I just noticed Silumgar's Scorn . it's obviously unplayable because no deck with blue is ever going to be running a dragon in Modern, but between that and revolt, I do thinkit's just a matter of time before we get a really good counterspell, with some thematic drawback that let's tempo and control use it with minimal difficult but makes it something combo doesn't want.
U- counter target spell, reveal your hand, the controller of the countered spell chooses a card from your hand, discard that card
I didn't include a nonland clause in the interest of it being able to hit Tron as well. no combo deck that can be obliterated by the loss of a crucial card will want to run it, but a more controlling deck would have enough draw to mitigate it. too weak, too strong?
I also welcome Memory Lapse, it would be able to shut down all in combo, but I'm not convinced how powerful it would be in the decks that already have Remand
This hypothetical card is completely unplayable - 2-for-1'ing yourself to counter a spell is never acceptable unless against something like Ad Nauseam. This card is like Swan Song except instead of getting a 2/2 bird they get to Thoughtseize you.
I'm also not sure, despite the common opinion, that better generic, 2-mana countermagic is really what control decks need to improve in Modern. Mana Leak is much-maligned but it's actually a pretty decent counterspell. The difference between Counterspell and Mana Leak, IMO, is overstated.
The real issue with countermagic in Modern is that the good ones cost 2 mana or more, which is bad for tempo when most decks have so many 1-drops.
Ok, so with Bloo's real turn 3 rate being around 12.5%, I think we can safely assume that Infect is definitely higher than the 10-12% Bfrie proposed, unless I'm just wrong about Infect having a higher rate than Bloo, which I don't think I am.
Using an n=1500 sample of MTGO data from 2015 (old data but it did include Become Immense), I can find Infect's old T2-T3 win-rate.
In that dataset, Infect won 12% of its total games on T2-T3. This breaks down 1% T2 and 11% T3. As a function of wins, 22% of Infect's total wins were T2-T3 wins. The rest were T4+. This latter number is more academic than useful: the number we need to focus on is the 12% T2-T3 win-rate.
By contrast, Amulet Bloom rocked a 23% T2-T3 win-rate as a function of total games. That's almost double Infect's 12% T2-T3 win-rate.
I have no idea how these numbers changed in 2016, but it aligns with the analysis other players performed in this thread. Because of that, I suspect Infect really does have a 12% T2-T3 win-rate in real games.
If I may ask, how did you get all of this data from MTGO and sort through it to get those #'s?
It would have to be subjective, I think. You could easily take Y = 90% or 95%, but there are infinitely many possible board states/situations so it seems like it would be too hard to do the calculation for every "virtually winning" board state and its corresponding value of X.
How do you define a virtual win then? Well, subjectively, but being smart about it. For example, you could use Huey Jensen's commonly used criterion of "even if Player A could stack his/her deck, Player B would still be very likely win this game". So watch a bunch of matches featuring Dredge, and record the number of times that (a) Dredge actually kills the opponent before turn 3 and (b) the above criterion applies (to the best of your estimate).
You might not be in love with this method, but lets face it: most bannings in Modern have been largely based on subjective arguments. I think that's pretty much necessary.
I'm sure Wizards uses a mix of data-driven and theory-based arguments when doing a ban, and I'm also sure they have no exact cutoffs for anything. They do probably have indicators, many of which are probably numeric. But again, no hard cutoffs; we really only use the cutoffs as players to try and describe/predict metagames, bans, and unbans. I'm totally on board with that method. Many complicated policies shouldn't have explicit cutoffs, and Modern banning policy is one such example.
My issue with Dredge is that I don't think anyone actually has the numbers you or I are speculating about. People keep claiming it's a T3 deck but with no data to support that. I'm not saying data is the only thing they need, but you also can't make a firm ban argument without at least some data to back it up. If someone has that data I'd love to see it, but from what I've seen, it's a lot of anecdote without any numbers.
Also, in case it wasn't clear, you don't need numbers alone. You need a mix of theory and numbers to support a ban. Right now, I've seen a lot of theory and very few numbers, so that pillar needs to be added to create a strong case for any bans.
100% agreed. Which is why I think my suggested method of (1) establishing a subjective, but reasonable, criterion for a "virtual win" and then (2) use existing recorded games, or test games, to come up with a number of many virtual wins Dredge gets before turn 3 each game, would be a reasonable way for Wizards to determine whether Dredge is really breaking the t4 rule or not.
Purely basing it on actual wins before turn 4 is too short-sighted, I think, for the reasons I stated in my earlier post.
I don't see why "virtual" turn < 4 wins shouldn't be considered.
