Datamining MTGO is about to get a bunch harder.

Do we even know if Wizards has the ability to data mine their own results?

Half kidding, but the latest @diophan repost of the P9 challenge data makes you wonder if they even have the ability to look at more than one month worth of data.

Also: their Post can basically be read as "we can't make this game interesting enough for a large group of people on the Internet" which a.) is likely untrue and b.) if true, it's their fault and not ours.

last edited by ribby

WotC is doing this for the same reason they used to restrict publishing fewer decklists in The Duelist and The Sideboard magazines, the same reason they don't post the winning decklists from ALL grinders at the GP/Pro Tours (they only do a portion), the same reason they don't post results and winning decklists for each tournament on MTGO every day (they have dramatically narrowed what they publish from where MTGO originally started), and the same reason they proactively contact sites like MTGGoldfish and others to prevent them from scraping data.

They are intentionally trying to slow formats from becoming "solved" by those with the time and inclination to mine big data. The less data there is, the more uncertainty there is about what the best cards, decks, and strategies are in a given format. It is certainly frustrating for anyone who wants to collect data to measure card and deck performance, but this speaks to the "discovery" line about gaming, and they will continue to do this going forward. We are supposed to trust that they'll make the best decisions with the data, but I don't know many people who trust WotC to do this with any good faith, especially in light of what we've recently seen.

"We noticed people are coming to different conclusions than us due to our incompetence. Therefore we have decided to censor information so you have to take what we say as truth." -Wizards

I would really like to have wizards decision making process explained to me. It seems like one very stubborn person makes all decisions (and maybe a couple of yes-men).
I can't see how a group of people through intelligent discussion could find this to be a smart move.

@shawnthehero said:

They seem to think that the metagame will become stale if the players have all the statistics, but to me Magic is more than statistics; it always seems possible to break the meta by discovering that perfect card.

The statistics help the meta get to where its going. If we see that Shops, for example, has a 60% win then more people will start to play Shops. This opens up opportunities for players to come along and innovate to prey on Shops decks. Overtime, I would suspect the win% of Shops to then decrease.

Isn't it just constantly changing? Stale? Stale like how?

This statement works great for Vintage, but is less true for Standard because there's a lot lesser tools for metagaming against specific decks. Bare in mind that I'm not disagreeing that this is dumb.

@JACO said:

WotC is doing this for the same reason they used to restrict publishing fewer decklists in The Duelist and The Sideboard magazines, the same reason they don't post the winning decklists from ALL grinders at the GP/Pro Tours (they only do a portion), the same reason they don't post results and winning decklists for each tournament on MTGO every day (they have dramatically narrowed what they publish from where MTGO originally started), and the same reason they proactively contact sites like MTGGoldfish and others to prevent them from scraping data.

They are intentionally trying to slow formats from becoming "solved" by those with the time and inclination to mine big data. The less data there is, the more uncertainty there is about what the best cards, decks, and strategies are in a given format. It is certainly frustrating for anyone who wants to collect data to measure card and deck performance, but this speaks to the "discovery" line about gaming, and they will continue to do this going forward. We are supposed to trust that they'll make the best decisions with the data, but I don't know many people who trust WotC to do this with any good faith, especially in light of what we've recently seen.

The problem with this is it seems to assume that metagames are constant and there is one best deck for a given format, and that is almost never true. There is usually a best deck for a single tournament, and then because that deck is well positioned, the other best decks all shift based on if they beat that deck, and maybe lose to decks that deck beats, etc. The meta game is pretty much constantly in flux for most formats, that aren't terribly unbalanced (see modern with eldrazi, or affinity in long ago standard). This really only turns bad when there is a mistake in printings, and then you don't generally need a whole lot of data to see what is going wrong, two weeks of SCG opens for modern is enough to tell us how crappy modern with eldrazi is, for example.

This is an attempt to make us less intelligent.

