Unlocking Legacy – A Mystifying Decision

Grand Prix GP Columbus July 30-August 1, 2010
Thursday, July 1st – Legacy veteran Christopher Coppola questions the recent banning of Mystical Tutor and refutes the arguments given in the official explanation.

I. The Appeal of Legacy

One of the hallmarks of Legacy, one of the unique qualities which has made this format continuously interesting for six years, is a principle that was used in creating the original banned list, and which has so far been followed well (with one terrible accident in April 2007). This principle is that all of the cards that make both new and classic decks powerful, fun, and sometimes competitive will be allowed in this format unless they are so egregiously degenerate that there is no way to balance them in a diverse environment. This means they must generate massive card advantage, easily abusable amounts of mana, or some other effect which would eclipse the normal interactivity and pace of a game of Magic. This is a subtle but amazing principle which is basically responsible for the popularity of Legacy; it does not mean that cards will be banned on power level alone, but a much stronger condition – their tendency to warp individual decks and as a result the entire format. It means that players can play with almost every card that they like, and therefore almost any deck that they have built, watched, or imagined in their time as a Magic player. Being able to play any card you want is an enormous attraction for a format and has made Legacy a success. This policy means that yes, you can play with Goblin Lackey, Dark Ritual, Wasteland, Aluren, Force of Will, Psychatog, Tendrils of Agony, Cursed Scroll, Illusions of Grandeur, Arcbound Ravager, and a host of other cards which established high power levels in older, weaker formats, but are completely safe in Legacy (and often are not even good enough to compete against the best decks). Legacy is a format which depends heavily on synergy and efficiency, and without broken engines to abuse good mana sources or cheap disruption, it is balanced and interactive. These very popular cards work well in Legacy because there is a deep card pool that provides efficient disruption, removal, and threats, very reliable manabases, and a nearly continuous spectrum of archetypes from which to choose in building decks. Despite what their associations might be from other formats, these cards are not broken, and in being played in thousands of large tournaments, many with ambitious, competitive players, have never shown any signs of causing problems.

Mystical Tutor is one of these cards.

II. DCI Objectives

Since sanctioned Magic tournaments have existed, there has been a need to regulate which cards are allowed to be played and in what numbers. The old method of restricting cards was retired long ago and is now only used in Vintage, and modern set development has improved to the point where the DCI has rarely interfered in Standard in the last decade. Since the program of balancing cards before they are printed has been such a success, many players may not be familiar with DCI precedents and responsibilities, especially those players that have come to play Legacy from Standard or the current Extended. I will give some background information to put this latest decision in context.

Format management has been going on since about 1996, so I will stick only to the important ideas in this article. For power level purposes, the only useful comparison to Legacy is the original Extended format, but all DCI actions across all formats are useful for understanding DCI policy and the methods they use.

At a very basic level, the object in managing formats is to remove cards which tend to warp them – cards that have such strong effects that merely not putting them in your deck puts you at a competitive disadvantage. These cards almost always break game rules in powerful ways, often by drawing a large number of cards or creating a large amount of mana from a small investment. Whether or not a card is broken in a certain format is not an easy thing to measure, and traditionally the DCI has relied on data to make these decisions, using metrics such as how many decks played certain cards, how well those decks did, and how desperate the anti-decks were. In addition, the DCI usually invokes a very high standard of evidence, relying on the ambition of tournament players to really abuse cards and prove that they are problematic before removing them. This is a good principle, since the total intellectual resource of all the competitive Magic players is orders of magnitude larger than any official body designated to “solve” a format, and if there are effective solutions in the card pool then players will find them and use them to win prizes. There are extremely rare circumstances where it would be justified to ban a card without a large amount of tournament data, like when Development operates without adequate oversight (Memory Jar), or when the rules managers operate without any oversight at all (Flash). The DCI chose not to preemptively ban Flash despite the unique circumstances of the errata and risk to the imminent Grand Prix, which demonstrates just how high the standard of evidence has to be to determine whether or not a card has distorting effects.

