Magic Arena's Matchmaking EXPOSED - How to Cheat the Algorithm Using Data

The inner secrets of Arena's matchmaking ranking (MMR) for Brawl (and other formats). Yes, Hell Queue is real. But we already knew that.
Links to Reddit Post / Spreadsheet:
UPDATED SPREADSHEET: docs.google.co...
OLDER SPREADSHEETS:
Commander Cards: docs.google.co...
Main Deck Cards: docs.google.co...
MMR Deck Calculator: / i_made_a_calculator_fo...
The Post That Started It: / everyday_is_momir
Follow me on Twitch ► / amazonian
Follow me on Twitter ► / col_amazonian
Subscribe on KZbin ► bit.ly/Amazonia...
Support the Channel ► ko-fi.com/amaz...
#historicbrawl #magicthegathering #MTGA

[@LegenVD]

Now that the data is out they should form a Brawl Advisory Group to adjust the numbers.

[@Amazonian]

I think your opinion would be very valuable in adjusting said numbers.

[@LegenVD]

@@Amazonian Your breakdown of the commanders and cards was completely in line with my experiences playing Brawl!

[@hadalabyss]

really hope they'll take advice like this

[@suNn.K.O]

Cat's out of the BAG.

[@KL-ie3dj]

Said advisory group should consider many things, such as creating a version of Brawl that doesn't include alchemy cards AT ALL ... rethinking the banned list (Paradox Engine, really?!?), why no commander staples like Blasphemous Act?, etc.

[@Dargaran]

Wrath of God is 45 Points. Day of Judgment is 18 Points. So if you are not too scared of the three regenerate creatures on arena, play Day of Judgement. :D

[@Amazonian]

It's so silly.

[@TheStanishStudios]

Wow! I was playing it for the art lmao

[@notreallyhere67]

I run both :D

[@McGrath435]

this algo is really sad... you get punished for evaluating cards well and choosing the strictly better option

[@Thickolas]

@@McGrath435 Knowing how cards work? That's some suspiciously "Hell-Queue" behavior there fella

[@yotuel9064]

I'm gonna resume the video in a couple sentences: Amy peeks behind the curtain. Amy realizes that all her conspiracy theories are true and lose her sanity.

[@metalhev3989]

Anyone who plays arena enough knows these "conspiracy theories" are all true. Even worse if you like to tank your mmr on purpose, then you'll start seeing some truly wild shit.

[@EionBlue]

@@metalhev3989 My favorite example is Crim/AsianAvenger playing a mill deck and literally getting matched with something like four 100+ card deck players back to back.

[@Jeffko78]

@@EionBlue I had that happen to my own blue/black mill deck (though 3 in a row, not 4) AND two decks with Gaea's Blessing.

[@dalexa]

@@EionBlueI’m playing a deck that exiles all copies of stuff the opponent has… suddenly I begun to play against people with 200 cards or people that has only one copy of their cards.

[@zlkanglwrth2776]

Its not conspiracy theory, it's just obvious. How are people this naive and ignorant???

[@Amazonian]

This was recorded live on my stream, I just turned on the camera and ranted for 40 minutes. Editing? What editing.

[@MH-pe8wj]

Love it

[@Azrichiel]

Editing is for Editors. We're here to play magic and complain about the matchmaking. On a real note, I'm a huge fan of stream of thought videos like this rather than overly edited "perfect" videos where nearly every single sentence cuts between the next because someone is doing retakes until they're blue in the face. To be clear I'm not saying that's you, just throwing words into the wind. In regards to the leaked data, I'm glad that this is public. While I won't go to insane lengths to "game" the system, I'll definitely be using this to help avoid some of the insane match ups that I get because I inadvertently add too many heavily weighted cards into say my Geralf "let's turn cheap fat crabs into thicc zomboys."

[@stephengrant3859]

Here for the rant, always love the content!

[@HTYM]

Editing, shmediting. Nobody cares because you are Amy the Awesome-zonian!

[@FluffySpikeM]

You're way more natural here, and your emotions look much more appealing. In more edited content it sometimes comes off as overproduced and thus inauthentic. Here it doesn't, it's charming.

[@kelzamgamingdesign]

No surprise that the Arena team is years behind on updating the matchmaking weight. They haven't even added battleground maps to the Brawl queue rotation since AFR, and still haven't fixed the direct challenge bugs 🤡🤡🤡

[@megamarine]

For some reaosn when challenging friends I can use only 2 decks out of 10 brawl decks i have

[@krimhorn]

@@megamarine They separate Brawl and Friendly Brawl because the ban list is different. I've had to switch regular Brawl decks over to Friendly to get them to be selectable in direct challenge and then swap it back afterward to play in the regular queue.

[@cardigansrule]

Small indie developer.....

[@megamarine]

@@krimhorn Okay. Why I can't play normal brawl but I am forced to play this friendly version for some reason? It's idiotic?

[@CorneliusQvist]

@@megamarineScroll down and edit your decks and immediately press Done on the ones that it says you can’t use and they pop back up to the top

[@alistairetheblu]

Arcane Signet, the card introduced in Brawl decks that had people immediately complaining is another auto-include, is 9 points. The Irencrag, which everyone plays as a "bad 2nd Arcane Signet" because even a bad Arcane Signet is still good, is 36. This list is bonkers.

[@bru4773]

This actually makes perfect sense if it's based on winrate. If everyone runs an Arcane Signet, then it's win rate is about 50%, since both the winner and loser had one. But if the people who run Irencrag as a second arcane signet are winning more often than those who choose not to, it's winrate would be higher than arcane signet, and would gain a higher weight, even if the card is technically worse.

[@ayylith]

​@@bru4773 I can confirm that not everyone runs Arcane Signet. Sometimes I'm only looking for cards in the colours of my commander and literally just forget to add any colourless cards. Several times I've been watching MTG games on KZbin and seen someone play Arcane Signet and gone "oh! wait! forgot to put that in my new deck!" & I have to pause the video and open Arena just to put Arcane Signet in my deck while I remember

[@apelsejati3261]

@@ayylithpoint still stand tho right? A lot more people will still play that card and it will win and lost many game tho it is not going to be exactly 50% it will hover around that number, also it is an uncommon so many people wold has access to it. Tho still cant deny the system is kinda broken

[@KK-pm7ud]

Speaking of Hell Queue, thank you for going onto Reddit so I don't have to.

[@Amazonian]

and double thank you to the people in the discord for going on reddit so I don't have to (they saw this before I did)

[@kronozord8346]

People there would say its fake... lol

[@yaytan2]

Some assessments: Commanders seem to have a grade between -1 and 5, then weighted by 360. There not being a zero does make it sound like it accidentally got changed to -1, and the weights were never expected to be negative. All 99 cards are also non-negative, so this is almost definitely the issue. The 99 cards have a median of 9, but almost all lands are weighted zero. Taking 36 lands and 63 non-lands, the weighted value of the 99 is often ~567. The median of commanders is 360, giving a total of ~927, so it seems they want an average weight of about 1000. The maximum deck has a weight of 6561, but we've seen that isn't playable. Some commanders seem to get 5, not because they are strong, but because the cards they go with are very low weighted. Take Fynn for example: a weak card, playing other weak cards, that need to be weighted together so that it doesn't only get matched against weak decks, that it would crush through synergy. Others possibly because whoever does the manual values is salty and doesn't want to play against it. The biggest thing that should change in my opinion, is that many cards should have much higher weight. Paradox Engine should have a weight of around 1k on its own. As a matter of fact, I think the strongest card in the 99 should be a stronger factor in the deck. As another commenter mentioned, some cards, like slivers, should possibly have negative values, with a final weight having a minimum of zero regardless. Here's an example formula: Deck Weight = Commander weight + Sum of 99 weights + (weight of highest card in 99) * 10.

