You do realize the same data that was available a year ago is available today?
The change in the tier structure has nothing to do with data availability. As other people have hinted at... it looks like a change of mods.
So you're saying that the data Ktkenshinx worked painfully to acquire, assess and contextualize was plainly available for all to see, and continues to be plainly available despite the fact that he isn't doing that anymore.
Yeah I can see why we disagree. You are underestimating how much work Ktkenshinx put in (for free, mind you) and are now demanding that someone else do that job for free. Ok.
Again, I ask: did you volunteer?
A) data collection is ALWAYS time consuming, regardless of how accessible the information is.
B) I dislike your straw-man argument. You are basically saying if I don't want to volunteer don't complain if it becomes worthless.
C) I am not complaining. Just pointing out if the changes go through, the tier system becomes far less relevant.
My only issue with the tier system is its weighed heavily on large weekend events that the majority of players will never play in that meta. Its very misleading to someone looking for a deck to play. I cant tell you how many times I have seen some newer player go out and buy the flavor of the month deck and show up to a local event and go 0-2 drop and be pissed this top deck failed them. Its one thing for a company to mislead players, its something much different (and in my opinion worse) when the player base who creates these tier systems mislead fellow players.
Wow, this seems really off base. There's no feasible way we can compile, nor account for local metas. What goes 4-0 at Jim Bob's Card shop and what top 8's a GP are two very different fields. Local metas can drastically different from the metagame at large. The only way we could even incorporate something that encompassed each and every local metagame. Possibility Storm is a deck that's taken FNMs in my local meta. Does that mean we should include that as a Tier 1 deck?
Another wide misconception seems to be that building a "Tier 1" deck should equate to immediate success at any event. Modern is a deeper format than that, and even building a top Tier deck requires practice and knowledge of metas (both local and overall) to have success. Tiers should be established based on proven results... which means measuring the stats that we can measure.
I think the methods used to determine Tiering during Ktkenshin's run of things was perhaps as good as we're going to get without more readily available data.
I think this is a fight we are just going to have to give up on. If someone thinks the tier system is a bad way to judge decks because it doesn't accurately describe their local meta then there isn't any combination/order of words we can put together to change their perspective. Their thought is so fundamentally far from what we are actually discussing to even try to engage it.
If you only play a local meta, then why do you ever look at tiers? And if you are a new player, you should lose more often than not regardless of what deck you play; it is the nature of the format. Again, I see no point in making the tier system in order to appeal to locals and newer players because the information is of the least use to those groups even when the information is geared specifically toward them.
We are changing something to better suite the people who don't want/need it but making it worse for the people it actually applies to. It's dumb.
I'd just like to see a consistent set of standards applied to tiering decks that won't change on a whim when a new person is in charge. So until we get some kind of explanation from Lantern for the change we can all enjoy the speculation game. I'm sure the powers that be behind this website love it b/c it generates comments and views for their ads so maybe it's being done on purpose for that reason alone.
I'm not actually sure that is what you want though, because that is exactly what we had and you attacked Ktkenshinx for it. The previous system was very good at accurately determining what decks you were most likely to face, and it was pretty resilient to fad spikes like Jeskai Nahiri as well as decently resilient to the "expensive to change your deck overnight" mentality. Tiers are fluid and constantly changing shape but that doesn't mean we should have the forum change every day or minute.
What you describe as wanting is what we already had, a structured, formulated, albeit far from perfect, way of using data to measure decks that removed preference bias, player skill, and match ups. I'm mostly referring to this part of your quote
I'd just like to see a consistent set of standards applied to tiering decks that won't change on a whim when a new person is in charge.
Ktkenshinx most likely had nothing to do with how the new system was implemented or what changes took place, how is he at all responsible for how the tiers look now because the rules were changed afterwards? whether or not a tier system changes when a new person is in charge is completely up to the new person in charge and what changes they make. I just don't see what you are actually after here. What could ktkenshinx, or anyone else for that matter, possibly done to prevent future people from changing their rules? Have a read only excel file?
Seriously, how is anyone supposed to enforce his set of standards on new people that have the power to change or create any standard they want?
Previous system was Sheridan's opinion of how the tiers should be arranged. Current system (until an explanation is given) is Lantern's opinion on how they should be arranged. A website or forum standard coming down from the top means that the tiering system does not see such a significant change when one person steps down and another steps into their place. It's fine now I see the tiers are nothing more than one person's opinion. Who that opinion belongs to doesn't concern me I know I'm only getting a picture as only one person wants me to see it.
Private Mod Note
():
Rollback Post to RevisionRollBack
FREE MODERN. Break the Standard link.
I play Magic: the Gathering, not Magic: the Commandering.
Looking at real life finishes. I have a hunch that the current tier system is putting more emphasis on MTGO.
Nearly all the new decks that have made Tier 1 status have seen a recently spiked on MTGO only.
If my hunch is correct... that is potentially very dangerous. Using MTGO to define "tiers" will lead to a higher number of tier 1.5 - 2 decks being classified as tier 1.
No, using MTGO won't lead to higher number of tiered decks, using a lower cutoff will. The problem is not using MTGO data itself, the problem is changing the cutoff in such a way that we have +10 decks in each tier. Since Lantern isn't giving any explanation, I'll give it a shot. From the Sultai thread:
This deck has been popping up enough online to barely make tier 2. As such we should have a thread for it.
We know that MtgGoldfish has it at 0.74%, MtgTop8 has it at 1% and my own data has it at 1% too, so probably Lantern has it around those numbers. We can assume the Tier 2 cutoff is ~1% now, which is a problem. Even if all our data was from paper and it had Sultai at 1%, it would be a problem having it at Tier 2 given that cutoff. Regarding the other decks you mention, I have:
Dredge - 7.7%
UR Storm - 6.7%
Bant Company - 4.7%
Wx Taxes - 3.7%
WU Control - 3.3%
Bant Eldrazi - 3.0%
While I agree with you Wx Taxes and WU Control have no business being Tier 1, your hunch may be wrong about Storm and the Bant decks. Also, this would mean the new Tier 1 cutoff is just over ~3.0%. So, I guess that the new system is, roughly, Tier 1 for +3% decks and Tier 2 for +1% decks. I'd personally use 4.4% and 2.7%, but oh well
Given this current discussion, I figured I might as well just publicly share my stats, and let everyone use their own cutoff with the data, instead of posting the numbers every ~10 days. You can Click Here anytime (also on my signature). Feedback is welcome guys!
