Login | Register


All times are UTC - 7 hours


It is currently 2017-Sep-25 8:29 am




Post new topic Reply to topic  [ 25 posts ]  Go to page 1, 2  Next
Author Message
 Post subject: EDHREC
AgePosted: 2017-May-17 10:10 am 
User avatar

Joined: 2008-Mar-24 12:14 am
Age: Elder Dragon
Location: Oakland, CA
This came up in another thread and I thought it was interesting but didn't want to even further derail.

Uktabi_Kong wrote:
Could we for the rest of time on these boards not use EDHREC as a source? It's a pretty decent resource for deckbuilding ideas, but it's terrible to try to describe EDH as a whole. Its sample size is too small, and the results are confounded by the fact that every deck on the site comes from someone who specifically put it there. I've built, edited, played, rebuilt, and dismantled probably somewhere around 50 different decks (including having the same deck but changing its focus over and over). I have yet to post on EDHREC once, and as far as I know nobody I've played with in the last couple years has either.

My experience with EDHREC has been that it pretty accurately represents the majority of decks I'm likely to run into at a new LGS. I don't know whether you need to post specifically to EDHREC or if just posting to tappedout or deckstats under the EDH/Commander tag is enough to be picked up by the EDHREC engine, but it sure has a lot of decks. I guess you could call it a small sample size in comparison to all the decks (and deck ideas) ever built, but as long as it keeps growing it'll be the biggest sample size I think we'll see, and I think it's big enough already to call its data meaningful at least.

The fact that there is a certain type of builder who is more likely to post their decklist online is relevant. I think Uktabi_Kong is right that it skews the data somewhat. I think this will be an issue if anyone ever wants to get a collection of non-anecdotal data for this format. Maybe the best approach is to ignore the desire for valid, concrete data? Just stick to gut feeling and anecdotes?

What do y'all feel about it? Does the data in it mostly match what you would expect from a new LGS? Or is it especially skewed? Is it more casual? More Spikey? Less quirky/innovative?


Top
 Offline Profile  
 
 Post subject: Re: EDHREC
AgePosted: 2017-May-17 10:46 am 
User avatar

Joined: 2012-Feb-07 4:15 pm
Age: Elder Dragon
I actually happen to like EDHREC a ton and use it all the time for ideas for new decks and/or a database to find new cards for the decks I already have.

However, I think the vast majority of the card choices and decks it shows are skewed toward the group I'd like to call "LGS Kings". Basically, they tend to have decks that largely outperform most "casual" decks yet aren't actually strong enough to be considered competitive. This can be for any number of a couple reasons. Maybe they're just really good deckbuilders (or netdeckers) but still are playing more for fun than for winning. Maybe they are attempting to build an actual competitive deck but aren't actually good at it. Or (I think by far the most common) they understand how to build and play really OP decks but (intentionally or not) choose a strategy/general that isn't viable in a true competitive meta.

The other thing that is both a benefit and problem with the site is that about 90% of the cards you'll see for any popular general are staples. For some generals you can get some cool/interesting choices in the "signature cards" section, but the vast majority of the rest of the cards shown are more or less common among everything else in the archetype. Any given big tribe will show the same 15 lords alongside Coat of Arms and Door of Destinies, removal spells will always include the same couple staples, any Voltron general will always recommend the Swords of Meat and Potatoes, ect. It also tends to promote more of a "friendly" meta in general, with unfun staples like Armageddon and Humility being notably absent from most lists.

_________________
thaumaturge wrote:
Actually, Uktabi_Kong and Kemev are practically the royal family of deck doctoring on this forum, and it's generally a good idea to listen to just about anything they have to say.


Treamayne wrote:
Deck DR Kong


1st Place in the Gabriel Angelfire Iron Chef EDH Competition

3rd Place in the Zedruu Iron Chef EDH Competition


Top
 Offline Profile  
 
 Post subject: Re: How do you y'all deal with Cyclonic Rift?
AgePosted: 2017-May-17 11:09 am 

Joined: 2017-May-09 1:04 pm
Age: Wyvern
Uktabi_Kong wrote:
I have yet to post on EDHREC once, and as far as I know nobody I've played with in the last couple years has either.


