Starting Over

Posted in Making Magic on February 20, 2004

By Mark Rosewater

Working in R&D since '95, Mark became Magic head designer in '03. His hobbies: spending time with family, writing about Magic in all mediums, and creating short bios.

“Starting Over”

Welcome to Mulligan Week! (For those that care, I actually called “dibs” on that joke in e-mail when Scott announced to the group of writers that Mulligan Week was coming up. Playfully quirky or terribly sad, you decide.) It’s time to talk about mulligans. So I thought I’d use my column this week to fill you in on a little history, a behind-the-scenes mulligan story, and then end with a little discussion about a myth concerning mulligans.

Before I jump in, let me define the term “mulligan” for my readers that are unfamiliar with the term. At the beginning of the game, you draw your hand of seven cards. If you are unhappy with those cards, the rules allow you to shuffle your hand back into your library and draw one less card. (And as you will see, this is actually the latest in a line of mulligans.) In competitive play, mulligans are very important. Pro Tours have swung on players properly or improperly mulliganing. In casual play, things often are much looser. (“Don’t like your hand? Ah, just shuffle it back in and draw seven.”)

Drawing From The Past

Let’s start by talking about where the mulligan came from. In the beginning, there was Alpha. And it was good. But players quickly realized that any game that forced you to shuffle your deck before you began playing was bound to fall victim to… random chance. And with sixty cards (or possibly forty if you were actually following the Alpha rule book), you had a lot of different draw possibilities.

Quick aside. When Magic Online was planning their first add campaign, they thought it might be neat to quote a bunch of cool stats in their ad. One statistic they wanted was the number of iterations that a sixty card deck might have with the available 2000+ cards that the online game would have available. As Marketing isn’t known for their higher math skills, the question got brought to R&D. (Don’t let my oddball writing background throw you, R&D is full of math and science folk.) Assuming you had a sixty card deck and access to 2000 cards, how many different decks could be built? I don’t remember the answer, but I do remember that it was significantly greater than the number of atoms in the universe.

Back to probability of the opening draw. Magic is a fun game. But it’s only fun if you get to actually play. Players drew unplayable hands often enough that informal rules started popping up. Now remember in the early days, there were no sanctioned tournaments (the DCI, then just the DC – Duelist Convocation – would not sanction tournaments until early 1994). So, rules varied from place to place. Here is a short list of the number of different mulligans that appeared in the early days: (note that all the early mulligans had you drawing a full seven card replacement)

  • mulligan once per game for any reason
  • mulligan if you have no land
  • mulligan if you have one land or no land
  • mulligan if you have no land or all land
  • mulligan if your opponent mulligans
  • Every combination of any of the above

Trying to discover the origin of the mulligan, I started asking Richard Garfield and some of the other Alpha playtesters if they had any memory of when mulligans began. Richard knew that the idea of mulligans was talked about before the game came out, but the original playtesters never adopted a mulligan rule. The major reason was the belief by some of the playtesters that mulligans rewarded the players who hadn’t spent the time figuring out the best mana base.

Draw of the Crowd

So Magic was published with no mulligan rule. And as with all early organized play, players started creating their own rules. Eventually the DC(I) came along and started creating official rules for mulligans. To the best of my knowledge, the earliest DCI sanctioned mulligan rule was the “all land or no land” mulligan. The way this worked was that a player could shuffle his or her hand into his library and draw seven cards if he revealed his opening hand and had either zero or seven land.

If you had a hand of cards with only non-mana producing lands (we used to make those way back when), it wasn’t mulligan worthy. But if you drew a hand full of non-land mana sources (such as Lotus and Moxes), you did have the option of mulliganing. And once you mulliganed, you could no longer mulligan again that game.

Maze of Ith says no mulligan for you...

One of my favorite stories involves a mulligan gone awry. This story happened before I came to Wizards, so a long, long time ago (probably mid 1994). I shuffle and draw a hand of seven cards. I have no land. I mulligan. I draw a new hand. Still no land. But I’m in trouble now as I can’t mulligan a second time. Turn one, I do nothing. Turn two, I do nothing. Turn three, I draw my Black Lotus (yes, this is back in the days when there was only one constructed format). I use my Lotus to play two Birds and an Elf.

I never draw a land or a Mox, but I’m able to use the Birds and Elf to create enough mana to win a few turns later. With the final blow, my opponent looks at me and says, “You didn’t have land. Or any of your moxes. I’m supposed to win when that happens.”

We’ll Always Have Paris

So how did the mulligan of old become the mulligan we all love and adore? Where did the “Paris” mulligan come from? The answer is a man named Matt Hyra. Matt was a Pro Tour player who got hired by Wizards. (The first one for the trivia buffs out there.) But unlike most other Pros, he didn’t come to work in R&D. Matt started out as the Head Judge for the Wizards Tournament Center.

The old mulligan rule was clunky, so R&D had a standing order to try and come up with something better. Meanwhile, Matt had come up with his own mulligan rule. He started using it at unsanctioned tournaments that he judged. And people seemed to like it. So he brought the idea to R&D.

R&D’s first reaction was rather positive. It seemed quite elegant and it did a much better job of making mulliganing an interesting strategic choice. But R&D was afraid of what impact it might have on deck construction. Would combo decks, for example, could thrive because they had the ability to get a couple extra shots at drawing the key cards?

Repeat PT champion Tommi HoviTo test this out, R&D decided to use some money from our discretionary funds to run two constructed tournaments. Each tournament would have a large cash prize on it to lure out good players to come and try to break the system. One event was to be held in Boston and the other in Los Angeles (or more accurately, Costa Mesa). Now, I went to school in Boston (at the wonderful Boston University – Go Terriers!) so I was eager for a chance to get back to my old stomping grounds.

