[Back] [Home] [Up] [Next]

Salmon Snack.

Description of deck by it's author (quoted):
"Combination decks are decks of details. This fact is true with many deck however it's ever more important for combination deck than aggressive decks (which don't really care which threat they play.) Combinations decks require a very exact mix of solutions, acceleration and combination cards. After months of testing I feel comfortable stating that this is the correct build for a Snack based combo deck and that in the hands of an expert it is the best deck in OD/Torment Type II.

Please note the caveat ""in the hands on an expert."" This deck is incredibly difficult to play properly. I hesitated in publishing this deck as I'm not confident of my ability to describe how to properly play Salmon Snack. Even once you understand the mechanics you'll find yourself repeatedly making playing errors with this deck. In many fights the deck is forgiving of minor playing errors. However others require precise play and in those the smallest of playing mistakes will cost you games and matches.

To compound this difficulty, due to the non-interactivity in playing Snack, I find it socially unpleasant to test Snack even with my team of fairly dedicated and serious players. When playing against Snack my opponents eye's often glaze over with boredom. This is not a sign that they're having a good time :-). It's even less fun for opponents on MODO. However, if you want to play Salmon Snack you have to practice it. A lot.

Salmon Snack also has a severe shuffling issue, more so than any other deck I've dealt with, including other combination decks. This is due to the deck sorting itself during game play. At the end of most games with Salmon Snack you'll have separated out all the land from all the spells and even subdivided the spells somewhat. This means Salmon Snack requires through shuffling and although MODO relieves this difficulty in real life matches it's a pain.

I've noticed that most players do not sufficiently shuffle their decks in tournament play and certainly not while practicing. This is in part due to the serious number of riffle shuffles required to ""sufficiently randomize"" a deck. From a mathematical perspective a strong argument can be made for requiring several million riffle shuffles in order to ""sufficiently randomize."" Another factor is the poor technique that many players use in the limited shuffling they do. What makes up for this deficit and keeps the system from breaking down is almost all player use instinctual knowledge to judge the ""goodness"" of a deck's shuffle. Most players shuffle until their deck feels shuffled. Feels shuffled reflects this ""goodness"" quality that is more closely related to ""sufficiently distributed"" than the ""sufficiently randomized"" that the DCI floor rules call for. However, as people don't have the ability to ""sufficiently randomize"" a deck via manual shuffling under the available time allotments, ""sufficiently distributed"" is the adaptation that seems to be used.

As mentioned, normal shuffling towards ""sufficiently distributed"" relies on a feeling of ""goodness"". This doesn't work for Salmon Snack. I believe this occurs because of the following: You'll notice that players mulligan more frequently when their deck is shuffled from a completely sorted state than after it's been thoroughly mixed . This can be most readily seen in the above average number of mulligans during game 1 of a Limited event or immediately following a deck check. This increase in mulligans is due to the poor starting basis of the deck and this poor basis throws off players feeling in determining the ""goodness"" of a shuffled deck. As Salmon Snack starts every shuffling period with a poor basis, shuffling Snack until it feels shuffled will cause far more mulligans than should occur randomly.

To help with this I strongly suggest using the following protocol for shuffling Salmon Snack (I actually recommend it for any between match shuffling, it's necessary however for Snack). Given a deck that you desire to randomize:
1) Pick a number to use for the number of riffle shuffles throughout this protocol (some number from 1-6 say 3)
2) Riffle shuffle the deck that number of times
3) Roll a 10 sided die and add 1 (giving a number from 2-11). Flat/Pile shuffle the deck on that number of piles.
4) Repeat step 1-3, using the same number in step 1 and rerolling in step 3.
5) As above repeat steps 1-3 for a third time.
Note: make sure you roll a distinct number every time in step 3, rerolling where necessary.

This should be the entirety of the shuffling done with the deck. The goal of this shuffling protocol is to quickly & randomly ""sufficiently distribute"" the deck and remove all human instinct from the shuffling process. This protocol has been heavily tested and in practice seems to come up with hands that require mulliganing about 1 in 13, very close to the expected value if the deck was truly randomized. Let me reiterate please use this shuffling protocol whenever playing the deck even in casual practice with it. Otherwise you're very likely to practice with very unusual draws from the deck and not gain meaningful practice (and it's unlikely you're just playing Salmon Snack for fun.)

So if I haven't chased you off yet, then onto a discussion of the deck. The first difference between Salmon Snack and other related combination decks is it's pace. Salmon Snack has a beat and to win with Snack you play to the beat. The beat is: Land; Land, Mana Accel; Land, Defensive Spell; Draw Cards, Land; Go off. The deck follows this pattern a lot. About the only variation is sometimes you ""Draw Cards, Land; Defensive Spell"", before going off. The deck goes off on Turn 4 occasionally and Turn 5 regularly - even opposed.

Going off means either taking a whole bunch of turns in a row or putting a bunch of mana in the pool and Blaze/Death Grasping the opponent to death. Going off usually begins with a Time Stretch. If you can get a Stretch to resolve you WILL win. Snack is very trustworthy at the point in time that you can cast a Stretch. Following a Stretch you will top deck either another Stretch, a recursion card or a draw spell. This reason is once you can generate the 10 mana needed you've searched out over half of the lands in the deck. There's not much left but ""good stuff"" to draw.