It is also vital that banned lists are kept as small as possible. Banning cards unnecessarily reduces design space, limiting the possibilities for deck construction, and as a consequence is incredibly unpopular among players. The original motivation for banning cards was to keep formats balanced so that everyone would have a fair chance to win, which in turn would keep players interested. Senseless bannings which do not improve the balance of the format removes interesting deck design and gameplay choices and reduces players’ incentive to play at all. The ability to play all but a handful of cards out of the entire Magic library is an extremely important selling point of Legacy, and any decision to remove cards has to be strongly justified with realistic, intelligent, and empirical arguments.

The overall philosophy of format management is intuitive and sensible, but there are sometimes inscrutable choices, and of course not every decision has been ideal. The DCI’s actions with respect to Legacy have sometimes had inconsistent methodology. I think the original banned list for Legacy was far too conservative, and I have previously discussed cards that I think do not deserve their status on the banned list (in particular, Land Tax). Except for Hermit Druid, the rest of the cards I originally recommended for unbanning have become unbanned and are quite safe in Legacy (Mind Over Matter, Replenish, Entomb, and Grim Monolith. I wrote that article in September 2006, before Bridge from Below, Narcomoeba, and Dread Return had been printed; because of those cards, it is much more difficult to make a case for unbanning it, and it is right to keep it banned at this time). I also think that the decision to leave Flash in the card pool before Grand Prix: Columbus was a mistake, considering the unique circumstances of the errata, the vastly higher power level of the card, and the data from Grand Prix Trials and medium-sized tournaments, but I agree with the principles used to defend that decision. I only bring up this controversy to point out that even a card as obviously ban-worthy as Flash had to prove its degeneracy to a very high standard in order to get the DCI to take action against it.

III. The Decks that Play Mystical Tutor

Ad Nauseam is an eminently fair card, as is the Legacy deck with the same name. In fact, Ad Nauseam is so fair that it is not even a tier one deck. This is mainly because it does not reliably beat Blue disruption, which is ubiquitous at the top tier of Legacy in decks such as Countertop, Merfolk, New Horizons, and Reanimator. It even interacts with Zoo and other fast aggro decks that can reduce the effectiveness of the spell by dealing damage with creatures, and forcing the combo player to draw fewer cards so they do not die to a burn spell. Ad Nauseam still has a good matchup against aggro, but the fact that a typical aggro deck with some direct damage can actually interact with the combo engine is a strong indicator that the deck is at a healthy power level. Ad Nauseam actually runs less than four copies of its main engine card so that it doesn’t deal a lot of damage to itself when it resolves — the fact that it uses four Mystical Tutor to get them is an example of good deck design, not degenerate behavior. Spending two cards (and usually a turn to draw the tutor target) on a single spell is a perfectly fair transaction in Legacy where, as I explained earlier, there are no targets powerful enough to make tutoring a problem. Mystical Tutor only allows decks to draw balanced cards more consistently; in this case, it means that the deck “draws” Ad Nauseam with a moderately increased frequently, and is significantly less likely to kill the pilot when it resolves.

Reanimator, however, is among the highest tier of competitive Legacy decks, but like Ad Nauseam has never been a problem. Despite being a better deck overall, it is considerably less common than Ad Nauseam decks, which is due in part to the extremely effective sideboards that players have now learned to use (it may also be more difficult to play optimally). Around the time that Reanimator was being developed and started appearing in tournaments, Legacy hit a critical mass of graveyard-dependent strategies and the number of slots used in the sideboard to answer these strategies increased significantly. Reanimator is vulnerable to Counterbalance, Merfolk, and New Horizons, although not quite as much as Ad Nauseam is, but it is easily disrupted by graveyard hate, and even removal like Swords to Plowshares or edict effects, which are common enough to be relevant answers. Two of the best disruption cards to play against Reanimator are Tormod’s Crypt and Faerie Macabre, both of which are free and have no color restrictions. Reanimator is still a strong deck that can overcome these hate cards if played well and with a strong sideboard of its own, but in larger tournaments this accumulated uphill battle makes it a less ideal deck choice. This is clearly apparent from the performance of Reanimator in the Legacy tour that StarCityGames.com has been running since last year, where, on average, it only makes one deck in the Top 8 every other event (or half an appearance per Top 8). There were zero Reanimator or Ad Nauseam decks in the top sixteen of the StarCityGames.com St. Louis open this past weekend, where the prizes included $5,000 in cash and points towards the SCG Invitational. If these decks had an unfair advantage that players could use to win this much money, this is not what the format would look like.

