Feature Article – My Extended Testing Process

Buy, Sell and Trade with StarCityGames.com at Grand Prix Los Angeles!
Thursday, January 8th – Aaron Nicastri, 2008 Rookie of the Year, returns to set out a comprehensive testing process for the approaching Extended PTQ season. He presents us with his personal Extended gauntlet, and details an intensive testing plan that should maximize your testing success in the coming weeks and months…

Today, I would like to share an insight into testing relative to the Extended PTQ season, into which the world is about to take a nose dive. First, I’d like to suggest that Magic players don’t test enough, and even when we do, we do not test very well. Understanding this topic is paramount. The reason is simple: there is a huge margin to be gained over your opponents if you do test enough, and do test properly. I think I have the tools that can make your testing more productive and structured.

Before I say anything more, I’d like to mention that the following concepts and ideas are in direct relation to my own testing experiences. While I was in France before Pro Tour: Berlin, I had the privilege of staying and testing with Sylvain Lauriol (two-time French National Champion) at Equinox 7. I discovered through our conversations that we had similar interests in the development and applications of the testing process. These discussions have led me to believe that there is a better way to play Magic.

First things first: you need to build a gauntlet. Usually there is about six or so decks that warrant building for a gauntlet (however, a format like Extended can be broader). To be well informed, you need to know these decks, and need to have played games both with and against each of them. The Number 1 mistake I see people make is not keeping their gauntlet generic. It’s always modified, which defeats the purpose of the gauntlet. The gauntlet merely establishes, or tries to establish, how the decks play against each other. It is not the place for innovation. Basically, the rule is ‘keep it simple, stupid.’ When you have your own idea, crash it against a generic gauntlet. When you make modifications to a deck, it is different from your gauntlet deck, and it should be treated as such.

I’m going to give you a gauntlet of what I consider to be the essentials for the current Extended season (along with a few honorable mentions).

Gauntlet Deck 1: Zoo

This is Sebastian Thaler’s 6-0 Worlds list:

Gauntlet Deck 2: Faeries

While there are differences between lists — Japanese, Mono-Blue, and Bitterblossom/Dark Confidant builds – we can’t afford to waste multiple slots on decks that function the same and have the same game plan.

Masaya Kitayama’s 6-0 Worlds list:

Gauntlet Deck 3: Elves

Akira Asahara’s Worlds list:

Before I move onto decks four through six, I’d like to say that the three decks listed above are by far the most important for the upcoming PTQ events. They are the Tier 1 decks… furthermore, I would like to add that the following lists can only partially represent the diversity of Extended (you will need more knowledge than six lists, although six is a fine springboard).

Gauntlet Deck 4: Affinity

With almost no hate for this deck being played at present, I think Affinity is sizzling. However, this can, and probably will, change in the next month or so.

Young Han Choo’s Worlds list:

Gauntlet Deck 5: Red Deck Wins

Jon Finkel Worlds list:

Deck 6 Swans

LSV’s 6-0 Worlds list:

Here are some Honorable Mentions. Be sure to take the time to look into each of these decks in some capacity.

All-In Red (Deus plus Demigod Decks)
Death Cloud
Mind’s Desire

So, now you have a gauntlet. The next step is to get your testing group together (with, ideally, around seven players). Build up the gauntlet, sleeve it, and proxy what you don’t own. Yes, I know it’s a pain and a lot of effort.

Seven is a nice number for a testing team, as you can have two games consistently running at all times in the testing sessions, and three running for a fair amount too. A team is very important if you want to test properly, because no one can do all the work on their own. The reason I have stated an odd number is because you will find people cannot play all the time, and work needs to be done during testing sessions (such as building new decks, recording results, etc). Having an odd number in your team will give everyone a little breathing room. Another thing I would like to note is that you need, or would at least desire, all players to be on an even skill level. Happily, the testing method will help to stop the results skewing considerably.

The Testing Method

Test each matchup in a series of twenty games, alternating who plays first. When you get to game ten, switch decks.

The reasoning: twenty games is a nice sample space, and will really let you understand the matchup. In fact, you will find your play improves across the twenty games as you teach each other how to win with each deck respectively. By alternating play and draw, each deck will have ten games on the play and ten games on the draw; by switching decks at the halfway point, you will help eliminate skewed results. The better player will win more with both decks, but overall the better deck in the matchup should shine through.

