Hello, and welcome back to Burning Questions, your weekly exploration of the hottest topics in Marvel Snap. This week, we're talking theory-crafting, in-game events, and mill.
1) Is it harder to theory craft and predict the power of new cards in Snap than other games?
This response was inspired by an exchange between KMBest and Maverick on Twitter that I found fascinating. Of course, KM is correct; predictions are just guesses, and guesses can be wrong. However, I think Marvel Snap has a few unique aspects that make predictions more difficult.
Do you think it's harder to theory craft in #marvelsnap ?
— Maverick0023 (@Maverick0023_) April 21, 2024
It seems like every month most creators over rate one card and under rate another; just wondering why you think that is?
I kind of feel like the constant "change the meta just because" philosophy of balance makes it hard
First, the small deck size makes Snap behave fundamentally differently from other card games. Compared to games with 60 or 40-card decks, the opportunity cost eaten up by putting a card in your deck is massive. In a game like Magic: The Gathering, you need to ensure that a card (or 4x of a card) is actively helping you more than a different card you could run. In Snap, this magnified exponentially. The effect of a card like Howard the Duck is not negative. It's a valuable effect. But it's so much worse than simply putting in a different card that it makes Howard uniquely bad in a Snap context. The same is true for the original iteration of Snowguard. The rotation between Hawk and Bear meant you could only get value from the card in a handful of games from a large sample. It wasn't actively bad, it would even occasionally win a game with something like Bear onto Savage Land locking out the entire board. But the effect being usable so rarely made the card equivalent to running a blank Misty Knight in your deck. Running a vanilla 1-2, like Misty Knight, seems suboptimal at first glance, but it's far worse than that. You're passing up the opportunity to play a better card because you are playing that card in one of your valuable 12-deck slots. Opportunity cost is a core part of Snap deck building. This is also why Black Swan is less powerful than many people predicted. She has a powerful ability, but she comes at the cost of putting a bunch of 1-cost cards in your deck. Putting a lot of 1-cost cards in your deck is a massive downside from an opportunity cost perspective since you only get 12 cards, and higher-cost cards are more powerful. Black Swan will improve if 1-cost cards as a whole improve. Deck size and opportunity cost are a primary reason players misestimated a card before release. Then there's the ultimate lesson in opportunity cost being ignored during card evaluation, Kang.
Second, the relative power of differently costed cards is different than the curve in other games and behaves in some unexpected ways so far in the game's history. This is because the game is only 6 (or 7) turns and, again, because of the tiny deck size. 2-cost cards tend to be quite powerful since they can be played alongside 4s, which have also traditionally been powerful. 3s can feel very awkward if it's a card you want to play on curve since you have limited time to draw and play it. 1s have felt relatively weak, and perhaps the developers have been too cautious with their power levels. Zabu had a neutron star-sized warping effect on all of this, so it will be interesting to see if and how this all changes now that he is in hibernation. Tight curve considerations have led to predictions on cards such as War Machine and Spider-Man 2099 to be off target.
Third, locations are where a great deal of the randomness resides in Snap. Having control of the variance in a card game is incredibly powerful. In other card games, more randomness resides in your draws. Again, because decks are only 12 cards, you'll see most of your cards in any game, making decks fundamentally very consistent. To make the game fresh and fun (and occasionally frustrating), the game designers build a great amount of variance into the location mechanic. Most of the animus are directed at locations in Snap because it's where a lot of randomness comes from. This underrated concept led people to underestimate Legion in particular. I would argue this mechanic has a lot to do with Jeff's power level as well and he was underrated by some as well.
These are just a few ways Snap is unique and can lead players of other games to misjudge new Snap cards. Snap is still a young game, and these will fade as the game ages, but you should keep these concepts in mind when choosing which new cards to target.
Nocturne
A very interesting test case for this is the upcoming card Nocturne since she is very mixed based on the above criteria. People often cite “3-5 with an upside” as well-statted, but this falls into the trap of comparing Snap with other games. In other games, cards above the “vanilla” or textless stat lines are usually good. However, because of how the curve works out in Snap, 3-5 isn't very good if the ability isn't actively helping you. The ability seems pretty minor, combining Nightcrawler plus Scarlet Witch, which is two cards that are significantly lower on the cost curve. However, as pointed out above, location control and movement can mitigate much of the game's variance. Try to make your own prediction about Nocturne and check back after a few months to see how accurate you were.
2) What would you suggest for an “imbalance patch” event?
So far, we've had an Avengers vs X-Men event and the Valentine's event as our two initial in-game events. I would love for future events to take inspiration from comics storylines and epic moments. There's a lot of comics history to plumb for inspiration. But as a wacky, off-the-wall suggestion, I'd like to choose a lesser-known Marvel Comics Battle. How about for 24 hours, we all play out Galactus vs Squirrel Girl? Buff them both, but design it so Squirrel Girl is the perfect counter to Galactus!
Maybe their pixel variants can get an additional temporary boost?
3) Is Mill good?
