What Combo?
For anyone who does not know, "The Gitrog Monster Combo" is The Gitrog Monster + Putrid Imp + Dakmor Salvage + Kozilek, Butcher of Truth. You discard Dakmor Salvage, causing The Gitrog Monster to create a draw trigger. That trigger is used to dredge Dakmor Salvage back into your hand. If you dredged 1 or 2 lands, you get a free draw trigger. The idea is that you can keep doing this to draw an unlimited number of cards. However, you need Kozilek to shuffle your graveyard back into your library, and that complicates things significantly. If you simply let your graveyard shuffle whenever Kozilek was dredged, you could theoretically get Kozilek + some nonland card on the top of your deck an infinite number of times. To resolve this problem, I wrote a logical proof (*This proof needs to be rewritten to demonstrate new methodology*) to show that the combo can be performed in a way that guarantees you draw your entire deck in a finite number of steps. That is fun to think about, but it is still difficult to tell the actual steps that should be performed at any given point of the combo. For this, I present the following flow chart algorithm.
A flow chart algorithm that shows one way to perform The Gitrog Monster Combo in a determinate manner.
Changelog:
2016-05-18
Completely redone and simplified by viewing problem from direction of generating T rather than reducing N
An accompanying file for the algorithm that explains the shorthand and symbols used.
Changelog:
2016-05-18
Fixed terms to relate to new algorithm method
2016-05-17.3
Added minimum land count requirement
2016-05-17.2
Added definition for "Deck"
Added a pre-setup requirement for the algorithm
2016-05-17.1
Clarified that you dredge in response to the shuffle trigger
Future Files:
A logical proof that demonstrates how you can guarantee that each iteration through your deck is advancing you towards the end goal of getting your whole deck in hand.
Notes:
Please let me know if you find any errors in my logic (either the proof or its translation into flow chart form).
Are any confusing parts that I should revise the presentation or add more explanation to?
Currently, the proof and algorithm are looking at the problem from completely different perspectives, proof will be updated later
Are there any other things that would be useful to add as a future file?
I'm sorry. I read through the PDFs and the proof and I just can't understand what prevents the following from happening:
1. I have no draw triggers at the start of this, but I do have Gitrog and imp out, and Dakmor Salvage in my hand.
2. I discard Dakmor salvage, generating a draw trigger.
3. I use that trigger to dredge Dakmor salvage back to my hand.
4. The top two cards of my library happen to be Kozilek and Sol Ring.
5. I shuffle my graveyard back into my library from Kozilek's trigger.
6. Go to Step 1. (I keep shuffling such that, by pure chance, Kozilek and Sol Ring are the top two cards every time until the heat death of the universe)
I'm sorry. I read through the PDFs and the proof and I just can't understand what prevents the following from happening:
1. I have no draw triggers at the start of this, but I do have Gitrog and imp out, and Dakmor Salvage in my hand.
2. I discard Dakmor salvage, generating a draw trigger.
3. I use that trigger to dredge Dakmor salvage back to my hand.
4. The top two cards of my library happen to be Kozilek and Sol Ring.
5. I shuffle my graveyard back into my library from Kozilek's trigger.
Yes, this is the very core of the idea. The trick here is that you do not simply allow your deck to shuffle (that is only allowed right after Start when it says "Shuffle Graveyard"). Instead, you discard Dakmor Salvage in response to the shuffle trigger and keep going. I just updated the Guide to make a comment about that, good catch.
I'm sorry. I read through the PDFs and the proof and I just can't understand what prevents the following from happening:
1. I have no draw triggers at the start of this, but I do have Gitrog and imp out, and Dakmor Salvage in my hand.
2. I discard Dakmor salvage, generating a draw trigger.
3. I use that trigger to dredge Dakmor salvage back to my hand.
4. The top two cards of my library happen to be Kozilek and Sol Ring.
5. I shuffle my graveyard back into my library from Kozilek's trigger.
6. Go to Step 1. (I keep shuffling such that, by pure chance, Kozilek and Sol Ring are the top two cards every time until the heat death of the universe)
Yes, this is the very core of the idea. The trick here is that you do not simply allow your deck to shuffle (that is only allowed right after Start when it says "Shuffle Graveyard"). Instead, you discard Dakmor Salvage in response to the shuffle trigger and keep going. I should update the Guide to make a comment about that, good catch.
