I'm bracing myself for the coming of the Surveil Variant of Treasure Cruise or Dig Through Time...
And man, poor RW can't stop getting crap mechanics. Mentor, Radiance, Batallion... add Battle Cry since Mirrans were RW. I wish for the day the color pair is not seen as the weenie - burn color but the Armageddon and Stone Rain colors.
That amazing GP Prague metagame collection project is still going, and there are some preliminary Day 1 and Day 2 #s up with about 50% of all decks accounted for. Very interesting results, especially regarding the MWPs and overall win rates. I did a quick rank-based analysis of conversion rates from Day1 to Day2, Day2 to T32, Day 2 to T16, and MWP to try and determine the "best" performing decks in that event across all conversion rates and MWP. Based on that preliminary analysis, here's my list of the best-performing Modern decks based on GP Prague data. Note that these stats might change as more data is added. As a rough summary of the numbers, the ranking points just represent how well the deck did relative to other decks in the pool. So Bant Spirits had the best performance at a rank 1. Jeskai Control performed almost twice as poorly as Bant Spirits at 2.08 and comparably to Humans at 2.19. Titanshift was the worst performer, doing almost four times as poorly on all measures as Bant Spirits (and twice as poorly as Humans/Jeskai Control). That's just a sample way to read the data but it's really interchangeable for all deck positioning.
Big takeaways based solely on this data (these may or may not be extrapolatable to non-GP Prague events):
Bant Spirits, Hardened Scales, and UW Control performed the best.
Gx Tron, Storm, and Titanshift performed the worst.
Humans, arguably the "best" deck in Modern, actually had the most middling performance. This may be because many pilots tried it, but probably not because UW actually had more pilots and performed a lot better.
Titanshift is unusually bad in this metagame. Not only is it the worst-performing deck but it also had the biggest difference between its performance and the performance of the previous deck.
Spirits seem really strong, with both decks performing in the top grouping.
Midrange struggled, with GDS, Jund, and Mardu all in the bottom 50% of performers.
I'll keep adding to this as the data increases, and let me know if you think of more performance metrics worth ranking and including.
surveil seems too simple to mess up. i assume its boring but subtlety powerful by intent. its even flavorful for being so. also scry already exists, so its not like they are going to suddenly print some 1 cmc cantrip that has Surveil 5 - Draw 1.
there was almost zero chance they could have gotten delve right.
also dont forget that the impact of the Play Design team should be in full force this set. maybe a noticeable difference, maybe not. at the very least they can safeguard against the larger mistakes.
i agree with boros getting the short stick though. those sort of 'attack together' mechanics pretty much set the bar at the cards having to be powerful enough to be played without the mechanic existing; at least for modern and beyond anyways. aggressive weenies makes sense though. its called the boros legion after all.
Private Mod Note
():
Rollback Post to RevisionRollBack
Modern: UWGSnow-Bant Control BURGrixis Death's Shadow GWBCoCo Elves WCDeath and Taxes (sold)
@ktk Thanks for the data. Titan Shift performing that poorly is really interesting to me, because I actually ran into it a higher percentage of the time this month than ever before, almost 5%. The average is closer to 2%. I thought that maybe it was more well positioned now, but obviously that's not the case.
And like I said in here a little while ago, Tron is not nearly as well positioned now as it was in the recent past, so that poor showing doesn't surprise me much. Storm did terribly as well. So I think it's safe to take those Damping Spheres out of your sideboards now guys
surveil seems too simple to mess up. i assume its boring but subtlety powerful by intent. its even flavorful for being so. also scry already exists, so its not like they are going to suddenly print some 1 cmc cantrip that has Surveil 5 - Draw 1.
Ooh, please gimme that instant speed U - Surveil 2, Draw a card so I can play 8 copies of Thought Scour in GDS, lol.
Edit: damn, I just realized how busted that would be, it would basically be an instant speed Preordain. There's no way they're making a one mana Surveil cantrip, it would be really good.