Taking away information makes us all dumber. Whatever upsides they feel exist from this, I dislike being made dumber.

@garbageaggro Yes, this was my question exactly. I don't play Standard; is it true that in Standard, there is always a "best deck," and that "solving" the format means determining the decklist of this best deck? Isn't that more an indictment of Wizards R&D, than of data miners?

Mathematically speaking, with enough data it should be possible to determine equilibrium metagames: for instance, one might (hypothetically) conclude that a metagame with 40% Shops, 30% Mentor, 10% Landstill, 10% Storm, and 10% Dredge is in equilibrium, so that if all Vintage players are currently using these percentages to determine their deck selection, no single player can benefit by going rogue.

However it's not clear to me 1) that this kind of analysis is actually useful in the wild, or 2) that having access to this kind of analysis is harmful or makes Magic "less interesting."

@CHA1N5 said:

Hopefully at least Matt and Ryan can/will keep participating and mining for the Premier events for us.

@diophan said:

Anyway, I assume that "if you participate in a tournament you'll be able to replay any of the matches from that event" means we are fine as long as one of us continues to play in the P9 challenges. If we're both busy we might need to recruit some help though.

Yea, that part wasn’t super clear to me. If you just have to “participate” in an event, I’m willing to chip in a couple bucks to have someone join and drop just to have “participated.” If that works, someone will probably come up with a bot and a subscription site that monitors all that crap (which they’ll then disable)

@benjamin_berry said:

If they made it public after a delay, say a month, I think it would be a better compromise.

Heck, even more than a month would be fine. Perhaps after the next premiere event so that you can’t use last month’s event data to metagame.

@lots_of_people mentioned:

This is to keep the format from getting solved/stale

I don’t think that there’s ever a best 60-card deck but there certainly gets to be points where there’s a best “deck” where there’s a best 50-55 cards. Personally, I find the meta-gaming that occurs in those remaining slots to be every bit as interesting. WotC would rather have major changes that are easier for newer players to understand.

I think this is especially true in Vintage where two “decks” might have only a handful of cards different between them but they are considered completely different decks. Take the various Vault or Workshop decks as an example.

@evouga said:

However it's not clear to me 1) that this kind of analysis is actually useful in the wild,

It’s certainly plausible that it’s useful. Actually determining whether it’s useful would require something like building a deck in a complete information vacuum and comparing its results to a deck built after examining recent decks.

As an example:

If I see that no one played Dredge in the last P9 event, I might be inclined to skimp a little on my Dredge hate. Another player might look at the data and determine that players are likely to skimp on Dredge hate and therefore Dredge is a good plan. Then you get the “next level” where I know that you know that I know that Dredge is underplayed!

This next bit, I’m going to take an extremely negative view, so be prepared to possibly be annoyed by the pessimism of my conclusions.

WotC has said in the past that many B&R decisions are primarily driven by tournament results. While we cannot know the precise metrics that they use with this data, an overview of the data should at least support their conclusions. Having more tournament data allows players to more accurately audit the DCI’s B&R decisions.

A few recent B&R changes (or lack thereof) across multiple formats indicate that this data-drive decision making is not the case. Rather, WotC, through their control of the DCI (a whole other thing I’m not a fan of) is attempting to sculpt formats to their vision of the formats, rather than what is “healthy” or “diverse.”

To be clear, by WotC’s own announcements or lack thereof, a spell may put 32 out of 32 possible copies in the top 8 of a major event and not be banned. But Lodestone Golem, with around half those numbers, causes “Mishra's Workshop–based decks [to] continue to be significantly overrepresented, reducing the competitive metagame.” [source]

My conclusion: At least a part of the reason for further reducing the amount of information available to players is to conceal WotC's manipulation of the formats while claiming to be data-driven.

last edited by thecravenone

Actual title of a post on the Magic website today:

HOW TO SOLVE NEW STANDARD FORMATS

  • 21
    Posts
  • 11170
    Views