I mean, the whole reason the rule even exists is to prevent "non-games", or to be more precise, games in which one or both players didn't feel like they had the chance to make enough meaningful decisions or plays.
A game that "virtually" ends on turn 3 is as much of a non-game as one that actually ends on turn 3. If I'm so behind on board to Dredge by turn 3 that I just scoop, that should count as a turn 3 win.
Imagine if there were a deck that could, consistently, leave the opponent with no cards in hand and no permanents on the board by the end of turn 3. I think it would be silly to argue against this deck receiving a ban, even though it's not technically ending the game turn 3.
But how are you measuring this? It can't be arbitrary because Modern is full of decks that "virtually win" in certain early board states. That's true of many formats. You'd have to find some measurement that says "Dredge has this board state in X% of games" and another measurement that says "Dredge ends up winning Y% of those games with that board state." If someone had those numbers, I think that could be a very compelling case. I also haven't seen anything like that so far and suspect no one but Wizards has that data.
It would have to be subjective, I think. You could easily take Y = 90% or 95%, but there are infinitely many possible board states/situations so it seems like it would be too hard to do the calculation for every "virtually winning" board state and its corresponding value of X.
How do you define a virtual win then? Well, subjectively, but being smart about it. For example, you could use Huey Jensen's commonly used criterion of "even if Player A could stack his/her deck, Player B would still be very likely win this game". So watch a bunch of matches featuring Dredge, and record the number of times that (a) Dredge actually kills the opponent before turn 3 and (b) the above criterion applies (to the best of your estimate).
You might not be in love with this method, but lets face it: most bannings in Modern have been largely based on subjective arguments. I think that's pretty much necessary.
I don't see why "virtual" turn < 4 wins shouldn't be considered.
I mean, the whole reason the rule even exists is to prevent "non-games", or to be more precise, games in which one or both players didn't feel like they had the chance to make enough meaningful decisions or plays.
A game that "virtually" ends on turn 3 is as much of a non-game as one that actually ends on turn 3. If I'm so behind on board to Dredge by turn 3 that I just scoop, that should count as a turn 3 win.
Imagine if there were a deck that could, consistently, leave the opponent with no cards in hand and no permanents on the board by the end of turn 3. I think it would be silly to argue against this deck receiving a ban, even though it's not technically ending the game turn 3.
I could discuss JTMS unban. I would not be willing to discuss an SFM unban. But why not just unban Preordain first to see what happens? And on top of that, wouldn't JTMS force more people to leave Jund and play more linear decks to just kill before he resolves? BGx would take BBE as well?
With Jace being such a compact win-con in any deck (4 slots win the game!), it would leave decks to dedicate almost the entirety of their strategy to interaction. Grixis decks would no longer have to worry about not having a clock, UB would have a purpose (maindeck damnations seem just fine alongside a Jace win-con and UB stomps combo pretty hard), and etc. The only analogous card that even comes close to Jace is Nahiri, and she's in an awkward color combination and requires 5/6 spots (and her non-'win the game' modes aren't nearly as useful as Jace's - her ultimate wins just the same but that's about it... most of the time). But honestly - I'm not sure what happens to the meta. I know it would change a ton, but I'm not sure what happens to every individual deck as it would make serious waves in the meta.
Jace would not eliminate the need for a clock in Grixis, not at all. Grixis needs a clock to beat Tron, Scapeshift, and other uninteractive decks that have inevitability in the late game. Jace is not going to do enough against those deck. In fact, when Jace was in the process of breaking standard one of the best counters was Valakut. Unbanning Jace would just encourage people to play those types of decks even more, as well as fast combo, because there's no point in trying to play fair against Jace.
A Jace unban would decrease the meta share of BGx and just increase the meta share of Valakut and Bant Eldrazi. Good luck using your Jace to bounce reality smasher or drowner of hope...
All that's really happening to BGx is that Junk is cannibalizing some of Jund's metashare. They may be down a collective 3.2% but when it's still over 10% total it's not a huge decline. This does, though, go against the narrative that if SFM were to be unbanned people would still play Jund and Junk wouldn't cannibalize Jund's metashare. We're already seeing that happen, and if SFM were to be unleashed it would be much more drastic.
BGx decks are certainly on the decline due to ramp decks and Dredge. Thair metagame combined all together is still solid but if you look at Jund' and Abzan's numbers on their own they aren't very good. This might change in the future but these Eldrazi and Valakut decks along with Dredge are really painful for BGx. I guess we can hope this will change and perhaps it will but till then there's nothing you can do and it's probably better to play some proactive deck if you intend to go to a larger tournament at least.
