A year or two ago, I was terrible at sideboarding. I could never figure out what cards you wanted to sideboard out of a deck, and I missed a lot of little subtlety. I could pick up a complicated deck and play it, but I had no idea how to metagame it or tweak it when I needed to. In other words I could play a deck, but I had not learned the skills it takes to be a better player. This kind of player can do well at tournaments, but only to a certain extent. The right deck can be incredibly favorable in a metagame, and any player can win with it. But when the metagame becomes unfavorable, then only the better players will be able to adapt and continue to win. I am presenting this article because I hope it will make it easier for other people to acquire these skills and succeed.
The most important piece of a deck is the manabase; the wrong manabase will kill you very quickly and silently. And yet often bad manabases persist, or even get copied. With Legacy’s dual lands and fetchlands, you can often have moderate success with a bad manabase just by filling your deck full of fetches and duals. You’ll hit your colors most of the time and just struggle with a few decks that take advantage of this. But with the Wasteland decks seemingly at an ebb, a lot of awful manabases are doing well in the metagame. I’ve written about this cycle and I hoped that the Wasteland-packing Threshold decks would help stop this trend, but players have been lured to four- and five-color Threshold decks that pack to Wasteland instead of packing Wastelands. It is possible to run 4 Wasteland and 4 Stifle in a Threshold deck, even while playing Counterbalance. For everyone else, basic lands are a must. Most decks start with six fetchlands in their primary color and then four of every dual land that links their primary color with their secondary colors. They add one or two basic lands and fill the rest of their slots with more dual lands. Then they’re surprised when they lose to Blood Moon. Why is four of each dual land considered necessary? It is incredibly rare to actually have all four dual lands in play at any time, so why run 4? Dual lands do two things very well: they let a fetchland fetch an enemy color, and they provide access to two colors of mana at once. They are however incredibly fragile; fetching a turn 1 dual land for a spell like Thoughtseize is likely to be met with a turn 1 Wasteland. While most of the time that Wasteland will be the wrong play, sometimes it will catch you with your pants down.
Dragon Stompy exists. You absolutely need a basic land in every color you need to win the game, and you would be better off getting more than one. Counter-intuitively, the colors least represented by your fetchlands are the ones you need more basic lands in order to see them often enough. In UGW Threshold I am fine with only having one Plains because I have at least four fetchlands that can find it. I do need two Forests because I will have at most three fetchlands that can find it. Getting basic lands into play is superior whenever possible (who knows when that guy boarded Price of Progress just to spite you), so dual lands should be used only when you absolutely need them. In a perfect world, we would only draw fetchlands and the appropriate colored basic lands; when you needed a dual land you would fetch it. We don’t live in that world, but we can get close by running one to three of each dual land instead of four. Generally you want as many duals in a color as times you want to tap that color, plus a basic land if you want constant access to a color. If have one to five cards in a color with no abilities, and it’s a spell you want to cast once and not on turn 1, you only need one dual land in that color. If you have between six and eight or nine, you want two dual lands. Any more than that and you want three, and you can afford four dual lands if you have more than eight or nine basic lands. Remember to count sideboard cards; you might have just four Swords to Plowshares in the maindeck, but if sideboard contains Serenity and Orim’s Chant, you’ll need three Tundras instead of two. Take into account that you need more sources against decks with land disruption; in UGB Threshold you’ll want the third Underground Sea to cast Engineered Plagues against Goblins. But if you run UGW Threshold with few White sideboard cards and a slow number of decks packing Wasteland, you can probably afford to drop to 2 Tundras. You do have to be much more conservative with your dual lands this way. If you are used to overwhelming Wastelands by just fetching out exclusively duals, and if none of the players in your local metagame run Blood Moon or Price of Progress (or I suppose Ruination), you may be able to get away with more dual lands. I prefer to be a little safer though.
And how many fetchlands should you run? Tuning that aspect of your manabase is based on intuition and experience more than science and formula. I start my decklists with four fetchlands and work my way up. If testing shows that I tend to have the right number of lands in my opening hands but I flood later in the game, I add more fetchlands. At this stage in the game I can also add more lands if necessary. Right before I test the deck, I identify the weaker or more cuttable cards, and I keep an eye on those cards as I play the game. If I constantly find myself wishing those cards were something else (for example a land) I make the switch. You can maximize your testing this way by tracking how often you prefer one card to another. This method is a good way to evaluate similar cards like Smother versus Diabolic Edict or Counterspell versus Mana Leak.
A deck you get from the Internet was built the way it was for a reason. And yet people are often tempted to make changes without realizing the consequences of those changes. The cards into a deck fall into a few categories: card draw/tutoring, disruption, removal, threats, mana, combo pieces. In almost every circumstance it is wrong to change cards between categories without testing the deck. Turning one type of removal into another is fine, but how can you know that the deck does not have enough threats without testing it? Once you divide the deck into these categories then you can figure out how many of each you want. If you find yourself not having enough removal, threats, mana or combo pieces, then you know what you need to add. Knowing when you need card draw is a bit trickier. Generally I know that I need to add more card draw when the deck is too inconsistent. If I find that I have too much removal in about half my games and not enough in the other half, I will try to make some cuts in the other categories evenly in order to add some filtering power. For non-Blue decks the best place to start is often Sensei’s Divining Top, since it basically lets you choose which class of card you want to draw. Then again, I tend towards controlling or midrange style decks, and my decks often have more draw power than they need. I have this feeling (don’t most people share it?) that I am better than most players at any given tournament. If the only way you can lose is by drawing the wrong cards at the wrong time, it can be correct to err on the side of extra draw power.