You don't post on EDHREC, it pulls its information from other places where you upload decklists. And its sample size is not insignificant. 150,000 decks, why don't you take a look at how many decks are running cyclonic rift. https://edhrec.com/top/

EDHREC is absolutely a source. I don't like using it, and when I do I mostly use the advanced filter to find decks that are going to be similar to what I'm trying to build. I prefer advanced search on gatherer more. But regardless, EDHREC is a huge source, if not the biggest source in the EDH community.


Top
 Offline Profile  
 
 Post subject: Re: EDHREC
AgePosted: 2017-May-17 12:07 pm 
User avatar

Joined: 2017-May-01 12:07 am
Age: Hatchling
The EDHREC filter system is very good.
If you know what youime, but it want it is very simple to build a deck with this idea and improve it just by inserting the correct filters.

I usually prepar three or four lists, step them into some excel macros that i did and then make the fine adjustments.

It takes me some time but it is funny


Top
 Offline Profile  
 
 Post subject: Re: EDHREC
AgePosted: 2017-May-19 1:04 am 

Joined: 2008-Sep-12 8:43 pm
Age: Drake
Location: Madison, WI
"Swords of Meat and Potatoes" is my new favorite card nickname. Thanks for that.

I like browsing EDHREC to get an idea of cards that may synergize well with the general that I may not have thought of before. I find it faster sometimes than trying to massage Gatherer to get the right keywords to come up when I'm looking for all cards that do X. EDHREC may not catch all of those, but there's a good chance it will show the most commonly used variants.

The usefulness of the site really comes down to what you're trying to accomplish when building a deck. Take something like Atraxa for example. The cards it puts up right now are basically every good stuff card that works with proliferate, and it's more than you could conceivably build a single deck with. However, whether you're looking to go super friends, poison, or dudes with counters on them, the site gives you a pretty good idea of what works well in the deck. Apply the same logic to pretty much any common tribal/blink/pillowfort general out there.

I really don't see what's wrong with that. You can still find your own hidden gems. It's just a jumping off point when brainstorming ideas for a decklist.

In terms of the data question in the OP - I don't think there's any one type of person who posts lists online. I've seen some truly horrible EDH lists posted in various forums over the years, and I've seen perfectly normal and totally cutthroat ones as well. Some people generally are looking for help, some are just wanting to show off the cool list they put together, and some are just putting them online so they have an easy place go reference the list they came up with.

All that said, I feel like the cards I see pop up on EDHREC are exactly what I expect to see when I sit down with randoms to play a game. That's not a bad thing. It's just a sign that the format does have staples and people who are serious about the format will work to acquire and build with them.

_________________
My Decks
Damia - Maga - Bolas - Karrthus - Karador


Top
 Offline Profile  
 
 Post subject: Re: EDHREC
AgePosted: 2017-May-19 5:51 am 

Joined: 2009-Apr-21 3:38 pm
Age: Elder Dragon
Location: Palm Springs Area, CA
I don't think I've ever seen EDHREC list a card for a particular commander that I thought was innovative or quirky. I think the nature of the site prevents that.

Its a great resource I suppose, just not one that I need.

_________________
3DH4L1F3


Top
 Offline Profile  
 
 Post subject: Re: EDHREC
AgePosted: 2017-May-20 7:18 pm 
User avatar

Joined: 2012-Dec-25 1:37 am
Age: Elder Dragon
Uktabi_Kong wrote:
However, I think the vast majority of the card choices and decks it shows are skewed toward the group I'd like to call "LGS Kings". Basically, they tend to have decks that largely outperform most "casual" decks yet aren't actually strong enough to be considered competitive. This can be for any number of a couple reasons. Maybe they're just really good deckbuilders (or netdeckers) but still are playing more for fun than for winning. Maybe they are attempting to build an actual competitive deck but aren't actually good at it. Or (I think by far the most common) they understand how to build and play really OP decks but (intentionally or not) choose a strategy/general that isn't viable in a true competitive meta.