Private Mod Note
():
Rollback Post to RevisionRollBack
Modern:WU WU Control | WBG Abzan Company Frontier:UBR Grixis Control | BRG Jund Delirium
Looking at real life finishes. I have a hunch that the current tier system is putting more emphasis on MTGO.
Nearly all the new decks that have made Tier 1 status have seen a recently spiked on MTGO only.
If my hunch is correct... that is potentially very dangerous. Using MTGO to define "tiers" will lead to a higher number of tier 1.5 - 2 decks being classified as tier 1.
No, using MTGO won't lead to higher number of tiered decks, using a lower cutoff will. The problem is not using MTGO data itself, the problem is changing the cutoff in such a way that we have +10 decks in each tier. Since Lantern isn't giving any explanation, I'll give it a shot. From the Sultai thread:
This deck has been popping up enough online to barely make tier 2. As such we should have a thread for it.
We know that MtgGoldfish has it at 0.74%, MtgTop8 has it at 1% and my own data has it at 1% too, so probably Lantern has it around those numbers. We can assume the Tier 2 cutoff is ~1% now, which is a problem. Even if all our data was from paper and it had Sultai at 1%, it would be a problem having it at Tier 2 given that cutoff. Regarding the other decks you mention, I have:
Dredge - 7.7%
UR Storm - 6.7%
Bant Company - 4.7%
Wx Taxes - 3.7%
WU Control - 3.3%
Bant Eldrazi - 3.0%
While I agree with you Wx Taxes and WU Control have no business being Tier 1, your hunch may be wrong about Storm and the Bant decks. Also, this would mean the new Tier 1 cutoff is just over ~3.0%. So, I guess that the new system is, roughly, Tier 1 for +3% decks and Tier 2 for +1% decks. I'd personally use 4.4% and 2.7%, but oh well
Given this current discussion, I figured I might as well just publicly share my stats, and let everyone use their own cutoff with the data, instead of posting the numbers every ~10 days. You can Click Here anytime (also on my signature). Feedback is welcome guys!
Like I said: that's once you factor in MTGO. Should MTGO define the tiers then? that's basically what is being argued.
Dredge, Storm, and Bant company fall to 3% or less when you look at real life results only.
For example, UR Storm has a total goldfish representation of 6.7%. On average 53% of those decks are MTGO.
How is that not skewing the data?
And if the person in charge of the site changes then they are free to change it again... There is literally, absolutely, positively, unattainably, no way to implement what you are suggesting. No one can force, or enforce, their metrics for defining a meta on the future. From what I can tell it was much closer to that before than it is now.
The tiers were not his or anyone else's opinion, and it isn't that now. People use metrics to determine percentages and then have cut offs of different percentages to rank tiers. When those metrics and cut offs are clearly defined and not changing then you cannot seriously call it an opinion. The only "opinion" was on how to collect data and what cut offs to use and you can argue that those weren't done well, but when those metrics have been in place for years you simply cannot state that anyone had a bias toward a certain deck when the deck continues in tier 1 because the metrics were defined without future results being available. It's pretty hard to build a statistical set to fit your agenda when you don't know what numbers you need to build around.
I hope the site has what it needs to continue the great job that has been done here.
My only issue with the tier system is its weighed heavily on large weekend events that the majority of players will never play in that meta. Its very misleading to someone looking for a deck to play. I cant tell you how many times I have seen some newer player go out and buy the flavor of the month deck and show up to a local event and go 0-2 drop and be pissed this top deck failed them. Its one thing for a company to mislead players, its something much different (and in my opinion worse) when the player base who creates these tier systems mislead fellow players.
Wow, this seems really off base. There's no feasible way we can compile, nor account for local metas. What goes 4-0 at Jim Bob's Card shop and what top 8's a GP are two very different fields. Local metas can drastically different from the metagame at large. The only way we could even incorporate something that encompassed each and every local metagame. Possibility Storm is a deck that's taken FNMs in my local meta. Does that mean we should include that as a Tier 1 deck?
Another wide misconception seems to be that building a "Tier 1" deck should equate to immediate success at any event. Modern is a deeper format than that, and even building a top Tier deck requires practice and knowledge of metas (both local and overall) to have success. Tiers should be established based on proven results... which means measuring the stats that we can measure.
I think the methods used to determine Tiering during Ktkenshin's run of things was perhaps as good as we're going to get without more readily available data.
I think this is a fight we are just going to have to give up on. If someone thinks the tier system is a bad way to judge decks because it doesn't accurately describe their local meta then there isn't any combination/order of words we can put together to change their perspective. Their thought is so fundamentally far from what we are actually discussing to even try to engage it.
If you only play a local meta, then why do you ever look at tiers? And if you are a new player, you should lose more often than not regardless of what deck you play; it is the nature of the format. Again, I see no point in making the tier system in order to appeal to locals and newer players because the information is of the least use to those groups even when the information is geared specifically toward them.
We are changing something to better suite the people who don't want/need it but making it worse for the people it actually applies to. It's dumb.
The Tier system should be focused on the metas the largest group of players will play in, not the least.
As I have said, basing the tier system on the finishes on SCG and other big weekend events, the events the majority of players will NEVER compete in is misleading to those who play in a different meta.
All I am asking for is a weighted system and an explanation to NEW players what the tier system is and how the numbers are calculated. So they dont go out and spend $1000+ on a deck and get crushed in a meta set up to destroy said deck and get salty.
You may be a grinder or someone who plays for a PT or GP spot, but you are in the minority of the player base.
What's the problem with MTGO data ? Large paper events are few and far between and as such using exclusively paper data doesn't do anything by itself. The last modern GP excluding the team event was in mid February, and the next one is at the end of May. The last SCG modern open was in mid March and the most recent big paper modern event was the MKM modern event at the end of April.
All those paper event data would be long obsolete by the time the next big paper Modern event rolls around and this is where the MTGO data comes in. Neither are conclusive by themselves, they have to complement each other in order to produce the clearest picture.