[@laytonjr6601]

To do my dailies I tried Ragavan + 60 commons "deal damage to target creature", "enchanted/equipped/target creature has menace/trample/first strike (until end of turn)" under 3 mana and I'd say people forfeit way too often against it. Of course if I played good spells the hell queue would be deserved Edit: my 99 actually has a weight of 1,017 I didn't expect that

[@satibel]

as a dev, I bet they did something like changing a return value from 0 to -1 maybe they simplified the code and removed the other cards, and used the default value, which instead of being set to 0 returns -1

[@Kushiel13]

About the “weaker” commanders with high points, I doubt that the values were assigned with deckbuilding in mind. I would guess that metadata compiled the winrates of decks with commander A vs B, then assigned tiers based on relatively balanced win rates. That would explain why new commanders are present in every tier for 4-6 months until enough data is available to show which group the decks are more “fair” in.

[@hereisanoriginalname]

There are commanders with values of 9/18, they were left out of the person who collated Amy's data probably due to how odd it was that they weren't a multiple of 360. They're basically 0 for the purposes of matchmaking.

[@Senumunu]

you sound like a guy that is not being payed by WOTC. i can tell since what you say makes sense.

[@DerekS-kq3zh]

The only reason this is such a big deal is because WotC has been so willfully opaque about how matchmaking in the Brawl queue actually works. One of the things that keeps Commander fair is a "Rule Zero" about power levels and which kinds of decks should be played against others. The Brawl queue takes that away from the players entirely, and for the longest time we had no concrete way of knowing exactly how it was matching us. Now we do. And it's going to have dramatic affects on the way players construct their decks. Which is, I imagine, the entire reason WotC didn't want to reveal this information in the first place. But they've painted themselves into a corner here. Brawl deck construction is going to be heavily influenced by these numbers for a long time to come because this is the first and only insight we've ever had into how Rule Zero works in this game we call Brawl.

[@tonysmith9905]

Only the sweatiest of the sweats will worry about this. Most players will never even know this is a thing.

[@Rampleraft]

We should be allowed to select which commanders we do not want to play against, but at a limited rate. Maybe three. Because sometimes, you just get tired of playing the same few commanders. Not to mention that they disgustingly rise in popularity for a period of time when a content creator makes a video and everyone netdecks. erm hrmm CGB *cough* .

[@guico33]

That's assuming wotz does nothing about it. The simplest for them could simply be to ignore cards weight and simply match players based on mmr until they figure out a more robust system.

[@DerekS-kq3zh]

@@tonysmith9905 It affects every single game of Brawl you will ever play whether or not you're aware of it.

[@emoman7777]

​@@RampleraftI feel this on a whole other level. I don't know how many times I have nearly broke my pc playing against Rusko and op control decks while I am trying to have fun with dragons with miirym I get it trying to match equal power with equal power and opposite strategy, but it gets aggravating and boring real quick when they constantly line me up against hellqueue commanders

[@TheSmartCinema]

My General Kudro always seemed to get queued with much more powerful decks, and I never quite understood why that deck in particular suffered this much of it. Now with this list, I realize a bunch of random, replaceable humans in my deck had stupidly high values for no apparent reason, like Benalish Marshal and Dauntless Body Guard. Can't believe they have the same value as Ragavan lol

[@MaleusMaleficarum]

That is... I believe... because of the synergy between the Humans. A lot of the card values seem to reflect relevant strategies

[@TheSmartCinema]

@@MaleusMaleficarum Recruitment Officer, a random one drop, was at 45, just like Ragavan or Omnicience. I removed 10 underpowered humans with values of 45 and switched them for values of like 9, and I already could see the difference. Crazy shit.

[@MaleusMaleficarum]

@@TheSmartCinema yeah. I suspect, like many others, the values are at least partially based on high win rate standard decks, regardless of individual card value.

[@CharlotteMimic]

Same with me and Nadaar, Selfless Paladin :(

[@bodaciouschad]

Kaya the inexorable is an orzhov control commander that predates upon esika decks and is essentially running 30 removal spells and planning on ulting kaya for a win.

[@ImNotFine]

Hey, thats my Tajic deck! Pretty funny experiment. And yes, that reddit account is very old.

[@raulpacheco7488]

I knew it!

[@Amazonian]

Oh! Nice to see a friendly face out here shitposting

[@ImNotFine]

@@Amazonian Also the queue times for that deck are insanely long, usually around 2-3 minutes since the matchmaking has trouble finding an opponent.

[@Amazonian]

Lmaooo

[@doktorwyvern2883]

​@@ImNotFine Any wins with it?

[@cole3823]

someone put in -360 in some field to try and subtract 360 from them, but that made their weight -360 instead of subtracting 360

[@HJOTech]

holy shit you're so right

[@thomasmiller8289]

Fucking brilliant

[@Kaltes1]

Wait so some programmer at wizards never learned from the Nuclear option Gandhi?

[@Hamun002]

@@Kaltes1no, the Gandhi issue is a memory issue, the -2 causing an issue with a circular array instead of either throwing an error in C or requiring an int that is positive and negative. The error is hard to diagnose because everything isn’t visible, because the values are changed in memory. This would be not understanding your database, SQL or NoSQL in this case. The result is similar to making an assignment and never making sure when you set a = 200 that you didn’t type -200. I think this isn’t a mistake because it could be a way to offset other strong cars to avoid having a special hell queue above hell queue

[@metalhev3989]

Just keep in mind, all this only works if you keep your queue time low. After 45 seconds in queue, the matchmaker says "fuck it" and will match anything with anything for speed, same as any other matchmaker ever.

[@MisterAssasine]

I mean, you can stop queueing up or not?

[@metalhev3989]

@@MisterAssasine aye, you need to go up to ~40 seconds at max, then requeue. If you let it go longer you'll get hell queue 100% of the time because they're so overplayed compared to everything else.

[@Bub009]

Wrong! The queue makes my goblin deck wait a full minute every single time. All my other decks can find a player within 5 to 20 seconds.But I kid you not, the goblin deck has to wait for a full minute.Then I always get matched against white life gain.ALWAYS.It does not just give up.

[@symmetrybokte]

I'm absolutely sure some of the values assigned were done either algorithmically, or just with a broad stroke. Vault Skirge has a score of 45, same as the Praetors, presumably due to Phyrexian mana

[@symmetrybokte]

it's also interesting to note that, because the system used is such a simplistic 0-5 point system, there's zero reason not to run Mana Drain over Counterspell, for example. That's fascinating, I would've expected them to be ranked differently

[@vladsergref]

my guess is that it's there because of the all that glitters historic deck. Turn 2 Flying 4/4 Lifelink that grows with every artifact you play is (used to be?) a popular budget deck

[@symmetrybokte]

@@vladsergref I don't think a synergy from a format with a lower starting life total and decks with 40 fewer cards that can play full playsets of cards should at all influence Brawl matchmaking. There is no world in which Vault Skirge should get the same score as Mana Drain or Sheoldred, The Apocalypse, and a higher score than Paradox Engine

[@Pantjay]

@@symmetrybokte Oh they definitely shouldn't matter for brawl, but they do. I'm seeing a lot of constructed cards that don't see much play in Brawl. Monastery Swiftspear for example was recently banned in pauper and likely gets it's rating from there. Fanatical Firebrand iirc was part of some meta burn decks in standard a few years ago as well.