I’ve seen some concern on social media that mill decks are too strong, and the archetype needs to be nerfed. That didn't match my anecdotal experience, so I decided to look into the numbers.
The above numbers are filtered for the last week, infinite only. The numbers rise slightly when filtering to 80-100, indicating that the archetype is better against bots.
All five cards have pretty middling stats. Yondu’s numbers are so bad that he's likely actively hurting even within this archetype. The best card is Doctor Octopus. It seems more likely that any success Mill is having is due to people finally seeing how powerful Doctor Octopus can be rather than any of the actual Mill cards.
When looking at deck stats, the highest win percentage Mill deck with Baron Zemo is this list.
Baron Zemo
The stats are mediocre, with a win percentage of 52.33 and a cube rate of -.11.
There was a related question in the ask-the-team channel on the official Marvel Snap discord, so let's dive in and see what questions are being asked.
4) Q: On 6/13/23, Glenn said in this very forum, "Our decks are too small and our games too short to meaningfully support a mill strategy. It’s not very fun to be milled, and removes a lot of tension around snapping. Given how infrequently we want the deck to be empty, we don’t need to add a feature for when it is."
I'd like to know what's changed since then to enable this extraordinarily non-interactive and unfun archetype that now appears to be the majority of decks I play against every game to now be promoted by the game developers.
A: The answer you’re quoting was specifically in regard to a question about adding an auxiliary mechanic to support mill as a strategy, and my point in the reply was that such an addition would be unnecessary because it’s already impactful enough in a short game just to empty the deck. Plus, emptying the deck was fairly difficult to do.
Since then, we’ve found a few fun designs that could hook into such a strategy we didn’t have back then—namely Gladiator and Zemo, both of which can be interacted with via deckbuilding. We also supported many decks that can basically ignore their deck being depleted (such as Loki, Ramp, Destroy, Hela, and Thanos).
There are clearly more than a few players excited to engage with mill as a strategy, which is why you’re playing against it—it’s definitely not because it’s the best deck, because it isn’t. Finding designs that can satisfy players’ gameplay interests is our job. If those interests have some rough edges, we explore whether we can smooth them out and find something fun rather than abandon them.
-Glenn
**This leads to the following two questions: why do players like to play Mill despite it not being particularly good, and why do people hate to play against it despite it being beatable?
For the first, I think the answer is that players like doing stuff. They like taking a lot of game actions. They like seeing multiple tangible effects. I think this subconsciously tricks people into thinking a) they have agency and b) they are interacting with their opponent. I think other decks like this have attracted players in the past. I think Thanos was, at several points, the best deck in the game. However, at least some of his popularity was due to the illusion of agency that the deck had. You felt like you were doing a lot. As for why players dislike Mill, I think part of that concerns how unpleasant Mill is in other card games. In Snap, an opposing Yondu is helping you rather than hurting you because of the small deck size. Additionally, Mill feasts on glass cannon combo decks, a style that many players enjoy. As Glenn notes above, there are a lot of decks that are completely unbothered by Mill.
Mill’s playrate will likely naturally wane over time, especially as we enter “clog spring.” I think Mill is fun and doesn't have a strongly negative impact on the game. But can we at least agree to cut Yondu from those lists so it's a little better?
Speaking of Doctor Octopus…
5) Q: Rate at which Doc Ock pulls Shang Chi.
Hi, as I’m sure you are aware this is sort of meme for how often it occurs. I am just wondering if you have anyway of telling the percent at which this occurs. It feels like Doc dies most of the time.
Thank you
A: If the opponent has Shang-Chi in a 7-card hand and you play Ock at an empty location, my napkin math for HYPGEOMDIST indicates you're >50% to pull the Shang. Given any opponent with Shand-Chi in their deck is also ~67% to draw it by turn 5, it's just a mathematically likely outcome.
-Glenn
**Glenn is spot on with his assessment above. It's even more likely since players often have fewer than 7 cards in hand, and players generally hold Shang for the last turn. It all adds up to a situation where Shang is both likely to be in hand and likely to be pulled.
But this leads to a greater discussion on player conspiracy theories. So many of these are explainable with some logic and math.
Here are a few I've heard:
- Iceman always hits the card you least want him to hit. We've all had this feeling when an opponent plays Iceman: “Please just don't hit _____, anything but _____!” Then that card gets hit. What most people forget is that Iceman can't hit 6 cost cards. This automatically increases the odds that the card you wish to stay ice-free gets hit.
- Spider-Ham always hits my most important card! This is kind of the opposite of the Iceman situation. As you play cards out, the left side of your hand is more and more likely to contain your big finishing cards like Hela and Doctor Doom (since you didn't play them earlier).
- Spider-Man always moves my card to the least convenient place. Many people don't realize that Spider-Man isn't 100% random and is hard-coded to grab an opposing card and take it with him if he can. That means he will never fail to pull one of your cards and go to a location you have already filled out. This small lack of total randomness can make Peter Parker feel more random than he is.
That's all for this week. If you have any burning questions, feel free to hit me up on Twitter or provide your feedback in the comments section below!