AH! I get it now. Yes, given that the absolute key decision point that makes the algorithm possible is choosing whether or not to discard Dakmor while Kozilek's trigger is on the stack, I think that needs to be spelled out a little more clearly for dense people like me.
I'm sorry. I read through the PDFs and the proof and I just can't understand what prevents the following from happening:
1. I have no draw triggers at the start of this, but I do have Gitrog and imp out, and Dakmor Salvage in my hand.
2. I discard Dakmor salvage, generating a draw trigger.
3. I use that trigger to dredge Dakmor salvage back to my hand.
4. The top two cards of my library happen to be Kozilek and Sol Ring.
5. I shuffle my graveyard back into my library from Kozilek's trigger.
6. Go to Step 1. (I keep shuffling such that, by pure chance, Kozilek and Sol Ring are the top two cards every time until the heat death of the universe)
You respond to the shuffle trigger, and keep going.
In the proof there is the description of another algorythm. Maybe, together we can arrive to a final version that is both easy to understand and easy to perform
for now, I only read the PDFs and I have some little suggestion:
1) in the guide, maybe you should describe the minimum requirements of a starting state; for example, if I apply the algorythm, the first action that I should do is 'Shuffle Graveyard', but this can be done only if, in the starting state, there is a Kizilek Trigger in the stack; otherwise, I cannot shuffle graveyard without a reason
2) In the guide, you should maybe describe more precisely the meaning of the other terms used in the algorythm. For example, the term 'Deck' in the first diamond is referring to the actual number of cards in my Deck, or the number of cards in my Deck as the algorythm started?
There is a small error with you calculation. The location of Emrakul in your deck has an effect on the minimum number of draws you can get because you need to unwind the stack (so you do not lose the game from drawing on an empty deck). The worst case scenario would be [Emrakul], [all non-lands], [all lands]. In that case you will get about 1 draw for each 3 lands. The reason my version is so complicated is that I wanted to prove that it can be done with as few lands as possible (I demonstrated it down to 5, but someone might be able to revise it down to 4). My original intention was to just write the proof and then say, "if it is possible with just 5, it should be easy with a normal number like 15-20". I might want to try writing a simple version later.
You are right, I forgot that the algorithm assumes a pre-startup of dredging until you can shuffle. Deck should be defined, it is "number of cards in deck right now" because it means you can draw your whole deck. Updated those two points.
very interesting read, OP. I appreciate the effort you put into writing down all this information. I hadn't included this combo in my own gitrog deck because I don't play with people who respond well to such an intricate combo involving so many iterations, but I did enjoy reading about it.
however, the vorthos in me compels me to say that all the algorithms and math are basically the opposite of everything the gitrog stands for
It is impossible with 2 lands and kozi and any number of non-lands: (if even number of cards, then if library ends with land, land, kozi you are in trouble - if odd you are in trouble if it ends with kozi, land, land).
It can not be done with 1 land and kozi and some (non-zero) number of non-lands: (if odd number of cards in library, then if the land end up at the buttom always you can never draw. If even, then if the library ends with land, non-land, kozi, you cannot draw).
(of couse, if it is 1 land and kozi, you can just dredge, add the reshuffle over the draw trigger, let the reshuffle trigger and then dredge again and again putting reshuffle over the 2 draws and then draw the last 2 cards).
*EDIT:
Argument for how to draw deck with 3 lands in deck surely.
It is fairly easy to see how to draw your library if you only got kozi and either 1 land or 3 or more lands (and no non-lands), so I will ignore the case.
I will add 1 draw on stack or add 1 non-land to hand after at most every second reshuffle. After there is above n draws (where n is the number of cards in deck after reshuffling) draw all of them and repeatedly discard kozi before each remaining draw.
First some special cases that will be useful:
If you dredge kozi and a land at the same time, you can get 1 more draw next time, by first reshuffling and then drawing. I will therefore focus on the case where you only get 1 effect at a time.
If you got at least 1 land and kozi in deck you can reshuffle: dredge until you find kozi (or if you got a odd number and kozi is last, you must dredge the land at some point and can then draw the last card (kozi) by keeping the draw on the stack until you reach the end. In that case discard kozi).
