Doing the overnight shift feedings for the newborn means I won't get to see the results till around noon PST. Looking forward to never having to play against KCI, brewing terrible decks with Stoneforge after selling my extra copies for exorbitant markup, and complaining that Twin is still unjustly banned. See you all after the announcement!
Turns out body doesn't want to sleep.
Anyway, what a depressingly underwhelming update. Lots of info on KCI, and absolutely nothing on the status of the format or any indication of unbans.
Emphasis added. A lack of unbans is depressing, but the update has critical information about the status of the format. See just a selection of quotes below:
"Given that Modern has looked healthy and diverse at many levels of play outside of Ironworks's dominance at the Grand Prix level"
(R&D believes that Modern is currently healthy and diverse, except for GP, at multiple levels. This likely includes MTGO, GP, and smaller venues)
"As a category, we think these are generally healthy provided they appear in small doses and have reasonable win rates."
(Opal and Stirrings decks are not overperforming by any metric Wizards bases B&R decisions on. This notably includes metagame share and MWP)
"As Modern stands, our metagame data does not indicate a need to impact the other Ancient Stirrings or Mox Opal decks."
(Stirrings and Opal decks are currently at acceptable levels in Modern)
In summary, Wizards said that Modern is healthy and where they want it to be, except for KCI. Plus KCI and Stirrings are being monitored. That's quite a bit of information when all Wizards needed to do was just post the rationale for banning KCI itself.
So next B&R is March 11th, this is followed by 3 gps in the same month for modern. Not saying it means anything but just throwing it out there since an unban is a way to upset the meta without killing peoples decks outright
Because of those 3 GPs and the fact that today there were changes, I could certainly bet that the next announcement is going to be "No changes" unless some absurd broken deck surges like Eldrazi Winter did
I 100% agree. In fact, I would say that basically no amount of perceived brokenness would result in changes occurring in that update. Remember that in Eldrazi Winter we had a triple GP March and Wizards still didn't ban Eye/Temple/Eldrazi-card going into that month. This was a slightly different scenario because the planned B&R update was already set for April, not the middle of the GP stretch, but I think the precedent still holds. I would expect Wizards to look at format health after March and then make unbans (and bans, if anything really goes crazy) at that time.
I am also very happy that this B&R list vindicates the data-driven approach I and others used over the past year. Not only was this ultimately used as part of the B&R rationale, but so was its high T8 percentage relative to its overall GP prevalence ("While the primary reasons for banning a card from the Ironworks deck are its raw win rate and high GP Top 8 conversion rate..."). These were both publicly available information sources that we used to successfully predict a ban at the deck, and show that our MWP and GP T8 methodology is pretty sound for predicting format health; these methods also suggested Opal and Stirrings were safe.
I don’t really get how twin is “unfair” sure it combos off but never before t4. Also decks are currently already starting to run rending volley to deal with Titi so it’s starting to already have hate cards built into some decks (granted not common currently).
I don't think the argument for Twin is that it's a fair deck. At least, I hope not. It's definitely not fair by any definition I've ever seen or tried to craft. The argument is that Twin, although an unfair deck, decreases overall format speed because it disincentivizes players from running all-in T2, T3, or T4 decks. These fast decks tend to lose to Bolt, Remand, and Exarch/Mite tapdown followed by a lethal T4 combo. Twin also slows down games with the tempo gain of threatening the combo. As such, players might be more likely to run a grindier, more interactive deck that doesn't flat out lose to Twin. Or play their fast deck slower (e.g. Spellskite in Infect or holding up removal). This is the old 2015 "Twin test." The Twin proponents will argue that this test/effect increases the format's fundamental turn, generating overall more interactive games of Modern.
One problem with this argument is that the format isn't overall significantly faster now than in 2015. Only about 2% more games end before T4 now than in 2015, and we have no clue what share of those 2% of games are due to Twin's absence vs. new cards. It's hard to sell a Twin unban on the stake of 2% of games.
