(Kithkin Healer | Art by Rebecca Guay)

Information Overload

Welcome to EDH Political Science, the series where we discuss the multiplayer madness of EDH through the lens of history, science, or politics.

Florence Nightingale was a scientific hero truly before her time. She strongarmed the medical community into accepting modern hand-washing practices. She saved lives daily in her wartime nursing posts. She was a women’s rights advocate decades before the global women's movement gained momentum. Her genius breakthrough came when she used statistical analysis to prove her points in easy-to-read graphs. Waves of understanding about hygiene practices crashed upon the medical community. All in an era before germ theory!  

Nightingale is the sort of statistical success story we at EDHREC look up to. We see how data has changed the understanding of so many things in the world that it becomes easy to forget how difficult some things can be when you don’t have it, especially in today’s world where the data is twisted. Data is often mistrusted because the source comes from an opposing political faction. The act of looking for data can be politically motivated or be pushed in a way that is beneficial to one faction or another.  

There is a persisting fear that, as more people look at EDHREC to build their deck and put that decklist online, the more they'll be influenced by the data showing up on EDHREC. This feedback loop theoretically creates unwarranted consistency in the data.

There is some truth here: if you want to build an Atraxa, Praetor’s Voice EDH deck that looks like the average Atraxa deck, then that’s easy. If you build a powerful list here based off the main page of a commander and then upload your list online, that could get picked up by the data. These choices do happen, and they feed on themselves creating higher inclusion counts for those cards than might happen without the EDHREC data, but not only is this bias less significant than many assume, there are also ways to interpret data to take this into account. 

Let’s wield this data properly. Let’s consider what Florence Nightingale learned, what that can tell us about deckbuilding, and how that data effects play. When working with large amounts of raw data, like we have here on EDHREC, or when working from unreliable sources, we should do a few things: understand statistical biases, understand your biases, and consider statistical averages. 

Statistical Biases 

When viewing a single commander’s page, you might run into the following statistical biases: the Precon Effect, card availability, recency bias, budget bias, popularity bias, same-set bias... I could go on. There's an entire podcast devoted to these biases. At the top of every EDHRECast show, Joey reminds the listener, “…what we like to do is give all that [EDHREC] data a little more context.” Finding new ways to interpret the data of EDHREC is kind of the whole reason we have articles. Jevin’s Challenge the Stats series is about finding flaws in the data. Jaelyn’s Stapler Remover series is about highlighting alternatives to cards with bloated statistics. These biases not only effect your deckbuilding, but the entire meta you play in. Not all these biases are in play with every commander, and not all of them are as popular as the last. What they do remind us of is that data always needs to be coupled with context. 

When looking for data to prove that hygiene was tied to disease, Florence Nightingale didn’t present every piece of data she could find. She took only the data on deaths before and after hygiene practices were introduced into army hospitals stationed in Crimea, which limited data that could be related to alternative factors. It refined her point and showed exactly what she needed. Deaths were reduced when you implement good hygiene.  

Similarly, what I enjoy doing is finding pockets of EDHREC where data is not sparse, but thin enough to be distanced from many popular biases. Traveling to the pages of commanders that were not featured as the face commander of a pre-constructed deck and whose page features between 200 and 600 lists is a good place to be. Try for something that's recent enough to be available, yet not new enough that the lists are going to change much. Something like Zirda, the Dawnwaker is a good pick. She fits the bill perfectly, and another good thing about her is that if we're feeling spicy enough, we can use her commander page to inform decisions about making her a Companion instead of our commander.  

You don’t have to travel far down the Zirda page to find that she makes for a mean combo-oriented commander. This wouldn’t be surprising to a veteran EDH player. Cost reduction effects, like that on Zirda, supply great opportunities to play Basalt Monolith for infinite mana, Auriok Salvagers to partner with mana-producing artifacts like Lotus Petal, and plenty of places to use the infinite mana in Walking Ballista, Diviner’s Wand, and Staff of Domination. Using this information can bring us closer to something not bloated with decklists based on pre-constructed decks or lists all based on one another. Something with a good statistical variety in its card choices.

Personal Biases 

Plenty of people disagreed with Nightingale. John Simon, a contemporary of Nightingale’s and England’s chief medical officer, disputed her claims explicitly. “Dr. Simon wrote in 1858 that deaths from contagious disease were ‘practically speaking, unavoidable’--that there was nothing to be done to prevent future deaths,” wrote Tim Harford in his book The Data Detective. This is the sort of thing we see every day. Every single person makes judgements about data through their biases. Some of them are legendarily bad takes, like Dr. Simons'. Others may be benign insights into how to build a better EDH deck.  