[@stevebeaver6106]

vault skirge is one of the key cratures to my auras deck I use for pauper. flying and lifelink...

[@samuelzachary8723]

this kind of matchmaking needs to be data driven. they have enough games played to live update the weights algorithmically.

[@cardigansrule]

Yes if they actually cared about the game and spent any amount of money on it which they clearly do not. I mean there are literally 6 year old bugs in Arena

[@Senumunu]

if they actually made that it would be crystal clear that cards like paradox engine, emergent ultimatum or rusko are taking a massive shit on the whole format. they know. they are obfuscating this. at the core they want you stuck in standard. every other format is a thorn in their sides.

[@taylon5200]

​@cardigansrule right, if they actually cared they could actually give us real commander through arena.

[@thomgizziz]

@@cardigansrule The devs they use are second rate trash devs and that is saying something because 80% of devs at good companies are incompetent.

[@Whitewingdevil]

To to create an algorithmically perfect deck with exactly 0 deck score and see what the bottom rank of decks feel like

[@cardigansrule]

My assumption is that if you try this you won't find any other matches very quickly and after 10 or 20 seconds it will start expanding the range, and pretty soon you'll be matched against a deck that's ranked 1,000... if you wait long enough.

[@thomasmiller8289]

@@cardigansrulenah, they'll be matched against the flood of other players trying this exact same thing - let the basement wars begin

[@xchronox0]

@@thomasmiller8289 scuffed brawl deck fight club

[@notreallyhere67]

Saheeli, Sublime Artificer weighs in at -360 so that means I can run Mana Drain, Counterspell, Ragavan, and other Izzet goodstuff and avoid the hellqueue :D

[@laytonjr6601]

That's almost certainly a bug, no cards should have a negative value (especially it the matchmaking breaks if the overall value is negative)

[@notreallyhere67]

@@laytonjr6601 I would agree with you that it's a bug but it's one amongst many cards that have a -360 weight.

[@laytonjr6601]

@@notreallyhere67 Commanders have a rating of -1, 1, 2, 3, 4 or 5 (times 360) and regular cards have a rating of 0, 1, 2, 3, 4 or 5 (times 9). (A card can have a different rating as commander and in the 99). The bug is that it should be a 0 in place of the -1

[@anergyboy2192]

You can play those cards without getting hell queued regardless lol

[@Thonieler]

I need more of unedited amy on espresso! This was amazonian to watch and listen to 👏👏

[@Amazonian]

come over to the stream, I'm always talking about stuff

[@juliamoor6544]

This explains why my kaito deck, that is mostly leftover standard cards from around theros matches against calix, sithis and etali all the time. Took all the fun of playing brawl, so i stoped.

[@RobertEnleed]

I had same problem on Historic and standard. For me as a free player, the only place I can play without too much A.I. interferences is Historic. Standard is a mana screw or flood nightmare. Mtga is rigged and fucked up beyond belief.

[@TheZouKal]

Thank you for defending my one and only dear Ayula, Queen among Bears, against that meanie First Sliver pile....

[@_Ve_98]

As a game dev I'm honestly baffled that there's a lower boundary at all for this. I think that the mistake has to be this lower bound existing at all. Specially when whoever designed it didn't make sure to not leak internal data to the client. They could have clamped the value to zero, thrown a generic "invalid deck weight" or anything that didn't just give the client the exact values. I personally think players deserve to know in a general way how the matchmaking works to avoid things that will ruin your deck, like the Paradox Engine thing, but this is too much. I hope this at least forces them to improve some of those weights that make no sense.

[@sylverfyre]

yknow what This makes me think the algorithm for matchmaking based on "deck score" uses a logarithmic function or root. Like, something you can't do with a negative number.

[@JasonIdkau-pw8hi]

Def should have clamped

[@cardigansrule]

This is definitely a big oopsie, like a "You done f***** up A-A-Ron!!" level situation. Like, why in the hell would you ever put that number in a log, and specifically call it "deck rating"?? If anything it should have just said "Error code -1 invalid deck". The level of incompetence is quite truly baffling, but then again this is Arena, where incompetence is the norm.

[@brakalanga7667]

Looool, the weights are assigned by hand!? I would have sworn they were either assigned by some kind of neural network, or at least by a hand made algorhithm that automatically self regulates over time. Judging by how old unplayable cards are still so high ranked, and considering there are at least 12000 different cards in that list, the only possible explaination is that those cards need to be adjusted manually, one by one, and OF COURSE the whole procedure is so complex and time cunsuming it ends up in the "to do list" every single time. This video is hilarious, cheff's kiss, 10/10, IGN not even too much water.

[@williamdrum9899]

That would be funny if it weren't so pathetic

[@thomasmiller8289]

Neural networks are hard, I can pay someone to copy down in excel for 4 hours and voila job done

[@Senumunu]

if WOTC can pay some chode minimum wage to do this they will. neural network ? these guys cant even match the technical state of "Magic Duels", a steam magic game that is 10+y old now.

[@Pnic1193]

Yeah, this is the thing. To keep a 12000 item list up to date with weekly or even monthly meta shifts by hand would require a coordinated effort from an entire team of people. The reality is that there's probably 1 employee who's job it is to do that and it's probably not even their only responsibility. Truly baffling how they thought this would be a good way to do matchmaking. What we end up with is a system that is about as arbitrary as just finding matches randomly but with the addition of being extremely frustrating for people who want to use good cards in unconventional ways.

[@thomgizziz]

Why go to all that work when you can do it once and not have to worry about it breaking? This is literally what the devs were thinking. They didn't want to deal with it. There are some terrible devs working on Arena. Third party open source programs like cocatrice and xmage puts in whole sets with the new rules in a weekend and the devs claim it takes them months.

[@diegomarinho1322]

There are 4 queues in Brawl: 1- Joke Queue 2- Fun Queue 3- Hell Queue 4- Rusko Queue

[@sidneypowelstock6812]

how do I get into the meme queue?

[@simic0racle157]

@@sidneypowelstock6812 pick a -360 value commander and put exactly 360 value in the main deck, resulting in a 0 score deck.

[@ryorai5804]

No Queue

[@anannoyedpanda]

May rusko players step on Legos everyday.

[@Jejfimwianfn]

This explains so much about why I'm not having fun with Brawl anymore. Depending on the deck I play I have basically no variety in opponents, it is always the same decks over and over (and apparently I'm also part of the problem because that means I'm playing high power decks?) and I just got tired of it. Seeing this stuff is just validating my impression that I'm not going to have fun with the game because even if this system is fair it just ends up being repetitive.

[@lethargyzz]

Yeah, I'm hoping this fiasco leads to some kind of overhaul, hopefully with more transparency.