A more defensible argument is that Twin has an "appropriate power level" for the current Modern. That was the BB unban justification, and it's a very good standard for SFM. Of course, especially with Twin, the problem there is that it's hard to prove appropriateness without thousands of test games.
It was pretty much 100% Scrap Trawler and the slight innovations of Nass to include stuff like buried ruin, inventors fair and EE. Thats why I Think Scrap Trawler should be banned because it's the real reason the deck is busted. The reason why it goes infinite and the reason why it's resilient to all non persistent grave hate.
Old eggs could exist without both KCI and Scrap Trawler but is no where as good
I also think Trawler is the ban target. KCI itself is another option, but something like Wellspring is too cute and indirect.
That said, this isn't necessarily the banlist update where all these changes happen. The February update would also be possible. But I do wager we'll see some changes before March.
So tired of those posts getting made to Reddit and/or shared on social media. I disproved the last one about Ensnaring Bridge and it looks like this one is just as fraudulent. Posts like that just increase ban mania, uncertainty during an already uncertain period, distrust, and misinformation. They are terrible for the community and I wish Reddit was more responsible in shutting down or editing such information once it has clearly been proven false.
I'll also add that information consumers have a responsibility to be more discriminating. It takes about 15 seconds of running this through any free photo forensic software to identify glaring signs of a faked screenshot. It literally says it's a photoshopped image in the metadata. I don't know about you, but I don't run my screenshots through photoshop before posting them.
As I posted in that thread, I hope someone with more time/bandwidth than I can pool their GP Oakland sample with my sample to get N even larger. I give all the matchup and sample Ns plus percentages, so this should be methodically easy, albeit time consuming.
Traditionally, tempo decks deploy threats and then protect the threats and/or disrupt the opponent while the threat wins. Midrange decks tend to disrupt first and then deploy threats on a safe board, although some will still seek to protect a deployed threat later. See the difference between traditional Delver builds (T1 Delver/Swiftspear, hold up protection/removal) and old school BGx (T1 DS into T2 removal/Bob into T3 Lily into a clock). GDS can certainly role-shift in matchups, but it's more fundamentally a midrange deck than a tempo deck.
So traditionally gds is tempo. That’s why it plays stubborn denial to protect its big threat it played on t2. Gds isn’t midrange imo.
The line of T2 GDS/Angler/Tasigur into Denial is way less common than the line of T1 removal/discard into T2 removal/discard/cantrip into T3 threat. You just need to look at the decklists to verify this, with few top-placing lists running more than 3 Stubs. Instead, they run 6 discard and 6+ removal, plus all the cantrips. Even without watching a single game, we can tell that the odds of drawing deployable threat into Stubs is way lower than the other play patterns.
That doesn’t change the fact that DS plays a threat and then protects it. They don’t look to grind out games they look to end them. They don’t often need the t2 stub when they go disruption, disruption, and then threat. They look to fade 1 turn then protect the queen. While smashing face (sometimes killing in 1 turn with tbr).
It's possible you responded before my edit, in which case I encourage you to read Reid Duke's Thoughtseize and Fatal Push series on CF. He defines GDS as "black midrange," a definition I agree with based on both observed play patterns, deck composition, and historical classifications. Also, to put it simply, it's frikkin Reid Duke master of Midrange. If he calls it Midrange, it's Midrange.
I was a little more open to this archetype debate until I remembered this article and Duke's position. I'm just unwilling to argue with his classification given that man's knowledge of Magic, Modern, midrange, and these decks. Coupled with all the other elements I and others posted about, it's pretty clearly the proper classification.
Traditionally, tempo decks deploy threats and then protect the threats and/or disrupt the opponent while the threat wins. Midrange decks tend to disrupt first and then deploy threats on a safe board, although some will still seek to protect a deployed threat later. See the difference between traditional Delver builds (T1 Delver/Swiftspear, hold up protection/removal) and old school BGx (T1 DS into T2 removal/Bob into T3 Lily into a clock). GDS can certainly role-shift in matchups, but it's more fundamentally a midrange deck than a tempo deck.