Private Mod Note
():
Rollback Post to RevisionRollBack
Modern UBR Grixis Shadow UBR UR Izzet Phoenix UR UW UW Control UW GB GB Rock GB
Commander BG Meren of Clan Nel Toth BG BGUW Atraxa, Praetor's Voice BGUW
surveil seems too simple to mess up. i assume its boring but subtlety powerful by intent. its even flavorful for being so. also scry already exists, so its not like they are going to suddenly print some 1 cmc cantrip that has Surveil 5 - Draw 1.
Ooh, please gimme that instant speed U - Surveil 2, Draw a card so I can play 8 copies of Thought Scour in GDS, lol.
Edit: damn, I just realized how busted that would be, it would basically be an instant speed Preordain. There's no way they're making a one mana Surveil cantrip, it would be really good.
yeah thats what i mean. surveil is close enough to scry in both effect and power level that you immediately know such a cantrip would be too good.
tbh id be surprised if they just made opt v.2.0: U (Instant) - Surveil 1, Draw 1
sorta unrelated, but with more graveyard centric mechanics on the horizon i hope they print a new piece of hate. not sure what new form it could take, but that is wizards job. hell make it red and give it to burn.
Private Mod Note
():
Rollback Post to RevisionRollBack
Modern: UWGSnow-Bant Control BURGrixis Death's Shadow GWBCoCo Elves WCDeath and Taxes (sold)
surveil seems too simple to mess up. i assume its boring but subtlety powerful by intent. its even flavorful for being so. also scry already exists, so its not like they are going to suddenly print some 1 cmc cantrip that has Surveil 5 - Draw 1.
Ooh, please gimme that instant speed U - Surveil 2, Draw a card so I can play 8 copies of Thought Scour in GDS, lol.
Edit: damn, I just realized how busted that would be, it would basically be an instant speed Preordain. There's no way they're making a one mana Surveil cantrip, it would be really good.
yeah thats what i mean. surveil is close enough to scry in both effect and power level that you immediately know such a cantrip would be too good.
tbh id be surprised if they just made opt v.2.0: U (Instant) - Surveil 1, Draw 1
sorta unrelated, but with more graveyard centric mechanics on the horizon i hope they print a new piece of hate. not sure what new form it could take, but that is wizards job. hell make it red and give it to burn.
I'd honestly rather it be blue, if we give it to Red BridgeVine becomes more of a deck.
tbh id be surprised if they just made opt v.2.0: U (Instant) - Surveil 1, Draw 1
Yeah, that would basically just be a better Opt, because it lets you put a card in your graveyard for delving or for flashing back with Snapcaster, so whatever the Surveil cantrip is will probably cost at least 2 mana. If there's a 1 mana Surveil cantrip, it'll probably be Modern playable even if it's only Surveil 1.
Private Mod Note
():
Rollback Post to RevisionRollBack
Modern UBR Grixis Shadow UBR UR Izzet Phoenix UR UW UW Control UW GB GB Rock GB
Commander BG Meren of Clan Nel Toth BG BGUW Atraxa, Praetor's Voice BGUW
Ok not in standard but whatever you call it when a set is produced to promote a new set and show off it’s mechanics. I hate that some cards are legal and some aren’t personally though. Point is they know for these that they needed graveyard hate so why wouldn’t they know for the regular set. Drs reprinted is kinda a kick in the nuts though to eternal format players at this point.
That amazing GP Prague metagame collection project is still going, and there are some preliminary Day 1 and Day 2 #s up with about 50% of all decks accounted for. Very interesting results, especially regarding the MWPs and overall win rates. I did a quick rank-based analysis of conversion rates from Day1 to Day2, Day2 to T32, Day 2 to T16, and MWP to try and determine the "best" performing decks in that event across all conversion rates and MWP. Based on that preliminary analysis, here's my list of the best-performing Modern decks based on GP Prague data. Note that these stats might change as more data is added. As a rough summary of the numbers, the ranking points just represent how well the deck did relative to other decks in the pool. So Bant Spirits had the best performance at a rank 1. Jeskai Control performed almost twice as poorly as Bant Spirits at 2.08 and comparably to Humans at 2.19. Titanshift was the worst performer, doing almost four times as poorly on all measures as Bant Spirits (and twice as poorly as Humans/Jeskai Control). That's just a sample way to read the data but it's really interchangeable for all deck positioning.
