cfusionpm wrote: ↑4 years ago
I will reiterate my same concerns the last time this came up; namely the methodology and how byes/drops were factored into it. As well as which players were used in the data set (IE someone could come in with 3 byes, lose 3 matches, and drop with an extremely misleading 3-3 record. I don't remember how that was weighted or if the opportunity cost of matches post-drop factored either).
Byes were weighted and found to have no impact between formats. Byes absolutely, 100% mattered in reaching T32, but not more/less in any format compared to any other format.
But more specifically to the illustration I was trying to make, regardless of the methodology, any individual game or match can and will be heavily dependent on the matchup. That's why we specifically say things like "how do I sure up some of my bad matchups?" There are simply decks that you are statistically more likely to lose to than win against, due to the construction of your deck and their deck. And if that discrepancy is high enough, you can see people fumble and stumble their way through misplays and sloppy decisions completely crush someone playing a terrible deck at its highest capabilities. Additionally, players are
heavily rewarded for playing decks with a highly swingy set of matchups instead of something "50/50" as shown in
this recent CFB article, so long as you can predict/dodge/minimize the poor matchups.
I may have been unclear in my last post but know I made this caveat when I wrote up the original analysis. The analysis just found that matchup variance had no impact comparing formats. It did not necessarily show matchup variance having an impact winning an individual event, independent of the format. I expect the CFB article accurately describes how matchup variance benefits top players, but that's true in all formats, not just Modern. All I know is that in Modern, there is no difference in MWP and top player performance than in any other format. The same problems that apply in all tournament Magic (e.g. maybe swingier decks are slightly better, Byes are hugely important, matchup variance may or may not determine games) applies equally to players playing Modern as players playing Legacy or Standard. Limited was excluded from this and might be less variable. But the Modern vs. Standard vs. Legacy outcomes were largely the same.
Perhaps I am skewed by the stuff I teach middle schoolers, especially about choices and representation in data sets. But what specifically seems to bug me is that the average statistics don't take into account how the games played out. For example one could quantify each of the wins and losses on some metric that demonstrates how swingy the game was (IE was someone crushed out of existence quickly and decisively, or was it a long, slow, grindfest of difficult decisions; not sure how to do this other than something like weighing each win and loss with a turn count, and making some compound between match/game win % and turns per game). I could imagine how that data spread (MAD/IQR) would look for different formats. Perhaps Standard looks much more clumped together and low in this metric, where Modern would be more spread across the gamut. Even if their averages ended up the same, the data distribution could be wildly different. I would give my left arm to be able to pull these kinds of numbers from MTGO and crunch them myself. And as much as I would love to do a grass roots effort like the turn-count hero, school starts up again in a month and I simply couldn't undergo such a project.
There was a slight difference in the Standard vs. Modern data on this measure. Namely, when looking at recent MWP and tournament finishes (N=320 Standard events, N=130 Modern events) for the top 40 players over the last 2-3 years (effective 2018), the overall tournament MWP distribution had some differences. Here are the bins for Standard and Modern, read as "When the top 40 players played this format, Y% of them finished with an X% MWP."
Standard tournament MWP vs. % of players achieving that MWP
0% MWP: 3.1%
10% MWP: 0%
20% MWP: 2.8%
30% MWP: 2.5%
40% MWP: 10%
50% MWP: 16.6%
60% MWP: 19.4%
70% MWP: 18.8%
80% MWP: 19.1%
90% MWP: 6.6%
100% MWP: 1.3%
Modern tournament MWP vs. % of players achieving that MWP
0% MWP: 1.6%
10% MWP: 0%
20% MWP: 0%
30% MWP: 1.6%
40% MWP: 15%
50% MWP: 10.2%
60% MWP: 18.9%
70% MWP: 20.5%
80% MWP: 22.8%
90% MWP: 7.9%
100% MWP: 1.6%
The biggest immediate difference is in the 40% vs 50% MWP bins. For 40% it was 15% (M) vs. 10% (S), and for 50%, it was 10.2% (M) vs. 16.6% (S). At first glance, we might think this means Modern is slightly swingier than Standard, but the cumulative bins at the top and bottom show different stories. 18.4% of Standard players finished with 40% MWP or less in their events. For Modern, it was an identical 18.1%. This is because there were actually more 0%, 20%, and 30% MWP finishes in Standard events than in Modern events. So at the bottom of the performance spectrum, there was no difference between formats. There was a slight difference at the top. Standard players went 60%+ MWP plus in 65% of events. For Modern players, they performed 60%+ MWP in 71.7% of events. The best players were actually more likely to have a higher Modern performance than a Standard one. They were also likelier to go 50% in Standard than Modern (10% Modern vs. 16.6% Standard).
All of this shows Modern may be slightly swingier than Standard,
but it benefits the top players! Also, there was no difference between the formats for sub-40% performance. The top player analysis suggests a good player will see slightly more 50/50 performances in Standard, but a proportionate slight increase in 60%+ performances in Modern.
Finally, in case anyone cares about averages, the average MWP for top players in Standard events was 58.4%. For Modern, it was 60.9%.
tronix wrote: ↑4 years ago
how else do people think fair interactive decks are going to genuinely expand their foothold in modern? 'narrow-conditional-answer card #23436' certainly isnt the answer. its powerful multifunctional cards that are pushed to provide support on different fronts; generic answer, value, threat, defense, baked in synergy, etc. such things will inevitably leave less powerful tools behind. just contrast what it took to have UW(x) control climb its way to relevancy and the perpetual state of mediocrity for the entire 'blue control' archetype that existed before. mostly a cast of cards that bodied unfair and fair decks alike.