So traditionally gds is tempo. That’s why it plays stubborn denial to protect its big threat it played on t2. Gds isn’t midrange imo.
The line of T2 GDS/Angler/Tasigur into Denial is way less common than the line of T1 removal/discard into T2 removal/discard/cantrip into T3 threat. You just need to look at the decklists to verify this, with few top-placing lists running more than 3 Stubs. Instead, they run 6 discard and 6+ removal, plus all the cantrips. Even without watching a single game, we can tell that the odds of drawing deployable threat into Stubs is way lower than the other play patterns.
Traditionally, tempo decks deploy threats and then protect the threats and/or disrupt the opponent while the threat wins. Midrange decks tend to disrupt first and then deploy threats on a safe board, although some will still seek to protect a deployed threat later. See the difference between traditional Delver builds (T1 Delver/Swiftspear, hold up protection/removal) and old school BGx (T1 DS into T2 removal/Bob into T3 Lily into a clock). GDS can certainly role-shift in matchups, but it's more fundamentally a midrange deck than a tempo deck.
Reddit thread is up. I made some small edits to the post but didn't make a lot of changes overall from yesterday. Thanks to everyone who looked it over! I assumed, hopefully correctly, that a lack of major feedback here is because there weren't too many significant issues.
Re: SCG Day 2
Phoenix is a very strong deck. It runs a crap ton of cantrips and has extreme velocity, making it the exact kind of turbo-xerox consistency machine that pros gravitate towards. It has a relatively low floor to success and a very high ceiling. I think I talked about this before, but again, I expect these qualities will drive more and more top players towards this deck in the absence of other options. It doesn't hurt that the deck is strong against some other strategies in the field.
It's a 2018 summary of all the compiled GP/SCG data we have about matchups and MWP. I'll publish to Reddit on Monday after incorporating feedback. I request that people keep this under wraps until it's finalized. Please note the limitations section and don't just parrot about how N isn't 10,000. Nothing drives me crazier.
Here are the summarized MWPs for top decks with N>300 from the past year with sample size and 95% confidence interval for the MWP:
I'm not going to try and translate the matchup win% matrix here because there's too much going on there. Check out the post for details about those matchups.
Yeah I remember, and I think a lot of people around here at least hung on your website and words at the time, its why it was such a body blow when the hammer came down.
One of the issues is that I placed too much emphasis on overall metagame share to predict bans, when GP/PT T8s were a better predictor. I've adjusted since then, which is one reason I and others in this thread have successfully and confidently defended a "No changes/No bans" standpoint when it comes to metagame/format diversity violators since Eldrazi Winter. This, despite crap like people calling for bans on Humans, GDS, Gx Tron, and other decks that appeared to be the "best" but were unequivocally not violating the best ban indicators we had. T4 rule violators are harder to assess because it requires access to data we don't have. Other ban criteria, e.g. battle of sideboards, are also more opaque.
I didn't say that every ban was related to this. First of all, in the initial banlist they basically banned all the best decks from Extended. After the first Modern Pro Tour, they banned a bunch of cards from all the best performing decks. About a year later, the two best decks were Jund and Storm, so BBE and Seething Song got banned. A year after that, Jund was still the best deck, so DRS got banned. A year after that, putting aside the obviously broken Treasure Cruise, the best deck was Pod, so that got banned. The next year, Twin was the best deck, so it was banned. I'm not saying that all of these bans weren't justified, but WotC was obviously actively pruning the power of the format to keep it more modest. It wasn't really until Pod when people were grumbling about this, and until Twin when there was a sizeable backlash. The backlash of the Twin ban caused them to back off being so aggressive with their bans, and I would say that everything that's been banned since has been justified.
This is a more interesting discussion point. For reference, here's the original quote that Wraith appears to be clarifying:
Let's not kid ourselves here, Twin got banned because WotC was in the habit of banning the best deck every year to keep Modern's power level in check, and Twin just happened to be the best deck of 2015. The community outrage over the Twin banning is precisely why WotC has let up on their bans. I have no doubt that GDS would have been banned last year, and probably Humans and KCI this year, under their old way of doing things. And under their current more conservative approach, Twin probably wouldn't have been banned.