Big takeaways based solely on this data (these may or may not be extrapolatable to non-GP Prague events):
Bant Spirits, Hardened Scales, and UW Control performed the best.
Gx Tron, Storm, and Titanshift performed the worst.
Humans, arguably the "best" deck in Modern, actually had the most middling performance. This may be because many pilots tried it, but probably not because UW actually had more pilots and performed a lot better.
Titanshift is unusually bad in this metagame. Not only is it the worst-performing deck but it also had the biggest difference between its performance and the performance of the previous deck.
Spirits seem really strong, with both decks performing in the top grouping.
Midrange struggled, with GDS, Jund, and Mardu all in the bottom 50% of performers.
I'll keep adding to this as the data increases, and let me know if you think of more performance metrics worth ranking and including.
Following up from this, I wanted to add another point of clarification. It's tempting to view this exclusively as "Deck X is twice as bad as Deck Y," as in the Bant Spirits vs. Humans/Jeskai example I made in this post. This might be technically accurate but it misses on deck viability because it undersells the viability of all but the best performing decks.
A much more useful way to read this list is to compare relative viability of any given deck to a known tiered benchmark For example, we know Humans is an accepted and indisputable Tier 1 deck by almost every possible metric we can think of. On this list, it "only" scores 2.19. We could conclude that it's twice as "bad" as Bant Spirits, but that's pretty disingenuous because we already know Humans isn't bad at all. So this negative framing doesn't advance our understanding of the format. A much better conclusion would be that every deck that performed better than Humans is at least as viable as Humans itself. That means all the Top 10 decks could be considered "as good or better than" Humans in Modern, and would all be great Modern deck choices.
Similarly, the decks that are right below Humans are also probably just as viable as Humans (but probably not better). GDS, Hollow One, and Jund (Jund to a lesser extent) are pretty close to the Humans viability score, which suggests they are roughly as viable as Humans in a major event. The only decks that truly appear deficient in some way are the decks that scored 3 or lower. One of these is very surprising (Mardu Pyro), because there are a few pilots who did extremely well with Mardu over the past months. I have a theory about how to adjust for this by drawing on other GP and will do more work later.
I'm kind of annoyed at this because they did move away from duel decks because of mixing standard legal cards with rotated cards, resulting in casual players who don't know any better to get told their decks are not tournament legal. The only reason the guild decks aren't standard legal is probably Deathrite shaman being in the same format as Llanowar Elves.
Private Mod Note
():
Rollback Post to RevisionRollBack
1. (Ravnica Allegiance): You can't keep a good esper control deck down... Or Wilderness Reclamation... or Gates...
2. (War of the Spark): Guys, I know what we need! We need a cycle of really idiotic flavor text victory cards! Jace's Triumph...
3. (War of the Spark): Lets make the format with control have even more control!
We know Jumpstart Cards have the same cost + discard one card but that doesn't mean they have the same exact affect when you jumpstart them.
So you might get UR do two damage to any one target, jumpstart if jumpstarted it deals 3 damage to any one target instead. Or is it confirmed you get the same mana cost + discard and the exact same affect.
Updated analysis on deck positioning! I added in MWP from the previous GP events (per Erzel's Reddit analysis), plus T32 analysis from all GP in 2018. This includes both the straight average scores of decks in the T32, and their adjusted scores accounting for Byes (so decks with fewer Byes effectively get more weight). I then added those metrics to the existing analysis, ranked them, and spit out some new rankings. There aren't a lot of differences between this 2018 GP retrospective ranking and the one I posted earlier, except a few decks present as slightly better (e.g. Counters Company, Gx Tron) and a few are slightly worse (e.g. Jund, Hollow One).
Titanshift and Storm are still terrible in this metagame for some reason. KCI is slightly "better," UW Control is slightly "worse," etc. Remember that this isn't the best way to read the numbers. It is a fundamental misreading to say that UW Control is almost twice as "bad" as Hardened Scales, because the performance score itself doesn't really indicate a strictly objective performance. It's an average of the rankings of different performance metrics. Because of that, the best way to review these is to pick a known, viable, Tier 1 benchmark and compare relative performance. In this case, Humans is our benchmark in 12th place. Any deck that has a higher rating than Humans is at least as viable as Humans, likely slightly more. Any deck that has a lower rating than Humans is likely less viable, with a big dropoff between something like Gx Tron (which is barely less viable) to Titanshift (which is still presenting as really bad).
