Should we take good cards out for better cards while sideboarding (magic 2014 rules update)?

Question

It has been announced, among other changes, that you are allowed to increase the number of cards in your deck beyond 60 while side boarding. Meaning you can bring 5 cards in and not have to take 5 cards out.

There are decks and match ups where some cards are just bad, and you want to take them out (e.g. removal against legacy storm combo). Some match ups you have plenty of sweet cards in your deck, you just have sweet cards in your sideboard too. In this situation, it can be very difficult to decide which cards are "least relevant" to the matchup, to cut in favour of the sideboard cards which have a much bigger impact.

With the rules change, are we better off trying to keep our deck count to 60 wherever possible? or should the conventional wisdom be "cut the bad cards, bring in the good cards" and resultant card count being an afterthought?

Answer 1

Short answer

Yes. You should take cards out to keep a 60 card deck size.

Medium answer

Yes. It's always legal to play more than 60 cards. But when you design a deck, you don't - you keep it to 60 as much as possible. When you sideboard, you're redesigning your deck on the fly - and the same reasons that you kept it to 60 cards in the first place still apply.

Long answer

In a typical game (some highly specialised graveyard decks aside), you will only ever hold in hand and play a fraction of your deck - even in a game that goes to 13 turns (possible in casual but pretty unlikely in constructed), you will typically only have drawn 20 of your 60 cards - around a third of your deck.

The other two thirds, no matter how awesome the cards are, did you no good whatsoever.

So Magic is won on the draw. (After all, you win games not because you had great cards in your deck but because you had great cards on the battlefield.) If you've designed the deck properly, it's based around a winning strategy and some key cards that support that. You want to have as great a chance of drawing those winning cards as possible.

The mathematics of card draw and distribution for Magic have been extensively studied. (There's a pretty good article on the mathematics of card draw available.) And one of the reasons for the 60-card minimum is that it makes your deck unpredictable. Even if you have 4 of a key card in the deck, your chance to see one in the opening hand is just under 40%, and even by turn ten your chance is only 71.5%. (And if you haven't drawn it by turn ten, you're probably never going to play it.)

Increasing the deck size by only one card - to 61 cards - immediately takes about 1% off both those chances. So why harm your odds like that? If the new card is better for your strategy than the increased odds of drawing one that you have four of, your should remove one of those four. And if not, then you shouldn't be adding it.

(A telling point is that all of the top players field exactly 60 card decks. They don't have to; constructed rules permit more. They choose to.)

Some of the top players will occasionally play a 61 card deck. They do this to protect against decking and to subtly adjust their probabilities. But you can believe they plan for ages while considering the benefits of that 61st card, and they don't hesitate to remove it. If you're still at a level where you're asking this, you probably shouldn't try it.

So: Yes. Keep to 60 cards.

Answer 2

Tynam's answer covers the general case very well: the minimal deck size is optimal, and there are very few good reasons to deviate from that.

Nonetheless, here are some examples of corner cases where you may legitimately want to deviate from making one-for-one substitutions when you sideboard.

Storm-style combo decks

Unlike a more "traditional" two- or three- card combo deck, storm-style decks are essentially using every card in the deck as part of their combo. Good storm decks have a carefully-tuned ratio of cantrips, rituals, and kill cards, to avoid "fizzling" when they go off as much as possible (in the hands of a practiced pilot, at least). Their maindeck configurations tend to be very aggressive, with only a few slots set aside for disruption.

I think there will be some occasions when you'll want to bring a card into a storm deck without taking anything out, going to 61-14 post-board:

• When your plan in game 2 is to still go off as fast as possible, but you really want another hate card of some sort (e.g. you runs three Thoughtseize as your only maindeck combo-protection, with a fourth in the sideboard), going to 61-14 lets you bring in that extra card without messing up the ratios of the proactive part of your deck as much as a direct one-for-one substitution would.

• Against a slow deck with lots of disruption, like traditional control, the typical storm deck strategy is to avoid going off until you have a hand that can survive having a few of your cards countered. Since you're planning to go off much later than normal, it's reasonable to bring in a land or two to help hit your later land drops. If your deck deck has lots of fetchlands, they'll thin your deck out enough to counteract the extra land cards you've put in. Make sure those extra lands are really worth putting in your sideboard in the first place, however.

  ("When should you side lands in and out?" addressed trying to hit extra land drops in control mirrors. Note that there you are more likely to be making one-for-one substitutions, since your deck has a lot of reactive cards that are fairly weak against another control deck; whereas here your deck is mostly dedicated to your storm combo.)

All of these are basically about fine-tuning your percentages — how do I improve the chance of X while not screwing up Y too much? I don't see a strong case for going above 61-62 cards in this deck, though, because at that point you really ought to just cut a couple of your combo cards instead, and even these scenarios aren't going to apply to all storm decks or all matches.

Battle of Wits decks

When you're playing a Battle of Wits deck, you've already abandoned any reasonable chance of drawing a key card naturally. In exchange, you gain access to a hard-to-disrupt win condition and the ability to just maindeck the kinds of silver bullets a normal deck would only play in its sideboard.

Before the M14 sideboard rules change, the sideboard of a Battle of Wits deck was often considered perfunctory. You're already playing 230-250 card deck with a ton of tutors and card-filtering cards — what exactly do you want to put in your sideboard? Thus, in this kind of oversized deck, it's quite reasonable to have an empty or mostly-empty sideboard and use it as a dumping ground for the cards you don't want to be playing in game 2 and 3.

Shrinking your deck in this fashion is most effective against aggro and combo decks:

1. Because they tend to have one definite plan for how to win the game, it's usually very clear which cards in your deck can be removed safely. (Watch out for transforming combo decks, though.)
2. You don't have the luxury of time to sculpt your hand and tutor for stuff. So your dead cards are more of a liability as well.
3. To win off of BoW, you need 200 cards left in your library after you stick it. Aggro and combo opponents are unlikely to have a lot of answers to your BoW, so you can stick in earlier and the first one you play is likely to be a game-winner. In contrast, against a deck with strong disruption, you're likely have gone through more of your deck by the time you stick a BoW.

Answer 3

In general, the same reasoning as when constructing a deck can be used, i.e. if some cards are better than others, reducing the number of cards in the deck makes it more likely to draw the better cards. In the sideboarding case, we have the additional problem of ensuring a reasonable land/spell ratio, which means that just adding cards to the deck is unlikely to be optimal.

However, in some cases it might be beneficial to change the mana balance of the deck (see When should you side lands in and out?). It could also be argued that in some cases if you have cards that search for other cards, that the added flexibility of an extra cards outweighs the reduced consistency. In such cases it might make sense to add one or two cards without removing any.