations are a bit tougher to do since you need to account for pairs, but crossrides have enough pluses by themselves to make sure you don’t pass the 1 mark.
  • 8k 9k Grade 3’s: This is a severe handicap and in you could try a searcher. But decks that have those specific cards have better starters and superior builds.
  • Restraint: These are bad units in general. Most likely you’ll have multiple backup plans for something like this, reducing a searcher’s usefulness.
  • 6 grade 3’s: Using that many will artificially make the searcher seem more effective than it is. (But it is for that case)
  • 10+ grade 3’s: Presumably you want a checker to generate advantage. I’d like a very solid backup for this vanguard, so it gives searchers a run for its money.

So, is there any deck so far that actually need a grade 3 searcher? I can cheat a little bit and declare the Musketeers have a nice purpose for one. In the optimal deck you’ll find the generic 12k grade 3 and Cecilia. And her advantage is legendary, by having something that amounts to +2 as well as her efficiency roulette (which can also generate another +2 or +3). But otherwise, I highly advise against using searchers.

Every Trigger Ratio you’ve ever wanted

*This article has a tldr; tag

A: I run 6 crit and 6 draw
B: Nah man, that’s bad. Use 8 crit, 4 draw. Everyone uses it.
C: Why don’t you run 12 crit?
A&B: Are you a complete idiot? 12 crit is a horrible ratio to run.
C: Why?
A: Just ’cause.
B: Ya.

Not even math can deter these bigots. They usually have lame excuses (draw power gives you pluses!) if any. These are silly claims and I’m not concerned about convincing them. But for those who are seriously considering the why and are willing to accept the proof before your eyes then this amazing program will do that all for you:

Trigger comparison (Online)

But first, some logic. Why is 12 C/4 H the ideal trigger ratio?

Premise: 1 damage >> 1 card
Premise: Crit and Heal triggers increase/decrease damage by 1
Premise: Stand and Draw triggers increase/decrease cards by 1
Conclusion: Crit and Heal triggers >> Stand and Draw triggers

There is a small fallacy as a couple of decks get more then one card per trigger@LINK@, but those are exceptional cases. Okay so with that out of the way let’s get started!

The Interface


If you have any complaints about he interface, leave it. The source code contains a much snazzier interface but isn’t as convenient to set up.

We first want to figure out is the results of this program matches by hand-done calculations in my old standard form article, where I stated 12C/4H as the optimal line-up. Here were my results:

Plan Damage Shield Cards
First Attack Through 45/49 25 3 25/321
Second Attack Through 1 81/220 20 399/479 2 217/698
First Attack Through 1 81/220 18 178/435 2 217/698
Fewer C/L RG atkThrough 45/49 26 23/25 3 85/196

Converting the result from nice fractions to ugly decimals we get:

Attack Plan Defensive Measure Damage Shield Cards
R>V>R Let RG Through 0.91837 25000 3.0779
R>V>R Let VG Through 1.36818 20833 2.3109
V>R>R Let VG Through 1.36818 18409 2.3109
V>R>R Fewer C/L RG atk Through 0.91837 26920 3.4337

*just to note that I’m letting the lower crit rearguard through and not the higher crit one


Attack Plan Defensive Measure Damage Cards Shield
R>V>R Let VG Through 1.36818 2.3109 20833
R>V>R Let RG Through 0.91837 3.0779 25000
V>R>R Let VG Through 1.36818 2.3109 18409
V>R>R Let RG Through 0.91837 3.4337 26920

So now we compare with the numbers in the program and they should match up…


…They ALMOST match up. More specifically the only difference is the very last cell. The card loss and shield loss is slightly higher than before but this is on purpose: I made a small error in the other article in terms of attacking. Before the opponent would attack with the low crit unit before the high one, resulting in a potential counter-trigger. But what was more optimal was for the opponent to attack with the higher crit so that the counter-trigger can be avoided. Why no one noticed before is beyond my comprehension.

So now we can get started.

A solitary game

Let’s take the 12C/4H for both teams for starters.