[@reidyo5404]

The weighting system has always been in place though so that hasn’t been changed. There are more powerful cards in arena now so the brawl decks are more streamline.

[@SonnigesDeutschland]

I hated being paired against Rusko non-stop so much that I just built my own Rusko deck specialized on beating the mirror. That worked well but was not fun at all. I'm off Arena for a month now and don't really miss it. I miss playing Magic, but Arena is just so much worse than my live experience, that it's not fun anymore. The constant flood and mana screw in draft, the uneven matchmaking, the crashes, the economy... it's just not a fun place to play the game.

[@BrandonDenny-we1rw]

​@@SonnigesDeutschland i just came back to play arena for a bit and I almost swear the games are 100% rigged. My first Arena match I got the worst non pairing pulls and my opponent had a fluid perfectly syncable deck. I had no chance the first match, the second match played out so I had a comeback then a sound defeat and looking back at the cards a i drew and they drew I literally couldnt win. The game should be random but true random doesnt exist and when moneys involved im distrustful. Mtg online genuinely feels scripted.

[@Xorthis]

I agree partly, because think of paper magic. In a 4 player commander game, if you roll up with your Isshin, Two Heavens as One deck full of the most powerful cards and synergies in the game, and all your friends are playing their home brewed budget commanders, they're either going to ban your cracked deck, or they're going to gang up to ensure you're the first out as the biggest threat. You can't do that in Arena brawl. The system exists to attempt to emulate that 'Fuck that, my Greta, Sweettooth Scourge candy deck just got stomped by ANOTHER Calix enchantments, I refuse to play against your deck' (known as Rule Zero) that you can do in real life. What Amazonian got wrong is that the matchmaker broadens the search when it can't find a deck within the same power level. I sometimes play at unsociable hours and the matchmaker takes minutes trying to match some of my decks. They need to make it so it sometimes searches for rarely played decks, or decks outside of the range if your own, just now and then.

[@TheChiminguito]

This video should have more views. KZbin is not recommending it nor is it shown when searching for it until you get extremely specific

[@Amazonian]

It's recommending it at a far higher rate than my usual videos.

[@brennanclement8582]

I hadn't been recommended your channel before, this is the first one I've seen on my feed. Very clickable thumbnail, good job lol I really feel vindicated by this video. There have been several times I've tried to make meta-buster decks in other formats only to never see a single meta deck in like 10 games. Thanks for reporting!

[@ChrisOsberg]

It was recommended to me. Never seen a video, or thumbnail, from this channel before.

[@angelovillasanta2606]

Didn't expect a Jean-Paul Sartre quote

[@Amazonian]

hell is people that play decks I don't like

[@casparash5370]

@@Amazonian Hell is Alchemy people. Im looking at you grenzo-rusko you degenerates. Everyone should play Atris like me !

[@cardigansrule]

Hell is other people ....on Arena... who rope when they're losing ....with their mono red deck..... because they didn't play first

[@melind82]

Yeah, likely a programming bug. Every commander tier is 360 points apart except the tiers where they go from 360 to -360. I bet that -360 tier was supposed to add 0 to the score and negative scores weren't meant to be possible (hence the error).

[@Amazonian]

Same thing I'm speculating in the video. What a whoopsie

[@charanth182]

@melind82 considering how common an off by one error is especially when working with arrays in programming you gotta be right.

[@TheGprinziv]

Hell Queue is the reason why I stopped playing Historic Brawl. It's not that I want to stomp on people, but facing only the same 5 decks based on what I'm playing got so damned boring.

[@xddddddddddddd]

if you actually stuck on hell queue big chance you deserve playing vs these same decks

[@Darby_777]

@@xddddddddddddd this guy copies a busted deck, gets matched with busted decks and then complain lol

[@slaphappy-qb3jb]

​@@xdddddddddddddif you watch around 36 minutes on this video though, it disproves this assertion. The deck being shown is objectively dogshit, but it would rate as the best deck in the game according to the algorithm. That's a huge problem and why MM should be weighted off skill and win rate, NOT arbitrary numbers assigned to cards. What you're saying is a bad or even middling player needs to intentionally build shitty decks, or risk being stomped by the best of the best. If someone has to build a bad deck in order to have a fun and fair time, your system is fucked. If some person, playing MTG for 6 months decides that Nissa looks like a cool commander and builds a relatively normal deck, they're gonna have to play against good players who are gonna slap their shit. Not because they're good at the game, but because they had the audacity to think a card looked cool or fun and happened to think the wrong card was cool and fun. That's a shitty system.

[@egoalter1276]

There are literal literal papers written about why the chances of victory in mtg are mathematically uncalculable between decks, much less indivifual cards. Attempting to rank decknpower by static values manually assigned to cards therein is asinine.

[@bluegum6438]

​@@egoalter1276 irreducible complexity is part of the appeal of MTG, as you said there's no objective ranking of cards. There's a necessary involvement of "gentleman's rules" to keep the game balanced and fun when playing IRL that can't be applied to a digital format. I don't really think there's anything they can do to balance matchmaking in a computer game other than create increasingly specific formats, this card ranking thing is an attempt I guess but not very useful in practice

[@MTGGnome1]

Ive noticed that when i play jank like the momir roalesk, i experience longer queue times so this makes perfect sense. Great breakdown Amy!

[@sqrlmonger]

I strongly disagree that this info should not be public. In fact the information MUST be public. The problem is that smart and motivated people will ALWAYS find ways to access the information (as we see here) and they won't always make it public. This creates a scenario where a small number of people can gain a competitive advantage by gaming the system while their opponents are not able to. This is exactly why tournament winning decklists are published after large events for example. It creates an environment where players can react to other players so you end up with an adversarial process driven entirely by players. The bottom line is that if your matchmaking cannot survive when its inner workings are known publicly then it's simply a bad matchmaking system.

[@slaphappy-qb3jb]

The fact that some random new player with a very basic grasp of the game could look at a "hell queue" commander and say "hey thats neat" and then be butt fucked by amazing players playing hyper optimized decks is bullshit. Any MM not based on player skill and instead weighted on card value is inherently bullshit. It's basically saying a new player HAS to play intentionally shitty decks or just get their teeth kicked in constantly

[@guilhermegigeck5946]

The issue with this being public is the possibility of artificial engineered decks abusing the poorly rated cards to make a strong deck for a weak bracket and have one-sided matches against all but mirror. This is nothing like the tournament scenario you've described, they are not playing a good deck but they are playing the matchmaking score. Clash of Clans had this issue until they dumbed-down their matchmaking to be very superficial. Doing this here would just kill a myriad of decks.

[@slaphappy-qb3jb]

@@guilhermegigeck5946 their current mm already killed a myriad of decks. This deck weighting shit means if someone wants to try a given commander or card, if they pick the "wrong" one theyre now stuck in hell queue playing against the uber sweaty shit.

[@guilhermegigeck5946]

@@slaphappy-qb3jb this is another issue that rises from the devs trying to cheap out the matchmaking by looking at cards' expected power.

[@4everdex]

I disagree. As long as these values are changed by hand and don't represent the actual match and card values calculated by a really good algorithm, this data does not represent the actual strength of cards or decks. So people that invest a lot of time can make solid decks with the lowest score possible and crush everyone play their really shitty decks at that score range. They really need to adjust this method now and should have in the past. Manual adjustments Is just crazy in a game like this...

