Ya know, I was in a panic over this at first but now? Screw it, l'm in. What's the worst that could happen? New meta becomes Belcher and Storm vs. Chancellor and Leyline vs. Trinisphere? This actually sounds hilarious tbh. And if it somehow doesn't degenerate immediately into turn one wombo-combo then it sounds like a win.
All Spells Primer under construction: https://docs.google.com/document/d/e...Tl7utWpLo0/pub
PM me if you want to contribute!
Frank Karsten wrote an article with numbers regarding the new London mulligan
Definitely worth a read.
Well, I think this analysis is pretty solid. Not much to argue about with the numbers. On the basis of what the "first order" effects are and how they enumerate out, I think he is likely right on. Combo decks certainly gain more so than "fair" decks and lower land counts are plausible.
What is still unknown though, of course, is the resultant "second order" effects. To use a relatively simplistic scenario, let us say that decks do cut 1-2 lands or more as a result of the new rule. Now, in Modern, there aren't many ways for decks to punish this, because land-destruction comes at such a high price (even if that RG deck with Stone Rains does OK sometimes). However, in Legacy, we have a general prevalence of Wastelands already, plus things like Port and Thalia. These things already do punish decks for lower land counts and reliance on cantrips, so what effect would people pushing harder in that direction mean for these cards in the meta? And if that means "larger presence" what does to mean for the direction of the meta?
The second order effects are not knowable, mainly because we can't know the full extent of the practical implications of the first order effects let alone what the meta will do in response. There could well be a new "static" level of combo decks that is both higher and not oppressive. What Frank's article does is very nicely describe, numerically, how aggressive mulliganing can benefit certain strategies. What it does not do, and importantly, does not purport to do is speak of how well that will lead to actually winning games, or even correlate to winning games. It doesn't take much to figure that a better chance to "do your thing" likely means more wins, but the issue at hand is not gold-fishing, it is actual competative games of Magic, which means there is always an opponent on the other side of the table with a deck trying to do it's thing, which may well be to specifically hinder you from doing yours.
So, hypermetric geometry can tell us that taking a mull to 4 gives you a significantly higher change to hit A+B, but it does not tell us if aggressively mulliganing to 4 is a good strategy to win games over time. Again, because the meta may evolve to play "higher" levels of disruption, and aggressively going to 4 might be significantly worse if many decks are running things along the lines of 8 targeted discard spells and/or other disruption. In fact, it could be the case that mulling to 4 is so detrimental to your chances of winning versus anyone except a gold-fish that you likely would only ever do it out of sheer necessity (for example you really are in a gold-fish scenario, like fast, non-disruptive combo vs. fast, non-disruptive combo).
"The Ancients teach us that if we can but last, we shall prevail."
—Kaysa, Elder Druid of the Juniper Order
Right, again, this is why I qualify my statement, later in that paragraph, with the comment about cases where you are essentially goldfishing. I am not saying these situations don't happen, I am making the point that such games are relatively "rare" in Legacy. Most common Legacy decks are made to generally push some sort of interaction/disruption.
Even in the case of a combo deck going to 4 and a "fair" deck keeping 7, something as simple as Thoughseize or Force of Will can absolutely wreck the combo deck, which will then be at the mercy of the top of the deck, where the "fair" deck likely has 3-5 more "action" cards, or cantrips, at hand.
"The Ancients teach us that if we can but last, we shall prevail."
—Kaysa, Elder Druid of the Juniper Order
Yeah, all the hand-wringing about the new mulligan is substantially overwrought.
People aren't going to be throwing away keepable sevens because their six might be the nuts, and going to five is still going to suck. Hopefully, it'll suck marginally less than it currently does, because at least with the London mulligan you might have a couple lands and a couple spells, instead of no lands or all lands and having to keep it because mulling to four (and let's be honest, even 5) is basically just conceding with extra steps a huge portion of the time.
It's not the end of Magic or anything, but there exists feelbad moments when a Burn player keeps an above average 7 and is watched his opponent mull from 7 to 6 to 5 to 4 and then sigh and say sure, then proceeds to Tomb Petal Show Emrakul. Those already happen and suck, but they're uncommon. The Burn player who has build his deck as a paragon for consistency wins in the long run because godhands do exist, but that's just how uniform distribution works. And all this is for those players is they gain practically nothing, and get to watch their opponents reap uncommon wins.
Every breakdown has been accurate, this change shifts the nature of the mulligan a few degrees. In Standard you take a few steps out, the line hasn't moved any. In Modern a few yards out, sure, the drift is a couple inches, noticeable but whatever. By the time our out to Vintage distances, you're miles from where you once were.