Meanwhile, R&D felt confident that the new mulligan worked fine for limited so we used it at Pro Tour Los Angeles, the second one won by Tommi Hovi. (Ask me someday to retell the story of the near player’s riot at that Pro Tour.)

Back in Boston, the experimental tournament had a small turn-out. And no one was really building decks trying to abuse the new rule. So all in all, kind of a waste. I was told L.A. went a little better, but as you will see, the real experimental tournament was yet to come. Some of you might know it as Pro Tour Paris.

Wait a minute. If R&D was so unsure about the new mulligan in constructed, why would we test it out at a Pro Tour? Well, we didn’t. At least not on purpose. Remember how I said that R&D okayed the new mulligan for Pro Tour Los Angeles? You see, at every Pro Tour we send out a packet to the competitors outlining the rules for that particular tournament. PT L.A.’s floor rules explained the tournament would use the new mulligan rule. The person who made the floor rules for Pro Tour Paris simply copied large sections of the floor rules for the player’s packet not realizing that the new mulligan rule was supposed to be removed. So why did PT Paris use the new mulligan? Oh, because we forget to take the paragraph about it out of the player’s packet.

And as fate would have it PT Paris proved to be an excellent test. See, R&D was worried how the new mulligan rule worked with combo decks. And what was the defining deck of PT Paris? A little combo deck called Prosperous Bloom (ProsBloom for short). But in the end, everything worked out and the rule was adopted for tournament play. Years later we added the mulligan to the game’s rules.

One last thing before we move on. The new mulligan is known as the “Paris” mulligan. Note that the first Pro Tour to use the new mulligan rule was PT L.A. not Paris. I guess the “L.A.” mulligan just doesn’t sound as cool.

“That’s a myth, myth!”

There’s one last thing I wanted to talk about before I called this column a wrap. I believe that there is a common misconception that mulligans are some sort of fix for an inherent flaw in the game. I’ve read many bulletin board threads where players talk about how Magic would be perfect if only there wasn’t the mana screw. (For those out of the Magic slang loop, “mana screw” refers to the condition in the game where a player gets trapped at little or no land and thus cannot do anything.) While I agree that mana screw is annoying, it is the result of one of the most important aspects of the game.

Richard GarfieldIt is my belief that Richard Garfield had three great innovations when he created Magic. First was the creation of the trading card game genre. It’s what I call a “seamless” invention in that it seems so natural and obvious, yet no one had thought to do it. Second was the invention of the color wheel. As my regular readers know, I believe the color wheel is fundamental to how the game functions. But there was a third equally important discovery.

That final discovery was the mana-resource system. This invention includes such concepts as land, mana, mana costs, the mana pool, and mana burn. The whole system works so smoothly that few people stop to admire all it does for the game.

Trading card games present some real problems for a game designer. The biggest issue is this: the game exceeds the box. That is, most card games are balanced in the fact that the game designer gives the player all the pieces up front. Sure, some cards are better than others, but the game forces you to play with them all so it doesn’t matter. But trading card games allow players to pick and choose what cards they want to play with. Thus, cards of a different power level pose a real problem. Why will players ever use the weak cards?

Richard came up with a very elegant answer: What if cards’ power scaled conversely with a second attribute? Or in English, what if the more powerful cards had some handicap that made them not necessarily better than the weaker cards. The handicap was a resource (land) that was monitored by time (you can only play one land a turn). This system made it so that an individual card’s power varies over time. A big, powerful effect, for example, is useless until a certain point in the game, the point at which you have enough mana to play it.

In addition, the land resource solved a second problem. It allowed the power level of the cards to notch up over the length of the game. This is important for two very different reasons. One, it guaranteed the game would end more quickly. If the effects keep scaling up in power, it’s only a matter of time until one player defeats the other. Second, the power shift makes the game more flavorful and exciting. Just as in any dramatic fight (and by this I mean a fight one would see in an entertainment vehicle) the fight snowballs in intensity.

Finally, and this one’s more important than you might think at first blush, it adds randomness to the game. Why is this good? One, it guarantees that every game is different. Two, it allows for dynamic situations. (“So, I was stuck at two land and three life.” “So you lost?” “No.”) Three, it ensures that no match-up is a foregone conclusion. To take the extreme, the best player in the world playing the strongest deck in the most powerful format could conceivably lose to a novice playing a pre-constructed deck. The fact that such a thing is possible, even if highly unlikely, speaks volumes.

In short I’m saying that a world in which Magic didn’t need mulligans would be far, far different, and in my opinion, far, far inferior.


Before I end my column I wanted to talk about two upcoming columns. First, my follow-up to last week’s 100 Questions column is coming but it’s going to take a little while - I got thousands of replies (which is great), but it’s going to take some time to read. But I promise, I will do the follow-up.

Second, my column next week is about tying up loose ends. So, if there is anything I ever referenced but did not explain (often because I wasn’t allowed at the time) that you want to know about, let me know. You want to actually see the Tom Swifties? Let me know. You want to know what Mirrodin cards I was referencing? Let me know. You want to hear the story that I said I’d someday tell you? Let me know.

Join me next week when… well, I gave it away last paragraph.

Until then, may you know the joy of winning the unwinnable game.

Mark Rosewater

Mark may be reached at

Latest Making Magic Articles


August 8, 2022

Ravnica Design Handoff Document by, Mark Rosewater

An important part of Magic design is that the set passes through multiple leads. When Vision Design finishes and hands off the set to Set Design, we create a document known as a vision de...

Learn More


August 1, 2022

State of Design 2022 by, Mark Rosewater

Back in December of 2003, I became the head designer for Magic. One of the things I decided I wanted to do was write an article at the end of every Magic "year" where I discuss all the se...

Learn More



Making Magic Archive

Consult the archives for more articles!

See All