It is best to play all twenty games in one session, because both players become very focused. Once you get going, it shouldn’t take longer than two hours, depending on the matchup; in fact, Extended is a fast format, and you should knock over some matchups much faster. If it is a really slow matchup, or a really clear matchup, then you don’t need to overdo it with the full twenty games. Remember, time is always limited. All players should try to play quickly, for multiple reasons.

1: You need to get through games.
2: You’re on the clock in tournaments, and draws are generally really bad.
3: Your goal is not the same as in a tournament. You’re not trying to outplay the opponent, you’re trying to establish the better deck.

If both players play a little faster, they will make some mistakes, but these should level out and not really affect the results. Plus, with a pair of eyes looking at each play in the spirit of discovery rather than competition, the mistakes will likely be spotted by one of the two players at the table.

Play the games seriously, with pen and paper to record life, and make sure you record the game count. Don’t get to the end of a two-hour session without having records. If you do this, then it is not really worth putting in the effort to play so many games, or build a gauntlet, etc. Make some written comments on how you feel about the matchup from both sides, to go along with the results. Generally, both players should come to the same conclusions as to which is the better deck and what cards are of particular importance. Be aware that not everyone will get to play all the matchups with each of the decks, and so they will rely on your information.

With the six decks in the gauntlet, there should be fifteen possible matchups. If you bring your pet deck, or want to look at mirror matchups, it will be more. This is a lot of Magic. Assuming two hours for each matchup, it is thirty hours across three tables, thus ten hours per table. My suggestion would be that, with a group of seven, you probably need two or three hours to build everything, and one full day to test and record the results of all the matchups. To avoid making this hectically rushed, set out a plan or do it across a few days. [Another idea is, with seven players, have six of them make up one of the gauntlet lists and bring it to the session — Craig.] I’m obviously all for the “build and test for infinite hours in one single day” plan, but I am a little crazy. However, I would like to point out that doing this might be equivalent to the amount of time you would spend at a PTQ, if you were there to win. I think it is actually good preparation for all big events.

Sideboarding… now, if you’re doing what I suggest, then you’ll probably be testing more than you ever have before for any given event. If you only have limited time, then you should look at making sideboard strategies for each matchup, what comes in and what goes out etc. You have time to debate and discuss boarding decisions when you build the decks (personally, I find it easier to first pull the bad cards out of my deck before bringing in new ones… this helps identify how many slots are really necessary).

For those with more time, who can afford another day of testing, I promise you it will help immensely if you run everything sideboarded, as most of the games we play are actually sideboarded games. Correct sideboarding is insanely advantageous, maybe more than any other non pre-decided factor in the game. Most people, even those that do play a lot, don’t test many sideboarded games. Yes, we all know we should… it doesn’t mean we do. Instead, we make hypothetical decisions which can be very misguided. This is why correct sideboarding can create an enormous edge.

Sideboarded games will usually have a smaller sample space than your main deck games, because you will spend time working on how to sideboard, which can be deceivingly difficult. You may change the way you sideboard after a few games. You might see your opponent’s counteractive sideboard cards as a way to attack their deck. Usually, when you begin playing sideboarded games, it is because you have chosen a deck and need to know how to make it perform at its peak. When you begin gauntlet testing, you are merely trying to determine which deck to play.

Getting real results from sideboarded games is difficult, but once you have determined what exact sideboards are made for each deck, then you can play a group of games and record the results the same as you do for main deck. Also, once you know the sideboard plan, write it down. You can review what you sideboard in against certain decks, and even look at some of the cards they are likely to side in against you. Before your tournament begins, you should quickly quiz yourself on how you sideboard for each matchup, checking your answers against the notes.

After running all these games, you will have a lot of information. In your pool of seven people, I hope one of you knows how to use Excel… if not, pen and paper will work fine. Make a spreadsheet that allows you to see the win/loss ratios for all the decks. With good knowledge of Excel, you can compute field percentages for each deck. If you do this all correctly, you will actually be able to see which deck is best in the format relative to how correct your percentages of the field actually were. Also, you will have the ability to change the field percentages to indicate what deck would be best in certain metagames.

I’d like to close by saying that running a strict testing plan such as this is taking your Magic game to a new level, and perhaps a level that Magic has not really existed for many in the past. Such thorough testing and preparation is where I see Magic heading, and my Magic in particular. I want the edge, and I believe that, with hard work, it’s everybody’s edge to have. It’s better to teach a man to fish so he can eat forever than give him fish so he can eat today.

I hope this information helps you to maximize your testing potential!

Aaron Nicastri