Decks and top cards to collect/craft doesn't seem to accurately reflect the current Standard meta (tournament data)

[oysteinga]

I don't understand exactly what decides which decks appear in the deck tabs on the left, how often they are updated, and which of the deck configurations is selected as the main configuration, but from the description on mtgazone.com/deck-advisor it seems to be the most popular decks from mtggoldfish for Standard.

What I find strange is that I don't see Korvold, Fae-Cursed King in my list of Top Cards to Collect/Craft. It is currently being played in Jund Sacrifice/Food which is the most popular among all the metadecks on mtggoldfish, and is also currently both in the Sample Deck for this archetype on mtggoldfish and it's listed that 94% of the Jund Food decks plays two copies of the card. I don't have a single copy, and would assume that this would make it one of the top cards for me to craft.

I further find it strange that I find Mono-Blue Mill at the top of the Standard decks, and Jund Food at no. 14. The main configuration selected for Jund Food doesn't contain Korvold, which could explain why it doesn't show up in Top Cards, but why is this the main configuration when most successfull (and all of the alternative) configurations contain this card?

Also worth noting, Korvold is the 43rd most played card according to mtgtop8.

I like the idea behind Deck Advisor, but I currently don't trust either the decks selected on the left and the order they are displayed, or the list of top cards to craft.

Further questions/suggestions:

• Have you considered using data from mtgdecks.net/Standard or mtgtop8.com? I mostly like the Average Deck on mtgdecks to see what cards are mostly used for the archetype, and similarly the "Compare Decks" function on mtgtop8. I also like the "Deck Performance" chart on mtgdecks, assigning points to each archetype based on tournament results. It could be the deck lists on mtggoldfish are just as good or better, and my main concern is with the selection of the main deck configuration.

For selecting the main deck configuration of an archetype, some kind of average of the different configurations could be used. A strict "average deck" across all configurations could create a good foundation for calculating Top Cards, but could result in strange decks if you're trying to build the actual average deck. Selecting an actual deck that most closely resembles an average configuration could be a better choice. "The best versions of the decks" (from the description on mtgazone) could skew the decks towards a configuration with strong individual results, but versions that don't really reflect the current meta. If "the best version of the deck" continues to be "the best version of the deck" across multiple tournaments, this version should after a short while end up as an average configuration.

Screenshots

[jceddy]

Just a couple of quick notes before I go into this in more detail:

The Standard deck data comes from mtggoldfish and is updated once a day (so there can be up to a day lag from when decks are updated on mtggoldfish).

The main deck configuration for a given archetype is the one linked to from the mtggoldfish standard decks main page.

The back-end script that pulls this data pulls the main deck, as well as other tournament configurations of the deck linked to from the main deck page on mtggoldfish (the links on the right of the page there).

The back-end script also pulls in the "budget" standard decks from mtggoldfish.

The "top cards to craft" are determined based on all of those decks combined.

The default ordering of the "no wildcards needed" sorts decks that require the fewest wildcards to the top of the tab strip.

It sounds like possible enhancements you are looking for, perhaps, are:

• an ordering option that ignores deck cost
• a filter/option to ignore "budget" decks in standard (or at least sort them to the bottom)
• different logic for determining "meta share" for cards, maybe by including "% of the meta" for decks from mtggoldfish, or even just grabbing straight-up meta share percentages from a different source

Does that all make sense/sound about right?

[oysteinga]

Thanks for the explanation.

The only thing I have really used Deck Advisor for is the Top Cards. What I really would want this list to contain is a list of the most "successful" cards. And what I really like is that cards drop down in the list when I get them in my collection.

This list has mostly seemed to correlate to the most played cards, minus the ones I already have.

My concern was that Korvold never appeared in this list, although it has been widely played in Jund Sacrifice. It could be that Korvold hasn't been in the main deck configuration before today though.

I guess what I want for the Standard format section is the cards shown in that list to reflect (like you are saying) the meta share, based on the number of tournament decks they appear in. Budget decks shouldn't really play into this. Budget decks could be another selection entirely (in addition to Standard and Arena Standard), but Arena Standard should probably pick up more of the budget cards already, as most players don't have complete collections. Which sources and how they are used to determine the meta share could be done in different ways, but I would expect it to either be based entirely on the number of tournament decks (top finishes) the cards appear in, or that the decks are also somehow weighted according to their results (To win a tournament is better than top 16).

PS: I just tried out mtgahelper (mtgahelper.com) and it's "Cards to craft" section which currently defaults to track 269 decks from MtgGoldfish. It seems to produce a very similar kind of list to yours and in the resulting list here, Korvold appears at number 47 for me. Interestingly it also tells me that Korvold appears in 4 decks, although Korvold appears in more than 4 of the Jund Food decks on MtgGoldfish. They must be filtering the decks somehow as well.