The only strategy that appears to be struggling in this metagame is midrange, with both Jund and Mardu performing worse than Humans. GDS, at least, appears to be at least as viable as Humans.
Meanwhile, UR Moon is nowhere to be seen (in any variant). LOL. Still playing mine, for better or for worse.
Blue Moon does show up in the datasets, but N is too small to make any conclusions about the deck. This is actually true of a lot of different decks and is a major limitation of this analysis project.
great work ktkenshinx. thanks for taking the time to do that.
so just to clarify, these rankings are a composite based on the following(?):
-match win percentage
-overall win rate
-day 2 conversion (from day 1)
-top 32 conversion (from day 2)
-top 16 conversion (from day 2)
with some adjustments for byes. its worth pointing out that this ranking has nothing to do with representation. so you are placing decks with multiple high profile finishes next to ones with a single finish. the distinction between something like overall deck performance, and performance at certain event(s) might not be clear.
for example elves. your analysis would conclude that elves is at least a good/viable/whatever as humans, if not more. you dont exactly have to look far to see that this is demonstrably not true.
cant say im fond of so much use of the V-word. mostly because its connotation as a buzz word to placate the community regarding relatively poor performing decks (ie. 'oh your deck might not be tier 1, but its viable).
Private Mod Note
():
Rollback Post to RevisionRollBack
Modern: UWGSnow-Bant Control BURGrixis Death's Shadow GWBCoCo Elves WCDeath and Taxes (sold)
great work ktkenshinx. thanks for taking the time to do that.
so just to clarify, these rankings are a composite based on the following(?):
-match win percentage
-overall win rate
-day 2 conversion (from day 1)
-top 32 conversion (from day 2)
-top 16 conversion (from day 2)
Specifically, MWP at GP Prague, MWP pooled between GP Barcelona/Sao Paulo/Vegas (it's really hard to separate events for distinct GP MWPs in Erzel's data). For the conversions, it's just GP Prague because that's the only GP where we have that data from Day 1 to Day 2. It also had some representation-based T32 data, but I'm finding that data to be troubling so I am removing it.
with some adjustments for byes. its worth pointing out that this ranking has nothing to do with representation. so you are placing decks with multiple high profile finishes next to ones with a single finish. the distinction between something like overall deck performance, and performance at certain event(s) might not be clear.
Yes and no. It's not representation based in a direct sense, but it only includes decks that have a big enough N to work with. So it's indirectly representation based so we can have ~100 matches in a sample.
for example elves. your analysis would conclude that elves is at least a good/viable/whatever as humans, if not more. you dont exactly have to look far to see that this is demonstrably not true.
This is an issue and suggests the analysis method may have issues. I'm trying to separate out the other GP events for distinct conversions.
Private Mod Note
():
Rollback Post to RevisionRollBack
Over-Extended/Modern Since 2010
To post a comment, please login or register a new account.
And man, poor RW can't stop getting crap mechanics. Mentor, Radiance, Batallion... add Battle Cry since Mirrans were RW. I wish for the day the color pair is not seen as the weenie - burn color but the Armageddon and Stone Rain colors.
1. Bant Spirits: 1
2. Hardened Scales: 1.06
3. UW Control: 1.22
4. Counters Company: 1.5
5. KCI: 1.53
6. UW Spirits: 1.53
7. Bridgevine: 1.58
8. Infect: 1.86
9. Jeskai Control: 2.08
10. Humans: 2.19
11. GDS: 2.47
12. Hollow One: 2.64
13. Jund: 2.78
14. Mardu Pyromancer: 3
15. Elves: 3.08
16. Burn: 3.36
17. Affinity: 3.44
18. Bogles: 3.47
19. Gx Tron: 3.58
20. Storm: 3.69
21. Titanshift: 4.25
Big takeaways based solely on this data (these may or may not be extrapolatable to non-GP Prague events):
there was almost zero chance they could have gotten delve right.
also dont forget that the impact of the Play Design team should be in full force this set. maybe a noticeable difference, maybe not. at the very least they can safeguard against the larger mistakes.
i agree with boros getting the short stick though. those sort of 'attack together' mechanics pretty much set the bar at the cards having to be powerful enough to be played without the mechanic existing; at least for modern and beyond anyways. aggressive weenies makes sense though. its called the boros legion after all.
