6  Best decks

7 Archetype

This analysis attempts to use regression to determine the decks with the best performance inside archetype or base archetype.

A binomial regression is initially trained on a set of decks. In order to be included in this analysis the archetype must be present at least 50 times in the dataset.

In order to be considered a card must be included at least 25 times in either the main deck or the sideboards, one or the other being considered separately. In models comparing the number of copies of each card, when a number of copies is less than 25 it is grouped with an adjacent number of copies. For example, a card that is present 32 times in 1 copy 200 times in 2 copies, 15 times in 3 copies and 47 times in 4 copies would lead to the following result 1/2 : 232 and 3/4 : 62. The formulation 2-4 indicates that the numbers of copies 2, 3 and 4 have been grouped together. Be careful, this part leads to results that I’m not really sure of. The interpretation of the regression coefficients seems really questionable, particularly in relation to the collinearity problem and the very large number of variables with sometimes small sample sizes. I would therefore encourage you to be very careful.

A total of 6 quasibinomial regression models are created using the wins and losses of each deck:.

  • Two models using the deck as a whole (maindeck and sideboard)

    • Comparing for each card presence +1 vs absence 0.

    • Comparing each card count with a sufficient sample size 0 vs 1 vs 3-4 for example

  • Four separate models 2 for maindeck and 2 for sideboard

    • Comparing for each card presence +1 vs absence 0

    • Comparing each card count with a sufficient sample size 0 vs 1 vs 3-4 for example

These different models are then used to determine the 7 complete decks (maindeck and sideboard) with the highest probability of victory for each archetype (weeks are expressed in 2 last digits years.weeks of the year).

As well as the 7 maindecks and 7 sideboards with the highest probability of victory are presented for each archetype.Warning: this second part can lead to inconsistent combinations. It seemed useful if you want explore the maindecks and sides separately.

Table shows the top7 decks:

  • Firstly base cards (present in all decklist).

  • Variables cards are present as card name average number of cards[minimum; maximum number of cards]number of base cards* (if this card is also in base cards)

