This project provides a set of tools for optimizing a Magic: The Gathering deck.
Goldfishing is one of the possible approaches to optimize a deck.
Not the only one, not the best one, just one of the ways. Quite relevant though with aggro or combo decks, where speed is a key factor.
Perfect Deck is a tool (developed in Java) that helps you simulate hundreds of games in just milliseconds, tremendously accelerating the optimization process.
Perfect Deck is intended for developers and non-developers, but with a different scope of use. Here is a table which represents the 3 factors allowing to optimize a deck.
| optimization factor | explanation | in Perfect Deck | scope |
|---|---|---|---|
| game strategy | How do you play your cards? When do you land, which land first? How do you optimize your curve? Sometimes it's obvious, sometimes it's not (ex: how would you play Magma Jet in a burn deck? Some argue it's optimal on T3 upkeep to filter out lands for the next two turns) | This has to be developed (this is called the deck pilot) | Requires developer skills |
| mulligans strategy | What are the objective criteria to keep or reject a hand ? Too restrictive criteria will cause a large rate of mulligans, while too loose criteria will probably leave you with slow hands. |
This has to be developed (in the deck pilot too) | Requires developer skills |
| the list | This is a very obvious optimization factor. Changing the cards, the number of mana sources, ... has a direct impact of your deck speed. | Deck editor in the tool | For everyone |
The tool is packaged in an auto executable jar, so all you need is:
- Java installed (v>=8),
- download the tool (jar with dependencies) from the releases page,
- double click the jar file.
Once started, select the deck pilot you want to use, and type-in a deck list in the editor.
From that point you can observe some games (that is run some games and see the logs) in order to make sure the pilot is making wise decisions (if not the case, you'll have either change the deck pilot code and propose me a pull request, or create an issue with all required details).
You can also (and btw it's the very goal of Perfect Deck) simulate a large number of games and compute statistics.
At any time you can mark the current statistics as a reference (thus all other statistics will be compared to this one), and you can also copy the current deck list (create a cloned tab) to work on an alternate version of the list.
This tool supports MWS, Apprentice and Cockatrice deck file formats.
Here is an example of a deck file:
// discard
4 Faithless Looting
4 Putrid Imp
// reanimation
4 Animate Dead
4 Exhume
// steaks
4 Hand of Emrakul
2 Greater Sandwurm
1 Pathrazer of Ulamog
4 Ulamog's Crusher
// mana
3 Dark Ritual
4 Lotus Petal
1 Crumbling Vestige
7 Mountain
10 Swamp
// others
3 Dragon Breath
3 Gitaxian Probe
2 Shred Memory
// Sideboard:
SB: 3 Apostle's Blessing
SB: 4 Duress
SB: 2 Electrickery
SB: 2 Mind Rake
SB: 2 Pyroblast
SB: 2 Shenanigans
Syntax is quite straightforward:
- an empty line is ignored
- a line starting with
#or//is ignored (comment) - a line starting with
SB:orSBis considered as a card from sideboard - a card can be defined as:
{number} {name}(ex:3 Swamp){number}x {name}(ex:3x Swamp){name}(ex:Swamp); counts only one card
I developed this tool in the first place because I wanted to optimize my Peasant reanimator deck, with an on the draw strategy, as explained in my Coupe de France 2019 report (fr).
As a Peasant player, all deck pilots developed so far are for this format. But the good news is that they should be compatible for Pauper format (as it is a subset of Peasant).
Here are articles that inspired me a lot while developing this tool:
- Frank Analysis – Finding the Optimal Aggro Deck via Computer Simulation
- Frank Analysis – Optimal Decks for Ten New Goldfish Formats
This code is under Apache-2.0 License