This is one of the pitfulls of using metagame representation data to estimate the power level of decks. There are other important factors that have an effect. In the case of BGx, it's always going to be a little overrepresented based on the fact that there are a ton of players who play BGx and only BGx; after all, the archetype is very expensive and a lot of players who buy into it don't have remaining funds to switch to something else when the going gets rough. The same goes for URx decks, to a lesser extent. I doubt there would be as many people playing Grixis Delver right now for example if so many players weren't already invested in the URx archetype from playing Twin before.
I think most players who have experience with BGx would agree that it is not well-positioned in the current metagame. As Ayiluss said, Bant Eldrazi is a really bad matchup and it's currently top dog in the format. Dredge and Valakut are also really miserable matchups. On the flip side we don't really have any great matchups except Infect. There are times where Jund in practice is more of a 45/55 deck than a 50/50 deck, and this is one of those times.
Metagame data is of course very helpful but it has to be taken with a grain of salt, and one should always be aware of underlying factors that might affect a deck's meta share other than power level.
Or, they could just unban Splinter Twin and make URx playable again. It's a far better solution than unbanning JTMS and making blue decks cost $600 more and still suck, or unbanning Preordain which likely would not have enough impact.
The reasoning behind the ban is clearly flawed, and we've had enough time to see now that Twin wasn't the reason Jeskai Control, Temur Tempo (lol) etc weren't being heavily played. Those decks just aren't strong enough, and now we have definitive proof.
URx Snapcaster + Bolt decks have been a pillar of the format since the beginning. It would be a shame if they just allow the entire archetype to die out like this.
How is Jeskai not good when it's the best Uxx control deck since the beginning of the format?
Compare to the esper or cruel control variants which are way harder to pilot or build correctly for any big tournament. Or the underdogs Temur, Bant and Sultai...
Well I personally would consider Twin a control deck, so from that perspective it's not.
Besides, I think if a deck that's not even Tier 1 anymore is the best control deck in the history of the format, that says a lot about how underpowered the control archetype is in general.
Or, they could just unban Splinter Twin and make URx playable again. It's a far better solution than unbanning JTMS and making blue decks cost $600 more and still suck, or unbanning Preordain which likely would not have enough impact.
The reasoning behind the ban is clearly flawed, and we've had enough time to see now that Twin wasn't the reason Jeskai Control, Temur Tempo (lol) etc weren't being heavily played. Those decks just aren't strong enough, and now we have definitive proof.
URx Snapcaster + Bolt decks have been a pillar of the format since the beginning. It would be a shame if they just allow the entire archetype to die out like this.
What disappoints me is the lack of explanation. Instead of just writing "no changes", how about explaining why? They do that when they ban or unban cards, but in some cases a lack of action should also be explained.
If Preordain actually breaks Goryo's, that's not really an issue with Preordain.
Preordain is worse than Faithless Looting in that deck. It's actually sort of silly that Faithless Looting is legal and Preordain is banned, since Faithless Looting is only used to enable GY-based combos and does so better than Preordain does.
The two decks that were the reason for Preordain's banning at the time (Twin and UR Storm) have since been hit with bans, and it's extremely unlikely that Storm would suddenly be too consistent with Preordain, since on top of the Seething Song ban Eidolon of the Great Revel and Eidolon of Rhetoric have both since been printed.
If Preordain breaks Goryo's, ban SSG or Looting or Goryo's itself.
Preordain at least deserves a shot in the current format.
Back to the topic, I don't like JTMS as an unban target. It doesn't empower control to handle fast, linear strategies. It just improves the control vs. grindy deck matchup (similar to AV's effect). JTMS is a slow, specialized answer for certain matchups. Preordain lets you dig for cheap, generic answers as early as turn one.
Why keep him banned then?
If JTMS proves the ultimate grindy trump, then players who were trying to run slower strategies will just shift to even faster, more linear strategies to try and get under JTMS and win before he becomes relevant. Preordain punishes that by finding generic answers quicker. If you overcommit to an early line and Preordain digs for the T2 answer, you're in trouble.
Bingo.
Jace in the format would just make fair magic more miserable, and not actually help blue decks fight linear strategies.
Well, we'd need to know under which conditions Infect wins T1-3. It can't rely on any individual 4-of, and any "8-ofs"(a card, and then a functional copy of it), there can't be more than two.