There are other decks in Legacy that have Mystical Tutor in them, but they have not been specifically accused of being problematic, and their presence is so small compared to these two decks that they are not relevant to the discussion of banning cards.

IV. Unacceptable Methodology

With some important context established, let us now turn to the current situation. On Friday, Tom LaPille wrote an article related to the controversial June announcement to ban Mystical Tutor in Legacy. I was stunned at this announcement, as were all of the Legacy players I know. Tom LaPille article did not offer any real justification for the decision, which in addition to being quite disappointing, raises some questions about the motivations of the banned list managers. Given the brief explanations I have already given about Legacy and standard DCI policy, it should be somewhat apparent why this explanation was a failure and the banning decision a mistake, but I would like to go over the statements made and try to clarify why they are so inadequate.

First, Tom argues that Mystical Tutor is a problem because it was used to search for Flash:

Three Flash decks made the Top 8; one played zero Mystical Tutors, one played two Mystical Tutors, and the last one played four Mystical Tutors. Can you guess which deck won? That’s right, the one with four Mystical Tutors. Having played in that Grand Prix myself, I can also report that there was no shortage of Flash players who did not make the Top 8, and many of them had also adopted Mystical Tutor as one of their tutors of choice. Somehow I did not play Flash at that tournament, and a Mystical Tutor often heralded my impending demise.

This argument is nothing more than an observation that Mystical Tutor makes the Flash deck better. Unfortunately, that is not applicable to the current situation — there were no problems with Mystical Tutor before Flash was created, and there were no problems with Mystical Tutor after Flash was banned. The DCI has mercifully kept Flash banned for the time being, so this story contributes nothing to answering the question of why Mystical Tutor is being banned.

Continuing on, Tom admits that Flash was the problem, but then goes back to his original ploy and tries to associate some of that blame with Mystical Tutor again:

It’s easy to say that the bigger problem here is Flash, and I would agree with you. The people behind the Legacy banned list also agreed with you, as they banned Flash after this tournament. I think that we also learned something about Mystical Tutor here, though, as it was an important part of what made the Flash deck so consistent across the entire tournament.

Tom does not say what specifically he learned about Mystical Tutor that was unknown when it was originally allowed into Legacy, but I don’t think it would be helpful in justifying its banning. Mystical Tutor is a utility card for combo decks that was deliberately left off the banned list because it is a fair card.

Continuing on, Tom then makes a legitimate attempt to explain this decision with current tournament data…

Let’s flash forward to Grand Prix—Madrid earlier this year. Three of the Top 8 decks were Mystical Tutor-based decks: one Reanimator and two Ad Nauseam Tendrils. Each of those decks defeated a non-Mystical Tutor deck in the quarterfinals, and two of those three decks faced off in the finals.

Neither Reanimator nor Ad Nauseam Tendrils is anywhere near as onerous as Flash, but both are strong spell-based combination decks that have Mystical Tutor standing behind them. We interpreted this as a pattern. We also revised our belief that Mystical Tutor was a second-tier tutor. We now think it’s tier one, and we think that any time a new spell-based combination deck arrives, Mystical Tutor would have been excited to be part of it.

… But then fails to interpret this information appropriately. He also tries to compare Mystical Tutor to Demonic Tutor and Demonic Consultation by labeling it a “tier one” tutor, but this comparison is unfair. Mystical Tutor is several power levels below those cards, and should be treated much differently. Decks built with either of those cards would warp Legacy in a matter of days, while Mystical Tutor has already been legal for almost six years without any incidents.