The other thing that is both a benefit and problem with the site is that about 90% of the cards you'll see for any popular general are staples. For some generals you can get some cool/interesting choices in the "signature cards" section, but the vast majority of the rest of the cards shown are more or less common among everything else in the archetype. Any given big tribe will show the same 15 lords alongside Coat of Arms and Door of Destinies, removal spells will always include the same couple staples, any Voltron general will always recommend the Swords of Meat and Potatoes, ect. It also tends to promote more of a "friendly" meta in general, with unfun staples like Armageddon and Humility being notably absent from most lists.

Give me data that shows EDHREC decks are not representative of the kinds of decks you would expect to find in Commander and I'll happily stop using it as a citable resource.

You are making plenty of statements about the website and the decks that people post "there" (they don't, the data is withdrawn from exterior websites, as said before) without counter-arguing how and why these do not represent the deckbuilding philosophy of the average playgroup. What makes you say that decks worldwide are not filled with staples? What proof do you have that everyone who runs white actively plays stax effects, unlike what the data shows on EDHREC?

Also, if you ever meet a biologist such as myself, absolutely NEVER tell them that a 80.000+ sample size is too small.

_________________
Name: Forged in Stone
General: Nahiri, the Lithomancer
Archetype: Aggro

Name: Night of the Ninja
General: Vela the Night-Clad
Archetype: Aggro-Control

Name: Boros Legion
General: Aurelia, the Warleader
Archetype: Aggro

Name: Dragon Fire
General: Karrthus, Tyrant of Jund
Archetype: Midrange


Top
 Offline Profile  
 
 Post subject: Re: EDHREC
AgePosted: 2017-May-21 2:24 am 
User avatar

Joined: 2012-Mar-31 11:52 am
Age: Elder Dragon
This subject came up on MTGS as well, in response to a point/counterpoint article some site ran. I agree with the sentiment that EDHREC is a useful tool for getting a statistical average, but it doesn't represent a complete built deck, because players will always make meta changes or have financial restraints. MTGS also has a similar thread where one user developed a script to compile decklists for particular generals, and the results are largely similar.

One other thing to note which I saw another person mention, is that if you build a deck and play with it and tweak it, your deck will probably be pretty close to a "generic" EDHREC deck.


Top
 Offline Profile  
 
 Post subject: Re: EDHREC
AgePosted: 2017-May-22 7:37 am 

Joined: 2009-Apr-21 3:38 pm
Age: Elder Dragon
Location: Palm Springs Area, CA
Online postings are inherently biased. Doesnt mean they're wrong, but its a bias.

Based on what I've seen EDHREC is pretty generic with its card suggestions.

_________________
3DH4L1F3


Top
 Offline Profile  
 
 Post subject: Re: EDHREC
AgePosted: 2017-May-22 7:39 am 
User avatar

Joined: 2012-Feb-07 4:15 pm
Age: Elder Dragon
niheloim wrote:
Online postings are inherently biased. Doesnt mean they're wrong, but its a bias.

Based on what I've seen EDHREC is pretty generic with its card suggestions.

I may have spoken in a hyperbole, but this is pretty much what I meant in a nutshell.

_________________
thaumaturge wrote:
Actually, Uktabi_Kong and Kemev are practically the royal family of deck doctoring on this forum, and it's generally a good idea to listen to just about anything they have to say.


Treamayne wrote:
Deck DR Kong


1st Place in the Gabriel Angelfire Iron Chef EDH Competition

3rd Place in the Zedruu Iron Chef EDH Competition


Top
 Offline Profile  
 
 Post subject: Re: EDHREC
AgePosted: 2017-May-22 7:58 am 
User avatar

Joined: 2012-Dec-25 1:37 am
Age: Elder Dragon
niheloim wrote:
Online postings are inherently biased. Doesnt mean they're wrong, but its a bias.

Based on what I've seen EDHREC is pretty generic with its card suggestions.