But as always different≠worse
Well, I think Frank shows that, mathematically speaking, mulliganing now does it need suck significantly less from the standpoint of "find a hand that can play a somewhat meaningful game of Magic" but indeed the qualitative analysis of translating that to actual win percentages is not really determinate and so can't really be computed in a simple probabilistic manner.
So, said simplistically, we can think of it in the follow manner: the London Mulligan gives you a significantly higher chance to find a functional hand and functional hands will have a significant higher chance of winning a game of Magic than do non-fuctional hands. So, therefore, it seems that clearly the London Mulligan should result in more wins. So, A leads to B and B leads to C, therefor A leads to C.
However, this sort of propositional logic does not accurately describe what goes on in many game of Magic. While it seems determinate that A would lead to C, that is A to C in isolation. The confounding factor is that, one, lower resources (cards in hand) leads to lower probability of winning. So, "functional hand" is really not a sufficient criteria to presuppose you win, when "functional hand" comes at a cost of resources. So, what any of these probabilistic models lack is accounting for the detrimental effect that lower resources can have on actually winning.
In Frank's example, you want A+B plus 2 lands. But in these cases, perhaps the combo costs 3 mana. So, now you need to factor in the chance of blind top deck, plus the confounding factor of any possible disruption the opponent may have.
Let's use something like, Sneak and Show vs some kind Grixis Delver, for example:
SnS aggressively mulls to Fetch, Ancient Tomb, Show and Tell, G-man.
Delver keeps Fetch, Wasteland, Daze, Delver.
(Note, I am not a mathematician, but both these hands seem roughly equally probabilistic to me.)
In any case, on the play or the draw, and excluding confounding top-decks (because that will enter in many different factors) are we really to say that the Delver deck is in terrible position? In fact, the Show and Tell player, while in a position to possibly win the game, if that Delver flips (probable) there is a good deal of pressure at hand. If they don't play Tomb and miss the land-drop, they are more playing directly to Daze, if they play the Tomb and get it Wastelanded, that also helps Daze's case. In either way, the Delver deck is not in terrible position.
We can complicate this freely if we consider Delver's range of "keepable" hands by adding even something like 1 mana discard spells from the sideboard (or perhaps some where in the main), in addition to Daze (I won't complicate it by considering Force). In this case, Land (19/60), Wasteland (4/60), Disputation (8/60) and Delver (4/60), will lead to some greater chance of finding a keepable hand in each successive mulligan, which means they likely need to mull less for a fuctional, disruptive hand. Where the SnS deck needs exactly Show and Tell in this case, since Sneak Attack needs even more mana to work (and so is more open to the Wasteland being even better). This might (again, I'm not a mathematician) lead to, over time, fair decks flatly need to mull less over time.
So, all I am trying to illustrate is that while this rule no doubt favors combo decks in general, that does not mean that "fair" decks will stand no chance. In fact, it's even plausible they will stand a better chance over time, because they can keep hands with higher potential resources that are still functional, because disruptive elements are more "universal" than specific discrete combo pieces.
All of this is to essentially say that considering just A+B as a flat win though is likely not a good indicator of how actual games of Magic will play out.
"The Ancients teach us that if we can but last, we shall prevail."
—Kaysa, Elder Druid of the Juniper Order
Huh? Is it not a fact that most Delver decks run Wasteland? Which do not? All i can think of is UR Delver. Am I missing something here?
And, on your final point, my illustration above is only to point out that "fair" decks aren't necessarily at a vast disadvantage as a matter of fact under the London Mulligan, not specifically that Delver decks get explicitly better.
"The Ancients teach us that if we can but last, we shall prevail."
—Kaysa, Elder Druid of the Juniper Order
I agree with everything except his analysis that bazaar should be restricted. That would be the most non-sensical thing to do, akin to banning brainstorm in legacy. He kind of botched the numbers in his analysis because he assumes that Nature's Claim is the only card you would need to draw. However, vintage dredge plays 5 sideboard manasources so it is a 3 card combo (bazaar+mana source+nature's claim), and then the numbers come out very much in favor of Leyline.
Not to mention, the Leyline deck needs to not do anything else as well. Leyline + Misstep and any action could well be the game right there. Or Leyline + Sphere/Thorn/Trini.
As I mention above, the issue at hand is that this level of analysis is not really incorrect in it's evaluation, it's just incomplete.
"The Ancients teach us that if we can but last, we shall prevail."
—Kaysa, Elder Druid of the Juniper Order
There are currently 1 users browsing this thread. (0 members and 1 guests)