7.1 All Data Model best deck prediction

7.1.1 Affinity

7.1.1.1 Global deck models

7.1.1.2 Separate model for main and side

7.1.1.2.1 Mainboard
7.1.1.2.2 Sideboard

7.1.2 Belcher

7.1.2.1 Global deck models

7.1.2.2 Separate model for main and side

7.1.2.2.1 Mainboard
7.1.2.2.2 Sideboard

7.1.3 Domain Zoo

7.1.3.1 Global deck models

7.1.3.2 Separate model for main and side

7.1.3.2.1 Mainboard
7.1.3.2.2 Sideboard

7.1.4 Eldrazi : Broodscale Combo

7.1.4.1 Global deck models

7.1.4.2 Separate model for main and side

7.1.4.2.1 Mainboard
7.1.4.2.2 Sideboard

7.1.5 Eldrazi : Ramp Eldrazi

7.1.5.1 Global deck models

7.1.5.2 Separate model for main and side

7.1.5.2.1 Mainboard
7.1.5.2.2 Sideboard

7.1.6 Energy

7.1.6.1 Global deck models

7.1.6.2 Separate model for main and side

7.1.6.2.1 Mainboard
7.1.6.2.2 Sideboard

7.1.7 Goryo Reanimator

7.1.7.1 Global deck models

7.1.7.2 Separate model for main and side

7.1.7.2.1 Mainboard
7.1.7.2.2 Sideboard

7.1.8 Hollow One

7.1.8.1 Global deck models

7.1.8.2 Separate model for main and side

7.1.8.2.1 Mainboard
7.1.8.2.2 Sideboard

7.1.9 Mill

7.1.9.1 Global deck models

7.1.9.2 Separate model for main and side

7.1.9.2.1 Mainboard
7.1.9.2.2 Sideboard

7.1.10 Red Aggro : Burn

7.1.10.1 Global deck models

7.1.10.2 Separate model for main and side

7.1.10.2.1 Mainboard
7.1.10.2.2 Sideboard

7.1.11 Red Aggro : Prowess

7.1.11.1 Global deck models

7.1.11.2 Separate model for main and side

7.1.11.2.1 Mainboard
7.1.11.2.2 Sideboard

7.1.12 Ruby Storm

7.1.12.1 Global deck models

7.1.12.2 Separate model for main and side

7.1.12.2.1 Mainboard
7.1.12.2.2 Sideboard

7.1.13 Titan : Amulet Titan

7.1.13.1 Global deck models

7.1.13.2 Separate model for main and side

7.1.13.2.1 Mainboard
7.1.13.2.2 Sideboard

7.1.14 Tron : Eldrazi Tron

7.1.14.1 Global deck models

7.1.14.2 Separate model for main and side

7.1.14.2.1 Mainboard
7.1.14.2.2 Sideboard

7.1.15 UB Midrange

7.1.15.1 Global deck models

7.1.15.2 Separate model for main and side

7.1.15.2.1 Mainboard
7.1.15.2.2 Sideboard

7.1.16 UW Control

7.1.16.1 Global deck models

7.1.16.2 Separate model for main and side

7.1.16.2.1 Mainboard
7.1.16.2.2 Sideboard

7.1.18 Wizards

7.1.18.1 Global deck models

7.1.18.2 Separate model for main and side

7.1.18.2.1 Mainboard
7.1.18.2.2 Sideboard

7.1.19 Yawgmoth

7.1.19.1 Global deck models

7.1.19.2 Separate model for main and side

7.1.19.2.1 Mainboard
7.1.19.2.2 Sideboard

7.2 Two weeks Data Model best deck prediction

7.2.1 Energy

7.2.1.1 Global deck models

7.2.1.2 Separate model for main and side

7.2.1.2.1 Mainboard
7.2.1.2.2 Sideboard

7.2.2 Red Aggro : Prowess

7.2.2.1 Global deck models

7.2.2.2 Separate model for main and side

7.2.2.2.1 Mainboard
7.2.2.2.2 Sideboard

7.2.3 Titan : Amulet Titan

7.2.3.1 Global deck models

7.2.3.2 Separate model for main and side

7.2.3.2.1 Mainboard
7.2.3.2.2 Sideboard

7.2.4 UB Midrange

7.2.4.1 Global deck models

7.2.4.2 Separate model for main and side

7.2.4.2.1 Mainboard
7.2.4.2.2 Sideboard

7.2.6 Wizards

7.2.6.1 Global deck models

7.2.6.2 Separate model for main and side

7.2.6.2.1 Mainboard
7.2.6.2.2 Sideboard

8 Base Archetype

This analysis attempts to use regression to determine the decks with the best performance inside archetype or base archetype.

A binomial regression is initially trained on a set of decks. In order to be included in this analysis the archetype must be present at least 50 times in the dataset.

In order to be considered a card must be included at least 25 times in either the main deck or the sideboards, one or the other being considered separately. In models comparing the number of copies of each card, when a number of copies is less than 25 it is grouped with an adjacent number of copies. For example, a card that is present 32 times in 1 copy 200 times in 2 copies, 15 times in 3 copies and 47 times in 4 copies would lead to the following result 1/2 : 232 and 3/4 : 62. The formulation 2-4 indicates that the numbers of copies 2, 3 and 4 have been grouped together. Be careful, this part leads to results that I’m not really sure of. The interpretation of the regression coefficients seems really questionable, particularly in relation to the collinearity problem and the very large number of variables with sometimes small sample sizes. I would therefore encourage you to be very careful.

A total of 6 quasibinomial regression models are created using the wins and losses of each deck:.

  • Two models using the deck as a whole (maindeck and sideboard)

    • Comparing for each card presence +1 vs absence 0.

    • Comparing each card count with a sufficient sample size 0 vs 1 vs 3-4 for example

  • Four separate models 2 for maindeck and 2 for sideboard

    • Comparing for each card presence +1 vs absence 0

    • Comparing each card count with a sufficient sample size 0 vs 1 vs 3-4 for example

These different models are then used to determine the 7 complete decks (maindeck and sideboard) with the highest probability of victory for each archetype (weeks are expressed in 2 last digits years.weeks of the year).

As well as the 7 maindecks and 7 sideboards with the highest probability of victory are presented for each archetype.Warning: this second part can lead to inconsistent combinations. It seemed useful if you want explore the maindecks and sides separately.

Table shows the top7 decks:

  • Firstly base cards (present in all decklist).