Dana Roach, co-host of EDHRECast, authored an article series called In the Margins which detailed cards that players play more often than their strictly better counterparts. He took an analytical sword to the throat of cards like Eternal Witness, Manalith, and Putrefy. This reminded people to check their own biases as well as the biases of the EDHREC system. Shining light on these ideas, they become a shortcut for people to think to themselves, "Oh yeah, there are better things to be playing than this card." Shortcuts are powerful and give your brain the ability to concentrate on more important things. 

However, when you create these shortcuts, you're inherently creating another bias. A more useful one than the last, but still a bias. In creating this bias, you need to remind yourself that this shouldn't overwhelm every other idea. If you're in a pickle where you can't find your Mind Stone, and all your Signets are in other decks, and you absolutely cannot find anything better in your collection than Manalith, then it's more important to run bad ramp than be short on mana acceleration. Manalith is still a bad card. That doesn't change just because you found a circumstance where it works. The point here isn't that Manalith can be good or that Dana can be wrong or that biases are bad. What's important to realize here is how powerful doubt is. When your biases skew towards doubt, those are the ones that breed and spread. You begin sowing the seeds for uncertainty towards everything. 

This is a powerful effect. Doubt is one of the most important parts of misinformation campaigns. They are often focused less on "this idea versus that one" but on doubting everything. Misinformation, like what Nightingale was up against, is about breeding skepticism for basic facts and reasoning.

This aspect is so powerful that it can happen regardless of if you recognize it or not: Cursed Mirror is a card from the new Commander product that I immediately disregarded as being another Manalith. That doubt almost caused me to miss one of the most powerful cards from the new set. It acts much closer to singular Sneak Attack than it does to a random mana rock. Point being, keep your skepticism in check, still listen to what Dana Roach has to say, and usually don't run Manalith.

Statistical Averages 

Florence Nightingale didn't have a large amount of statistical data, but she did have enough to understand that, on average, more people died when hygiene was poor. This was the beauty of looking at the rose graph she crafted. You can look at that and see the data lining up with her point. She had a hunch about her idea and needed to look at the statistics to prove that.

When sitting down for a game of Commander, we don’t have access to the statistics of our opponents' lists. We can’t see the wonderful EDHREC pie graph depicting the creature spread or bar graph showing the mana value curve. We don’t get to pick apart each card in the list and see how often other popular cards are ran with that commander. There are limitations we must abide by. 

That doesn’t mean you have to be blind to the possibilities. Scouring EDHREC can give you valuable insight into what you can expect to see at the table. The more you intake information about the format, the better you'll understand the composition of every list you face. This can be used to your advantage in a couple ways. You can skew expectations about what’s in your list by understanding baseline expectations.  

Skewing other’s expectations is as simple as finding a new way to represent your commander. If you have a novel idea for running a particular commander, you can skew other’s understanding of your list. For instance, my Ephara, God of the Polis list (featured in an article I did last year) is often met with, “Oh neat, I love blink decks.” This is commonplace with every archetype and popular commander. People look for the familiar; they love to see a commander they've built before across the table and see what it can do in the hands of someone else. While the “blink” comment isn’t entirely wrong, they likely don’t expect Torpor Orb effects to be an important piece to running the list. Astute Commander players may understand when you combine the effect with blink cards, like Venser, the Sojourner, you can make Wormfang Manta grant infinite turns. Wormfang Manta is only in 104 lists on EDHREC. That would be a pretty deep cut for someone to pull without prior knowledge. Obscuring your win-conditions with arcane information is a great way to gain advantage in a game. 

I've already shown that list in a prior article, so for this article, I'd like to showcase another deck that fits this line of thought:

The List

This list is something that's fun because of the way it can play with expectations. Livio, Oathsworn Sentinel and Alena, Kessig Trapper with Zirda, the Dawnwaker as Companion sets expectations for a high likelihood of infinite-mana combos. There’s no tempering that. However, it is slightly tempered by the fact that you are restricted by Companion. I enjoy the puzzle-like experience of finding new ways to combo with the pieces given.

Playing this list, you're looking to gain value by blinking creatures with powerful tap abilities so that you can use them again in combination with other creatures with tap abilities or by themselves over and over again. You can repeatedly ping creatures and players for three damage by blinking Kamahl, Pit Fighter and Inferno Titan. Flickering Pentarch Paladin or Mangara of Corondor to remove problematic targets is valuable. This is not to mention the bevy of combo potential. Sword of the Paruns, Basalt Monolith, and Aggravated Assault all have potential to go infinite with Zirda and/or Alena. 

That’s all I have for you today. I hope you had fun looking at the world though how Florence Nightingale understood statistics. If you have thoughts on the subject or would like to suggest topics for future articles let me know in the comments. Otherwise, you can talk to me on twitter @RickWorldNews or on Reddit at u/xebra7.  

Rick Amundson

Rick is a professional bookseller, writer and gamer. He could be battling international trade atrocities with his Bachelors Degree in International Studies, but he would rather be playing games of magic, running game night at the bookstore, playing an insidious dungeon master at a convention, or talking and writing about any of the above.