Emphasis added. This bolded statement is sort-of, kind-of true, but not in any meaningful sense. It's true in the sense that the real ban criteria of "T4 rule violation" and "metagame diversity violation" tend to also be associated with best decks. That is, if your deck is strong enough to be 25% of the metagame, it's probably the best at the time. But it's not true in the sense that a best deck at any time is in danger of a ban, with the sole exception of Twin. You can't use "best deck" and "metagame diversity violator"/"T4 rule violator" as synonymous.
Up until Twin, all of the metagame diversity violators were pretty stark and extreme. BBE Jund exceeded 25% of the metagame. DRS BGx was 20%-25% depending on what stats you looked at. Pod and TC were in the 20% range. Comparing prevalence exceeding 20% to deck prevalence in normal, healthy Modern (where nothing sustains >15% for longer than a few months), and it's clear this is a few deviations over the norm. Twin was obviously questionable from a metagame share perspective, and I've already argued multiple times with extensive evidence, whether or not one agrees with it, that the true reason to ban Twin was to shake up the PT. But as you said, since then, Wizards has stopped this practice.
If we are going to talk about historical ban context, we need to use the known methods, quotes, metrics, and standards that Wizards used at the time. This means we should not infer that a "best deck" was banned when Wizards literally offered explanations for each ban that cited a series of reasons that did not include that deck being the best. The exception to this was Twin, when I and others did some intensive digging to build extensive cases for ulterior motives. But that case-building cannot be done with a single paragraph.
Private Mod Note
():
Rollback Post to RevisionRollBack
Over-Extended/Modern Since 2010
To post a comment, please login or register a new account.
Emphasis added. A lack of unbans is depressing, but the update has critical information about the status of the format. See just a selection of quotes below:
"Given that Modern has looked healthy and diverse at many levels of play outside of Ironworks's dominance at the Grand Prix level"
(R&D believes that Modern is currently healthy and diverse, except for GP, at multiple levels. This likely includes MTGO, GP, and smaller venues)
"As a category, we think these are generally healthy provided they appear in small doses and have reasonable win rates."
(Opal and Stirrings decks are not overperforming by any metric Wizards bases B&R decisions on. This notably includes metagame share and MWP)
"As Modern stands, our metagame data does not indicate a need to impact the other Ancient Stirrings or Mox Opal decks."
(Stirrings and Opal decks are currently at acceptable levels in Modern)
In summary, Wizards said that Modern is healthy and where they want it to be, except for KCI. Plus KCI and Stirrings are being monitored. That's quite a bit of information when all Wizards needed to do was just post the rationale for banning KCI itself.
I 100% agree. In fact, I would say that basically no amount of perceived brokenness would result in changes occurring in that update. Remember that in Eldrazi Winter we had a triple GP March and Wizards still didn't ban Eye/Temple/Eldrazi-card going into that month. This was a slightly different scenario because the planned B&R update was already set for April, not the middle of the GP stretch, but I think the precedent still holds. I would expect Wizards to look at format health after March and then make unbans (and bans, if anything really goes crazy) at that time.
I don't think the argument for Twin is that it's a fair deck. At least, I hope not. It's definitely not fair by any definition I've ever seen or tried to craft. The argument is that Twin, although an unfair deck, decreases overall format speed because it disincentivizes players from running all-in T2, T3, or T4 decks. These fast decks tend to lose to Bolt, Remand, and Exarch/Mite tapdown followed by a lethal T4 combo. Twin also slows down games with the tempo gain of threatening the combo. As such, players might be more likely to run a grindier, more interactive deck that doesn't flat out lose to Twin. Or play their fast deck slower (e.g. Spellskite in Infect or holding up removal). This is the old 2015 "Twin test." The Twin proponents will argue that this test/effect increases the format's fundamental turn, generating overall more interactive games of Modern.