I'm not questioning the existence of a bias. I am, however, arguing that nobody knows what that bias is, or how skewed the data is. There may not even be a bias at all. As long as there is nobody collecting data about the decks played by the average playgroup, EDHREC is the best source of data we currently have about the types of cards and decks that are played in the format. Until someone can prove how biased or skewed the data is, I see no valid reason to dismiss arguments based on EDHREC that assume it is an accurate representation of cards and decks played in Commander.

_________________
Name: Forged in Stone
General: Nahiri, the Lithomancer
Archetype: Aggro

Name: Night of the Ninja
General: Vela the Night-Clad
Archetype: Aggro-Control

Name: Boros Legion
General: Aurelia, the Warleader
Archetype: Aggro

Name: Dragon Fire
General: Karrthus, Tyrant of Jund
Archetype: Midrange


Top
 Offline Profile  
 
 Post subject: Re: EDHREC
AgePosted: 2017-May-22 8:21 am 
User avatar

Joined: 2017-Mar-11 6:43 am
Age: Drake
Maluko wrote:
niheloim wrote:
Online postings are inherently biased. Doesnt mean they're wrong, but its a bias.

Based on what I've seen EDHREC is pretty generic with its card suggestions.

I'm not questioning the existence of a bias. I am, however, arguing that nobody knows what that bias is, or how skewed the data is.


EDHREC pulls it's data from TappedOut postings. I tested this by seeing if my Ephara deck is present; and if it skewed the percentages at all (since there aren't all that many Ephara decks listed.)

https://edhrec.com/commanders/ephara-god-of-the-polis?f=in%3DSpringjack+Pasture

The biases come from TappedOut, which as a platform is used to share decklists, and EDHREC, as a platform is used to create suggestions based on sourcing those lists. This is why EDHREC tends to skew more towards higher-power builds. The site isn't trying to turn anybody into spikes, because if it were, Monastery Mentor would be on the "Signature Cards" for Ephara- but not enough people really try to push the General to understand what a linchpin it is for any build with Ephara- atleast not enough people who posted lists to TappedOut.

_________________
niheloim wrote:
Wall of Chat. 2U
Creature- Wall

Defender
Wall of chat exceeds at using a lot of words to mischaracterize opposing view points.

Warp Riders (Ephara Solar Flare)


Top
 Offline Profile  
 
 Post subject: Re: EDHREC
AgePosted: 2017-May-24 7:43 am 

Joined: 2009-Apr-21 3:38 pm
Age: Elder Dragon
Location: Palm Springs Area, CA
Maluko wrote:
niheloim wrote:
Online postings are inherently biased. Doesnt mean they're wrong, but its a bias.

Based on what I've seen EDHREC is pretty generic with its card suggestions.

I'm not questioning the existence of a bias. I am, however, arguing that nobody knows what that bias is, or how skewed the data is. There may not even be a bias at all. As long as there is nobody collecting data about the decks played by the average playgroup, EDHREC is the best source of data we currently have about the types of cards and decks that are played in the format. Until someone can prove how biased or skewed the data is, I see no valid reason to dismiss arguments based on EDHREC that assume it is an accurate representation of cards and decks played in Commander.

unless the sample is not a sample but the whole population there is bias.

Random sampling goes a long way to minimizing that so that we can use statistics, but there is no way a self-selecting collection of web postings will be random.

But if you want a valid reason why not to use the data... how about its unreliable. It cannot be shown to represent the population.

_________________
3DH4L1F3


Top
 Offline Profile  
 
 Post subject: Re: EDHREC
AgePosted: 2017-May-26 4:56 pm 
User avatar

Joined: 2015-Mar-18 12:55 pm
Age: Drake
I ran my deck (http://tappedout.net/mtg-decks/treva-enchanted/) through EDHRec to see what it would come up with. Unsurprisingly, most of it was Bant goodstuff. Here's the list of the cards in my deck, according to EDHRec, from most to least unique: https://edhrec.com/recs/?url=tappedout. ... niquecards

_________________
The QFT Section
Sheldon wrote:
The cards didn't just warp the way the games were played, they warped how I was conceiving and perceiving the format. That's the sign of a problem.