[@kleebgaming5209]

This is an excellent foundation to a matchmaking system that crunches numbers from matches and dynamically assigns weights for cards based on the results of those matches. For each card that's brawl-legal, there would be two lists of weights. "Synergy" weights that add/subtract weight based on the content of the other 99 cards in your deck, and "opposing" weights based on the contents of your opponents' decks. Add up all of these weights and you get the value of your deck, which models how good your deck is in a vacuum, but also how good it is vs. your opponent's 100. There are currently close to 10,000 cards legal in arena, so napkin math would say that this weight database would need to contain ~200 million numbers (10k ^2 * 2), which is feasible for a large dataset. ~2 gb if each one is an 8-bit number. Training would be interesting too. Maybe only adjust the weights of cards seen during the match. Would also make data-driven ban decisions more rigorous.

[@Amazonian]

It would be cool to have a living algorithm based on card pairs

[@Pnic1193]

They would need to graduate from manually assigning values for this. Perhaps this data becoming public and then getting clowned on like this can be the kick in the butt they need to actually do that, but considering they thought manually updating an ever-expanding list of all cards on arena was a good idea in the first place, my expectation is they'll change nothing and pretend this isn't a problem.

[@colinwilliams3459]

That would be an interesting way of doing it. But calculating "synergy" and "counter" would be difficult, and neural networks aren't some magic wand that could automatically find perfect weights. Also you could cut the size of the dataset in half by only storing connections. Card A will have the same synergy with card B as card B will have with card A so we don't need to store the card twice. You could use some balanced search tree as well to get fast lookup time

[@johnriley4971]

Commenting because I hope WOTC Sees this and moves Paradox into Hell Queue

[@jojodelacroix]

Or they could just ban Paradox Engine.

[@LucazNieddu]

@@jojodelacroix they said it's fine just a month ago or something... Are you expecting for them to care about a format that gives them zero profit lmao

[@user-cf3xm1cq4e]

They will make alchemy version. Like "Untap once each turn"

[@christopherjackson6345]

Captain sisay feels too mean to play because I rarely find myself in hell queue and just combo off on decks that are clearly “the best cards in my collection.deck”

[@dagamingsalmon7668]

I realized this was the case as there was a Steam review of MTG Arena that explained all of this, and how to take advantage of it. Once I gave it a try _it was like night and day._ My winrate skyrocketed as well since I could literally dodge (silver) bullets that the matchmaker used to throw at me all of the time.

[@bennettpalmer1741]

I think, if we're assuming you're correct about the values being the same for all formats, that paradox engine being so low actually makes a lot of sense from a certain line of logic. That card is extremely janky in all 60 card formats, if it were put in hell queue in those formats, people would probably complain. Having the "paradox engine commanders" be highly rated, but the engine itself lowly rated, means that people trying some jank in historic aren't punished for using a strong brawl card, and people trying to abuse the card in brawl still get matched in hell queue most of the time.

[@jmanwild87]

I know it would be a lot more work for whoever is working on the matchmaking algorithm but the cleanest way to solve this is to make the deck weight stuff format dependent

[@skykrow]

So give it a 27 then. Middle of the road. 9 doesn't make sense at all, its lower than a common

[@winkelfilms]

This video was so enlightening, I've been playing brawl for so long but I've never really been connected to the community, and you were just preaching to the choir the whole time!

[@winkelfilms]

It drives me crazy when people play the first sliver, because you're right, people don't play it right

[@reidslough7579]

The biggest thing about this is it’s proof that what people were saying deck weighted matchmaking was part of the game for a long time. Wizards gaslighted everyone saying it wasn’t. If they are willing to gaslight about a thing everyone knows exist in a casual queue what are they gaslighting about that actual effects ranked.

[@godlynewbie]

wizards saying it's all random is a trust me bro statement

[@victorph5680]

I play 99% games in casual bo1. Ranked only for gold or diamond. Opponent goes first three times in a row - EVERY DAY. Me go first three times in a row - once a week! Four or five times in a row - EASY for opponent, once a year or never for me 😂 I play since CBT and after the developers changed the matchmaking, my winrate was constantly dropping lower and lower from 53% to 43%. I've missed only one day since the release MTGA when I didn't make daily wins. I've been playing for 15 wins a day for many years. More than 30k games played, but it's getting harder and harder all the time

[@ducciotarno3900]

Wotc never denied weighted matchmaking in unranked modes, they confirmed it years ago

[@mibbzx1493]

Tbh i just want a commander format without alchemy cards. I usually just play arena to test out some decks/cards i dont have in paper or just to help them flow better or just try out. They should just make a jank game mode for them rusko or grenzo players cuz i just be trying to play regular stuff

[@ImperialVersian]

Absolutely agree. Magic has had many ups and downs throughout the years. But in my opinion, Alchemy is the worst thing by far.

[@abyssalAnalyst]

​@@ImperialVersian, alchemy is a nice concept, but the way they are implementing it just infuriates me. Especially in brawl. Queuing... Grenzo. Scoop. Queuing... Mythweaver PoG. That's poggers. Scoop. Queuing... Rusko. Scoop. Queuing... Grenzo. Ah, shit. Here we go again.

[@vermora356]

I disagree that we shouldn't be looking at this data, but only because it's been done incompetently, and we need to pressure Wotc to fix it. Card weights, and especially commander weights, need to be algorithmically driven for fair matchmaking.

[@bladorac]

How about do it like any competitive game and make mmr by player acc/player skills and not the cards they bought. this is a big fail from the get go by wizard and the person who came up with the card weight idea needs to be fired.

[@masonwhitehead7203]

⁠@@bladorac A player being locked into only being able to play decks of a certain power level isn’t desirable. Mmr is great if you’re only playing one deck, or if there’s a ladder which incentivizes playing optimal decks. It’s terrible in a format where you want to be able to use several different decks of varying power

[@reidyo5404]

@@bladoracthat is dumb unless it was a brawl ranked queue. The card weights is a fine idea if the system wasn’t implemented so horribly. Fun decks should face fun decks

[@bladorac]

@@masonwhitehead7203 Wrong. Simple reason is if your mmr goes down and you lose by changin your deck to a weaker deck is on you and you as a player with that choice should not be rewarded into easier matchups. Reason is this is easy abused by ppl who understand this type of mmr or coding. ive seen it on rank already been abused by like 2 months maybe 3 now. so I stay my case on make mmr skill only period. get gud dont make the cards carry the weight a pilot has to.

[@bladorac]

@@masonwhitehead7203 To be more straight into your point. mmr or card weight in casual non rank matchups in this case brawl is pointless useless and a waste of resurces. brawl is arena commander so is a casual format why the fuck is wizard adding mmr into it hahaha come on in those formats matchmaking should just be random/ This kind of weight as pulled in this video is been taken from standard formats meaning this is affecting all formats and what ruins your brawl game is ruining my standard rank format too. I say make mmr only pilot skills and no card weight at all. this is is been manipulated already. If i play and lose with my jank vs temur ramp is on me for playing jank. if i lose mmr by changing my 5c atraxa for a jank deck and lose is on me for choosing a different deck. Card mmr is dumb is only here cause some of players cant accept they suck and need to be helped. The negative is when a good pilot abuses it.