One problem with this argument is that the format isn't overall significantly faster now than in 2015. Only about 2% more games end before T4 now than in 2015, and we have no clue what share of those 2% of games are due to Twin's absence vs. new cards. It's hard to sell a Twin unban on the stake of 2% of games.
A more defensible argument is that Twin has an "appropriate power level" for the current Modern. That was the BB unban justification, and it's a very good standard for SFM. Of course, especially with Twin, the problem there is that it's hard to prove appropriateness without thousands of test games.
I also think Trawler is the ban target. KCI itself is another option, but something like Wellspring is too cute and indirect.
That said, this isn't necessarily the banlist update where all these changes happen. The February update would also be possible. But I do wager we'll see some changes before March.
I'll also add that information consumers have a responsibility to be more discriminating. It takes about 15 seconds of running this through any free photo forensic software to identify glaring signs of a faked screenshot. It literally says it's a photoshopped image in the metadata. I don't know about you, but I don't run my screenshots through photoshop before posting them.
As I posted in that thread, I hope someone with more time/bandwidth than I can pool their GP Oakland sample with my sample to get N even larger. I give all the matchup and sample Ns plus percentages, so this should be methodically easy, albeit time consuming.
It's possible you responded before my edit, in which case I encourage you to read Reid Duke's Thoughtseize and Fatal Push series on CF. He defines GDS as "black midrange," a definition I agree with based on both observed play patterns, deck composition, and historical classifications. Also, to put it simply, it's frikkin Reid Duke master of Midrange. If he calls it Midrange, it's Midrange.
See here for where he sets up GDS as Black Midrange and defines the archetype: https://www.channelfireball.com/articles/thoughtseizes-and-fatal-pushes-part-i/
See here for his dig into GDS as a Black Midrange deck specifically: https://www.channelfireball.com/articles/thoughtseizes-and-fatal-pushes-part-iii-grixis-deaths-shadow/
I was a little more open to this archetype debate until I remembered this article and Duke's position. I'm just unwilling to argue with his classification given that man's knowledge of Magic, Modern, midrange, and these decks. Coupled with all the other elements I and others posted about, it's pretty clearly the proper classification.
The line of T2 GDS/Angler/Tasigur into Denial is way less common than the line of T1 removal/discard into T2 removal/discard/cantrip into T3 threat. You just need to look at the decklists to verify this, with few top-placing lists running more than 3 Stubs. Instead, they run 6 discard and 6+ removal, plus all the cantrips. Even without watching a single game, we can tell that the odds of drawing deployable threat into Stubs is way lower than the other play patterns.
EDIT: In addition to these observations, I'll add that Reid Duke classifies it as midrange in his Thoughtseize/Fatal Push series. All those datapoints point to it being midrange, not tempo: https://www.channelfireball.com/articles/thoughtseizes-and-fatal-pushes-part-iii-grixis-deaths-shadow/
Traditionally, tempo decks deploy threats and then protect the threats and/or disrupt the opponent while the threat wins. Midrange decks tend to disrupt first and then deploy threats on a safe board, although some will still seek to protect a deployed threat later. See the difference between traditional Delver builds (T1 Delver/Swiftspear, hold up protection/removal) and old school BGx (T1 DS into T2 removal/Bob into T3 Lily into a clock). GDS can certainly role-shift in matchups, but it's more fundamentally a midrange deck than a tempo deck.
Reddit: https://www.reddit.com/r/ModernMagic/comments/afmpyd/2018_top_deck_matchup_win_percentage_review/
Blogpost: https://mtgmodernmetrics.wordpress.com/2019/01/12/2018-top-deck-performance-review/
Re: SCG Day 2
Phoenix is a very strong deck. It runs a crap ton of cantrips and has extreme velocity, making it the exact kind of turbo-xerox consistency machine that pros gravitate towards. It has a relatively low floor to success and a very high ceiling. I think I talked about this before, but again, I expect these qualities will drive more and more top players towards this deck in the absence of other options. It doesn't hurt that the deck is strong against some other strategies in the field.
https://mtgmodernmetrics.wordpress.com/2019/01/12/2018-top-deck-performance-review/
It's a 2018 summary of all the compiled GP/SCG data we have about matchups and MWP. I'll publish to Reddit on Monday after incorporating feedback. I request that people keep this under wraps until it's finalized. Please note the limitations section and don't just parrot about how N isn't 10,000. Nothing drives me crazier.
