Quote from cfusionpm »KCI is also a miserable, cancerous deck that creates horrible gameplay patterns and often causes awful experiences for both players, win or lose.
How does this relate to Mox Opal? Or is this a separate topic? Is this just a general frustration or a veiled ban suggestion? If the latter, no Modern deck has been banned for allegations of a bad, unpopular, or otherwise "miserable" game experience. Lantern notably avoided such a fate despite similar complaints. That doesn't mean it can't happen in the future, it just makes it much less likely.
I'll also add, in reference to an exchange you, Tronix, and I had, this quoted post is the exact kind of categorical judgment I will always push back against. It is not a feeling statement of personal experience. It is a sweeping, format-wide indictment without any evidence to make the claim on such a broad scale. This kind of assertion does not really improve our Modern understanding.
2
1
2
"Except for these decks it's used in, name another deck it's used in! Bet you can't."
2
Not to be obnoxious, but this shows a hilarious misunderstanding of the differences between UW and UWR control. The Jeskai version is not "splashing" red; try taking even a cursory glance at the deck lists. G Tron and GB Tron overlap much more.
2
I think your analysis of the top players in the two eternal formats led to some excellent insights. I really appreciated it. But you can't entirely rule out the effect of a "match-up lottery" in Modern from your comparison though.
Repeatable success isn't necessarily a way to rule out match-up variance because there's no way to control for the potential influence of top players ability to choose the right deck for a metagame. There's a reasonable argument to be made that Modern's wider metagame with fewer answers and more viable decks requires more "deck selection" skill than Legacy. If that's the case, then MWPs for both formats include the ability for top players to navigate the metagame; however Modern MWPs could reflect both increased match-up AND higher level of skill of the top players to select the right deck.
It may be impossible to entirely rule out that possibility, but an analysis of how often top players in each format rotate between decks may narrow it's likelihood. If "deck selection" is more important in Modern, you'd expect to see the Modern players with the highest MWP playing a wider variety of decks than their Legacy counterparts.
3
I'd generally agree with most of what you've stated here. I'd be amiss though if I didn't point out that there's a certain condescension that surfaces at times in your responses. I'm not trying to knock you down a peg, just share how it comes across. And I feel pretty comfortable identifying that tone since it's one I'm prone to myself.
Obviously several posters appreciate the intellectual bent you bring to the conversation (and it's sparked a interesting train of replies). But you have to understand that asking an anonymous internet poster to outline his preferred meta by deck and percentages is more than a little silly. Some players don't want intellectual rigor or theory, they want a fun game (with "fun" often defined in an ambiguous, personal way). Calling those players "selfish" is a conversational dead-end. Like a counter war on the stack, pick the exchanges where there's a chance of something productive.
1
You're right, that's my mistake. I meandered between 'linearity' and 'interactivity', using them interchangeably there. They're highly similar in my mind, but I think the difference actually helps clear up the scenario I'm pondering in my last paragraph. When looking at a decision tree of a match, the fewer the number of branches the greater the linearity. The decisions (represented by branches) can be an interactive responses -- e.g. Lightning Bolt a Dark Confidant -- or non-interactive responses -- e.g. the timing of a Lightning Bolt to face while an opponent is tapped out.
I'm in broad agreement with you. Not all branches on the decision tree are likely. We'd need weighted branches to accurately represent that Lightning Bolt and Lava Spike are, in practice, used very, very similarly (redundancy is one the strengths of Burn). You also won't see Grixis DS using Kolaghan's Command to damage themselves very often. Like you mention, putting together a full decision tree for every deck for every match-up pretty quickly becomes impractical. Really only useful as a way of thinking about the concepts, not actually creating a model unless you have a few spare thousand hours on your hands.
1
That's a fair point, although I think KTK's general direction is viable. You could weight the actions by category -- 4 points for interacting with what an opponent is trying to do (on the stack), 3 points for interacting with what an opponent has done (on the battlefield), 2 points for interacting with what they may immediately do (their hand), 1 point for interacting with what they don't know about but may eventually do (their library). There would be some disagreement about the values, but that might define relative interaction along intuitive lines. And of course the end result would have an element of subjectivity to it, although (theoretically at least) you could use regressions to tweak the values assigned and mitigate collinearity.
I love the idea of a weighted decision tree of actual matches. With enough information and data coding, you could assign a "decision" score to each deck. But I'm conflicted on whether that gives you interactivity or complexity. Decisions could be mild pivots based on an opponent's actions (do/don't attack) rather than interaction, although perhaps that's widely applicable enough to be canceled out in comparisons across decks.
1
Several people have conceptualized the idea of "linearity" as it pertains to this discussion. One rough version would be that "linearity" is the degree to which a deck's play pattern changes according to an opponent's plays, as compared to a baseline goldfish match. There are three places you can interact with your opponent, beyond their life total -- the board, their hand, and the stack. Depending on the volume and intensity of a deck's cards that interact in those ways, it falls somewhere along a spectrum of "linearity". Yes, all decks interact in some way; they have to, by definition, otherwise the game would never end. But no, they do not interact equally. That's not a matter of bias for or against it, simply an observation of play patterns that follows both wide anecdotal evidence and even cursory analysis much less what KTK has proposed.
1
Without intending to offend anyone, this is a very naive view regarding statistical analysis. A larger sample set provides a more accurate view of a population. No, it will not perfectly describe any one particular sample, past or future. But greater accuracy overall means more predictive value. The difference between 4% and 8% may not matter to you, but that means you're twice as likely to face a given deck. When choosing how to gear my sideboard or flex slots, that's welcome information. Not to mention the secondary analysis that's enabled by trustworthy numbers -- change rates, relative comparisons, ban likelihood, etc -- but that is garbage when the available information has such a wide margin of error.
The greater grumble is probably that over the last few years WoTC has intentionally downgraded the available data. Imagine trying to evaluate the quality of a road by driving on it. We started in a new BMW 3 series, able to call out the major potholes and uneven grade. Now, to obscure visibility of their own mistakes, we can only feel the road through the rusted suspension of a 1985 Honda Civic. These metagame analysis expectations aren't nonsensical -- they were mostly fulfilled less than 6 months ago.