Guys please opinions on japanese cards. Lost a 3/3 creature against Japan celestial colonade. This guy played all creatures and spells in english cards, but some cards in his manabase was japanese. I dont registrated this really ( my brain say its all fine and all english to me lets attack his empty board)...and i am sure it is a Kind of legal cheating. It is not ok, but i know legal. I Hate such people. I never forget colonade normally, but with this Tricks it can happen one time in 3 years and such people take advantage of this
If I am a customer spending premium amount of dollars, I expect a premium service. Jund falls into the category of a premium deck costing more dollars than a majority of the rest of the format. I'm not getting the desired performance ratio per dollars spent out of the Jund deck because WOTC decided to make the format more diverse.
What's the problem with MTGO data ? Large paper events are few and far between and as such using exclusively paper data doesn't do anything by itself. The last modern GP excluding the team event was in mid February, and the next one is at the end of May. The last SCG modern open was in mid March and the most recent big paper modern event was the MKM modern event at the end of April.
All those paper event data would be long obsolete by the time the next big paper Modern event rolls around and this is where the MTGO data comes in. Neither are conclusive by themselves, they have to complement each other in order to produce the clearest picture.
There are nothing wrong with MTGO, per say. Even the previous tier system used MTGO to define tiers.
There are a few concerns with using MTGO as a major tier indicator; however.
1) The MTGO meta can be vastly different than paper. This means what is good online... could be terrible in person.
2) The MTGO meta has more variance, flavor of the month movement. Additionally cheap "good decks" can occasionally spike making it look better than it actually is. (see UR storm).
3) MTGO data got reduced. It only shows 5-0 decks now. This paints a very narrow image which may not capture the entire meta.
As a result... if you give MTGO an equal wait... it will skew the tiers.
Like I said: that's once you factor in MTGO. Should MTGO define the tiers then? that's basically what is being argued.
I thought all this discussion was about the tiering system change and how now have +10 decks in each tier. You said that "using MTGO to define "tiers" will lead to a higher number of tier 1.5 - 2 decks being classified as tier 1." and I was explaining that this isn't a problem of MTGO data itself, but rather a problem with the tier cutoffs. LIke I said, if all our data was paper data and Sultai was at 1.0% and WU Control was 3.3%, it would still be problematic to have Sultai at Tier 2 and WU Control at Tier 1. What many people are criticizing (+10 decks in each tier "distort" the tiers) is happening because the new moderator in charge is likely using new, lower cutoffs for the tiers.
Regarding your other concern, you said yourself that you "want to use whatever tier system was used before". What we used before was Modern Nexus's tiers and the problem was that the forum's tiers depended on Nexus's updates. Then Sheridan sold/gave it to the QuietSpeculation guy and the Nexus site went downhill, including 2/3 weeks of "lag" and months with no update at all, which in turn affected Salvation's tiers as well. My guess is that Lantern is using another site to get the data AND using another tiering cutoff.
That tier system you want back already used MTGO data (Modern Nexus did and still does) and other sites also do use MTGO data (MtgGoldfish, MtgTop8, the MainPhase guys that tracks MtgGoldfish weekly, etc). MTGO has always defined the tiers to some extent, and there's no way around it. What data do you want to use? 8 GPs a year and SCG Opens and Classics? Do we have any other way to get data of successful decks on a daily basis? I personally think MTGO data addresses 2 issues I have with paper data: geographic metagame skew and the difficulty to change decks. I'll be the first to acknowledge its shortcomings as well, mostly the ~4.4% sample error we have to live with
Private Mod Note
():
Rollback Post to RevisionRollBack
Modern:WU WU Control | WBG Abzan Company Frontier:UBR Grixis Control | BRG Jund Delirium
Like I said: that's once you factor in MTGO. Should MTGO define the tiers then? that's basically what is being argued.
I thought all this discussion was about the tiering system change and how now have +10 decks in each tier. You said that "using MTGO to define "tiers" will lead to a higher number of tier 1.5 - 2 decks being classified as tier 1." and I was explaining that this isn't a problem of MTGO data itself, but rather a problem with the tier cutoffs. LIke I said, if all our data was paper data and Sultai was at 1.0% and WU Control was 3.3%, it would still be problematic to have Sultai at Tier 2 and WU Control at Tier 1. What many people are criticizing (+10 decks in each tier "distort" the tiers) is happening because the new moderator in charge is likely using new, lower cutoffs for the tiers.
Regarding your other concern, you said yourself that you "want to use whatever tier system was used before". What we used before was Modern Nexus's tiers and the problem was that the forum's tiers depended on Nexus's updates. Then Sheridan sold/gave it to the QuietSpeculation guy and the Nexus site went downhill, including 2/3 weeks of "lag" and months with no update at all, which in turn affected Salvation's tiers as well. My guess is that Lantern is using another site to get the data AND using another tiering cutoff.
That tier system you want back already used MTGO data (Modern Nexus did and still does) and other sites also do use MTGO data (MtgGoldfish, MtgTop8, the MainPhase guys that tracks MtgGoldfish weekly, etc). MTGO has always defined the tiers to some extent, and there's no way around it. What data do you want to use? 8 GPs a year and SCG Opens and Classics? Do we have any other way to get data of successful decks on a daily basis? I personally think MTGO data addresses 2 issues I have with paper data: geographic metagame skew and the difficulty to change decks. I'll be the first to acknowledge its shortcomings as well, mostly the ~4.4% sample error we have to live with
The previous tier system (correct me if I am wrong) weighted MTGO data. It did not simply take the overall % meta and compare it to a single cutoff.
A weighted system would be best. Otherwise FotM decks on MTGO (such as UW control and UR storm) will be classified as tier 1 always.
Also - the previous system also required good conversion rates in person before being tiered.
Annnnnd why is the new Mod in charge using new, lower cutoffs for the tiers? Because in his opinion that's what should be done, an opinion not shared obviously by the previous Mod. Hence you see how opinions, despite the use of statistical data, do in fact influence the picture of the format presented to us by the two respective Mods. I imagine the real truth, as it always is w/ diverging opinions, is somewhere in between.
Private Mod Note
():
Rollback Post to RevisionRollBack
FREE MODERN. Break the Standard link.
I play Magic: the Gathering, not Magic: the Commandering.
Annnnnd why is the new Mod in charge using new, lower cutoffs for the tiers? Because in his opinion that's what should be done, an opinion not shared obviously by the previous Mod. Hence you see how opinions, despite the use of statistical data, do in fact influence the picture of the format presented to us by the two respective Mods. I imagine the real truth, as it always is w/ diverging opinions, is somewhere in between.