The fact that Reanimator and Ad Nauseam made it to the finals does not indicate that anything is wrong with the format. Zoo had the strongest performance over the tournament, and had more decks than any other in the Top 8. Making banning decisions based on the way the Top 8 played out is equivalent to just using one eight-man tournament as a substitute for legitimate format analysis. This is a very poor policy, and is not up the DCI’s own standard of decision making. There were ten thousand matches played in Madrid, and at the very least the entire Day 2 needs to be considered if you are interested in patterns based in reality. The contents of the two decks in the finals are statistically insignificant compared to the total amount of information available for the second day of 2,227 player tournament. This is so basic to tournament analysis that it is hard to imagine that Tom and Aaron overlooked it. (This begins to raise the question of whether or not there are separate reasons for banning Mystical Tutor, which are not in line with the DCI’s stated objectives).

If they had used tournament data to evaluate the power of Ad Nauseam or Reanimator, they would have found that there is nothing dangerous going on at all. Players at a Grand Prix are going to bring the deck they believe has the best chance to win money, and the information on what they bring to the tournament has some value in determining what is a problem. Here are roughly how the more popular archetypes were represented on Day 1:

10.3% – Zoo
9.2% – Countertop
7.3% – Merfolk
5.7% – Ad Nauseam
5.7% – Goblins
4.7% – Ichorid
3.8% – Survival
3.4% – Threshold
2.3% – Reanimator

The data speaks for itself: 92% of players who entered the tournament in Madrid decided that a deck without Mystical Tutors offered them the best chance of winning some of the $30,000 in prizes given away at this event. This is a clear sign that the card does not offer any unfair advantages. The largest group of players even decided to play Zoo, which has poor matchups against Ad Nauseam and Reanimator, and as we will see Zoo in many ways was the most successful deck of the Grand Prix.

Looking at what decks succeeded on Day 2 is even more useful in determining the power level of decks. Here are roughly how the most popular archetypes were represented on Day 2:

17.7% – Zoo
9.7% – Countertop
7.1% – Merfolk
9.3% – Ad Nauseam
5.1% – Goblins
5.5% – Ichorid
3.8% – Survival
4.2% – Threshold
3.8% – Reanimator

Among the most popular decks, Zoo increased its field fraction by a huge margin. Normally, if an archetype is overrepresented in a given field (for example, if it were outclassed by a deck with an unfairly high power level), it will be thinned by exposure to better decks, but in this case Zoo succeeded because the format was well balanced throughout the entire tournament.

The field fractions of Reanimator and Ad Nauseam also increased somewhat, and the rest of the decks stayed at approximately the same ratio. If we conclude that the performance of Ad Nauseam or Reanimator justifies some type of DCI action, we will surely have to come to the same conclusion about Zoo, since it improved more. Note that Reanimator is ranked ninth in popularity on Day 2, barely distinguishing itself from the 89 decks below it which are essentially classified as “random.”

Zoo maintained the dominant position in the field all the way through to the Top 8. The fact that three Zoo decks made Top 8 in a field with almost an equal number of Ad Nauseam and Reanimator decks is another sign that the format is well-balanced.

None of these decks show problematic success in making Day 2 from their representation on Day 1. Each of them has weaknesses to other decks in this field, and overall they make a very diverse and balanced environment: Zoo has no consistent disruption against Ad Nauseam or Reanimator; Countertop and Survival both have trouble taking control against Merfolk and Ichorid; Merfolk does not have effective answers against Zoo and Ichorid; Goblins is too slow to handle Zoo, Ad Nauseam, and Reanimator; Ichorid can’t consistently interact with Ad Nauseam or Reanimator; Threshold has difficult strategic problems against Ichorid, Survival, and Countertop; and Ad Nauseam and Reanimator lose the disruption war against Countertop, Threshold, and Merfolk.

The official explanation for the banning of Mystical Tutor becomes inadequate in the next section. Tom begins with:

When we saw the Grand Prix—Madrid finals decks, a few of us got worried, jumped onto Magic Online, and started playing some Legacy with them.

This is inaccurate. There is no such thing as Legacy in Magic Online, mainly due to the fact that Magic Online is missing around fifteen expansions. So this anecdote has no relevance to what is going on in real Legacy. It is not a rigorous testing environment anyway, especially if you are playing with no tournament incentive and against unknown players with unknown skill levels.