Here are the summarized MWPs for top decks with N>300 from the past year with sample size and 95% confidence interval for the MWP:
KCI: 56.9% (N= 693, 53.2%-60.5%)
Dredge: 55.1% (N= 314, 49.6%-60.6%)
HS Affinity: 55% (N= 460, 50.5%-59.5%)
Bant Spirits: 53.9% (N= 323, 48.4%-59.3%)
Counters Company: 53.7% (N= 419, 48.9%-58.5%)
Humans: 51.8% (N= 1520, 49.3%-54.3%)
UW Control: 51.6% (N= 888, 48.3%-54.9%)
Gx Tron: 51.5% (N= 1026, 48.4%-54.5%)
Hollow One: 50.8% (N= 494, 46.4%-55.2%)
Storm: 50.1% (N= 425, 45.4%-54.9%)
Grixis Death’s Shadow: 50.1% (N= 485, 45.7%-54.6%)
Infect: 49.6% (N= 510, 45.3%-53.9%)
Burn: 49.6% (N= 1136, 46.7%-52.5%)
Jeskai Control: 48.4% (N= 833, 45%-51.8%)
Titanshift: 47% (N= 483, 42.5%-51.4%)
Jund: 46.3% (N= 734, 42.7%-49.9%)
Mardu Pyromancer: 46% (N= 678, 42.3%-49.8%)
Affinity: 44% (N= 609, 40.1%-47.9%)
I'm not going to try and translate the matchup win% matrix here because there's too much going on there. Check out the post for details about those matchups.
Looking forward to feedback!
One of the issues is that I placed too much emphasis on overall metagame share to predict bans, when GP/PT T8s were a better predictor. I've adjusted since then, which is one reason I and others in this thread have successfully and confidently defended a "No changes/No bans" standpoint when it comes to metagame/format diversity violators since Eldrazi Winter. This, despite crap like people calling for bans on Humans, GDS, Gx Tron, and other decks that appeared to be the "best" but were unequivocally not violating the best ban indicators we had. T4 rule violators are harder to assess because it requires access to data we don't have. Other ban criteria, e.g. battle of sideboards, are also more opaque.
This is a more interesting discussion point. For reference, here's the original quote that Wraith appears to be clarifying:
Emphasis added. This bolded statement is sort-of, kind-of true, but not in any meaningful sense. It's true in the sense that the real ban criteria of "T4 rule violation" and "metagame diversity violation" tend to also be associated with best decks. That is, if your deck is strong enough to be 25% of the metagame, it's probably the best at the time. But it's not true in the sense that a best deck at any time is in danger of a ban, with the sole exception of Twin. You can't use "best deck" and "metagame diversity violator"/"T4 rule violator" as synonymous.
Up until Twin, all of the metagame diversity violators were pretty stark and extreme. BBE Jund exceeded 25% of the metagame. DRS BGx was 20%-25% depending on what stats you looked at. Pod and TC were in the 20% range. Comparing prevalence exceeding 20% to deck prevalence in normal, healthy Modern (where nothing sustains >15% for longer than a few months), and it's clear this is a few deviations over the norm. Twin was obviously questionable from a metagame share perspective, and I've already argued multiple times with extensive evidence, whether or not one agrees with it, that the true reason to ban Twin was to shake up the PT. But as you said, since then, Wizards has stopped this practice.
If we are going to talk about historical ban context, we need to use the known methods, quotes, metrics, and standards that Wizards used at the time. This means we should not infer that a "best deck" was banned when Wizards literally offered explanations for each ban that cited a series of reasons that did not include that deck being the best. The exception to this was Twin, when I and others did some intensive digging to build extensive cases for ulterior motives. But that case-building cannot be done with a single paragraph.