[@Xgya2000]

Came here from a random stream start from CGB, and MAN am I glad I came. This is so interesting. I wonder if there's a similar manipulation in the play/ranked queues, mostly because most streamers seem to experience their opposition changing quite drastically whenever they change decks.

[@1Heirborn]

Just remember the #1 rule of online gaming: Exploit early, exploit often.

[@Goldy01]

if that's your #1 rule that's kinda sad dude

[@ColdCryomancer]

@@Goldy01He didn’t say it’s his rule. It is the rule.

[@Pnic1193]

​@@Goldy01it's an exploit or get exploited world out there unfortunately

[@brasswirebrush]

Someone made a whoopsie adjusting the number on a few cards, and now the result is that the weight of every single card in the match-making algorithm has become public knowledge. Welcome to hell.

[@lebeaumuni6247]

This is a fun way to check if your deck is really is a level 6 or level 8 deck :)

[@wlycdgrLuckyThanQuick]

The world of Magic is so ridiculous and fun. Also it's good to see you giving shoutouts to the people who extracted/compiled the data.

[@Rancidtunip]

basically Wizards is forcing you to rule zero on their terms.

[@daveshouse8105]

to be honest, there shouldn't even be any algorithm except to match you in a game to your rank, i hate playing a 250 card deck that has an answer to EVERYTHING while tossing you smut for draws, gold playing diamond etc, its a pattern i see too damn much in this game. not exactly "fair" i RARELY get on mtga because the algorithm is so busted.. its bogus. thats not realistic what so ever.

[@colyrne]

As someone who recently started streaming games with 99 commons and slowly adding in better cards as i win, I DEFINITELY noticed this. Right up until i had about 7 uncommons and 9 rares, Ojer Axonil had a much higher winrate and much higher commander variety. That may have been the first time I'd ever even PLAYED against a Hamza, Guardian of Arashin 😂 I was VERY vindicated and also enamored with this information as it comes to light. I honestly personally kinda wish that this information was public on client in a way to inform players (new or old) that they would be matchmade into low, mid, or high difficulty matches.

[@coldfusion230]

Wait, why is Poq so low? I thought he'd be hell queue...

[@collinbeal]

He's not Hell Queue, which is why my jank decks have to slam up against him and Grenzo.

[@1AlasBabylon]

Because... they don't think so. This is manually adjusted - we have no idea why they think the way they do about individual cards.

[@KL-ie3dj]

it's new... she covered that... she said a deck has to exist for a while before it gets high on the rankings

[@w.p.7129]

If you've ever played this one, or a variation of it, with other land-based commanders, you know why. Other Hell Queue commanders will run them over like a road roller. I was wondering why Poq actually had good statistics for me initially, but after a while they started to drop rapidly until I stopped playing him, but now I know what was the reason.

[@1AlasBabylon]

​@@KL-ie3dj Poq has been on the client since last December. It's been almost seven months at this point.

[@FluffySpikeM]

Also, I think at least some bottom-ranked commanders are there, because they're from precon decks. Sarkhan Dragonsoul, Oko the Trickster, Rin and Seri etc. It makes sense WotC would want somebody who bought a precon commander deck, registered on Arena and redeemed it via a code to have a good time playing it. I.e. win a little more. Nothing wrong with that, as those are usually much weaker than is standard. I think there is a positive side to this data being public. I know I'd be much more willing to play weak cards I like on flavor alone now.

[@Pnic1193]

Just consult the list first to make sure the card you think is weak doesn't have an arbitrarily high value for no reason

[@roshandala]

20:04 the pain in her voice when she said 'why' was heartbreaking 💔

[@ExaDat]

I hope this information coming out forces wizards to fix their match making. Now I get that my shitty commander decks gets matched against strong commanders because I have other strong cards in the deck, but I still think its stupid that my delina deck gets matched against atraxa or grenzo

[@7h3Y3ll0wD4r7]

I'll have to finish watching this another time, but I just have to say my god you are a FANTASTIC speaker. I don't know if you've done public speaking or education, but you are so articulate, your cadence is engaging and keeps the listener focused and attentive, and I haven't played Arena in years but I'm 100% fascinated in this discovery thanks in large part to your fantastic delivery of this information.

[@Amazonian]

I think streaming full time probably counts as public speaking

[@STS-qi1qy]

New commanders not getting ranked right away makes sense as to why my relatively mid-tier deck (I say with all love of it) was facing nothing but Voja for awhile and has now been replaced with facing nothing but Bonny Paul and Alchemy Grenzo. Was half expecting that I was misjudging how powerful some cards were on their own outside of the build but the calculator backs up my assessment of it.

[@Spotifist]

38:05 "...and a strong collection" is that really part of MMR? That's wild to me. Edit: Forgot to say thanks for the video, very interesting find and analysis, thanks!

[@Amazonian]

It's all speculation, but it seems to be that way.

[@williamdrum9899]

Incentivizing making new accounts en masse is the bane of any online game

[@cardigansrule]

As far as I know and I've been playing Arena since day one of open Beta they have publicly admitted that decks are given a power rating. They haven't said exactly how they're rated, and they specifically said they're not going to tell us how they're rated, otherwise we'd game the system, and of course we would... As far as anything else it's all speculation.

[@tinu7551]

Its not just power ranking though. I seem to get the same opponents whether im playing a méta or a même deck. I remember a few months After Arena launch, I already played meta decks from farming my collection in. limited. I ranked to platinum or something only playing méta deck. I drop by a friends flat, he 'started playing like the week before when I helped him pilot an draft to 7 wins. He climbed to plat too during that week. I was dumbfounded, I couldnt believe hé could win against méta deck with what hézhad build up. He accepts to launch a couple games just to show me and low and behold I was right, he was playing against similar decks to his, way power in power than thé méta.

[@abyssalAnalyst]

​@@tinu7551, wait, I thought platinum isn't for meta decks. For mee it seems like mythic plays competitive, diamond is overall solid, but not meta, and bronze-platinum is more or less the same.

[@raulpacheco7488]

10:05 Well to be fair,Selvala usually ranks S tier in EDH lists. Difficult to explain why, unless you play her

[@Amazonian]

yep, she's much easier to generate infinite mana with

[@BL-cc2nw]

Without this public knowledge, and your help, the system would continue to be broken. the public release of algorithms for stuff i pay for is essential to keeping companies honest

[@godlynewbie]

companies actually be honest 🤣🤣🤣. you darn well know it takes drama to make these companies to change things to be even a bit better

[@mandielyn0416]

This is amazing. Now I just need to know why I have an 83% standard on the draw rate for May, and a 71% on the draw rate for April. Untapped is great for clarifying suspicions for anyone that is new to the game.

[@paulcrisafi4415]

Same, *OPPONENT GOES FIRST*

[@JustinNovack]

What the fix should be: WotC adding columns to the "Weight" table per format, and letting win rate adjust the numbers algorithmically. What the fix will be: Remove DeckWeight from the log line and ignore the problem.

[@KL-ie3dj]

THANKS for this! I ONLY play Brawl on Arena. Aside from these rankings, I would love to hear a discussion about WHY (oh why) A CARD LIKE PARADOX ENGINE is even legal in Brawl when it's banned EDH? WHY CAN'T I PLAY BRAWL without f-ing Alchemy cards?!?!

