[Free Article] A Better Look At Tezzeret

[Smmenen]

Process matters. Preparation matters. Players tend to understand, in principle, the importance of both. Players understand the importance of playing a deck well and selecting the right deck. But, in most cases, the process of preparation and deck design/selection is unstructured. Players play a deck they have a good feel for, test it as much as they can, try to play it as well as they can, and hope for the best. Some players, the better organized, have playtest groups, record data, and select decks using a slightly more sophisticated version of the ‘feel’ system.

A few years ago, I wrote a step-by-step methodology for preparing a tournament winning decklist, using GroAtog.   In fact, that methodology was the one I used to win the 2007 Vintage Championship.  

In preparation for The Meandeck Open I wanted to use the same procedure to try to design a Tezzeret list.  Because it is so time intensive, this is the first time I used this procedure since 2007.  

This is a rather long article, reflecting the intensity of the process.  

I hope you enjoy it, and take something away from it.

http://www.starcitygames.com/magic/vintage/18347_So_Many_Insane_Plays_A_Better_Look_At_Tezzeret.html

EDIT:

I also offer this article as a counterpoint to the typical assumptions -- so common -- that there are 'objectively' best lists, and objectively optimal cards, and that deck construction should (if ever) proceed from such assumptions.   My methodology for deck construction makes virtually no such claims.  No card is sacred except those that are necessary to a particular strategy, your plan for winning.  

After reading this article, you should *very* skeptical of anyone who might claim such certainty, and, point them to this article as a way of understanding a better way to approach Vintage.   My teammates continue to make such claims, and I offer this article as a full throat rebuttal.  

I'm particularly interested in talking about the methodology, whether you think it could be improved.   I'm also interested in discussing particular choices I made at each step, and most importantly, the final step.    

[MirariKnight]

First off I'd like to say that I really enjoyed this article. This approach to deckbuilding is certainly a useful method. The last two week's articles have been excellent, keep it up.

Anyway, as far as methodology goes, I would like to raise one small concern:
In building the final list, you assembled multiple iterations of the Tezzeret deck that would be well suited to particular matchups. From there you complied a composite list of shared cards, and then went on from there to fill out the final slots.
While in theory this seems like an empirically sound method, I'm left wondering whether this actually detracts from the strength of the deck overall, even in the matchups that were considered in the deckbuilding process. The composite list cards show up in most/all of the "matchup" lists, but when made composite, do they still have the same synergy? I guess what I'm getting at is, is a Tezz list that's made of 40% Anti-Fish Tezz, 40% Anti-Mirror Tezz, and 20% Anti-Shop Tezz actually better than going for a 100% Anti-X Tezz list and making a board to cover your biggest weaknesses? I'm more sympathetic to your idea than mine, I just felt that this was a really important methodological article and would like to get discussion going on it.

[voltron00x]

I also offer this article as a counterpoint to the typical assumptions -- so common -- that there are 'objectively' best lists, and objectively optimal cards, and that deck construction should (if ever) proceed from such assumptions.   My methodology for deck construction makes virtually no such claims.  No card is sacred except those that are necessary to a particular strategy, your plan for winning.  

After reading this article, you should *very* skeptical of anyone who might claim such certainty, and, point them to this article as a way of understanding a better way to approach Vintage.   My teammates continue to make such claims, and I offer this article as a full throat rebuttal.  

Just felt like reposting this.  One of my favorite things you've ever written, actually.  It doesn't just apply to Vintage - but Vintage players are very guilty of this mindset.

Also, I feel this way about many things related to Magic that are said with absolute certainty.

[Smmenen]

First off I'd like to say that I really enjoyed this article. This approach to deckbuilding is certainly a useful method. The last two week's articles have been excellent, keep it up.

Wait 'til next week

Anyway, as far as methodology goes, I would like to raise one small concern:
In building the final list, you assembled multiple iterations of the Tezzeret deck that would be well suited to particular matchups. From there you complied a composite list of shared cards, and then went on from there to fill out the final slots.
While in theory this seems like an empirically sound method, I'm left wondering whether this actually detracts from the strength of the deck overall, even in the matchups that were considered in the deckbuilding process.

You are exactly right.    This is the weakness of my approach.   Internal synergies are given less of a priority than external synergies - that is, maximizing each cards effectiveness against expected opponents -- than is typically the case.    However, that doesn't mean that internal synergies aren't important or important to my approach.   My appoach is just more balanced.

Here's how my approach maintains high level of internal synergy.  First, a good number of the cards are present in all potential lists.  These cards are the 'core' of the list, since they appear in almost every version of the strategy.   They tend to reflect the strongest internal synergies, ensuring that most of the synergies that would appear in any list are present.   Secondly, there are decisions to be made in many steps, including decisions in the final step.   These decisions can be made with a recognition of possible internal synergy, as I did in the final step.  

But really, you have posed a higher order level question, of which there is no clear answer.   I have presented this methodology as a counterpoint to the all to common method of building decks exclusively or in too large of a measure based on internal synergies.    

It's a higher order question: what is the process of building a deck?   And, at a higher level, how do you evaluate what the best process for building a deck is?