For infect to win, it needs any combination of 3 pump spells (except 3x mutagenic). Noble hierarchs function as a pump spell here provided the other two arent also mutagenic or hierarchs. If become immense is used, it also needs two fetch lands
Basically, to turn two it needs either:
Glistener
2x mutagenic + a +4/+4 + hierarch or 2x +4/+4 + a mutagenic
2x green lands (pendelhaven removes need for hierarch)
OR
Glistener
2x fetches/probes
2x mutagenic
Become immense
When we get into turn 3 territory, the combination of cards becomes massive, many of them being capable of beating multiple pieces of interaction
Let's assume Infect is on the play.
What combination allows a turn 3 kill through a removal spell? Through 2 removal spells?
Infect has been an incredibly well performing deck since last year. Many pro's say it's the best deck. We all know its a turn 3 deck, and what i mean by this? I mean that it might not win on turn 3 most often, but if you couldn't interact on turn 3 you are basically dead on an average hand.
The deck is super fun to play, not so much to play against. I played around 50 games with Bant Eldrazi against this deck and Spellskite just doesn't do it, neither Chalice of the Void. Its super resilient.
I'm not the first one to say Immense might be on their sights for banning. Whether it's on this announcement or not is trivial. The deck is good, is Tier 1, it CAN win on turn 3 often.
Do i want it banned? No, i've never wanted any card banned, except for Eye of Ugin and Cruise.
It's definitely a card to consider and those who own Infect and/or DS Aggro, pray the night before each anouncement until WOTC states what are their thought's on the card.
I mean, you can't really complain about Infect being unfun to play against if you're playing something like Bant Eldrazi or Tron. Sure, if you don't have many ways to interact with Infect, it will usually kill you quickly. But that's the downside of playing a rather uninteractive deck. I'm glad that there's a fast, resilient deck that doesn't just fold to something like Chalice.
Playing Jeskai or Jund against Infect on the other hand is very fun, one of my favorite matchups in Modern on both sides.
If some users are serious about doing some testing to evaluate potential unbans, I'm all for it. I think that would be fun.
The problem is that you'd need a fairly large sample size to actually be able to say anything though, right? I'm no statistics expert but I think you can estimate the margin of error in an estimated win % "p" according to:
E = Z*SQRT(p*(1-p)/n)
Where n is the sample size (i.e., number of matches tested).
If you take Z as 1.96 (for 95% confidence), then you'd have to test more than 300 matches just to get an estimated win % with a margin of error of less than 5%!
Am I missing something there, ktkenshinx?
The math is fine but, as wpgstevo and bfrie said, we don't need nearly this much testing. Wizards is definitely not doing this much testing. If we want to predict an SFM unban, we just need to do as much testing as Wizards will probably do and then see if SFM appears safe in those tests. 10 pre-board and 20 post-board games would likely be MORE than enough
But if you test 10 pre-board and 20 post-board games with your friend, and I do the same with mine, we are likely to come to different conclusions based on the small sample size.
It'd be like flipping a coin 5 times to determine the probably of getting heads. Results will differ vastly based on variance.
It still sounds fun though, and I'm definitely up for it. What would be the best way to organize some testing between users?
This hypothetical card is completely unplayable - 2-for-1'ing yourself to counter a spell is never acceptable unless against something like Ad Nauseam. This card is like Swan Song except instead of getting a 2/2 bird they get to Thoughtseize you.
I'm also not sure, despite the common opinion, that better generic, 2-mana countermagic is really what control decks need to improve in Modern. Mana Leak is much-maligned but it's actually a pretty decent counterspell. The difference between Counterspell and Mana Leak, IMO, is overstated.
The real issue with countermagic in Modern is that the good ones cost 2 mana or more, which is bad for tempo when most decks have so many 1-drops.
If I may ask, how did you get all of this data from MTGO and sort through it to get those #'s?
100% agreed. Which is why I think my suggested method of (1) establishing a subjective, but reasonable, criterion for a "virtual win" and then (2) use existing recorded games, or test games, to come up with a number of many virtual wins Dredge gets before turn 3 each game, would be a reasonable way for Wizards to determine whether Dredge is really breaking the t4 rule or not.
Purely basing it on actual wins before turn 4 is too short-sighted, I think, for the reasons I stated in my earlier post.
It would have to be subjective, I think. You could easily take Y = 90% or 95%, but there are infinitely many possible board states/situations so it seems like it would be too hard to do the calculation for every "virtually winning" board state and its corresponding value of X.
How do you define a virtual win then? Well, subjectively, but being smart about it. For example, you could use Huey Jensen's commonly used criterion of "even if Player A could stack his/her deck, Player B would still be very likely win this game". So watch a bunch of matches featuring Dredge, and record the number of times that (a) Dredge actually kills the opponent before turn 3 and (b) the above criterion applies (to the best of your estimate).