[@mrJerro2112]

1. because in EDH it leads to 10 minutes turns without wincon forcing 3 people to sit and watch. While in Historic Brawl it is the wincon and you can just concede, same as you can concede when Ratadrabik went sacrificing Nazguls or Vito damaged you with Exquisite Blood on the board. Or maybe you want to also ban Ratadrabik and Vito? 2. for the same reason I cannot play most of the Commander cards or Fallout cards. Because the format is Historic Brawl, not Commander.

[@Oxygen1004]

​@@mrJerro2112I personally just pull up youtube on my 2nd monitor and make them play it out if they engine combo me. You can concede irl too once they show a loop.

[@xavier9319]

We really messed up when we started incorporating algorithms in match making instead of true random play. It really sucks and why I like in person play way more

[@FlawlessRythym]

Good to know my suspicions were correct. As a mostly free-to-play player, my decks are vastly different. Some of them are completely busted and not missing any pieces, and others are just an empty shell just waiting (ever so slowly) for more wildcards or lucky packs. My busted ones will take over a minute to queue up and only plays a very finite pool of enemy commanders. Those ones for me being Adeline, Ragavan, Chulane, and OG Jodah. Then I have decks that never have a problem queueing, and will often play against weaker/U commanders. Those for me being Iymrith, Grumgully, Drana, and Pippin. This feels like a crime. Just open the queue! Lol.

[@CourtesyCurtis]

Brawl Stars new difficulty system -360: Meh 360: You got it Buddy 720: Medium 1080: Oooh a Challenge 1440: Your On Your Own 1880: There’s no turning Back

[@18nightc322]

Oh no the shenanigans that will be happening here.

[@notreallyhere67]

Honestly, it's going to be too much of a hassle to gather the data when making a deck. I think 99% of people will just build decks as normal because who really wants to dig through spreadsheets?

[@bastius5930]

I love a good spreadsheet with thousands of rows.

[@dyne313]

@@bastius5930 But hide the children, because that kind of thing is not safe for children's eyes. It's a joke from a Victoria 3 youtuber. That game has a lot of spreadsheets in it and that's a recurring joke.

[@yotuel9064]

@@notreallyhere67 I dissent. Most of the arena fauna are gonna netdeck and the builders are the ones that are gonna check the spreadsheet. In less than a month most of arena brawl is gonna be following the "tricks" of the spreadsheet. And given Amy's insight its gonna need months till they do reevaluations given the new data (and it seems that there isn't lots of reevaluations given the zenith card).

[@raspernor11]

I sure feel bad for all the players that just happened to fall in love with a "hell" commander, but don't have the playskill or deck design skills to match. Really shows WotC's incompetence.

[@thomasmiller8289]

new meta is: zero weight decks will become optimized to win in that bracket, so what score above zero but below X gets you the ideal matches

[@thrillhouse4151]

I think it would be cool if you had more options for the unranked modes, stuff like rematches, weird home brew rules you can toggle like starting life totals and mulligan variations. I know that games are hard to program and even the simplest idea can prove to be too much work, but I feel like there’s so much *untapped* potential with Arena. It also melts my phone and computer despite being a card game for some reason.

[@thomaskellner8102]

to be honest. it would not be too difficult, if their code is properly modulized, but if you think about the amount of effort, around zero, they put into improving the player experience, i think there is no chance. i had stopped playing MTGA for around 4 years and after i came back everything still looked the same. They do not put alot of work into the client... which is very unfortunate.

[@MonochromaticPrism]

The main reason this isn’t done is that it splits the player base. The best you can hope for is a single mode with a rule set that rotates once a month.

[@lancesmith8298]

I’m just a Commander player off in my little corner, trying to make my proxied Tayam deck even better with the MH3 releases, and this video hits me like a brick from the blue. This is 40 minutes of me wildly oscillating between “Y’all live like this?” and “This is the torment that the average consumer of Magic the Gathering in 2024 has to endure”. Brawl already sounds like a truly miserable experience as is, but having some poor intern “help” steer the format with spaghetti code is a nightmare from top to bottom, an abomination only Ashiok could love. We deserve better.

[@Senumunu]

as long as this format competes with standard it will always be a neglected child. they want you in standard. everything else is there for the internal metrics.

[@caldazar7345]

That explain so much, I remember fighting so many calix decks which stopped suddenly, it was because the matchmaking wasn't uptaded yet

[@wintersoldier909]

Arena is trash. The shuffler isn't a true random, hand smoothing actively screws people who who build really powerful decks, and whatever deck you build if you're a free player the game will actively matchmake you against Decks that are likely to beat yours.

[@thekrakenwithin6048]

This is extremely interesting. I'm sure I could do something with this information if I had the brainpower, but I think instead I'll just keep building things as normal and hope that people trying to game the system doesn't impact my enjoyment too much.

[@AutkastKain]

The truth come out.

[@Hankathan]

I like to believe Ashiok, Nightmare Muse is up there because of me. A few years back I was playing my 80% winrate Ashiok surveil list at least 100 times a day and I guess someone at Wizards remembers.

[@abyssalAnalyst]

Whoah! This is your the most viewed video now! Congratulations, nice work!

[@gamingfossil6199]

Wow, that´s sad news... I loved Brawl, but for the last 5-6 months there were just too many broken ( mostly Alchemy ) cards to match against. And now it´s getting even worse ? Think I need a coffee,too. I´ll hope they will fix / change this ASAP. A big oopsie you say ? LOL. Sometimes WotC is just rediculous. This is sooo emberassing and unprofessional.

[@M3rett0]

This feature has existed for 3 years. It's your confirmation bias if you think Brawl suddenly got worse in last months.

[@CarlosEduardo-vy9qs]

Dovin 3 mana azorious Planeswalker for the win. Excellent content @Amazonian, I just got a 15 win streak with my azorious deck. No waiting lines of one minute trying to find matches in brawl anymore. It take less than 3 seconds to find a match. The matchmaking is one of the most hateful things on Arena, that is why MOL is so much better. No hand softner, no matchmaking, just pure random, like Magic is supposed to be.

[@zachariahalexander6371]

Jesus this explains a lot with my current match making with Atraxa. I was going up against decks I didn’t know would be in my queue.

[@paulcrisafi4415]

Same, playing mono black and getting matched against Turbo Jhoda Turns feels pretty bad. Especially losing turn 5-6 on four lands.

[@JoeyMossBBG]

This is a great breakdown of all the matchmaking nonsense. Glad I stumbled on this video. Subd

[@bearsaremonkeys]

it's baffling to me that they would have manual weights instead of just using an actual elo system

[@fastydave]

Brawl is unranked so it makes total sense

[@hereisanoriginalname]

You seem to have omitted Commanders that have a rating that's a multiple of 9 instead of a multiple of 360; those are basically 0 for the purposes of the algorithm. (I initially thought those were glitches too, but no, according to the spreadsheet provider). Alrund, Gishath, Cosima, Eriette sit there, it's probably the cutoff for "commanders that aren't bottom-tier awful."

[@Amazonian]

I don't seem to have those on the list I saw, but I also didn't collate the data.

[@Joseph125]

@@Amazonian There are several commanders that aren't on the commander weights spreadsheet. In that event they just use the matchmaker value of the card as if it were maindeck, typically 9 because they're weaker cards.

