Comprehensive Legacy Metagame Analysis and a Request for Data

[lordofthepit]

It would be useful to compile systematic data on Legacy format trends to help answer questions we've all discussed on the Source, such as the following:

• What are the most prevalent decks in the format?
• What are the strongest decks in the format?
• Are there decks that can prey on the current metagame?
• Do the decks that appear most frequently in Top 8's do so primarily because so many players are playing them, because the deck is inherently powerful, or because the deck has the tools to help stronger pilots outplay their opponents?

Unfortunately, none of these answers are really meaningful without large data sets, and ever since SCG stopped publishing the Too Much Information series, the Legacy community hasn't had access to these data. Furthermore, no one has the time to compile all these datasets, and a group effort would cause in inconsistencies in formatting that would make analysis difficult.

To address this problem, I've written computer scripts to help streamline the number crunching, and I need your help as tournament organizers to provide me with tournament results and decks associated with each participant. This would represent an attempt to capture every single match in these tournaments.

All I need from you for each tournaments are two things:

• Match results: I will take either WER (Wizard Event Reporter) files, or a spreadsheet with match results. If you provide me with a spreadsheet, please observe the formatting rules discussed in detail in the following section.
• Decks associated with each player: Again, please observe the formatting rules discussed below.

Data Submission

If you're interested in submitting data files to me, please send me a PM, and I will provide you with my email address. WER files are preferred for providing match results, because they are consistently formatted, but otherwise the following also works:



Note that there are generally four fields for each match result. The first column represents the round number, the second column the name of the first player, the third column the match result (in W-L-D format to represent games won, lost, and drawn by the first player), and the fourth column the opponent for the player in column 2. The screenshot above shows match results from a fictious "Duel of the Planeswalkers" (this actually is derived from real tournament data, but with participant names changed and decks arbitrarily assigned to each player). In Round 6, Oona defeated Mangara 2-0, whereas Elspeth Tirel defeated Venser in two games.

Note that there are only two fields when a player has a bye. In this case, Suleiman has a Round 6 bye.

In addition, there is a fifth field with "ID" to denote when a drawn match is done so intentionally. Note that I automatically assume any matches that are reported as 0-0-0 or 0-0-3 are intentional draws.



Please use Q, S, and F to denote the quarterfinal (top 8), semifinal (top 4), and final (top 2) rounds, as illustrated above.

Note that participant names can be given in any format, as long as they are consistent from round to round. That is, don't write "Elspeth Tirel" in one round then change it to "Tirel, Elspeth" in the next.



Please submit decks using an Excel spreadsheet or a CSV file. The first column denotes the player name and should be consistent with the names used in the match results. The second field represents the deck archetype associated with the player in column 1. If you want to include subarchetype information, please do so in the second field in parenthesis as shown above. In this case, Augustin IV played the Esper subarchetype of the Stoneblade deck.

There is a lot of debate on how to name decks, and there is just as much controversy on whether a deck qualifies as an archetype or a subarchetype. I am likely to rename a lot of archetypes from the submissions I receive for the sake of consistency.

Data Output

For every tournament data submission I receive, I will generate a spreadsheet (download) and a series of graphs to illustrate what happened at the tournament. To convince you tournament organizers to provide me with data, I'll let you in on how sweet this is going to be.



The first sheet of the spreadsheet will list every competitor and their associated deck by final Swiss standings. All fields are filterable and sortable, so you can easily look up how a player or an archetype performed. Tiebreakers are automatically calculated from match results.



The second sheet of the spreadsheet will list every match result (win, loss, draw, bye, ID) for each round, along with game results. Again, all fields are filterable and sortable.



The third sheet provides a breakdown of how each archetype performed, along with its prevalance in the field.



The fourth sheet provides a breakdown of how the most popular archetypes performed against each other.



A series of pie graphs illustrates the prevalance of each deck within the top 8, top 16, top 32, top 64, and in the general field.



A stacked bar plot will similarly convey the composition of archetypes in the top 8 or in other standing groups.



And a collection of histograms helps determine whether a particular deck over or underperformed. In this example, Jund was quite underwhelming, as all of its players finished outside of the top 64. (Note: decks were arbitrarily assigned to each player, don't read too much into this fake data!)

A final note: I assume any data that is shared with me is fair game for me to make public on the Source. If you have reservations about this arrangement, please let me know or I will assume otherwise.