You can get either: (1) reshuffle and a non-land drawn or (2) reshuffle and a draw on the stack (we might use 2 reshuffle for this latter effect) if you got 3 lands in library from the beginning as follows: Dredge until you get either kozi or a draw.
If draw, you got at least 1 land and kozi in deck so do as above to get at least a reshuffle in response to the draw .
If you dredge kozi, you have at least 3 lands left, so dredge until you get 1 draw, which you let resolve. Before drawing you got at least 1 land left in deck (and thus do not die to drawing the card).
If you draw a non-land you can let the shuffle resolve and you are up 1 non-land card.
If you got a land, you can reshuffle and you start with a land in hand. Discard it to put a draw on the stack. You have at least 1 land in the deck so get a reshuffle in response as above.
I have a general idea for how to do this in a simpler way. Basically, instead of worrying about reducing N, it will focus on build T. I will be doing some thinking to get the exact details and then either:
Update the algorithm if the land restriction is the same or better.
Add an extra simple version algorithm if the restriction is worse (>5).
If people are sick of reading about stuff just stop taking part. You have 100% control over what you read. Simic Ascendancy isn't going to get banned just because you didn't tell someone to shut up on the internet.
What is the point of all this? Without knowing the intermediate states and allowing removal or replys, its just math?
I assume that if someone is going to try to mess you up, they will either do it before you get the combo or during your first run through the deck. (forcing you to draw too many cards, exiling dakmor salvage, exiling kozilek, etc). If that happens, obviously you will do what you need to do.
The point then is that not every opponent will allow you to shortcut the combo (they might want to respond, they might not understand it fully, they might want to make you nervous so you possibly mess up, they might think you do not know how to do it properly, etc).
I am not sure what other intermediate state you could mean, this explicitly covers getting from start to finish.
Anyway: I think this is the simplest way to describe it:
Dredge repeatedly (in response to any draw triggers) until kozi is dredged (if you got an odd number of cards left, he might be last and you should resolve a draw trigger and discard him)
After dredging kozi, dredge and resolve draws while your library is strictly above 2.
After reshuffling, discard all lands drawn, in response to each other.
In response repeat this until you have more draws on stack than cards in library after reshuffle and draw all of them.
If you got atleast 3 lands this works and will draw your library after at most 4n/(l-1) reshuffles, because after every second reshuffle you get a draw from atleast (l-1)/2 lands (rounded down) - the division by 2 is because lands might come packed together. Observe that clearly you need at least 2n/l reshuffles in the worst case, because lands might come packed together.
Private Mod Note
():
Rollback Post to RevisionRollBack
"Hail to the speaker, hail to the knower; joy to he who has understood, delight to they who have listened." - Odin
Anyway: I think this is the simplest way to describe it:
Dredge repeatedly (in response to any draw triggers) until kozi is dredged (if you got an odd number of cards left, he might be last and you should resolve a draw trigger and discard him)
After dredging kozi, dredge and resolve draws while your library is strictly above 2.
After reshuffling, discard all lands drawn, in response to each other.
In response repeat this until you have more draws on stack than cards in library after reshuffle and draw all of them.
If you got atleast 3 lands this works and will draw your library after at most 4n/(l-1) reshuffles, because after every second reshuffle you get a draw from atleast (l-1)/2 lands (rounded down) - the division by 2 is because lands might come packed together. Observe that clearly you need at least 2n/l reshuffles in the worst case, because lands might come packed together.
I came to a similar conclusion while on a break, the flow I drew on a paper is pretty much exactly what you just said as it seems like the simplest route. Will be updating sometime soon, just have to make sure I completely understand the determinacy. I think the maximum iterations is a little bit more complicated than you said because there are occasions where you effectively net 0 cards, but those are always followed by occasions where you advance.
Edit: oh you said every other, yeh you probably saw the same thing I was looking at
I'm a tad confused as to why we need to respond to Kozi's trigger by discarding the Salvage. By then, isn't the Salvage in hand, so will not be shuffled into the library? If we are to discard the Salvage, conversely, doesn't it get shuffled in due to the trigger's resolution?
Furthermore, is there really a worry about decking in this case, when you can almost always just Dredge in response here in place of the draw (I guess there's an "almost" involved, but isn't focusing on it playing a bit too cautiously)?