UWGSnow-Bant Control
BURGrixis Death's Shadow
GWBCoCo Elves
WCDeath and Taxes(sold)And like I said in here a little while ago, Tron is not nearly as well positioned now as it was in the recent past, so that poor showing doesn't surprise me much. Storm did terribly as well. So I think it's safe to take those Damping Spheres out of your sideboards now guys
I'm not even sure Sphere's are good, but its all I can think to use in UR.
Spirits
Edit: damn, I just realized how busted that would be, it would basically be an instant speed Preordain. There's no way they're making a one mana Surveil cantrip, it would be really good.
UBR Grixis Shadow UBR
UR Izzet Phoenix UR
UW UW Control UW
GB GB Rock GB
Commander
BG Meren of Clan Nel Toth BG
BGUW Atraxa, Praetor's Voice BGUW
Spirits
Legacy: UWR Miracles [https://deckstats.net/decks/44442/1092831-uwr-miracles-2]
yeah thats what i mean. surveil is close enough to scry in both effect and power level that you immediately know such a cantrip would be too good.
tbh id be surprised if they just made opt v.2.0: U (Instant) - Surveil 1, Draw 1
sorta unrelated, but with more graveyard centric mechanics on the horizon i hope they print a new piece of hate. not sure what new form it could take, but that is wizards job. hell make it red and give it to burn.
UWGSnow-Bant Control
BURGrixis Death's Shadow
GWBCoCo Elves
WCDeath and Taxes(sold)I'd honestly rather it be blue, if we give it to Red BridgeVine becomes more of a deck.
UBR Grixis Shadow UBR
UR Izzet Phoenix UR
UW UW Control UW
GB GB Rock GB
Commander
BG Meren of Clan Nel Toth BG
BGUW Atraxa, Praetor's Voice BGUW
Sorry, what? Are you talking about the Guild-specific decks? Because those are not Standard legal, per Maro:
vaatiwithblue asked: Why are the cards in the Guilds of Ravnica guild kits NOT legal?
Because it allowed us to make more robust decks. It’s hard to make Standard-legal guild decks when the two previous guild blocks aren’t Standard-legal.
Following up from this, I wanted to add another point of clarification. It's tempting to view this exclusively as "Deck X is twice as bad as Deck Y," as in the Bant Spirits vs. Humans/Jeskai example I made in this post. This might be technically accurate but it misses on deck viability because it undersells the viability of all but the best performing decks.
A much more useful way to read this list is to compare relative viability of any given deck to a known tiered benchmark For example, we know Humans is an accepted and indisputable Tier 1 deck by almost every possible metric we can think of. On this list, it "only" scores 2.19. We could conclude that it's twice as "bad" as Bant Spirits, but that's pretty disingenuous because we already know Humans isn't bad at all. So this negative framing doesn't advance our understanding of the format. A much better conclusion would be that every deck that performed better than Humans is at least as viable as Humans itself. That means all the Top 10 decks could be considered "as good or better than" Humans in Modern, and would all be great Modern deck choices.
Similarly, the decks that are right below Humans are also probably just as viable as Humans (but probably not better). GDS, Hollow One, and Jund (Jund to a lesser extent) are pretty close to the Humans viability score, which suggests they are roughly as viable as Humans in a major event. The only decks that truly appear deficient in some way are the decks that scored 3 or lower. One of these is very surprising (Mardu Pyro), because there are a few pilots who did extremely well with Mardu over the past months. I have a theory about how to adjust for this by drawing on other GP and will do more work later.
I'm kind of annoyed at this because they did move away from duel decks because of mixing standard legal cards with rotated cards, resulting in casual players who don't know any better to get told their decks are not tournament legal. The only reason the guild decks aren't standard legal is probably Deathrite shaman being in the same format as Llanowar Elves.