  • Variables cards are present as card name average number of cards[minimum; maximum number of cards]number of base cards* (if this card is also in base cards)

8.1 All Data Model best deck prediction

8.1.1 Affinity

8.1.1.1 Global deck models

8.1.1.2 Separate model for main and side

8.1.1.2.1 Mainboard
8.1.1.2.2 Sideboard

8.1.2 Amulet Titan

8.1.2.1 Global deck models

8.1.2.2 Separate model for main and side

8.1.2.2.1 Mainboard
8.1.2.2.2 Sideboard

8.1.3 Ascendancy Combo

8.1.3.1 Global deck models

8.1.3.2 Separate model for main and side

8.1.3.2.1 Mainboard
8.1.3.2.2 Sideboard

8.1.4 Azorius Control

8.1.4.1 Global deck models

8.1.4.2 Separate model for main and side

8.1.4.2.1 Mainboard
8.1.4.2.2 Sideboard

8.1.5 Boros Energy

8.1.5.1 Global deck models

8.1.5.2 Separate model for main and side

8.1.5.2.1 Mainboard
8.1.5.2.2 Sideboard

8.1.6 Broodscale Combo

8.1.6.1 Global deck models

8.1.6.2 Separate model for main and side

8.1.6.2.1 Mainboard
8.1.6.2.2 Sideboard

8.1.7 Dimir Midrange

8.1.7.1 Global deck models

8.1.7.2 Separate model for main and side

8.1.7.2.1 Mainboard
8.1.7.2.2 Sideboard

8.1.8 Domain Zoo

8.1.8.1 Global deck models

8.1.8.2 Separate model for main and side

8.1.8.2.1 Mainboard
8.1.8.2.2 Sideboard

8.1.9 Eldrazi

8.1.9.1 Global deck models

8.1.9.2 Separate model for main and side

8.1.9.2.1 Mainboard
8.1.9.2.2 Sideboard

8.1.10 Eldrazi Tron

8.1.10.1 Global deck models

8.1.10.2 Separate model for main and side

8.1.10.2.1 Mainboard
8.1.10.2.2 Sideboard

8.1.11 Esper Midrange

8.1.11.1 Global deck models

8.1.11.2 Separate model for main and side

8.1.11.2.1 Mainboard
8.1.11.2.2 Sideboard

8.1.12 Goryo Reanimator

8.1.12.1 Global deck models

8.1.12.2 Separate model for main and side

8.1.12.2.1 Mainboard
8.1.12.2.2 Sideboard

8.1.13 Hollow One

8.1.13.1 Global deck models

8.1.13.2 Separate model for main and side

8.1.13.2.1 Mainboard
8.1.13.2.2 Sideboard

8.1.14 Jeskai Energy

8.1.14.1 Global deck models

8.1.14.2 Separate model for main and side

8.1.14.2.1 Mainboard
8.1.14.2.2 Sideboard

8.1.15 Mill

8.1.15.1 Global deck models

8.1.15.2 Separate model for main and side

8.1.15.2.1 Mainboard
8.1.15.2.2 Sideboard

8.1.16 Mono Blue Belcher

8.1.16.1 Global deck models

8.1.16.2 Separate model for main and side

8.1.16.2.1 Mainboard
8.1.16.2.2 Sideboard

8.1.18 Prowess

8.1.18.1 Global deck models

8.1.18.2 Separate model for main and side

8.1.18.2.1 Mainboard
8.1.18.2.2 Sideboard

8.1.19 Ramp Eldrazi

8.1.19.1 Global deck models

8.1.19.2 Separate model for main and side

8.1.19.2.1 Mainboard
8.1.19.2.2 Sideboard

8.1.20 Ruby Storm

8.1.20.1 Global deck models

8.1.20.2 Separate model for main and side

8.1.20.2.1 Mainboard
8.1.20.2.2 Sideboard

8.1.21 Yawgmoth

8.1.21.1 Global deck models

8.1.21.2 Separate model for main and side

8.1.21.2.1 Mainboard
8.1.21.2.2 Sideboard

8.2 Two weeks Data Model best deck prediction

8.2.1 Amulet Titan

8.2.1.1 Global deck models

8.2.1.2 Separate model for main and side

8.2.1.2.1 Mainboard
8.2.1.2.2 Sideboard

8.2.2 Boros Energy

8.2.2.1 Global deck models

8.2.2.2 Separate model for main and side

8.2.2.2.1 Mainboard
8.2.2.2.2 Sideboard