Finally, I know the Gitrog is a key piece in this play, but is there a means to make this deck X/B/G and still be able to reliably get this off to a higher degree than strictly B/G and led by Gitrog? Because I feel like Sphinx's Tutelage or Laboratory Maniac would be beautiful cards to include, if it were possible. By that token, what are some good "on draw" triggers to exploit in this deck without a third color?
Kalkris, the point of the thread is that formally speaking the shortcut rules (basically the ones for supporting combos) requires that you can express a sequence of steps and how the end state will look like (like, I tap my ctreatures for mana, I use the mana to make a token off some other creature, this triggers intruder alarm, untapping everything and I repeat this 1000 times, ending with 1000 tokens). You therefore cannot use things like "almost always", because you then cannot be sure if this is the one time it fails.
Concretly, the reason we need to respond to the shuffle trigger is that we could be unlucky and all the lands are after Kozi every time (it happens every time with probability 0% but that is still not good enough for the rules of shortcuts in Magic, where you must provide a guarantee on the number of steps). The idea is that you discard salvage, get a draw from the frog and dredge Salvage back off that draw and perhaps trigger another draw. You never leave Salvage in the graveyard. By doing it the right way, you can be sure that you will eventually, after a number of steps depending on the number of lands (at least 3 though) and non-lands in your deck, be able to draw all the cards. If you did not have enough lands, you cannot formally use the shortcut rules.
If you want something to do with the abillity to draw all cards, then you can play dark ritual any number of times and finish with a drain life or something.
Private Mod Note
():
Rollback Post to RevisionRollBack
"Hail to the speaker, hail to the knower; joy to he who has understood, delight to they who have listened." - Odin
I know the Gitrog is a key piece in this play, but is there a means to make this deck X/B/G and still be able to reliably get this off to a higher degree than strictly B/G and led by Gitrog? Because I feel like Sphinx's Tutelage or Laboratory Maniac would be beautiful cards to include, if it were possible. By that token, what are some good "on draw" triggers to exploit in this deck without a third color?
You could play this combo in a different deck, but the big problem is that one of the cards (the discarder) is pretty much useless without the combo. Because of this, you will have several usually dead cards in your deck. If you have a discarder that does something else in the deck, like Mind Over Matter or Land's Edge, it will be a bit better, but still seems like a stretch.
Are we afraid of decking here? Is that the sole reason for this statistical thread? Because the shortcuts are available to announce otherwise.
~Lil Kalki
We are afraid of calls for slow play, because just letting your deck shuffle whenever an eldrazi gets flipped can potentially warrant one.
More importantly, how can shortcuts be announced if you do not know how to perform the combo? What about if your opponent does not know how to perform the combo?
If you are afraid for a 'slow play' call, this wont help you. Since you can't tell the judge every single intermediate state, you will have to do each step.
Thats why the 4 Horsemen deck isnt legal any more.
If people are sick of reading about stuff just stop taking part. You have 100% control over what you read. Simic Ascendancy isn't going to get banned just because you didn't tell someone to shut up on the internet.
No, 4 horsemen has a problem different from this. Its problem is that it might not advance the state of the game, within any constant number of moves. Here we advance guarenteed. For instance, after 80000 moves you will draw your deck if you start with 3 lands and 100 cards in deck (more lands will speed it up). 4 horsemen cannot say the same, since they want to mill 3 specific cards before hitting a reshuffle. I.e. there is a probaility that they will do it in any loop, bjt no matter how many loops they do it will never be certain that they will have gotten it at some point.
I was wrong about the short cuts though since you cannot let the action depend on what happend.
Edit: simple proof that algorithm gives you atleast 1 more draw trigger (or a drawn non-land) every 2 reshuffles with 3 (or more) lands. Thus at most 2 times the number of cards reshuffles are needed.
Initial there are 3 lands in deck and kozi. There are 2 cases, either there are some lands before kozi or not.
If lands before kozi, you get at least 1 draw trigger from them (in case kozi is the last card and there is an odd total of cards you get at least 2 triggers and thus still have 1 after drawing him). Thus atleast 1 draw trigger before reshuffling.
If all lands after, then you resolve at least 1 draw trigger. If you draw a non-land you are good. Otherwise, if you draw a land, you can reshuffle and discard it to get a draw trigger and then get another reshuffle (in the worst case), since you got at least 2 lands left in deck (note that for every card drawn you dredged a land last reshuffle, which then get shuffled in).
