This is actually pretty cool. About 15 hours ago someone got a notification that their Brawl deck was invalid. So, they checked the logs and found that it was actually listed as being weighted well below 0 and the system would check that the deck needed to be at least 0. Then jsut 8 hours later, there’s now a complete list of all the weights. Links to the Reddit threads here and here respectively.

Again, this is cool. Someone just figured out how to do this and then did it in a pretty short time.

Supposedly, this weighting is for matching purposes. People should be matched against similarly powered decks. I find many problems with this, but at the same time, they are trying, which is good. Since my attempts to play Brawl are often cut short because I am always being amtched against a couple of very powerful commanders, I don’t think this approach is really working.

Anyhow, I have questions about this system. Again, it would appear that each deck is weighted based on the cards within it. Each card just has a value and how good a deck is based on the total value of those cards. Of course, this is just absurd, but at the same time, they need some kind of a metric and anything else would be very complicated. So, I would assume they have some kind of a learning system which comes up with these numbers.

If so, why are most of the weights divisible by 9? There are some outliers, but they are still divisible by 3. That is so weird. Someone must have hardcoded that. Someone must have thought that having values divisible by 3 is good. I just don’t see why. What would be the point? Since so many of them are divisible by 9, but there are some outliers, feels like that’s a bug. There’s some part of the system that pushes cards to those values and it just seems like it was unintended, but I can’t really know.

Another interesting thing is that 45 seems to be the upper limit, but two cards broke that limit: Tibalt's Trickery at 180 and Zenith Flare at 216. One thing this does is that this kind of automatically pushes the decks using these cards into the upper echelon, which is good, since so many of the cards around them are probably pretty low on totem pole.

But how did the system get there? Were these hard-coded as well? It would make sense, because those two cards can be bad experiences to play against and neither of them should exist (I guess the problem with Flare is more about the cards around it that were mishandled than the card itself). However, if the system is based on learning, why aren’t the cards around them similarly highly weighted. Tormod's Crypt (a common food for Trickery) is worth just 9, which is what most cards are weighted at.

So, what’s going on here? Is someone actually weighing everything manually? That would be so much work and would require so much maintainance that it would feel overwhelming. So, that can’t be it… but at the same time, that’s exactly what it seems to be.

This is all very interesting and gives us a lot of insight into what’s going on at WotC. There’s also a bunch of cards that apparently exist on Arena, but aren’t available. Some of these we know of (such as the Moxen and Black Lotus which are listed here at 0), while others I wasn’t aware of, like a number of cards from the original Theros block, like Doomwake Giant, Read the Bones and Bident of Thassa, which are all listed at 0.