What makes the engine go off is the massive mana generated by Far Wanderings. Far Wanderings has a difficulty however - it requires threshold to operate fully. Threshold is most easily achieved through Fact or Fiction. The last part of the triangle that runs the engine is Early Harvest. Often you'll cast threshold Far Wanderings have 3 left in your pool and then Early Harvest (giving you a minimal of 9 mana to play with.) You can run the engine without Early Harvest but it goes off a turn or two sooner with it.

Now to look at the build. First of all 18 lands. Casual observation would indicate that this insufficient. However, testing (both mine and other Snack builds) has shown that 18 is the correct number. Please see the discussion on mulliganing below.

Next are the Rampant Growths. Unless you need to fetch a Forest (if you Growthed off a Chromatic), you'll generally fetch an Island or a Mountain (depending on which defensive spell you intend to cast next), followed by a Swamp or Plain. The goal is to rapidly get one of each land types to support the Allied Strategies.

The Fertile Ground is also necessary for the deck, as it's beat desires a mana curve of 1,2,4. This is a very important pace for the deck and although I won't mulligan a hand just because it doesn't have a 2 drop mana accelerator, it's the single most important thing to have in your opening hand (aside from land.)

Far Wanderings are also a mana accelerator and while they don't speed you to 4, they often speed you from 3 to 5 and more importantly they let you jump from 5 on one turn to 9 on the next. As implied if you need to cast a Far on turn 3 pre-threshold you generally should do so. About the only times you won't are when you're going to cast a Fact that turn/the next or are holding back to counter. This is different from the Rampant Growth and Fertile Ground which you ALWAYS cast on turn 2. When given a choice midgame (turn 4/5) to cast a threshold Far Wanderings or a card draw spell generally you'll cast the Far Wanderings first. And lastly, make sure to look towards casting a Far Wanderings followed by an Early Harvest

Early Harvest is the mid-game Dark Ritual in Snack. With 6 lands it lets you free cast an Early Harvest and with 7 it allows you to immediately cast Time Stretch. Late game of course it allows you to cast really big X spells.

The primary defense spell in Salmon Snack is Earthquake. It is both necessary and desirable as it handles R/G efficiently (remember on turn 3 you can usually quake for 3) and it provides an answer to Braids (along with her assistants.) The counterspell of choice is Evasive Action. As a spell it's of minor importance to Snack unless the opponent has counters of their own. Then Evasive Action is a key spell in supporting your ability to go off. Ghitu Fire and Death Grasp are also Defensive Spells. Don't be afraid to use them for defense. You can always recur them for the kill.

Allied Strategies and Fact or Fiction provide the card power. The second most common way Salmon Snack loses is not drawing one of these spells over the course of a game (the most common way of losing is player error.) Both provide up to 5 card look ahead and over the course of a game (while going off) you'll regularly draw most if not all of the cards in your deck. Occasionally both of your non-global X spells are in the last few cards in your deck. Not to worry and no need to Quake the game into a draw. Instead once you're down to 5 or less cards cast a Fact or Fiction. It will resolve with however many cards remain in your library (without decking you.) You opponent will then have to make piles using the 2 or 3 cards (or however many are left in the deck). Then you can choose the X spell, cast your Harvest along with the chosen spell and start shuffling anew.

Lastly there are the recursion cards, Restock and Nostalgic Dreams. I wish I could provide more evidence other that to say ""thorough testing has shown that the numbers are correct."" However, that's all the evidence I can provide. I can say that the split between Dreams and Restock is significant in fights where the opponent is targeting you with Lobotomy style effects.

EDT wrote an excellent column on considerations of when to mulligan. In line with those concepts I'd like to discuss when to mulligan with Salmon Snack. With Snack you always mulligan a 0 land hand. However, you keep a 1 land hand if you have a Chromatic Sphere. Also, keep a 1 land if it's a Forest and you have a Rampant Growth or Fertile Ground in hand and you're going second. If you're going first, have a Forest and a Growth or Ground the decision whether to mulligan or not is based on the quality of the other 5 cards.

Playing Salmon Snack is like playing Magic the Puzzling every game - I was tempted to name the deck Vitamin I (Ibuprofen) from the headaches it can give you when playing it. However here's some general play advice:
When in doubt build your manabase.
Look for chances to generate threshold (sometimes you'll ""blow"" and Early Harvest to achieve threshold.
Don't go off till you have to.
Go off by Time Stretching.
If you've got nothing better to do, Restock/Dreams back something (defense, card power, even mana acceleration)
Tap the Swamp and Plains first (unless you have the Grasp in hand.)
Don't worry about Haunting Echoes. It doesn't hurt this deck that badly - unless you lose both the Fire and Grasp you can be Lobotomized for any two cards types and it won't even slow you down. Seriously - practice letting your opponent Echoes you to help build your confidence in the decks resiliency. "

Have a casual deck you want to submit to MDV?  Go here.  Do you see an error on this page?  Email it to this address.   

by Jay Schneider @ www.brainburst.com

UNIQUE: Domain - Far Wanderings / Time Stretch