If you are afraid for a 'slow play' call, this wont help you. Since you can't tell the judge every single intermediate state, you will have to do each step.
Thats why the 4 Horsemen deck isnt legal any more.
4 Horsemen only has a problem because Dread Return is a sorcery so they have to let the stack completely unwind before casting it. This is the same problem Gitrog would have if it tried to do a loop with Chains of Mephistopheles. However, since Putrid Imp type cards can respond to the shuffle trigger, every iteration through the deck can be guaranteed to demonstrate advancing the board state. That is what this algorithm is all about. A method of performing the combo in a determinate way. Slow play is a misnomer because it is really about repeating a board state rather than playing slowly or doing a combo that takes a long time.
Just to note, 4 Horsemen is legal, it is just really dangerous to use because you are probably going to get a slow play call. You could get super lucky and just happen to see everything in the right order, but it is unlikely.
Fair point, not illegal but unlikely to go un-penalized.
But the stack being empty has nothing to do with it according to the judges who have spoken on it. It is because you can't tell me what the board state looks like 13 loops in, in case I want to respond then. The board state advances in 4 horsemen, they just cant say how it will look. I don't think you can either.
If people are sick of reading about stuff just stop taking part. You have 100% control over what you read. Simic Ascendancy isn't going to get banned just because you didn't tell someone to shut up on the internet.
What you do is you explain the mechanics of how the combo functions, then show your entire deck to them (since the GY is a public zone). From there, you show Kozilek in your deck, thereby proceeding to explain that you intend to do this infinitely until you have enough on-draw triggers on whatever you're trying to draw for, or play infinite Dark Rituals into Exsanguinate, or whatever have you. If they do not then concede, they can be called out for slow play because they are prompting you to do this and refusing to acknowledge the motions of this combo and that it will inevitably get them.
It all boils down to communication. Or is that not how that could work?
For anyone who does not know, "The Gitrog Monster Combo" is The Gitrog Monster + Putrid Imp + Dakmor Salvage + Kozilek, Butcher of Truth. You discard Dakmor Salvage, causing The Gitrog Monster to create a draw trigger. That trigger is used to dredge Dakmor Salvage back into your hand. If you dredged 1 or 2 lands, you get a free draw trigger. The idea is that you can keep doing this to draw an unlimited number of cards. However, you need Kozilek to shuffle your graveyard back into your library, and that complicates things significantly. If you simply let your graveyard shuffle whenever Kozilek was dredged, you could theoretically get Kozilek + some nonland card on the top of your deck an infinite number of times. To resolve this problem, I wrote a logical proof (*This proof needs to be rewritten to demonstrate new methodology*) to show that the combo can be performed in a way that guarantees you draw your entire deck in a finite number of steps. That is fun to think about, but it is still difficult to tell the actual steps that should be performed at any given point of the combo. For this, I present the following flow chart algorithm.
Downloads:
The PDF files are saved in a folder on MEGA because MTGSalvation does not accept PDF.
Included Files:
A flow chart algorithm that shows one way to perform The Gitrog Monster Combo in a determinate manner.
Changelog:
An accompanying file for the algorithm that explains the shorthand and symbols used.
Changelog:
Added a pre-setup requirement for the algorithm
A logical proof that demonstrates how you can guarantee that each iteration through your deck is advancing you towards the end goal of getting your whole deck in hand.
Notes:
1. I have no draw triggers at the start of this, but I do have Gitrog and imp out, and Dakmor Salvage in my hand.
2. I discard Dakmor salvage, generating a draw trigger.
3. I use that trigger to dredge Dakmor salvage back to my hand.
4. The top two cards of my library happen to be Kozilek and Sol Ring.
5. I shuffle my graveyard back into my library from Kozilek's trigger.
6. Go to Step 1. (I keep shuffling such that, by pure chance, Kozilek and Sol Ring are the top two cards every time until the heat death of the universe)
What's that card, friend?
Fully-powered 600-Card "Dream Cube" https://cubecobra.com/cube/list/dreamcube
450-Card "Artificer's Cube" https://cubecobra.com/cube/list/artificer
Cubing in Indianapolis...send me a PM!!