If that is true... what is the point of a tier system?
We might as well just convert it into Proven, Established, and Developing like the other formats.
There are nothing wrong with MTGO, per say. Even the previous tier system used MTGO to define tiers.
There are a few concerns with using MTGO as a major tier indicator; however.
1) The MTGO meta can be vastly different than paper. This means what is good online... could be terrible in person.
2) The MTGO meta has more variance, flavor of the month movement. Additionally cheap "good decks" can occasionally spike making it look better than it actually is. (see UR storm).
3) MTGO data got reduced. It only shows 5-0 decks now. This paints a very narrow image which may not capture the entire meta.
As a result... if you give MTGO an equal wait... it will skew the tiers.
1)This is true, and it goes both ways. You can't go to +999 life with Abzan Company online before losing to time, for example.
2)I really don't know how you ever measured this. While cheap decks like Dredge, Storm and Burn have good numbers, expensive deck like Jund Shadow or Abzan Midrange also have good numbers. Decks do well regardless of pricetag, don't worry.
3) MTGO data actually increased. We used to have 4-0 and 3-1 data from dailies, around 6 decks a day. With competitive leagues now we have 10 decks a day. More importantly, the data always doesn't capture the entire meta. We don't get the online 3-2s and 1-4s as much as we don't get the paper 10-5s and 6-9s. The data always captures what I call successful metagame, rather than the actual, Round 0 metagame. And yes, the data always painst a very narrow image that does not capture the entire meta.
The previous tier system (correct me if I am wrong) weighted MTGO data. It did not simply take the overall % meta and compare it to a single cutoff.
A weighted system would be best. Otherwise FotM decks on MTGO (such as UW control and UR storm) will be classified as tier 1 always.
Also - the previous system also required good conversion rates in person before being tiered.
As far as I know, Modern Nexus weighs online, paper and big events TopX to get their "overall metagame". Decks are awarded points on each of those "sub-metagames" when compared to the cutoffs, which are the average interval + X standard deviations. So a deck could be Tier 1 points-wise even if its overal metagame number was lower than that of a Tier 2 deck. I don't know where does this conversion rate requirement comes from. Source?
Regarding WU Control and UR Storm, both decks have been going up steadily since February, rather than multiple ups and downs. To me, consistently improving success each month is hardly what you expect from a "flavor of the month" deck. Also, people shouldn't be so dimissive of decks with online success. Remember that the paper metagame, due to a couple reasons, is much slower at changing and adapting. Then people act suprised when Jund Shadow takes a GP by storm becasue a pro went to Twitch and talked to the grinders that had been playing the deck for months I'm sure the paper metagame will start adopting Storm, at least. Some numbers:
WU Control - Feb: 1.1%, Mar: 2.0%, Apr: 3.3%
UR Storm - Feb: 2.5%, Mar: 3.7%, Apr: 6.7%
Private Mod Note
():
Rollback Post to RevisionRollBack
Modern:WU WU Control | WBG Abzan Company Frontier:UBR Grixis Control | BRG Jund Delirium
Looking at real life finishes. I have a hunch that the current tier system is putting more emphasis on MTGO.
Nearly all the new decks that have made Tier 1 status have seen a recently spiked on MTGO only.
If my hunch is correct... that is potentially very dangerous. Using MTGO to define "tiers" will lead to a higher number of tier 1.5 - 2 decks being classified as tier 1.
No, using MTGO won't lead to higher number of tiered decks, using a lower cutoff will. The problem is not using MTGO data itself, the problem is changing the cutoff in such a way that we have +10 decks in each tier. Since Lantern isn't giving any explanation, I'll give it a shot. From the Sultai thread:
This deck has been popping up enough online to barely make tier 2. As such we should have a thread for it.
We know that MtgGoldfish has it at 0.74%, MtgTop8 has it at 1% and my own data has it at 1% too, so probably Lantern has it around those numbers. We can assume the Tier 2 cutoff is ~1% now, which is a problem. Even if all our data was from paper and it had Sultai at 1%, it would be a problem having it at Tier 2 given that cutoff. Regarding the other decks you mention, I have:
Dredge - 7.7%
UR Storm - 6.7%
Bant Company - 4.7%
Wx Taxes - 3.7%
WU Control - 3.3%
Bant Eldrazi - 3.0%
While I agree with you Wx Taxes and WU Control have no business being Tier 1, your hunch may be wrong about Storm and the Bant decks. Also, this would mean the new Tier 1 cutoff is just over ~3.0%. So, I guess that the new system is, roughly, Tier 1 for +3% decks and Tier 2 for +1% decks. I'd personally use 4.4% and 2.7%, but oh well
Given this current discussion, I figured I might as well just publicly share my stats, and let everyone use their own cutoff with the data, instead of posting the numbers every ~10 days. You can Click Here anytime (also on my signature). Feedback is welcome guys!
Like I said: that's once you factor in MTGO. Should MTGO define the tiers then? that's basically what is being argued.
Dredge, Storm, and Bant company fall to 3% or less when you look at real life results only.
For example, UR Storm has a total goldfish representation of 6.7%. On average 53% of those decks are MTGO.
How is that not skewing the data?
Why are online results invalid?
Why are real life results more valid than online ones?
Certainly, stats can be skewed in one environment versus the other (environmental bias). But it seems as though MTGO is useful as a data point combined with real life results. The biggest negative I can see would be tilting the results towards an online metagame since there are more events there, and the shorter rounds can favor inconsistent but powerful decks in the short term. (4c Loam is a solid deck for a 5 round tournament, substantially less so in a 12 round one).
With enough data points across both (or weighting for a more complex model), we should be able to minimize error bars from environmental bias.
I did not realize they changed the tier structure. The new structure... is abysmal.
Tiers should be based on both a combination of popularity + consistency + top table performance.
It looks like this site shifted to a popularity rating only. Which is BAD IMO.
UR storm is seriously a tier 1 deck while Jund is not? 0_o This new "ranking" system is awful.
It really should go back to the way it was before.
...Storm is putting up better results (online, at least) than Jund is. Why not have it tier 1?