1. (Ravnica Allegiance): You can't keep a good esper control deck down... Or Wilderness Reclamation... or Gates...
2. (War of the Spark): Guys, I know what we need! We need a cycle of really idiotic flavor text victory cards! Jace's Triumph...
3. (War of the Spark): Lets make the format with control have even more control!
So you might get UR do two damage to any one target, jumpstart if jumpstarted it deals 3 damage to any one target instead. Or is it confirmed you get the same mana cost + discard and the exact same affect.
1. Hardened Scales: 1
2. Counters Company: 1.26
3. KCI: 1.47
4. Bant Spirits: 1.65
5. UW Spirits: 1.71
6. Infect: 1.77
7. UW Control: 1.91
8. Jeskai Control: 1.98
9. Bridgevine: 2.24
10. Elves: 2.38
11. GDS: 2.44
12. Humans: 2.5
13. Gx Tron: 2.53
14. Hollow One: 2.71
15. Bogles: 2.74
16. Jund: 2.76
17. Burn: 2.77
18. Mardu Pyromancer: 3
19. Affinity: 3
20. Storm: 3.55
21. Titanshift: 3.62
Titanshift and Storm are still terrible in this metagame for some reason. KCI is slightly "better," UW Control is slightly "worse," etc. Remember that this isn't the best way to read the numbers. It is a fundamental misreading to say that UW Control is almost twice as "bad" as Hardened Scales, because the performance score itself doesn't really indicate a strictly objective performance. It's an average of the rankings of different performance metrics. Because of that, the best way to review these is to pick a known, viable, Tier 1 benchmark and compare relative performance. In this case, Humans is our benchmark in 12th place. Any deck that has a higher rating than Humans is at least as viable as Humans, likely slightly more. Any deck that has a lower rating than Humans is likely less viable, with a big dropoff between something like Gx Tron (which is barely less viable) to Titanshift (which is still presenting as really bad).
The only strategy that appears to be struggling in this metagame is midrange, with both Jund and Mardu performing worse than Humans. GDS, at least, appears to be at least as viable as Humans.
UR ....... WUBR ........... WB ............. RGW ........ UBR ....... WUB .... BGU
Spells / Blink & Combo / Token Grind / Dino Tribal / Draw Cards / Zombies / Reanimate
I was one by trade in my last job, now I'm just doing some fun stats as a hobby.
Blue Moon does show up in the datasets, but N is too small to make any conclusions about the deck. This is actually true of a lot of different decks and is a major limitation of this analysis project.
great work ktkenshinx. thanks for taking the time to do that.
so just to clarify, these rankings are a composite based on the following(?):
-match win percentage
-overall win rate
-day 2 conversion (from day 1)
-top 32 conversion (from day 2)
-top 16 conversion (from day 2)
with some adjustments for byes. its worth pointing out that this ranking has nothing to do with representation. so you are placing decks with multiple high profile finishes next to ones with a single finish. the distinction between something like overall deck performance, and performance at certain event(s) might not be clear.
for example elves. your analysis would conclude that elves is at least a good/viable/whatever as humans, if not more. you dont exactly have to look far to see that this is demonstrably not true.
cant say im fond of so much use of the V-word. mostly because its connotation as a buzz word to placate the community regarding relatively poor performing decks (ie. 'oh your deck might not be tier 1, but its viable).
UWGSnow-Bant Control
BURGrixis Death's Shadow
GWBCoCo Elves
WCDeath and Taxes(sold)I refuse to believe I'm as good at Magic, as my Blue Moon record would indicate. The deck is good.
Spirits
Specifically, MWP at GP Prague, MWP pooled between GP Barcelona/Sao Paulo/Vegas (it's really hard to separate events for distinct GP MWPs in Erzel's data). For the conversions, it's just GP Prague because that's the only GP where we have that data from Day 1 to Day 2. It also had some representation-based T32 data, but I'm finding that data to be troubling so I am removing it.
Yes and no. It's not representation based in a direct sense, but it only includes decks that have a big enough N to work with. So it's indirectly representation based so we can have ~100 matches in a sample.
This is an issue and suggests the analysis method may have issues. I'm trying to separate out the other GP events for distinct conversions.