Some Data Analysis of My Decks

User avatar
materpillar
the caterpillar
Posts: 1315
Joined: 4 years ago
Pronoun: Unlisted
Location: Ohio

Post by materpillar » 2 years ago

I've been learning python for data analysis purposes ever since I saw @pokken's auto-deck grader about a year ago. I was poking through my 10 decks and figured I'd share some of my preliminary results.

I complain a lot about powercreep, and I was curious to know how much it actually affects my deck builds. So I went through my 10 decks and calculated the age of the first printing of each card in all 10 of my decks. Just to see if the decks are weighted heavily towards the last 5 years or so. All things considered my decks seemed to be way less weighted towards the recent years than I was expecting although a good bit of that could be because I throw in a lot of older stuff purely for nostalgia.

Median card age of each deck
Golos Storm - 2004.0
Dredge - 2007.0
Chromium - 2008.0
Changeling - 2008.5
Jaya - 2009.0
Lasav - 2011.0
High CMC - 2012.0
Tivadar - 2013.0
Boros Polymorph - 2015.0
Dinos - 2017.0

Histograms. These aren't for unique cards, but none of them have basic lands included. So each of my 5 Sol Rings are counted.
All Non-Basic Cards
Show
Hide
Image
I was curious to see if there was a drastic difference between the creatures and non-creatures since the narrative is that more recent creatures are more pushed.
Creatures vs Non-Creatures, no-lands
Show
Hide
[Image
Image
Non-Basic Lands vs Non-Lands
Show
Hide
Image
Image

Most common cards in my decks
  • Sol Ring: 5
  • Worn Powerstone: 3
  • Anger: 3
  • Hedron Archive: 3
  • Sword of the Animist: 3
Every other card is used 2 or less times

User avatar
Dragoon
Posts: 417
Joined: 4 years ago
Pronoun: Unlisted

Post by Dragoon » 2 years ago

Interesting idea! I'm also learning Python, may I ask for your code? How did you fetch the card data? I'm curious to see how this can be implemented.

Why is your median age for your Changeling deck 2008.5 while everything else ends in .0? Shouldn't it be either 2008 or 2009? Or did you mean the average value?

The recent years (2016-2020) seem to have performed really well, but 2007 seems like a pretty big outlier for you (except for non-creatures where 2019 has the lion share). Did you start (or seriously restart) playing magic around that time?

User avatar
pokken
Posts: 6344
Joined: 4 years ago
Answers: 2
Pronoun: he / him

Post by pokken » 2 years ago

The easiest source of card data is mtgjson, I start with here: https://www.mtgjson.com/files/AllPrintingsCSVFiles.zip

User avatar
RxPhantom
Fully Vaxxed, Baby!
Posts: 1513
Joined: 4 years ago
Pronoun: Unlisted
Location: Southern Maryland

Post by RxPhantom » 2 years ago

So...ummm...I have a confession to make:
.
.
.
.
*whoosh*
Can you name all of the creature types with at least 20 cards? Try my Sporcle Quiz! Last Updated: 2/18/22 (Kamigawa: Neon Dynasty)

User avatar
materpillar
the caterpillar
Posts: 1315
Joined: 4 years ago
Pronoun: Unlisted
Location: Ohio

Post by materpillar » 2 years ago

RxPhantom wrote:
2 years ago
So...ummm...I have a confession to make:
.
.
.
.
*whoosh*
Well I assume that means you're vaguely curious. I find this immensely interesting so if you're curious here's a TL;DR explanation. I have 10 EDH decks, so 1000 cards. I made a list of all those cards and got my computer to look up the date that each card was printed. Then I messed around with the list a bunch to see if I'd learn anything.

The first thing I did was find the median age (basically just the average but calculated slightly differently) of the cards in my decks (not including basic lands because that'll abnormally drag the average age down). Those numbers are as follows.
Median card age of each deck
Golos Storm - 2004.0
Dredge - 2007.0
Chromium - 2008.0
Changeling - 2008.5
Jaya - 2009.0
Lasav - 2011.0
High CMC - 2012.0
Tivadar - 2013.0
Boros Polymorph - 2015.0
Dinos - 2017.0
So the average year the cards in my Golos Storm deck were printed was 2004.
The average year the cards in my Dinos deck were printed was 2017.
Basically my golos storm deck has been way way less affected by recent printings than my Dinos deck which makes sense because my golos deck runs a mana base of all duals and has a ton of old elves and spells in it. While there aren't really any good old dinos as that is a tribe that only got support around Ixalan (in 2017).

As for the graphs...
They're histograms, so just bar graphs that record the number of cards per length of time. Here's a slightly more digestible version. (Ignore the title of the graph I forgot to change it, the graph is all non-basic lands in my 10 decks).
Image
So I'm playing 47 cards that were first printed between 1993 and 1997. I'm playing 213 cards printed between 2017 and Strixhaven.
This graph is just divided by year instead of 4 years
Show
Hide
Image
So I'm playing 23 cards from 1993 and only 22 cards from 2021.
Dragoon wrote:
2 years ago
The recent years (2016-2020) seem to have performed really well, but 2007 seems like a pretty big outlier for you (except for non-creatures where 2019 has the lion share). Did you start (or seriously restart) playing magic around that time?
There's a bunch of promos that get filled under the year that the promos line got started. So Etali, Primal Storm was getting registered as 2007 instead of 2017. Annoying. I should have it fixed, but I don't have time to repost the graphs today. Likely can tomorrow.

As for my code. I'll try to post it in the near future. It's a hot indecipherable mess right now though. I'll want to pretty it up some first. Probably be a week or two. I'd recommend downloading Pokken's code. His is way more professional than mine since I'm still very new to this.

Post Reply Previous topicNext topic

Return to “Commander”