12C/4H/0D vs 12C/4H/0D
Attack Plan Defensive Measure Damage Cards Shield
R>V>R Let VG Through 1.36818 2.3109 20833
R>V>R Let RG Through 0.91837 3.0779 25000
V>R>R Let VG Through 1.36818 2.3109 18409
V>R>R Let RG Through 0.91837 3.4388 27653
Damage you heal 0.16
Cards you draw 0

Let’s start at say, 2 damage. If we take it like that:

P1 dmg P1 card P1 shield P2 dmg P2 card P2 shield
-0.16327 0.0000 0 1.36818 2.3109 20833
1.20491 2.3109 20833 1.20491 2.3109 20833
1.04164 2.3109 20833 2.57309 4.6218 41666
2.40982 4.6218 41666 2.40982 4.6218 41666

From here, the damage is just about 4 to 4. Now the plan switches to V>R>R to kill the hand:

2.24655 4.6218 41666 3.32819 8.0606 69319
3.16492 8.0606 69319 3.16492 8.0606 69319

And from here it becomes a battle of B.A.S.S.. From here we start turn #9. As before this is completely neglecting the various abilities of cards and assuming complete uniformity. If you are wondering what the 0.16492 portion mean, it is the chance you’ll die, before that turn. (since in that chance, you’ll have 6 damage instead of 5)

Trigger Dump

We’ll compare three common line-ups against each other and see if they hold merit (By common I mean those that I see frequently run.) Those are 12C, 8C/4D & 6C/6H. With my program the data is easy to acquire:

12C/4H/0D vs 8C/4H/4D
Attack Plan Defensive Measure Damage Cards Shield
R>V>R Let VG Through 1.36818 2.1893 20833
R>V>R Let RG Through 0.91837 2.9963 25000
V>R>R Let VG Through 1.36818 2.1893 18409
V>R>R Let RG Through 0.91837 3.3571 27653
Damage you heal 0.1633
Cards you draw 0.0000
8C/4H/4D vs 12C/4H/0D
Attack Plan Defensive Measure Damage Cards Shield
R>V>R Let VG Through 1.21824 2.3426 21100
R>V>R Let RG Through 0.91837 3.0779 25000
V>R>R Let VG Through 1.21824 2.3426 18937
V>R>R Let RG Through 0.91837 3.3265 26905
Damage you heal 0.1633
Cards you draw 0.1633
12C/4H/0D vs 6C/4H/6D
Attack Plan Defensive Measure Damage Cards Shield
R>V>R Let VG Through 1.36818 2.1285 20833
R>V>R Let RG Through 0.91837 2.9554 25000
V>R>R Let VG Through 1.36818 2.1285 18409
V>R>R Let RG Through 0.91837 3.3163 27653
Damage you heal 0.1633
Cards you draw 0.0000
6C/4H/6D vs 12C/4H/0D
Attack Plan Defensive Measure Damage Cards Shield
R>V>R Let VG Through 1.14327 2.3583 21233
R>V>R Let RG Through 0.91837 3.0779 25000
V>R>R Let VG Through 1.14327 2.3583 19202
V>R>R Let RG Through 0.91837 3.2704 26480
Damage you heal 0.1633
Cards you draw 0.2449

The analysis

That was the easy part. We’ll start with 12C vs 8C/4D and again we start at 2 damage and check the score with the 12C going first:

P1 dmg P1 card P1 shield P2 dmg P2 card P2 shield
-0.16327 0.0000 0 1.36818 2.1893 20833
1.05497 2.3426 21100 1.20491 2.0260 20833
0.89170 2.3426 21100 2.57309 4.2153 41666
2.10994 4.6852 42200 2.40982 4.0520 41666
1.94667 4.6852 42200 3.32819 7.4091 69319
3.16491 7.0278 63300 3.16492 7.2458 69319

WAIT WHAT? The damage is equal (to 0.00001)! Wow this was surprising for me too. This implies you’ll send an opponent to the endgame as the same rate regardless of what ratio you have. But look closely at turn 7. You see here the opponent switches gears to defend a V>R>R, while they still need to perform R>V>R for one more turn. In fact, this means a 12C player forces an opponent into the end game one entire turn earlier!

But that’s not all. Remember how the opponent was drawing all those cards? Well ironically they ended up using more cards then the 12C player! (7.2458>7.0278). They also lost more shield in the process. So, after seeing this, is it really necessary to continue?

But the 12C went first!

Oh, I see how it is.  Make me do more work. We’ll have the 8C go first this time:

P1 dmg P1 card P1 shield P2 dmg P2 card P2 shield
-0.16327 -0.1633 0 1.21824 2.3426 21100
1.20491 2.0260 20833 1.05497 2.3426 21100
1.04164 1.8627 20833 2.27321 4.6852 42200
2.40982 4.0520 41666 2.10994 4.6852 42200
2.24655 3.8887 41666 3.32818 7.0278 63300
3.16492 7.2458 69319 3.16491 7.0278 63300

Are you surprised that the end result is the same? Because I’m not.