Online is also theoretically a completely different meta than paper tournaments. Moreover - UR storm is SUPER cheap for a modern deck (on MTGO). Which means a higher frequency of players are playing it.
Now UR storm "could" be tier one again. But we don't actually have conversion statistics to back that up (unless we convert the tier system into simply a popularity contest). I would say UR storm is probably tier 1.5 at the moment. We need more data before we can actually say it's tier 1.
I would say that online is a better representation of what is actually good as it is trivial for most MTGO players to buy/sell out of and into a new deck at a over all cheaper rate than in paper. When DS broke out at the which ever GP its was( the event name escapes me atm) a deck which I had played against maybe 2 times in 3 weeks the very next day 6 out of 10 match ups where DSjund; it is essentially impossible to see this kind of shift in paper as the card price/scarcity is way higher and you have to wait for arrival of products which can take weeks while online it takes minutes.
I don't see how in the same post you can say that UR storm isn't tiered properly because its numbers are artificially bolstered due to its price tag, and then bring up conversion rates. They are connected if more people play 8-wack 8-wack will have higher conversion rates just due to variance in match ups.
There are nothing wrong with MTGO, per say. Even the previous tier system used MTGO to define tiers.
There are a few concerns with using MTGO as a major tier indicator; however.
1) The MTGO meta can be vastly different than paper. This means what is good online... could be terrible in person.
2) The MTGO meta has more variance, flavor of the month movement. Additionally cheap "good decks" can occasionally spike making it look better than it actually is. (see UR storm).
3) MTGO data got reduced. It only shows 5-0 decks now. This paints a very narrow image which may not capture the entire meta.
As a result... if you give MTGO an equal wait... it will skew the tiers.
1)This is true, and it goes both ways. You can't go to +999 life with Abzan Company online before losing to time, for example.
2)I really don't know how you ever measured this. While cheap decks like Dredge, Storm and Burn have good numbers, expensive deck like Jund Shadow or Abzan Midrange also have good numbers. Decks do well regardless of pricetag, don't worry.
3) MTGO data actually increased. We used to have 4-0 and 3-1 data from dailies, around 6 decks a day. With competitive leagues now we have 10 decks a day. More importantly, the data always doesn't capture the entire meta. We don't get the online 3-2s and 1-4s as much as we don't get the paper 10-5s and 6-9s. The data always captures what I call successful metagame, rather than the actual, Round 0 metagame. And yes, the data always painst a very narrow image that does not capture the entire meta.
The previous tier system (correct me if I am wrong) weighted MTGO data. It did not simply take the overall % meta and compare it to a single cutoff.
A weighted system would be best. Otherwise FotM decks on MTGO (such as UW control and UR storm) will be classified as tier 1 always.
Also - the previous system also required good conversion rates in person before being tiered.
As far as I know, Modern Nexus weighs online, paper and big events TopX to get their "overall metagame". Decks are awarded points on each of those "sub-metagames" when compared to the cutoffs, which are the average interval + X standard deviations. So a deck could be Tier 1 points-wise even if its overal metagame number was lower than that of a Tier 2 deck. I don't know where does this conversion rate requirement comes from. Source?
Regarding WU Control and UR Storm, both decks have been going up steadily since February, rather than multiple ups and downs. To me, consistently improving success each month is hardly what you expect from a "flavor of the month" deck. Also, people shouldn't be so dimissive of decks with online success. Remember that the paper metagame, due to a couple reasons, is much slower at changing and adapting. Then people act suprised when Jund Shadow takes a GP by storm becasue a pro went to Twitch and talked to the grinders that had been playing the deck for months I'm sure the paper metagame will start adopting Storm, at least. Some numbers:
Also just remembered that physical tournaments aren't themselves definitive sources of truth either. Regional or Geographic bias exists, and card availability is much more of a concern in real life. The broader the data set with the appropriate controls, the more reliable and less error prone it becomes.
I'm going to be blunt: I have serious concerns of how this site is now classifying what is tier 1 and tier 2.
There are several decks on the tier 1 list that are "questionable". This includes the following
- UW Control
- UR Storm
- Death and Taxes
- Knightfall
- Dredge
This is a list of questionable decks on tier 2
- Bant Eldrazi (probably should still be tier one)
- Sultai Delirium (not enough tournament finishes to be truly classified as a tier deck yet)
For me: Tier 1 has always been the decks to beat. While Tier 2 are viable (but under played) decks.
This new tier system... does not capture that at all.
I'm curious as to how the tierings are being done as well. Lantern has stated they are working on the tierings and and explanation to be published.
That being said, I'm also curious what metrics you are using to determine they do not belong in the tierings they are in now. How did you come to the determination that those specific decks do not belong.
Looking at real life finishes. I have a hunch that the current tier system is putting more emphasis on MTGO.
Nearly all the new decks that have made Tier 1 status have seen a recently spiked on MTGO only.
If my hunch is correct... that is potentially very dangerous. Using MTGO to define "tiers" will lead to a higher number of tier 1.5 - 2 decks being classified as tier 1.
Or it will drive the paper meta-game to play decks with more consistent results.
DSJund was a deck online way before it was ever a thing in paper, the online meta-game does have some oddities such as a stricter timer etc... but generally if a deck can put up consistent 5-0's it is a legit deck to be Tiered at 1. Online events are very heavily tilted towards spikes you don't grind leagues/events on MTGO with a deck that doesn't perform well unless someone else is paying your way.
It is a very trivial affair to trade out of one deck and buy into another online, this is not the case in paper. I know plenty of people who play irl magic and go to GP's etc... and many of them have 1 Modern deck and that is the deck they play regardless of how it is positioned because irl if is far more expensive and timely to change out of your deck. When DSjund broke out it was probably less than a half hour later that it went from a once every couple of days deck you would run into to a majority of match up deck, that isn't going to happen in paper ever.
I'm going to be blunt: I have serious concerns of how this site is now classifying what is tier 1 and tier 2.
There are several decks on the tier 1 list that are "questionable". This includes the following
- UW Control
- UR Storm
- Death and Taxes
- Knightfall
- Dredge
This is a list of questionable decks on tier 2
- Bant Eldrazi (probably should still be tier one)
- Sultai Delirium (not enough tournament finishes to be truly classified as a tier deck yet)
For me: Tier 1 has always been the decks to beat. While Tier 2 are viable (but under played) decks.