We were terrified by what we found. Although we were playing in the tournament practice room, which is hardly the same thing as a real tournament environment, we weren’t losing very many matches with either Reanimator or Ad Nauseam. In my case, I don’t recall losing any matches with either deck outside of a misclick while I was still learning the Ad Nauseam deck. The decks were just so strong that opponents not set up in their maindecks or that didn’t sideboard heavily against us couldn’t compete.

The conclusion that this has anything to do with Mystical Tutor is a significant error in reasoning. In more objective terms, here is Tom’s argument: Two former pro players, Aaron Forsythe and Tom LaPille, take fast combo decks onto Magic Online, play random players who are probably not very good, and they are surprised when they win most of their games? Is this really the methodology used to determine what is broken? If they had decided to play Zoo instead, would we be facing the banning of Wild Nacatl?

As a player who has played weekly Legacy tournaments on and off since 2004, I found the conclusion in the next section to be absurd:

Our research took another turn, however, when we investigated how Legacy is played in the real world. We discovered something rather interesting, and that is that Mystical Tutor decks were quite rare at Legacy tournaments that did not have tons of money on the line. At Grand Prix and other cash tournaments, people were happy to bust out their Mystical Tutors. However, in the comfort of their home stores they seemed to prefer doing other things that were more fun, if perhaps less powerful. This struck me as being a sort of gentleman’s agreement; everyone knew what sick decks were out there, but they chose not to play them.

Tom is saying that Magic players know that Ad Nauseam and Reanimator are so good that no one else has a chance (demonstrably false from tournament data), and that further, they willingly agree to give up the purported advantages of those decks (inconceivable for a game as competitive as Magic). I can assure both Tom and the people who read his article that this is not happening at any tournaments.

If players are not bringing Ad Nauseam and Reanimator to tournaments, it’s because they either don’t have the cards, or they don’t want to have to play against the more popular decks such as Merfolk, Threshold, Countertop, or New Horizons, which all play a strong set of Blue disruption cards and can consistently beat those decks. The group tasked with analyzing formats and judging card balance should not speculate in conspiracy theories when they are expected to use methods based on evidence.

If, while attending a tournament, a Magic player observes that a particular deck doesn’t place well, or it doesn’t even show up at all, he or she will consider that evidence that the deck is not a strong choice (assuming everyone in the area has all the cards needed to play that deck). In bewildering fashion, Tom is trying to conjure the opposite conclusion from the same evidence by saying that Magic players are so uncompetitive that they would agree to not play certain decks. The idea that players would act in this way is unrealistic. It is superficially ridiculous because Magic is a competitive game and most tournaments do have prize incentives, and it is logically impossible because these decks are not good enough to exert any undue influence on the format. All competitive metagames have a reliable amount of Blue disruption and access to efficient sideboard cards, which is a barrier that Combo has never consistently overcome (except during the Flash fiasco). Not only do Ad Nauseam and Reanimator not exert any undue influence on the format, but their actual tournament performance is weaker than many popular decks.

V. A Fixable Mistake

Until today, Mystical Tutor was a really important card for Legacy. It provides a way for Combo decks to play more consistently and have more flexible strategies. It is important for Combo decks to be viable in Legacy, in order to have a diverse and rich format, one that is alluringly distinct from the Aggro-saturated formats that arise from all card pools smaller than Legacy. Legacy has no broken tutor targets, and having Mystical Tutor available for Combo decks to utilize makes the format much more interesting. A format with unique and viable Combo decks is way more fun than one without them, and the proof is in the numbers — attendance for Legacy events has been record-breaking.

If the day comes that decks running a particular card create a lasting, unbalancing effect on Legacy, it will be time to examine that deck and determine if DCI actions are necessary to fix the situation. At the moment, there is no such deck, and the irrational attack on Mystical Tutor from the DCI is capricious and unjustified. The removal of Mystical Tutor from Legacy does much harm and no good, and it should be returned to the format as soon as possible.

Christopher Coppola