You will perform far better with a deck if you build the sideboard yourself. When you use someone else’s sideboard it is far easier to under-sideboard or over-sideboard. Either way will leave you hurting in some key matchups. The first problem with using someone else’s sideboarding plans is memorization. I don’t know if everyone has this problem, but I can’t remember exact sideboarding plans that someone else came up with. It’s rote memorization, and I tend to be bad at that. If I build the sideboard myself, I remember it far better. This leads to fewer mistakes and potential sideboarding errors.
If you have to sideboard on the spot, be very careful. I’ve already suggested shuffling in your whole sideboard and taking fifteen cards out. Aside from the benefits I already espoused, it helps the sideboard sleeves wear evenly and helps you go longer without a marked deck. Otherwise, your sideboard tends to be less worn than your maindeck and this can cause a marked deck. After I’ve shuffled my sideboard into my deck, I flip through it and remove cards into a few piles. The first pile is the cards that definitely come out of the deck, like Blue Elemental Blasts against Threshold. The cards that will probably come out of the deck get sorted into their own piles by cards, so I have the Thoughtseizes in one pile and the Engineered Plagues in a separate pile. Once I have the cards out of the deck that are definitely sideboarded out, I can pull the weakest cards from the probably piles until I have removed 15 cards from the deck. In that method I am less likely to miss sideboarding in a card that might be good (like an extra Engineered Explosives), but I am still very likely to take the wrong card out.
About a year ago, I suddenly became pretty good at building sideboards. I credit this very systematic method. I’m not claiming to invent it; it’s not very original, but it does help me a lot. The method takes three steps. For the first step once I have my maindeck mostly figured out, I write down the matchups I expect to see the most or that I most want to sideboard for. For each one, I write down the cards in the maindeck that are most likely to come out in the deck. I also try to include the cards that can come out if there are more slots. So for UGW Threshold versus Goblins, I’d write down that 2 Spell Snare 4 Counterbalance are definitely coming out. Once I have created that list for every deck, I have an idea of how many cards I need for each matchup. So step two is where I actually craft the sideboard. I look at the total number of sideboard slots needed. If that number is less than fifteen, go back to step one and include more decks. If that number is fifteen, then I can just start writing down sideboard cards. There are two keys here: combination and prioritization. As much as possible I try to get similar decks to use similar sideboard cards. This might mean bringing in a sub-par card for one matchup if that card is superior in another matchup. It really is an optimization problem, especially when you consider part two: prioritization. The decks you expect to see the most, and the decks that you need the most help in, get the most and the best sideboard cards. If you are most worried about Dragon Stompy, pack whatever cards help you the most there and then accommodate the rest of the sideboard around those cards. This might mean you start sideboarding against Goblins with 4 Blue Elemental Blast, or it just might mean that your Tormod’s Crypts and Planar Voids mean that you only have 10 slots to use the rest of the sideboard. You do whatever it takes to get to just 15 cards. Finally, once you have the fifteen cards set, go back through and rebuild the lists from step one but while filling in the cards you want to take out. This is where you discover that the Wrath of Gods you are sideboarding against Goblins are better than a random counter against Ichorid. More famously this is also the step where you discover that your Blue Elemental Blasts you are sideboarding against Goblins are better than Swords to Plowshares against Reset High Tide. And when you’ve done the work to craft the sideboard, you’ll remember it better.
It used to be on The Mana Drain that no one posted sideboards for decks because the sideboard was tuned for the local metagame, and the particular set of fifteen cards would not be optimal for everyone else. I’m not saying we need to get back to that practice, but it’s always worth tweaking or building your own sideboard even for a netdeck.
PS: Remember Doug Linn Whiskey Tango Foxtrot deck from two weeks ago? I was playing around with it, and I got the deck fairly well tuned. The Salvagers combo and the Enlightened Tutor engine felt out of place; it’s always easier to just set up Painter’s Servant plus Grindstone than anything else. Using that experience and tools from the other build, I put together a Survival of the Fittest plus Painter’s Servant deck. I’ll be testing this and I’d like to report back on it in a month, but I wanted to put the deck out and get some responses from readers:
Creatures (21)
- 4 Rofellos, Llanowar Emissary
- 1 Genesis
- 2 Eternal Witness
- 4 Trinket Mage
- 1 Squee, Goblin Nabob
- 1 Mesmeric Fiend
- 1 Harmonic Sliver
- 4 Tarmogoyf
- 2 Shriekmaw
- 1 Painter's Servant
Lands (21)
Spells (18)