This new tier system... does not capture that at all.
I'm curious as to how the tierings are being done as well. Lantern has stated they are working on the tierings and and explanation to be published.
That being said, I'm also curious what metrics you are using to determine they do not belong in the tierings they are in now. How did you come to the determination that those specific decks do not belong.
Looking at real life finishes. I have a hunch that the current tier system is putting more emphasis on MTGO.
Nearly all the new decks that have made Tier 1 status have seen a recently spiked on MTGO only.
If my hunch is correct... that is potentially very dangerous. Using MTGO to define "tiers" will lead to a higher number of tier 1.5 - 2 decks being classified as tier 1.
Or it will drive the paper meta-game to play decks with more consistent results.
DSJund was a deck online way before it was ever a thing in paper, the online meta-game does have some oddities such as a stricter timer etc... but generally if a deck can put up consistent 5-0's it is a legit deck to be Tiered at 1. Online events are very heavily tilted towards spikes you don't grind leagues/events on MTGO with a deck that doesn't perform well unless someone else is paying your way.
It is a very trivial affair to trade out of one deck and buy into another online, this is not the case in paper. I know plenty of people who play irl magic and go to GP's etc... and many of them have 1 Modern deck and that is the deck they play regardless of how it is positioned because irl if is far more expensive and timely to change out of your deck. When DSjund broke out it was probably less than a half hour later that it went from a once every couple of days deck you would run into to a majority of match up deck, that isn't going to happen in paper ever.
Suicide Zoo was always a tier 2 deck in paper. Suicide is a really bad example because the spike is due to bannings... not the MTGO meta shifting over.
The Tier system should be focused on the metas the largest group of players will play in, not the least.
Why? If that is the case you are free to make a tier system for your local meta and explain it as your heart desires. There is absolutely no way to make a tier system that can accurately gauge the experience of every player. The best we can do is take an aggregate of those to build an overall picture. Which is what was being done.
Casual modern players don't care at all about tiers, they play what they want with little consideration for metagaming. Local only players don't need tiers because they already know what they are likely to face. If this "largest group" of players gains nothing from the change and it only hurts the smaller group, then why would you do it?
As I have said, basing the tier system on the finishes on SCG and other big weekend events, the events the majority of players will NEVER compete in is misleading to those who play in a different meta.
So what? If you already know the meta you play in then you don't need to consider tiers. I really want you to explain this bold part (I will probably regret it). What possible metric could you give someone that would accurately describe the meta THEY are going to walk into? Expecting a tier system that can change on a whim based on your location is absurd and the most unrealistic thing I have seen on this thread, and this is shortly after discussing a mental misstep unban. The tier system isn't meant for timmy and johnny and there is nothing that can be done to the system to make it so that it is. Feel free to tier your meta and spread it all you want, but the entire point of the tiers was to simulate if every person did that and you combined them.
All I am asking for is a weighted system and an explanation to NEW players what the tier system is and how the numbers are calculated. So they don't go out and spend $1000+ on a deck and get crushed in a meta set up to destroy said deck and get salty.
That is what we had. I don't remember the specifics but it was weighted and it was very transparent. ktkenshinx was always willing to discuss his parameters. The fact that his tiers changed more slowly actually prevented this "get crushed in this meta and be salty" thing you keep bringing up (which is a dumb argument in its own right) because if you attempt to change the tiers every day or minute or second you will constantly be lagging behind the current trends and actually be giving an even more false sense of the meta since you are claiming it is fluid and up to date.
New players are going to lose. I don't care what deck they buy. It's very noble of you but I have no idea why you are so concerned with the idiot that goes out and spends $1000 on a deck without doing a single bit of research. Even if you prevent that, those people are going to make way worse decisions than that in their lives. Good luck trying to stop all of them.
You may be a grinder or someone who plays for a PT or GP spot, but you are in the minority of the player base.
Casuals do not care about your tier system, nor will they ever. It cannot do them any good if they don't pay attention to it. It would be like trying to make bath and body works more appealing to football players; it doesn't make any sense.
What you are asking for is completely pointless, worse than what we had, and unattainable.
Can anyone explain to me how the Tier system has changed on this site? Does it have to do with Ktkenshinx leaving?
Apparently, Lantern has changed both the data source and the tiering cutoffs. Now it seems +1% decks are Tier 2 and +3% decks are Tier 1, roughyl. No official layout and/or explanation of changes by any mod so far.
Private Mod Note
():
Rollback Post to RevisionRollBack
Modern:WU WU Control | WBG Abzan Company Frontier:UBR Grixis Control | BRG Jund Delirium
I think this new tiering system better reflects the uncertainty we have of all decks true positioning in the meta. This will displease those who want a clear picture of the best of the best for better spiking purposes, I believe that the way Wizards obscures data prevents us from having the resources to give iron clad tiers the way people want.
How did Sheridan manage to do this before? I believe he used more strict cut-offs and created his own weighted system of balancing online to paper data, but with so much other data being withheld (true winrates and online metashares) this was essentially the most statistically sound and objective way of creating the illusion of solid clear cut tiers. Trying to do too much with too little.
It's been pointed out a few times but I still find it funny that Divas wake up call about mtgo dailies (she herself contributed 2% to UR aggro) went largely unnoticed.
The meta information is flawed, the emperor has no clothes.
Private Mod Note
():
Rollback Post to RevisionRollBack
Modern U Merfolk UB Tezzerator UB Mill
To post a comment, please login or register a new account.
A) data collection is ALWAYS time consuming, regardless of how accessible the information is.
B) I dislike your straw-man argument. You are basically saying if I don't want to volunteer don't complain if it becomes worthless.
C) I am not complaining. Just pointing out if the changes go through, the tier system becomes far less relevant.
Twitter: twitter.com/axmanonline
Stream: twitch.tv/axman
Current Decks
Modern: Affinity
Standard: BW Control
Legacy: Death and Taxes :symw::symr:
Vintage: NA
Previous system was Sheridan's opinion of how the tiers should be arranged. Current system (until an explanation is given) is Lantern's opinion on how they should be arranged. A website or forum standard coming down from the top means that the tiering system does not see such a significant change when one person steps down and another steps into their place. It's fine now I see the tiers are nothing more than one person's opinion. Who that opinion belongs to doesn't concern me I know I'm only getting a picture as only one person wants me to see it.