Carthain wrote:
The idea that you should be able to build your deck however you want and still be competitive is false, and a bad idea to have. Taken to the extreme, that's like making a deck with no removal in it, and then complaining that you can't win because stuff your opponents play gets in your way.


Top
 Offline Profile  
 
 Post subject: Re: EDHREC
AgePosted: 2017-May-27 3:40 am 
User avatar

Joined: 2012-Dec-25 1:37 am
Age: Elder Dragon
niheloim wrote:
unless the sample is not a sample but the whole population there is bias.

Random sampling goes a long way to minimizing that so that we can use statistics, but there is no way a self-selecting collection of web postings will be random.

But if you want a valid reason why not to use the data... how about its unreliable. It cannot be shown to represent the population.

Your post suggests you don't have much knowledge about how data is collected for scientific studies in order to reach a conclusion, so I'll try my best to enlighten you.

First of all, having a true random sampling of a population is incredibly difficult. There are very few studies, if any, that can claim to have a good enough random sampling of a population to represent it's reality. You could collect data about certain variables from the people in your country, but it would be just that, a random sampling from people in your country. It would never be representative of the worldwide population. And then, you need to account other factors, such as age, sex, level of education, and so on, in order to reach a general conclusion about the people in your country regarding the variables you measured. Now repeat this for every other country in the world. You can certainly imagine the huge, HUGE effort that would take to perform this task.

This is why nobody does this. Scientists collect data according to the resources they have available, including money, time and effort, and this data is never representative of the whole population because these resources are always very limited. That doesn't mean this data shouldn't be cited and used by other people and other scientists, because we have to start somewhere. If we were using truly random sampling from the start, then science would progress incredibly slowly, because studies like these would take DECADES to see the light of day.

Now, to stay on topic and give you a concrete example, suppose you want to collect data about which cards are being used in which EDH decks, something that has never been done before. Where do you start? You probably start locally and collect data from the people that play in your local game store. Now, you could stop here and, based on the evidence that you collected, be able to reach a conclusion for some cards. Would this sample be representative of the whole EDH population in the world? Absolutely not. BUT it's a start! Nobody ever did a study like this before. So now you have two choices: either you stop collecting data now and publish your study for everyone to see, or you continue to collect data from other stores and other players. What would happen if you were to post your study now? People would comment. People would look at your data and draw conclusions. Maybe even the RC could look at it and use it as a resource to understand how the format is developing, and make changes. Maybe someone looks at your data and decides to replicate your study in their own store, in their own country.

Now you have the other end of the spectrum, which is you think this is too low and you really want a true random sampling from every country, every LGS, every player below 20 years old, every woman who plays at least once a week, etc. What happens is you would die without ever reaching your goal, your research would never get published, and the players and the format gained nothing from your long hard work.

You see where I'm going with this? Yes, maybe the people who post decks online have a certain type of personality that is not representative of all the people in the world who play EDH. But you don't have evidence that support this, and you never will, because collecting the amount of data from people worldwide who play EDH to match the amount of data that EDHREC collects from people who post their decks online would take you decades.

So please stop with the absurd claims that EDHREC data is unreliable. You don't have any evidence even remotely suggesting this is true other than your own opinion. At the very least, show me a study that suggests people who post online are biased towards a certain end of the personality spectrum and I'll gladly give you the benefit of the doubt. Until we have a better data set from where we can extrapolate general conclusions about EDH decks, EDHREC is the best we have.

_________________
Name: Forged in Stone
General: Nahiri, the Lithomancer
Archetype: Aggro

Name: Night of the Ninja
General: Vela the Night-Clad
Archetype: Aggro-Control

Name: Boros Legion
General: Aurelia, the Warleader
Archetype: Aggro

Name: Dragon Fire
General: Karrthus, Tyrant of Jund
Archetype: Midrange


Top
 Offline Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 25 posts ]  Go to page 1, 2  Next

All times are UTC - 7 hours


Who is online

Users browsing this forum: Google [Bot] and 33 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
Jump to: