Superior Numbers – Land Counts

[WWolfe]

It would make sense that the partial decklists that don't include a full mana base or any mana base would adversely skew the numbers. I know my girlfriend for instance has a deck she's building using the site but not adding the cards to it until she has physically acquired the card which includes only basic lands and the few dual lands she already had.

[Morganator 2.0]

If only I would put this much effort into something useful. Like cooking.

EDHREC still hasn't hit it's rollover, but I've discovered some interesting things. For the mundane stuff, I found that:

• EDHREC filters out decks with less than or more than 100 cards
• EDHREC does include decks that are listed as a work in progress

Now for the good stuff.

EDHREC sucks at calculating averages. Remember Marwyn? Well apparently, when you ask EDHREC for all Marwyn decks with less than 20 lands, it says the average number of lands is 28.

So I decided to check with a few different commanders (Aurelia, Exemplar of Justice, Grunn, the Lonely King). Aurelia had an accurate average, but Grunn was off (the actual average was 32, whereas EDHREC listed it as 35). But while looking at some of the decklists for Grunn, I discovered something else.

EDHREC sucks at scanning decks. I looked at the decks with low land counts... and found that they were wrong.
https://edhrec.com/deckpreview/ef/67/ba1d9afe https://www.mtggoldfish.com/deck/1161413#paper
EDHREC says they have 0 lands, they actually have 36
https://edhrec.com/deckpreview/48/71/5b27c2d1 https://www.mtggoldfish.com/deck/1078378#paper
EDHREC says they have 18 lands, they actually have 31
https://edhrec.com/deckpreview/30/c8/e0775573 https://deckstats.net/decks/104426/979969-grunn-n-friends/en
EDHREC says they have 11 lands, they actually have 30
https://edhrec.com/deckpreview/81/ff/367a98bf https://www.mtggoldfish.com/deck/1052424#paper
EDHREC says they have 15 lands, they actually have 30

Now, going back to Marwyn, the land counts of the decks that were supposed to have less than 20 lands actually had an average of... 28.

So I think I've figured it out. EDHREC is inaccurately reading the number of lands that decks have. They seem to read all other supertypes correctly, it's just lands that are off.

[Soren841]

So now email EDHREC with this info.

[robort]

Thanks morgantor for that insight, I just checked something else out in your findings. In the 18 land count vs the 31. I went into the EDHRec link you provided for the 18 land count, scrolled to see the lands shown and it has 14 including the forest. 13 of the lands are non basic lands with the forest. Giving us a total of 14 lands, so did it count 3 extra forest as lands bringing the total to 18 as it suggests? If that be the case why only 3? why not the entire 18 forests that are in the deck?

Now I did the same thing to the 15 land edhrec link you provided, scrolled to the bottom and it shows 16 lands there. Clearly showing that 15 lands are supposively there when clearly the count is at least 16. Which land did it not count? Could it be forest again? is the land count that EDHrec suspectable on counting basic lands?
If it is suspectable on counting basic lands, could that be the reason why the Cedh decks skew so low because it doesn't count the basic lands very well?

But back to Not having enough lands and I will use myself as the example and just use 3 formula's as well to give me a general avg of if I have enough or not..

First is the one in this thread.. 

Number of lands = 31 + amount of colors in commander’s color identity + commander CMC, with 0-CMC mana rocks counting as lands.

Then 2 others which are almost the same just about.. 1st is 50% must be mana producers from lands/rocks/ramp..
but 37-38 cards are the deck itself and the other 12-13 are interactions, 4 of the interactions is draw while 8-9 is some sort of removal.

2nd is 40 lands/10 rocks or ramp/5 singleton removal/5 mass removal/10 draws/30 is the deck itself..

Taking my Thantis deck based apon the lands itself not if they are the correct/wrong ones but it has 37 lands with 9 forms of ramp. According to any of those 3 ideas/suggestions I am off

I should have 40 lands in the 1st one because 31+3(amount of colors)+6(Commander's cmc)= 40 but I have only 37 with 0 rocks in the deck with cmc cost of 0. Off by 3 lands

Second one I should have 50% producing some type of mana and I have only 48% of producing that mana. Off by 2% However I have Rites of Flourishing in the draw category but it allows to play more lands and not counted in the ramp category.

Third one I should also have 50% as well but again only have 48% and off by 2%. Again Rites of Flourishing to play more lands and not counted in the ramp category.

Running avg costs for spells is 3.86 so in essence I should have the land count increased based on what is being revealed. 

Now looking at edhrec on the avg lands thantis has shows avg distribution of lands is 36. I am higher than the avg distribution but lower then the suggestions. So an argument can be made for me to include a land or two especially based apon my avg cost of casting spells. The land counts are important especially in this example of Thantis.