Yes, this is the very core of the idea. The trick here is that you do not simply allow your deck to shuffle (that is only allowed right after Start when it says "Shuffle Graveyard"). Instead, you discard Dakmor Salvage in response to the shuffle trigger and keep going. I just updated the Guide to make a comment about that, good catch.
AH! I get it now. Yes, given that the absolute key decision point that makes the algorithm possible is choosing whether or not to discard Dakmor while Kozilek's trigger is on the stack, I think that needs to be spelled out a little more clearly for dense people like me.
Great work on this. It's fascinating.
http://www.mtgsalvation.com/forums/the-game/commander-edh/multiplayer-commander-decklists/680594-the-gitrog-monster-kermit-druid-combo
You respond to the shuffle trigger, and keep going.
Retired EDH - Tibor and Lumia | [PR]Nemata |Ramirez dePietro | [C]Edric | Riku | Jenara | Lazav | Heliod | Daxos | Roon | Kozilek
There is a small error with you calculation. The location of Emrakul in your deck has an effect on the minimum number of draws you can get because you need to unwind the stack (so you do not lose the game from drawing on an empty deck). The worst case scenario would be [Emrakul], [all non-lands], [all lands]. In that case you will get about 1 draw for each 3 lands. The reason my version is so complicated is that I wanted to prove that it can be done with as few lands as possible (I demonstrated it down to 5, but someone might be able to revise it down to 4). My original intention was to just write the proof and then say, "if it is possible with just 5, it should be easy with a normal number like 15-20". I might want to try writing a simple version later.
You are right, I forgot that the algorithm assumes a pre-startup of dredging until you can shuffle. Deck should be defined, it is "number of cards in deck right now" because it means you can draw your whole deck. Updated those two points.
however, the vorthos in me compels me to say that all the algorithms and math are basically the opposite of everything the gitrog stands for
It can not be done with 1 land and kozi and some (non-zero) number of non-lands: (if odd number of cards in library, then if the land end up at the buttom always you can never draw. If even, then if the library ends with land, non-land, kozi, you cannot draw).
(of couse, if it is 1 land and kozi, you can just dredge, add the reshuffle over the draw trigger, let the reshuffle trigger and then dredge again and again putting reshuffle over the 2 draws and then draw the last 2 cards).
*EDIT:
Argument for how to draw deck with 3 lands in deck surely.
It is fairly easy to see how to draw your library if you only got kozi and either 1 land or 3 or more lands (and no non-lands), so I will ignore the case.
I will add 1 draw on stack or add 1 non-land to hand after at most every second reshuffle. After there is above n draws (where n is the number of cards in deck after reshuffling) draw all of them and repeatedly discard kozi before each remaining draw.
First some special cases that will be useful:
You can get either: (1) reshuffle and a non-land drawn or (2) reshuffle and a draw on the stack (we might use 2 reshuffle for this latter effect) if you got 3 lands in library from the beginning as follows: Dredge until you get either kozi or a draw.
Also, yeah, it is interesting math.
Draft my Mono-Blue Cube!
lichess.org | chess.com
I assume that if someone is going to try to mess you up, they will either do it before you get the combo or during your first run through the deck. (forcing you to draw too many cards, exiling dakmor salvage, exiling kozilek, etc). If that happens, obviously you will do what you need to do.
The point then is that not every opponent will allow you to shortcut the combo (they might want to respond, they might not understand it fully, they might want to make you nervous so you possibly mess up, they might think you do not know how to do it properly, etc).
I am not sure what other intermediate state you could mean, this explicitly covers getting from start to finish.
Anyway: I think this is the simplest way to describe it:
Dredge repeatedly (in response to any draw triggers) until kozi is dredged (if you got an odd number of cards left, he might be last and you should resolve a draw trigger and discard him)
After dredging kozi, dredge and resolve draws while your library is strictly above 2.
After reshuffling, discard all lands drawn, in response to each other.
In response repeat this until you have more draws on stack than cards in library after reshuffle and draw all of them.
If you got atleast 3 lands this works and will draw your library after at most 4n/(l-1) reshuffles, because after every second reshuffle you get a draw from atleast (l-1)/2 lands (rounded down) - the division by 2 is because lands might come packed together. Observe that clearly you need at least 2n/l reshuffles in the worst case, because lands might come packed together.