I play Magic: the Gathering, not Magic: the Commandering.
We know that MtgGoldfish has it at 0.74%, MtgTop8 has it at 1% and my own data has it at 1% too, so probably Lantern has it around those numbers. We can assume the Tier 2 cutoff is ~1% now, which is a problem. Even if all our data was from paper and it had Sultai at 1%, it would be a problem having it at Tier 2 given that cutoff. Regarding the other decks you mention, I have:
Given this current discussion, I figured I might as well just publicly share my stats, and let everyone use their own cutoff with the data, instead of posting the numbers every ~10 days. You can Click Here anytime (also on my signature). Feedback is welcome guys!
Frontier: UBR Grixis Control | BRG Jund Delirium
Like I said: that's once you factor in MTGO. Should MTGO define the tiers then? that's basically what is being argued.
Dredge, Storm, and Bant company fall to 3% or less when you look at real life results only.
For example, UR Storm has a total goldfish representation of 6.7%. On average 53% of those decks are MTGO.
How is that not skewing the data?
Twitter: twitter.com/axmanonline
Stream: twitch.tv/axman
Current Decks
Modern: Affinity
Standard: BW Control
Legacy: Death and Taxes :symw::symr:
Vintage: NA
The tiers were not his or anyone else's opinion, and it isn't that now. People use metrics to determine percentages and then have cut offs of different percentages to rank tiers. When those metrics and cut offs are clearly defined and not changing then you cannot seriously call it an opinion. The only "opinion" was on how to collect data and what cut offs to use and you can argue that those weren't done well, but when those metrics have been in place for years you simply cannot state that anyone had a bias toward a certain deck when the deck continues in tier 1 because the metrics were defined without future results being available. It's pretty hard to build a statistical set to fit your agenda when you don't know what numbers you need to build around.
I hope the site has what it needs to continue the great job that has been done here.
The Tier system should be focused on the metas the largest group of players will play in, not the least.
As I have said, basing the tier system on the finishes on SCG and other big weekend events, the events the majority of players will NEVER compete in is misleading to those who play in a different meta.
All I am asking for is a weighted system and an explanation to NEW players what the tier system is and how the numbers are calculated. So they dont go out and spend $1000+ on a deck and get crushed in a meta set up to destroy said deck and get salty.
You may be a grinder or someone who plays for a PT or GP spot, but you are in the minority of the player base.
All those paper event data would be long obsolete by the time the next big paper Modern event rolls around and this is where the MTGO data comes in. Neither are conclusive by themselves, they have to complement each other in order to produce the clearest picture.
There are nothing wrong with MTGO, per say. Even the previous tier system used MTGO to define tiers.
There are a few concerns with using MTGO as a major tier indicator; however.
1) The MTGO meta can be vastly different than paper. This means what is good online... could be terrible in person.
2) The MTGO meta has more variance, flavor of the month movement. Additionally cheap "good decks" can occasionally spike making it look better than it actually is. (see UR storm).
3) MTGO data got reduced. It only shows 5-0 decks now. This paints a very narrow image which may not capture the entire meta.
As a result... if you give MTGO an equal wait... it will skew the tiers.
Twitter: twitter.com/axmanonline
Stream: twitch.tv/axman
Current Decks
Modern: Affinity
Standard: BW Control
Legacy: Death and Taxes :symw::symr:
Vintage: NA
Regarding your other concern, you said yourself that you "want to use whatever tier system was used before". What we used before was Modern Nexus's tiers and the problem was that the forum's tiers depended on Nexus's updates. Then Sheridan sold/gave it to the QuietSpeculation guy and the Nexus site went downhill, including 2/3 weeks of "lag" and months with no update at all, which in turn affected Salvation's tiers as well. My guess is that Lantern is using another site to get the data AND using another tiering cutoff.
That tier system you want back already used MTGO data (Modern Nexus did and still does) and other sites also do use MTGO data (MtgGoldfish, MtgTop8, the MainPhase guys that tracks MtgGoldfish weekly, etc). MTGO has always defined the tiers to some extent, and there's no way around it. What data do you want to use? 8 GPs a year and SCG Opens and Classics? Do we have any other way to get data of successful decks on a daily basis? I personally think MTGO data addresses 2 issues I have with paper data: geographic metagame skew and the difficulty to change decks. I'll be the first to acknowledge its shortcomings as well, mostly the ~4.4% sample error we have to live with
Frontier: UBR Grixis Control | BRG Jund Delirium
The previous tier system (correct me if I am wrong) weighted MTGO data. It did not simply take the overall % meta and compare it to a single cutoff.
A weighted system would be best. Otherwise FotM decks on MTGO (such as UW control and UR storm) will be classified as tier 1 always.
Also - the previous system also required good conversion rates in person before being tiered.
Twitter: twitter.com/axmanonline
Stream: twitch.tv/axman
Current Decks
Modern: Affinity
Standard: BW Control
Legacy: Death and Taxes :symw::symr:
Vintage: NA
I play Magic: the Gathering, not Magic: the Commandering.
If that is true... what is the point of a tier system?
We might as well just convert it into Proven, Established, and Developing like the other formats.
Twitter: twitter.com/axmanonline
Stream: twitch.tv/axman
Current Decks
Modern: Affinity
Standard: BW Control
Legacy: Death and Taxes :symw::symr:
Vintage: NA
1)This is true, and it goes both ways. You can't go to +999 life with Abzan Company online before losing to time, for example.
2)I really don't know how you ever measured this. While cheap decks like Dredge, Storm and Burn have good numbers, expensive deck like Jund Shadow or Abzan Midrange also have good numbers. Decks do well regardless of pricetag, don't worry.
3) MTGO data actually increased. We used to have 4-0 and 3-1 data from dailies, around 6 decks a day. With competitive leagues now we have 10 decks a day. More importantly, the data always doesn't capture the entire meta. We don't get the online 3-2s and 1-4s as much as we don't get the paper 10-5s and 6-9s. The data always captures what I call successful metagame, rather than the actual, Round 0 metagame. And yes, the data always painst a very narrow image that does not capture the entire meta. As far as I know, Modern Nexus weighs online, paper and big events TopX to get their "overall metagame". Decks are awarded points on each of those "sub-metagames" when compared to the cutoffs, which are the average interval + X standard deviations. So a deck could be Tier 1 points-wise even if its overal metagame number was lower than that of a Tier 2 deck. I don't know where does this conversion rate requirement comes from. Source?