I agree with the entirety of your post, but this quote is particularly important. Fair decks need "powerful, mutli-functional cards" to support the decks on "different fronts." Otherwise there is no way for these decks to compete with the wide variety of proactive decks that simply ask too many questions than fair decks can answer. This becomes a problem when the fair deck stops having bad matchups and can just attack every proactive deck on every front, pushing all its matchups to the 50/50+ range. As Stoddard said and the bans continue to support, if your worst matchup is the mirror, chances are you're getting banned. But to say W6 or any other card that has been around for 1.5 months is doing that, when we've had one GP and a metagame that was horribly warped by a legitimately busted deck in Hogaak Bridgevine? That is outrageous and unsupportable hyperbole. Ban arguments need to be supported with data and evidence. Wizards' own method has followed this for years, and our own predictions have been most accurate when they followed the same method. The whole "my subjective gut experience says this is busted" argument is just blindly throwing darts at a board. If you miss ten times (Coco busted! Temple busted! GDS busted! Humans busted! etc.) and then hit once (Hogaak/W6/T3feri/etc. busted!), that doesn't mean one's gut instinct is meaningful or accurate. It just means if you cry wolf enough times in a format with as many top dogs as Modern, eventually one of those top dogs is going to be a wolf.
Fair decks need cards like W6. These types of engines are why decks like UW Control are even viable. If we prefer a format where it's all proactive strategies posing questions/threats on multiple fronts with fair decks completely unable to keep up and choked out of the top, by all means, hate on the cards like W6. But as tronix suggests, if we prefer a balanced format where lots of strategies are viable (HINT: THIS IS WHAT WIZARDS WANTS), then we need these cards.
iTaLenTZ wrote: ↑4 years ago
If you look at the metagame picture Wrenn only improved Junds matchup vs UW Control, Humans and Company decks. So out of 20 tier 1 decks only 3 really care about Wrenn, the rest bypasses its CA because it seeks to win on turn 3 or goes big. Meanwhile if you look at tier 2 they all fold to Wrenn. Modern's meta cycle will look like this: People play Jund and get pushed out of the meta by goldfish decks --> without Jund other midrange and creature decks become viable --> with other midrange decks entering the format pushing away the goldfish decks Jund becomes viable again --> rinse and repeat.
So yes, Wrenn warps the meta around itself but Jund doesn't function like a police deck like 4-5 years ago because it doesn't do anything vs the uninteractive tier 1. It only made the gap between tier 1 and tier 2 a lot bigger and pushed a lot of strategies and decks out of playability.
These types of indirect arguments rarely pan out in practice or for any meaningful length of time. A great example, which I believe someone else already mentioned, is the 2018 Tron situation. Tron is a notorious midrange, control, and fair deck killer. In theory, when Tron proliferates, that shift incentivizes people to move away from fair decks and towards decks that go under Tron. Then when Tron gets suppressed by these goldfish decks that go under Tron, you might see fairer decks rise to beat them. Then Tron comes back and the cycle restarts. Sound like a familiar theory? It's exactly what you just posed for W6 Jund except replacing it with Tron. And as with the Tron argument, it is likely entirely speculative.
People play TRON and get pushed out of the meta by goldfish decks --> without TRON other midrange and creature decks become viable --> with other midrange decks entering the format pushing away the goldfish decks TRON becomes viable again --> rinse and repeat.
We can replace this kind of hypothetical, reductionist construction with any number of perceived best decks in Modern. For instance, one could do the same with Humans' impact on the metagame cycle due to Humans allegedly pushing out aggro decks. And whenever we do this, we find that history shows bans never happen as a result of that model. Tron is a particularly glaring example of this, because the italicized model above gets predicted every year and has been predicted every year for most of Modern's history. And yet, Tron does not get banned and routinely gets checked down to size while other decks emerge on top. It's just part of metagame cycling and it's just part of Modern. If Tron with 11 GP finishes in 2018 (8.6%) and Humans with 12 finishes (9.4%) were not bannable, there is simply no way W6 Jund is even on the table when we're 1.5 months out from W6's legalization. This is ban mania at its most hyperbolic. Claims need to be proportionate to the available and presented evidence of that claim. Your current W6 ban suggestion is completely out of proportion to any of that.
BloodyRabbit wrote: ↑4 years ago
I do think people are just too lazy to try alternatives. I know UW is great (and I play it since Modern's inception) but in the last two months I've been trashing my local leagues with Blue Moon, and yet nobody cares about thinking it may be good again (with the addition of new cards) because, simply, no one is playing it.
And this is just an example, of course.
People don't want to invest in archetypes that didn't put results in big tournaments, and Magic isn't the cheapest of the games.
There is absolutely something to be said for netdecking and laziness in contemporary tournament Magic. I also think it's not entirely irrational, and may even be smart for most players. Magic is a hard game. Deck building is hard, testing and tuning is time consuming, building good decks requires a grasp of numerous complicated Magic and format-specific dynamics, and the entire process is very resource-intensive. You are financially punished for failing by losing tournament entry money, and you often aren't financially rewarded with success any more than if you just used an established deck to get the same result. There are definitely some incentives to innovate (it's cool to be the next Chapin, next-leveling the metagame can result in lots of rewards if you succeed, brewing is fun, etc.), but those rewards are often lighter than the risks.
I will say that the MTGO 5-0 League dumps are a great way of encouraging brewing, and the sheer diversity of decks in any given dump is always awesome. This highlights a deck's success without necessarily showing how many tries it took for that deck to get there. Or that the metagame is .1% Goblins and 20% Dredge variants. But none of this is necessarily a Modern-specific issue, and applies to all formats more or less equally. If anything, it is better in Modern because there are so many diverse ways to 5-0 a League. And it's probably getting better in Standard because Arena gives you so many free opportunities to test and iterate. Overall, however, I agree with you that Magic would benefit from a little more experimentation and innovation, but also understand the reasons people don't do that.