Try Again

Okay maybe I picked a perfectly chosen ratio. Or maybe…you need to draw more! Cause that’s going to alleviate the problem right?

I’ll let the 6C/6D have the first blow

P1 dmg P1 card P1 shield P2 dmg P2 card P2 shield
-0.16327 -0.2449 0 1.14327 2.3583 21233
1.20491 1.8836 20833 0.98000 2.3583 21233
1.04164 1.6387 20833 2.12327 4.7166 42466
2.40982 3.7672 41666 1.96000 4.7166 42466
2.24655 3.5223 41666 3.10327 7.0749 63699
3.16492 6.8386 69319 2.94000 7.0749 63699

It is interesting to note that the amount of damage we reduced is about the same as the card advantage they opponent gains and since damage > cards they got the bad deal out of it.
There is one other catch-22 you need to pay attention to when using draws; they have cut shield. So in fact, the opponent is playing with a 30000 shield handicap. If you factor that in it will only make a 6C/6D even worse so I left it out.


tldr; 12C pwns

Now that I’ve sweetened your knowledge, you can determine other things like:

  • I need more draws! 8D/4C works right? (no)
  • Will 8C/4D or 6C/6D win if they faced each other? (8c)
  • Is 16 Critical is better than 16 Heal? (no)
  • Now we’ve seen 12C in action, but what if it was fighting a pure 16C? (exercise)
  • What if a player always goes V>R>R?
  • Who will win: 12C only R>R>V or 16D and optimal? (16d)
  • Would criticals be worse if damage went up to 10? (no)
  • What if the effects of triggers doubled?
  • The optimal grade ratio? [brute force may be useful] (16h)
  • Trigger line-up that isn’t 16? [requires a code change]
  • Deck has more then 50 cards? [requires a code change]
  • How does stand triggers mix it up? [requires a code change]

If you have a fabulous result to pass to me please let me know!

Nubatama’s Philosophy :: Set 13

*This article has a tldr; tag

Ancestral Origin

Nubatama is not the most technical deck I have used.

Granblue, Spike Brothers and Great Nature have a vast amount of options that can make or break their game, but Nubatama is really straightforward. Counterblast 1, they lose 1 card form their hand. Repeat until dead. Such a simple conversion rings an alarm to those who look at it: it’s too broken! Blaster Blade (Liberator) has all these conditions and still has a heavier cost*! What gives? *[As I wrote all of it a month prior, I had no idea they would create a cheaper Dorint clone that can actually unflip 1 to retire 1.] But let’s looks at the theory behind Nubatama’s card advantage. What seems like a straightforward +1 will become a maze of math and logic that will show the truth of Nubatama’s deceit:

First, some pre-reading.
[1] A Simple Construct by LittleFighterFox (that’s me!)
[2] First or Second Advantage  by TehNacho
[3] Nubatama Speculation by 3XXXDDD

The Matrix

Let’s look at this figure below from [1]. From this, Nubatama has the most difficult condition tree to learn:

let x = opponent’s hand size – your hand size D be your opponent’s damage C is the number of rearguard columns with Dreadmaster boosting

If D >= 4 then do V>R>R, no questions asked
If D < 4 and x > 3 and the VG is void/boost is dread then do V>R>R
If D < 4 and x<=0 then do V>R>R
If D < 4 and 0<x<3 and C = 0 then do R>V>R
If D < 4 and 0<x<3 and C > 0 then do R>

If the attack was let through then do >V>R
If the attack was blocked and C = 1 then do R>R>V
If the attack was blocked and C=2 then decide;

the best card:damage ratio for both R>V>R and R>R>V is to let the vanguard hit but;
If the opponent is willing to lose more cards to minimize damage then do R>R>V
If the opponent wants to endure the damage then do R>V>R

And that covers most of the branches…of attacking a vanguard only…without them getting any triggers. The solution to the rest matrix is left as an exercise.

First and Last

From [2], at the start of player 1’s turn they will start with the same number of cards as the opponent, while player 2 starts off with one less the player one. As a Nubatama player, going second is more advantageous for you. So how do you deal with being first? As you are the first to ‘unlock’ your grades you also get to play the down harder than them. This is nice to apply a nice amount of pressure turn 3, but unfortunately isn’t as useful as going second. But what happens after you connect a blow? You pad out the advantage to 0. sometimes you see that you have around 4 more cards on hand then the opponent, and the only way to get below them by your turn is the over-guard all attacks. D’oh! What are you doing? If you have the advantage, KEEP IT. A pretty common misconception is that playing Nubatama is akin to Russian Roulette. Not so. Here is a list Nubatama players can use to reactivate their discard ability:

  • Over-guard. Well that’s usually dumb.
  • Replace the field. Even dumber.
  • Wait for your opponent to gain advantage. This is the best thing to do, though there are some faults which I’ll address in the very next section.
  • Promise Daughter/Holy Disaster Dragon discard skills.

Discard skills serve as a cost to force an opponent to drop the same amount. In the late game these skills usually pump the attack by one column, forcing lo and behold, one extra card. If they let the attack through and live, you’ve just sent a card into the entropic void. So as you can tell, Nubatama is not an advantage nor a disadvantage engine. They serve as a vanilla-izer, keeping both sides with the same advantage. Whenever your opponent gets a leg up, you drag them back down by the same amount they went up. So do not consider Nubatama has a need for self-discarding tools.

You can tell from 3XXXDDD’s analysis on [3] most clans have ADVANTAGE engines. Midnight Bunny clearly performs a better plus then Dreadmaster’s “plus” for the same cost. But Midnight Bunny needs energy to bring her out again while Dreadmaster stays put. But I’ve already shown you it isn’t really useful to stay out any longer since the preceding turn will likely be equal.

Failed Assassination

In addition to the lack of Nuabatama’s effect on the opponent, there are very easy ways to escape from Nubatama’s neutralization trap. You either need to get more advantage than Nubatama can ever take down, or resort to generating a large amount of card advantage in a short period of time.We have cards like this:

  1. Crossrides: These little buggers can keep Nubatama at bay without a slightest regard for fairness.
  2. Break rides: And everyone is getting these visibly powerful units one and shoving into their deck.

Nubatama can’t use an anti-advantage arsenal anymore after an arrival of the re-King of Knights. They must gain advantage instead, like the rest of the clans. We’ll get back to this section in the near future but for now, I’ll discuss the three techniques Nubatama uses to secure a win when they acquired a new set of cards in Catastrophic Outbreak.

Forget about the early and middle game

The first technique is tempo increment. Mentioned briefly at the end of [1] The tempo of a R>V>R formation is 0.9 Damage or 1.4 Damage, and both have the same damage:card ratio. So normally, the decision falls on the opponent which is not something you want to give. With a pair of on-hit rearguards, it pressures the opponent into guarding the rearguard and produces a tempo of 1.4D. Over a couple of turns the state of the game reaches the end faster than they might realize. This is the same as a V>R>R formation in a different light. Simply, this means an opponent has less time to accumulate pluses and bring the state of the game to the endgame quickly. Much like Capablanca.

Optional Reading: Grandmaster José Raúl Capablanca

Capablanca is a chess champion and is considered one the great minds of chess (1988-42). His style of play is quite frustrating to some. His strategy is to get a very small positional advantage, then trade all the pieces quickly forcing the game right into the endgame in which he excelled at. Removing these pieces also removed a lot of chess complexity like pins and discoveries and brilliant moves in general (And may attribute to the fact that Capablanca has the lowest blunder %). In effect, he eliminated the need to think about both the early and middle game. It served well to his purposes, because by discarding the other two subgames he was able to fully develop his energy into just the endgame. While there has been several players that adopt positional tactics, playing chess in such a straightforward approach is usually frowned upon. Perhaps the most amusing conjecture that arises is that Capablanca complained that chess was too simple, after so many years of repeating the same endgame manoeuvres, and he proposed a revision into the rules of chess to keep it interesting. Go figure. Though it was never accepted you can still try it online.

After reading this it seems like vanguard went down this path, releasing a ton of endgame specific units, clumped advantage engines and the like. Whether or not it will ‘move away’ is something to be seen…

Bear necessities

The second advantage arrives in a pair of units: a starting vanguard with the ability to kill binded cards, and a rearguard that binds. As severely limited as our ability to generate advantage is, we must put everything we can to do justice. Sure Kabukicongo, Kujikiricongo and Fuuki binds, but note that Fuuki is a direct minus while the other former two activate at the 4th damage which is quite late for an on-hit attack to land. Thus, we can only say something good about Tamahagane and Kuroko. Regardless of how the events turn out, calling a Tamahagane will cause you either +0.5(By guarding the vanguard) or +1(By letting the vanguard hit). So you should run as many Stealth Beast Tamahagane as you can, because this is close to the only source of advantage you get.

You might be tempted to put Kabukicongo on the advantage list, but it is not so straightforward. If you just so happen to keep a Stealth Dragon Kokujou on the field, not having it blown up, then you can get something close to a +2 with a full field. Kabukicongo binds all the cards on the field, and gets 10000 power. Although you might be inclined to hope that binding does something, it doesn’t. (Only in the rare cases where the opponent teched in a special case unit can it is worth something, but we need to assume most decks are not reliant on them). Add in the fact that the opponent would most likely be perfect guarding the attack congo doesn’t amount to much. Next, we need to figure out just how much advantage they can generate when you return cards to the hand. Congo practically gives opponent options in the endgame where there aren’t any (Entropy is high). Its counterpart performs a better job by limiting the choices of what you can call, does not let you use up useless cards for shield and be able to retire a unit. While there is no such number for what returning a card exactly is, you can approximate by giving every clan a Bermuda Triangle bouncer. The equation speaks for itself:
[Kabukicongo:AUTO] {Bind field + POWER+10000} – {LB4 + CB1 + returning} = ? 0 + 2 – 1 – 0.5 – ~1.5~ = -1
All in all, it takes the joint effort of Kabukicongo, Kokujou and Kuroko to generate a tiny advantage (+1) that is not suffice for the other big endgame units out there that can do so much more. If the opponent survives, you’ve got to deal with their rejuvenated advantage engine[link:BDR], as well as the more focused field. If they just so happen to call out only 4 units instead of five then Kokujou becomes a +1 for the rest of its short-lived existence. Kokujou useful power boost fits closer to a different form of advantage.

The Hidden Crush Rush

As the title is the true face of Nubatama. There is no unit that truly adds advantage by itself; we’ve seen Tamahagane pull off a small plus. But there is another way to win, by reducing both players advantage at the same time. A classic Spike Brothers approach is to explode and waste all of its resources in one turn. Both Juggernaut maximum and Brakki incur heavy minuses so using them early can result in lost advantage. Dudley Emperor takes the most minus, and if anyone survived a crush rush can tell you the opponents field, hand, and soul were practically non-existent. Yet, Spike Brothers are one of the most difficult decks to face, since they pound out every last bit of defence you have somehow squeezing by for the touchdown. You only have a limited amount of cards per turn.

This is the procedure Nubatama follows. We have only have two units that all minus, but we can put them together in one turn so that we can bombard the opponent’s defence. The first is Hagakure. Surprisingly, this is actually Nubatama’s best card. It plays perfectly into what you want – it uses itself as a guard, meaning you have virtually one less card in hand. Then it forces the opponent to lose a card. Add in the advantage all Grade 0’s have, a 2 stage skill, you have a the makings of an incredible advantage turner. The way to use it is to keep as many as possible in hand, and just before declaring final turn, you drop your stash on the foe so that you have one turn when your opponent has taken a -4. If you flub your final turn, you walk in with 40000 less shield.  It doesn’t matter if the opponent is only attacking with one stage: empty all 8 stages at once. It might be contradictory to what I’ve stated at the beginning of the article, but that applied to over-guarding before the final turn. The most convenient time to use it is when the opponent has something around 6 cards in hand, you are at your limits and the opponent has four or more.

Which leads nicely to the other boss of Nubatama: Shura Stealth Dragon Kujikiricongo. Any analyst that compares it both in a vacuum and together in the clan will put it off as a very middling card. You -1 by riding while forcing a -1 from the opponent. You get your stages, then one extra temporary bind. This almost makes it the worse break ride in the game. Ashley gets a one-up in synergy due to Julia and its reverse form. There is very little synergy this card provides to the rest of the deck. But when paired with Hagakure, you can manage to empty all of the opponent’s hand, if they have 6. That means they are a sitting duck to your onslaught. And this amounts to all they can perform.

Playing tag with ninjas

And his assumes perfection. It is more likely you are only going to draw into 2/4 of the Hagakure you desperately need. That means you need to have the opponent at 4 cards and 4+ damage. There are other combinations which work, but you are severely limited in regards.

Darkest hour All in all, Nubatama is somewhat technical but it boils down to just memory. Know the attack matrix, train your memory so you can tell what is in the opponent’s hand, and remember odds of winning for any given damage and cards in hand. There are no combos to speak of and you don’t need to understand how to play the early game or middle game. Can you say that about any other deck? The deck is similar to about every other break ride focused deck in the game. The main difference is their break-rides INCREASE advantage so they don’t lose if they can’t finish someone off that turn.

What? You want a decklist? You must be joking, I’m a (self-proclaimed) mathematical theorist and game theory analyst. I don’t do decklists..

tl;dr: Dudley Emperor > Bad End Dragger has the same strategy as Nubatama but harder, better, faster, stronger in each and every way.

Just put Nubatama aside for now…

Python on Modules up on Github

Well the title is self-explanatory. Link here:

This link can also be accessed from the Research tab. First thing is that, unless you’ve got a geniuses mind you’ll have no idea what to do. Even if you have done python. I’ve only gone through rudimentary cleaning and commenting, which is to say; very bad programming practice. Hopefully given a bit of spare time I’ll make a pseudo-interface and maybe a GUI eventually.


For the absolute beginner:

  1. Download and Install Portable Python 3.x (I’m using but any other extension should work). It has the IDE set up and everything, so you don’t have to hassle over setting up the interpreter and PATH.
  2. Either download each file from my github page separately OR hit the button on the right that says download zip to download them all.  You know what to do with zips.
  3. Open up PyScripter-Portable.exe in whatever folder you installed Portable Python in.
  4. File > Open and search for any of the .py files that you acquired from (2)
  5. Hit Run (Shortcut: Crtl+F9)
  6. If you don’t see any warning or red popping up you are good to go!
  7. Move cursor/caret to the python interpreter at the bottom.

So once you have whatever file you have running, there are special commands you can type. For convenience I have italicized and bulleted it . Following this setup, just type in function(arg), where function is the italic part in the list and arg is the arguments or numbers you put in. Here is an example function.

  • quickodds(a,b,c,d)

Will be shown like this once you hit [Enter]:


Pay attention to make sure the command you type is the same character by character (this includes capitals). Also be careful with your arguments.


So let’s look at the files I have so far:

Note: Almost nothing else will work if you don’t have this one file!

This file does ALL the heavy lifting. It sports all the probability modules. Things like permutations, combinations, and a function for calculating the hyper-geometric cumulative distribution calculator. Using the Fraction class makes things neat and very precise. You can still use this module for fun with two commands:

  • quickodds(a,b,c,d)

a: Population size OR number of cards in the deck
b: Possible successes OR number of specific cards you want to pull in the deck
c: Sample size OR number of cards to draw in a row from the current deck
d: # of successes OR number of cards drawn that are those specific cards you wanted

Gives the same result as if you used stattrek’s calculator

  • cascadeodds(a,b,c)

Using the same parameters/arguments as quickodds. This simply print the odds for every possible /d/ value.

Type in

  • foo()

Get back average percentage of getting a certain value of shield.
*foo() has no arguments

Excluding Generation II (Tsukuyomi, Galahad) this prints probabilities of getting up to grade 3 in three turns based on the input.

  • ChanceToGetMin1(ini,end,n,mull,first)

This is a masterpiece of pain. The module uses this function extensively. Note that you can skip the latter 2 arguments to save time: ChanceToGetMin1(1,6,4) is the same as ChanceToGetMin1(1,6,4,3,True). Rest of the documentation will appear if you type the function in.

Q: Why am I getting fractions?A: Cause it is accurate. I’m pretty sure you’ll learn how to use the print() command.

  • BasicRide(gr1,gr2,gr3)
  • AdvanceRide(gr1,gr2,gr3,gr4)
  • Gen1Ride(g1,g2,g3)
  • Gen4Ride(g1,g2,g3)
  • Gen2Ride(gr1,gr2,gr3,r)

All but this one expects you to have 4 of each copy already in the deck. This one though, take /r/ as the number of the grade one ride piece. I was doing calculations with this number but it was quite useless later. you can omit r altogether and it’ll treat is as 4 copies.


Once done, it’ll be the first one written with documentation and cleaning done as I go along. It even has an interface. Therefore, I don’t have to tell you as much.

What does it do? Something quite a few people are wondering, proof for the best trigger ratio. This program compares the hand, shield and change in damage based on the trigger lineups specified by the user. 16 draw? 9S/5C/2H ? Now you are no longer bound to solid multiples of 4 and 6.


That’s all I have. I’ll upload and update more of the modules I have to do my research. Hopefully I can have transparent data available for peer review, modifications and the like. Whenever you doubt me, you can check the code and see how it works!


Get every new post delivered to your Inbox.