You might not be in love with this method, but lets face it: most bannings in Modern have been largely based on subjective arguments. I think that's pretty much necessary.
I mean, the whole reason the rule even exists is to prevent "non-games", or to be more precise, games in which one or both players didn't feel like they had the chance to make enough meaningful decisions or plays.
A game that "virtually" ends on turn 3 is as much of a non-game as one that actually ends on turn 3. If I'm so behind on board to Dredge by turn 3 that I just scoop, that should count as a turn 3 win.
Imagine if there were a deck that could, consistently, leave the opponent with no cards in hand and no permanents on the board by the end of turn 3. I think it would be silly to argue against this deck receiving a ban, even though it's not technically ending the game turn 3.
Jace would not eliminate the need for a clock in Grixis, not at all. Grixis needs a clock to beat Tron, Scapeshift, and other uninteractive decks that have inevitability in the late game. Jace is not going to do enough against those deck. In fact, when Jace was in the process of breaking standard one of the best counters was Valakut. Unbanning Jace would just encourage people to play those types of decks even more, as well as fast combo, because there's no point in trying to play fair against Jace.
A Jace unban would decrease the meta share of BGx and just increase the meta share of Valakut and Bant Eldrazi. Good luck using your Jace to bounce reality smasher or drowner of hope...
This is one of the pitfulls of using metagame representation data to estimate the power level of decks. There are other important factors that have an effect. In the case of BGx, it's always going to be a little overrepresented based on the fact that there are a ton of players who play BGx and only BGx; after all, the archetype is very expensive and a lot of players who buy into it don't have remaining funds to switch to something else when the going gets rough. The same goes for URx decks, to a lesser extent. I doubt there would be as many people playing Grixis Delver right now for example if so many players weren't already invested in the URx archetype from playing Twin before.
I think most players who have experience with BGx would agree that it is not well-positioned in the current metagame. As Ayiluss said, Bant Eldrazi is a really bad matchup and it's currently top dog in the format. Dredge and Valakut are also really miserable matchups. On the flip side we don't really have any great matchups except Infect. There are times where Jund in practice is more of a 45/55 deck than a 50/50 deck, and this is one of those times.
Metagame data is of course very helpful but it has to be taken with a grain of salt, and one should always be aware of underlying factors that might affect a deck's meta share other than power level.
Well I personally would consider Twin a control deck, so from that perspective it's not.
Besides, I think if a deck that's not even Tier 1 anymore is the best control deck in the history of the format, that says a lot about how underpowered the control archetype is in general.
The reasoning behind the ban is clearly flawed, and we've had enough time to see now that Twin wasn't the reason Jeskai Control, Temur Tempo (lol) etc weren't being heavily played. Those decks just aren't strong enough, and now we have definitive proof.
URx Snapcaster + Bolt decks have been a pillar of the format since the beginning. It would be a shame if they just allow the entire archetype to die out like this.
Preordain is worse than Faithless Looting in that deck. It's actually sort of silly that Faithless Looting is legal and Preordain is banned, since Faithless Looting is only used to enable GY-based combos and does so better than Preordain does.
The two decks that were the reason for Preordain's banning at the time (Twin and UR Storm) have since been hit with bans, and it's extremely unlikely that Storm would suddenly be too consistent with Preordain, since on top of the Seething Song ban Eidolon of the Great Revel and Eidolon of Rhetoric have both since been printed.
If Preordain breaks Goryo's, ban SSG or Looting or Goryo's itself.
Preordain at least deserves a shot in the current format.
Bingo.
Jace in the format would just make fair magic more miserable, and not actually help blue decks fight linear strategies.
Let's assume Infect is on the play.
What combination allows a turn 3 kill through a removal spell? Through 2 removal spells?
I mean, you can't really complain about Infect being unfun to play against if you're playing something like Bant Eldrazi or Tron. Sure, if you don't have many ways to interact with Infect, it will usually kill you quickly. But that's the downside of playing a rather uninteractive deck. I'm glad that there's a fast, resilient deck that doesn't just fold to something like Chalice.
Playing Jeskai or Jund against Infect on the other hand is very fun, one of my favorite matchups in Modern on both sides.
But if you test 10 pre-board and 20 post-board games with your friend, and I do the same with mine, we are likely to come to different conclusions based on the small sample size.
It'd be like flipping a coin 5 times to determine the probably of getting heads. Results will differ vastly based on variance.
It still sounds fun though, and I'm definitely up for it. What would be the best way to organize some testing between users?