I came to a similar conclusion while on a break, the flow I drew on a paper is pretty much exactly what you just said as it seems like the simplest route. Will be updating sometime soon, just have to make sure I completely understand the determinacy. I think the maximum iterations is a little bit more complicated than you said because there are occasions where you effectively net 0 cards, but those are always followed by occasions where you advance.
Edit: oh you said every other, yeh you probably saw the same thing I was looking at
Furthermore, is there really a worry about decking in this case, when you can almost always just Dredge in response here in place of the draw (I guess there's an "almost" involved, but isn't focusing on it playing a bit too cautiously)?
Finally, I know the Gitrog is a key piece in this play, but is there a means to make this deck X/B/G and still be able to reliably get this off to a higher degree than strictly B/G and led by Gitrog? Because I feel like Sphinx's Tutelage or Laboratory Maniac would be beautiful cards to include, if it were possible. By that token, what are some good "on draw" triggers to exploit in this deck without a third color?
~Lil Kalki
Proud Disciple of the Church of the Wary
Concretly, the reason we need to respond to the shuffle trigger is that we could be unlucky and all the lands are after Kozi every time (it happens every time with probability 0% but that is still not good enough for the rules of shortcuts in Magic, where you must provide a guarantee on the number of steps). The idea is that you discard salvage, get a draw from the frog and dredge Salvage back off that draw and perhaps trigger another draw. You never leave Salvage in the graveyard. By doing it the right way, you can be sure that you will eventually, after a number of steps depending on the number of lands (at least 3 though) and non-lands in your deck, be able to draw all the cards. If you did not have enough lands, you cannot formally use the shortcut rules.
If you want something to do with the abillity to draw all cards, then you can play dark ritual any number of times and finish with a drain life or something.
You could play this combo in a different deck, but the big problem is that one of the cards (the discarder) is pretty much useless without the combo. Because of this, you will have several usually dead cards in your deck. If you have a discarder that does something else in the deck, like Mind Over Matter or Land's Edge, it will be a bit better, but still seems like a stretch.
~Lil Kalki
Proud Disciple of the Church of the Wary
More importantly, how can shortcuts be announced if you do not know how to perform the combo? What about if your opponent does not know how to perform the combo?
Thats why the 4 Horsemen deck isnt legal any more.
I was wrong about the short cuts though since you cannot let the action depend on what happend.
Edit: simple proof that algorithm gives you atleast 1 more draw trigger (or a drawn non-land) every 2 reshuffles with 3 (or more) lands. Thus at most 2 times the number of cards reshuffles are needed.
Initial there are 3 lands in deck and kozi. There are 2 cases, either there are some lands before kozi or not.
If lands before kozi, you get at least 1 draw trigger from them (in case kozi is the last card and there is an odd total of cards you get at least 2 triggers and thus still have 1 after drawing him). Thus atleast 1 draw trigger before reshuffling.
If all lands after, then you resolve at least 1 draw trigger. If you draw a non-land you are good. Otherwise, if you draw a land, you can reshuffle and discard it to get a draw trigger and then get another reshuffle (in the worst case), since you got at least 2 lands left in deck (note that for every card drawn you dredged a land last reshuffle, which then get shuffled in).
4 Horsemen only has a problem because Dread Return is a sorcery so they have to let the stack completely unwind before casting it. This is the same problem Gitrog would have if it tried to do a loop with Chains of Mephistopheles. However, since Putrid Imp type cards can respond to the shuffle trigger, every iteration through the deck can be guaranteed to demonstrate advancing the board state. That is what this algorithm is all about. A method of performing the combo in a determinate way. Slow play is a misnomer because it is really about repeating a board state rather than playing slowly or doing a combo that takes a long time.
Just to note, 4 Horsemen is legal, it is just really dangerous to use because you are probably going to get a slow play call. You could get super lucky and just happen to see everything in the right order, but it is unlikely.
But the stack being empty has nothing to do with it according to the judges who have spoken on it. It is because you can't tell me what the board state looks like 13 loops in, in case I want to respond then. The board state advances in 4 horsemen, they just cant say how it will look. I don't think you can either.
It all boils down to communication. Or is that not how that could work?
~Lil Kalki
Proud Disciple of the Church of the Wary