I wrote an article a few weeks ago on Legacy, where I built a Grid of one of hte Legacy SCG $5ks that had showed almost every match played in the tournament.   The results were striking.   Zoo always beat Goblins.  Ad Nauseam always Beat Zoo.  Zoo always beat Merforlk.  And so on.  

This caused a number of deep questions on the Source: given this knowledge, is it worth it to even try to win these matchups?   It appears not to matter how skillful you are.  You just can't win those matchups.   But if you try to win them, how?  Do you try to mangle your mainboard to win them?  Or do you build your sb to do so?  Or do you just sacrifice those matchups?  

The truth is that there is no clear answer.   These are the highest order questions that can just about be asked, and the most difficult, but interesting, to discuss.  

EDIT:

From the Source:

So here's a question: assume that, as the article states and defends, Legacy is highly matchup-dependent. What does this imply for how we build decks?

Consider these proposals:

A. Sideboards should be very focused on using hate to remove your worst matchups. You can't afford to make your fifteen from 3 cards against each of 5 decks. Players should pack 6+ cards for their worst two matchups, and maybe some general utility in the remainder. 43land should play lots and lots of combo hate, and fight "fair" against CounterTop.

B. Players should ignore their bad matchups as unwinnable by any sideboarding means. Construct your sideboard to gain an edge in your winnable matchups: 43land should just assume it's never going to face Belcher, and pack its sideboard with tools to fight CounterTop decks.

1. Which of the above is true, if either?
2. Are these statements necessarily contradictory? (perhaps A is true for the best players, who can win their not-terrible-matchups through skill, and B is true for mediocre players)
3. Does the choice of strategy A vs. strategy B differ based on decktype? If so, how do you make that choice for each of the format's best decks?

http://www.mtgthesource.com/forums/showthread.php?p=401444#post401444

[Diakonov]

I wrote an article a few weeks ago on Legacy, where I built a Grid of one of hte Legacy SCG $5ks that had showed almost every match played in the tournament.   The results were striking.   Zoo always beat Goblins.  Ad Nauseam always Beat Zoo.  Zoo always beat Merforlk.  And so on.  

This caused a number of deep questions on the Source: given this knowledge, is it worth it to even try to win these matchups?   It appears not to matter how skillful you are.  You just can't win those matchups.   But if you try to win them, how?  Do you try to mangle your mainboard to win them?  Or do you build your sb to do so?  Or do you just sacrifice those matchups?  

The truth is that there is no clear answer.   These are the highest order questions that can just about be asked, and the most difficult, but interesting, to discuss.  

EDIT:

From the Source:

So here's a question: assume that, as the article states and defends, Legacy is highly matchup-dependent. What does this imply for how we build decks?

Consider these proposals:

A. Sideboards should be very focused on using hate to remove your worst matchups. You can't afford to make your fifteen from 3 cards against each of 5 decks. Players should pack 6+ cards for their worst two matchups, and maybe some general utility in the remainder. 43land should play lots and lots of combo hate, and fight "fair" against CounterTop.

B. Players should ignore their bad matchups as unwinnable by any sideboarding means. Construct your sideboard to gain an edge in your winnable matchups: 43land should just assume it's never going to face Belcher, and pack its sideboard with tools to fight CounterTop decks.

1. Which of the above is true, if either?
2. Are these statements necessarily contradictory? (perhaps A is true for the best players, who can win their not-terrible-matchups through skill, and B is true for mediocre players)
3. Does the choice of strategy A vs. strategy B differ based on decktype? If so, how do you make that choice for each of the format's best decks?

http://www.mtgthesource.com/forums/showthread.php?p=401444#post401444

This is interesting stuff.  I don't have a premium account so I am unable to read the whole article, but I like this topic.  I think that, mathematically, there is an answer to how you should build your deck optimally in this case, but it would require extensive testing and rigorous statistics to perform perfectly.

First you would need to get accurate information on how the metagame is probably going to look, which is difficult enough to do, but for the sake of discussion let's say that you could.  Second, you would need to look at the probabilities of how well each archetype is going to do in the tournament.  Compiling that information, you could determine what the probabilities are going to be that you face each type of deck during the tournament, assuming you keep winning every round (or you could plan for one loss maybe...shoddy planning though probably ).  Then you would need to compare these numbers to your testing results of how likely you are to win each of these types of match-ups, and your deck choice--and sideboarding options--ought to be based on those numbers. 

This would need to be done for every deck choice possibility though: If I am evaluating Deck A, I should examine the likelihood of playing against Deck B multiplied with my match-win percentage against that deck.  After looking at each possible combination for Deck A, you add them all together to see your composite score.  Then, move on to the numbers for Deck B, C, and so on.  Theoretically, the deck with the highest number ought to be your choice, ignoring play skill of course.

This would be nearly impossible to actually perform, because the permutations of possible sideboards that you would need to try to gain accurate data on achieving the best overall score for a particular deck would cause the experiment to take forever.  (Although, mathematically, there should be an optimal sideboard strategy for each deck for each tournament to have the best possible score.)  So, to a certain extent, you do need to rely somewhat on intuition instead, which means that realistically there is no clear answer, as you said.