[@OmegaFerretMusic]

That explains why my Slimefoot and Squee deck has such a high win rate. The commander has a pretty low score despite being incredibly efficient.

[@thomasohara5926]

God, I played the shit out of that Zenith Flare deck back in the day. It's fun to see the card's Matchmaking score so high as a weird artifact of a very particular time.

[@Amazonian]

Such a fun deck

[@gn0s1s]

Judging by most peoples brawl experience, the deckweight, whatever the criteria, is awful. Half of brawl where the shuffler doesnt auto beat you, is against opponents with no syncing between planes of interaction, which end up in tedious boring games if you stick it out, or total blowouts. This is both on the wins and losses sides. I would say 10 to 15% max are satifying games, that in IRL people would sit down and play each other with. And really there isn't much excuse for it, they have every interaction from each deck mapped, win rates, hands and cards played when scoops happen. If your deck is a creature deck, that gets going turn 5/6, maybe the board wipe / counter deck with all of brawls fast mana, isn't good match. Or you have a removal light deck mono green , with only creature based interaction, cool lets match you against an interaction based indestructible commander. It's surprising that this exists, because honestly with how bad the matchmaking is, it felt like arbitrary conditions that they contracted out to fiver.

[@ABZANTurtle]

been playing 250+ Decks for about 3 years and noticed the riggedness; 1. some decks would be ok-ish, except the constant amount of "cards that intuitively counter your decks strategy" being in their top 7 cards almost every game. Playing ANY artifact deck = ALL anti artifact cards, even the less than 10 different "destroy/exile all artifacts cards" that would be useless against anything else GUARANTEED to be seen every 10 games without fail. 2. after slimming down to 200 cards ... life got easier but got f***ed up every time I would play ranked for about 2-4 weeks of extra "unfair" matches, making the "free" booster worthless to me to even risk doing that ever again 3. some color pairs are literally forced to play only select strategies because if they do not = about less than 33% HARD LOCKED win rate, if they try

[@yes.thatjanedoe]

I used this data sheet to create a python script that takes an arena list (copy from clipboard) and outputs the Historic Brawl and Standard Brawl score just as a fun coding exercise

[@araylian6681]

After checking the MMR score of some brawl decks I would like to apologize to all the people who have to face Crucias or Laelia with your 300ish point decks. I won't apologize for Imoti though.

[@zotha]

How about just have a brawl ranked queue instead of screwing around with an arbitrary rating of cards and let people setting at the ranking where most other people are playing decks appropriate.

[@bungalowbill521]

“This information never should’ve been public.” This is backwards, every MTG mechanic should be transparent even on digital platforms.

[@lucasenraraujo]

Immediately thought of you when I saw those posts on the subreddit hahaha. And as much as I agree that it's better for the health of the game that matchmaking algorithms remain hidden, the Arena team could actually afford to be more transparent with the weighting stuff. Heck, if anything it would be interesting to let the community help with it (some voting?). And to some extent people trying to trick the matchmaker might lead to a better matchmaker in the future.

[@deant01]

I made my Kaya deck specialy to beat everithing in helkqueu. It realy has a good winrate agaiinst mist of the 1800 score commanders. I guess thats why she is next to them - she is basicly a remove nonland in the comzone.

[@SnackCakes]

I originally quit arena cause they manipulate your opening hand to increase the chances of drawing mana, so a deck with 10 land was crushing the meta. A loading screen even stated this was done on purpose to make matches more fair. When I made a post people didn't believe me. So I'm not surprised that they still refuse to make this actually fair. Wizards hates that their card game often has way too much luck involved.

[@chammy2812]

If new commanders are not listed/placed at a default value for a period, then it does suggest that these scores are generate algorithmically and not just some number a random person put in. They probably see how a commander performs over that first 4-6 month timeframe and change its value to put it in the most appropriate range. That is to say, where the deck scores approximately 50/50 against other decks of similar score.

[@oinkleberry]

Ah. So that's why my budget Fynn deck got Prismatic Bridge players 4 times in a row. Okay.

[@Amazonian]

Yep, Fynn gets wins. So his matchmaking is high

[@xddddddddddddd]

ngl deserved on you for playing fynn

[@Zeckenschwarm]

I'd honestly prefer it if there was no matchmaker at all. Just randomly match up everyone against everyone.

[@bladorac]

Or do like most competitive games smart developer teams do, only do players mmr.

[@mrJerro2112]

@@bladorac then active and decent player will never win with jank / roleplay / fun deck and everyone who plays 6+ months will be forced to play Rusko or Ragavan in order to win.

[@bladorac]

@@mrJerro2112 Wrong. jank/ fun deck are that jank and fun if you want to win with them it will come just dont expect to be 75% win rate come on dont be dilusional. MMR should be only with rank formats but wotc fucked up by putting it everywhere and by using the cards as mmr not the pilot win rate. This post has half a good point mmr should not be on casual format, in this case brawl. random matchmaking on that is probably the safest most healthy way to do it.

[@anannoyedpanda]

​@@mrJerro2112excuse you sir. You forgot Voja in your comment.

[@stephenb5085]

The fact nobody thinks these points are in rank baffles me. If you don't think ranked is rigged, you haven't been following along.

[@doublethick5752]

This explains why calamity is getting hosed in MM Your deck is filled with really high impact creatures that likely have high scores associated But it doesnt really understand that you dont really have much draw and very minor ramp and you just fold to mild removal or counter spells

[@simic0racle157]

built a 0 score stickfinges combo deck with this and matchmaking takes noticeably longer...

[@megapussi]

Yeah thats cuz u have to intentionally go out of ur way to get a 0 score deck, nobody out there is running one other than the other people intentionally making one. If u want to game the system to play against bad decks ur better off scoring in smth like the 400s

[@raneemfathy2581]

You think they’re gonna send goons after the people who made these lists? 💀

[@carlhannah1884]

They might as well do the wargame thing and just assign known point values to everything and have a special queue with rotating rules/point cap.

[@jordanledoux197]

100% guarantee you as a programmer who has worked on games that in NO reality are most of these cards manually scored. Some of them are probably manually adjusted after the fact, but the scores are almost certainly generated based on the card effects, cost, etc. and then clamped into buckets (the 45 bucket, the 9 bucket, etc.) It is also not guaranteed that these scores are static. What we see is the mined value, which is the final value after adjustments (if there are any). Thus the deck makeup of "known weights" could easily be influencing these numbers.

[@hitlord]

Well, I'm glad we finally got the backstage intel on how this works. We all knew this was in place for ages, and knowing this at least gives us a better fighting chance to avoid those ultra boring decks in the upper echelon of hell queue. Knowing the weights now though, kinda explains to me why I had so much when I played Basri Ket as a commander (and even then, I dodged every single Teferi or 5 color commander as I couldn't be bothered to play against people that wouldn't actually be "playing"). The top of the list is restricted to 50 commanders, which mostly amount to 10 decks with one card or another changed which generates those doubles/triples in the list. As you move to 1K weight and below, you have over 200 commanders already. Add in the negative ones that managed to get higher through the other 99, and you have a shit ton of weird match ups.

[@brylythhighlights4335]

This would be a really interesting mode where the information is (intentionally) public and lets you play objectively weaker cards that are 'lower budget'