Regarding WU Control and UR Storm, both decks have been going up steadily since February, rather than multiple ups and downs. To me, consistently improving success each month is hardly what you expect from a "flavor of the month" deck. Also, people shouldn't be so dimissive of decks with online success. Remember that the paper metagame, due to a couple reasons, is much slower at changing and adapting. Then people act suprised when Jund Shadow takes a GP by storm becasue a pro went to Twitch and talked to the grinders that had been playing the deck for months I'm sure the paper metagame will start adopting Storm, at least. Some numbers:
Frontier: UBR Grixis Control | BRG Jund Delirium
Why are online results invalid?
Why are real life results more valid than online ones?
Certainly, stats can be skewed in one environment versus the other (environmental bias). But it seems as though MTGO is useful as a data point combined with real life results. The biggest negative I can see would be tilting the results towards an online metagame since there are more events there, and the shorter rounds can favor inconsistent but powerful decks in the short term. (4c Loam is a solid deck for a 5 round tournament, substantially less so in a 12 round one).
With enough data points across both (or weighting for a more complex model), we should be able to minimize error bars from environmental bias.
I would say that online is a better representation of what is actually good as it is trivial for most MTGO players to buy/sell out of and into a new deck at a over all cheaper rate than in paper. When DS broke out at the which ever GP its was( the event name escapes me atm) a deck which I had played against maybe 2 times in 3 weeks the very next day 6 out of 10 match ups where DSjund; it is essentially impossible to see this kind of shift in paper as the card price/scarcity is way higher and you have to wait for arrival of products which can take weeks while online it takes minutes.
I don't see how in the same post you can say that UR storm isn't tiered properly because its numbers are artificially bolstered due to its price tag, and then bring up conversion rates. They are connected if more people play 8-wack 8-wack will have higher conversion rates just due to variance in match ups.
The "conversion rate" was the TopX calculation.
Twitter: twitter.com/axmanonline
Stream: twitch.tv/axman
Current Decks
Modern: Affinity
Standard: BW Control
Legacy: Death and Taxes :symw::symr:
Vintage: NA
Or it will drive the paper meta-game to play decks with more consistent results.
DSJund was a deck online way before it was ever a thing in paper, the online meta-game does have some oddities such as a stricter timer etc... but generally if a deck can put up consistent 5-0's it is a legit deck to be Tiered at 1. Online events are very heavily tilted towards spikes you don't grind leagues/events on MTGO with a deck that doesn't perform well unless someone else is paying your way.
It is a very trivial affair to trade out of one deck and buy into another online, this is not the case in paper. I know plenty of people who play irl magic and go to GP's etc... and many of them have 1 Modern deck and that is the deck they play regardless of how it is positioned because irl if is far more expensive and timely to change out of your deck. When DSjund broke out it was probably less than a half hour later that it went from a once every couple of days deck you would run into to a majority of match up deck, that isn't going to happen in paper ever.
Suicide Zoo was always a tier 2 deck in paper. Suicide is a really bad example because the spike is due to bannings... not the MTGO meta shifting over.
Twitter: twitter.com/axmanonline
Stream: twitch.tv/axman
Current Decks
Modern: Affinity
Standard: BW Control
Legacy: Death and Taxes :symw::symr:
Vintage: NA
Why? If that is the case you are free to make a tier system for your local meta and explain it as your heart desires. There is absolutely no way to make a tier system that can accurately gauge the experience of every player. The best we can do is take an aggregate of those to build an overall picture. Which is what was being done.
Casual modern players don't care at all about tiers, they play what they want with little consideration for metagaming. Local only players don't need tiers because they already know what they are likely to face. If this "largest group" of players gains nothing from the change and it only hurts the smaller group, then why would you do it?
So what? If you already know the meta you play in then you don't need to consider tiers. I really want you to explain this bold part (I will probably regret it). What possible metric could you give someone that would accurately describe the meta THEY are going to walk into? Expecting a tier system that can change on a whim based on your location is absurd and the most unrealistic thing I have seen on this thread, and this is shortly after discussing a mental misstep unban. The tier system isn't meant for timmy and johnny and there is nothing that can be done to the system to make it so that it is. Feel free to tier your meta and spread it all you want, but the entire point of the tiers was to simulate if every person did that and you combined them.
That is what we had. I don't remember the specifics but it was weighted and it was very transparent. ktkenshinx was always willing to discuss his parameters. The fact that his tiers changed more slowly actually prevented this "get crushed in this meta and be salty" thing you keep bringing up (which is a dumb argument in its own right) because if you attempt to change the tiers every day or minute or second you will constantly be lagging behind the current trends and actually be giving an even more false sense of the meta since you are claiming it is fluid and up to date.
New players are going to lose. I don't care what deck they buy. It's very noble of you but I have no idea why you are so concerned with the idiot that goes out and spends $1000 on a deck without doing a single bit of research. Even if you prevent that, those people are going to make way worse decisions than that in their lives. Good luck trying to stop all of them.
Casuals do not care about your tier system, nor will they ever. It cannot do them any good if they don't pay attention to it. It would be like trying to make bath and body works more appealing to football players; it doesn't make any sense.
What you are asking for is completely pointless, worse than what we had, and unattainable.
RGTron
UGInfect
URStorm
WUBRAd Nauseam
BRGrishoalbrand
URGScapeshift
WBGAbzan Company
WUBRGAmulet Titan
BRGLiving End
WGBogles
Frontier: UBR Grixis Control | BRG Jund Delirium
If that is the case it may just be random which are shown, but a cautious observer could easily accuse them of cherry picking to falsify diversity.
How did Sheridan manage to do this before? I believe he used more strict cut-offs and created his own weighted system of balancing online to paper data, but with so much other data being withheld (true winrates and online metashares) this was essentially the most statistically sound and objective way of creating the illusion of solid clear cut tiers. Trying to do too much with too little.
It's been pointed out a few times but I still find it funny that Divas wake up call about mtgo dailies (she herself contributed 2% to UR aggro) went largely unnoticed.
The meta information is flawed, the emperor has no clothes.
U Merfolk
UB Tezzerator
UB Mill