Building A Better (Insert Thing Here)

Ari writes about a bunch of things he’s been thinking about, from why Brimaz is one of the three most important cards in Standard to how to improve Limited testing for Pro Tours.

I wish I could write about the awesome work my team has put into Pro Tour Born of the Gods, but that will have to wait until next week.

Instead, I’ve fleshed out some of the non sequitur thoughts I’ve had over the past few weeks.

Brimaz Is The New Nantuko Shade

Michael Jacob wrote a long explanation last week of how he feels Brimaz, King of Oreskos matches up poorly against a large percentage of the format.

Long story short, I disagree. I think that Brimaz is one of the three most important creatures in Standard right now (next to Master of Waves and Pack Rat).

Let’s talk about why.

To do so, we have to take a trip back to 2002. Cabal Coffers, Mutilate, Innocent Blood, and Chainer’s Edict are all Standard legal. This is the Mono-Black Control deck that everyone became attached to and unsuccessfully tried to recreate for years.

Your maindeck is creatureless, with basically all removal and a Diabolic Tutor package. Your sideboard contains four Nantuko Shades, and the card comes in against everything except one deck (Astral Slide).

Why not just maindeck them at that point?

There are plenty of reasons, and I think that is the exact spot Brimaz has in control sideboards right now.

The first and most obvious one is that your opponent will be overloaded on dead removal game 1. Odds are you have your own dead cards as well. Game 1s with U/W Control against anything midrange or bigger often come down to you exploiting the fact that their cards are more dead to run them out of action and get a window to hammer the nail in with a planeswalker or Sphinx’s Revelation. Just having Brimaz in your deck lets them regain parity.

In game 2, both players board appropriately, and threats become more of an issue. Even if they play the "appropriate" amount of removal to handle Brimaz, he plays a big role in this fight. There are tons of oddball threats people can sideboard in to counteract your base control plan. Think Possibility Storm; Ruric Thar, the Unbowed; or even the more common Assemble the Legion. What’s the right solution to a wide swath of narrow threats?

To steal from AJ Sacher’s book and quote Day9: "Just go effing kill them."

Brimaz disrupts all of these carefully crafted game plans. Land an early Domri Rade or Xenagos, the Reveler? Guess what, attack it for four. Try to bait counters and resolve one of the oddball enchantments? Guess what, you are still dead on board to Cat tokens. It’s not like you can’t build your deck skewed toward proactive spot removal that forces Brimaz through. I’m especially interested in Far // Away in this slot.

That said, the way to juke the sideboard option of leaving in removal is have only Brimaz and Aetherling for creatures. No Blood Baron of Vizkopas since they trade for Lifebane Zombie. Stormbreath Dragon is matchup conditional.

I almost tried to make an argument here that Brimaz isn’t actually better than your game 1 cards because a lot of them are specialized to create blowouts, but then again so is Brimaz. Doesn’t die? Kill their blocker, kill them.

Also, the whole "gets beat by X" in combat thing really doesn’t matter much. Can most of those creatures attack into Brimaz? How many cost three mana? Doesn’t this parity just walk the deck playing creatures into the control deck’s game plan of playing defensive or extending into Supreme Verdict?

Seriously, this may be a classic Patrick Sullivan Gitaxian Probe case. If your deck makes white mana and doesn’t have Brimaz, replace three cards at random with them and it probably gets better.

The other thing you may notice is that I haven’t mentioned Brimaz outside of control decks. I’m pretty sure the card is more of a Tarmogoyf than a Hero of Bladehold. It’s cheap and efficient at killing people. You don’t need to be a beatdown deck to go to town with him.

Ultimate Price

This past week when I was speaking to Jackie Lee about Standard, the idea of splashing in G/R Monsters came up. White and blue have little to offer, but black actually solves most of the current problems the deck has. Specifically, Ultimate Price kills almost every creature the deck cares about. Stormbreath Dragon and Polukranos, World Eater in the mirror, Pack Rat and Desecration Demon versus Mono-Black Devotion and B/W Midrange, and Master of Waves and Tidebinder Mage against Mono-Blue Devotion. It even takes down a Brimaz, King of Oreskos or Archangel of Thune if U/W Control shows up with them maindeck, though Mizzium Mortars is likely better post-board since it kills Elspeth tokens.

Here is an approximate list that is ported almost directly from Kent Ketter’s second-place list from #SCGNASH.

Notice the disappearance of Plummet from the sideboard.

There is a cost of two additional enters the battlefield tapped lands and six additional Ravnica duals. I’m willing to pay that price to kill all the things I listed above.

The sideboard Flesh // Blood is only there to remind you that the added splash makes it easier to fuse or just cast the Golgari half.

Abrupt Decay may want to be Dreadbore. The idea is that you have a removal spell that hedges against Brimaz by being able to take out another permanent (Detention Sphere or a planeswalker). Either way, you only want one or two because it doesn’t kill Master of Waves or Desecration Demon in other important matchups.

I cut the Chandra, Pyromaster since it seemed like the odd card out, but you may want a planeswalker to maximize your resilient threat count. Vraska the Unseen is considerable and might let you move around some slots toward more Abrupt Decays or Mizzium Mortars because it kills Master of Waves and Desecration Demon.

Two Mutavaults might be pushing it with this deck. You do gain more scry lands, so losing the 2/2 body isn’t the end of the world. The card is very good though, and I would want to try to push the mana to support them.

I have been told Courser of Kruphix is way better than my initial assessment. I’ll trust people who have played the format more than I have on this one. Also worth mentioning is that Tom Ross wrote about a Jund Midrange deck last Friday that’s heavily based in the previously seen B/G Midrange decks. I’m not sure the deck is fully utilizing red (either within the color or over another one), but Courser of Kruphix seems insanely good there. More scry generated virtual card advantage added to all the black removal seems like a good way to make attrition worthwhile.

Redefining Limited Testing

I’m starting to realize that my self-perceived weakness in the Limited portion of Pro Tours has very little to do with being fundamentally bad at Limited itself.

If it isn’t being bad at the combat and in game threat/answer evaluation of Limited, then why do I keep underperforming in the six Draft rounds of Pro Tours?

One of my biggest strengths in Constructed is rapid evaluation of cards or decks in practice. Jam a list together and play a few quick games and I have a very good idea of what works or who is or isn’t favored in a matchup. In Limited, this is much more difficult. To test a strategy, you have to draft enough times for it to be open. Then there is the confounding factors of random packs and picks of the people around you. Did you 0-3 because your concept was bad or because you got cut out of blue on both sides? There’s also a much heavier time expenditure. In the time it takes to coordinate and run an eight-man draft, I could probably play ten to fifteen Constructed matches.

How many drafts does it take to reasonably evaluate each color combination? Ten? Twenty? More? I know there was exactly one format recently where I felt I did my own Draft preparation well, and that took 20-30 hours of solid double queuing Magic Online drafts plus being able to watch every match replay to recognize patterns.

This can’t be the right way to do things. The traditional "play more drafts" version of Limited testing is fundamentally flawed for Pro Tours. There is likely a way to increase the amount you learn in the minimal time you have to invest.

I would likely benefit from testing Limited more like I test Constructed. For the start of a format, I should be building "appropriately distributed" or "stock" Limited decks and battling against each other to determine if cards are actually as good as I feel they are. Assuming around 70% of your picks are on color, tier out the commons and uncommons, weight accordingly for how many of each category you see in that 70%, and go from there. This may overestimate pack 1 returns in multi-set formats as you bounce around more, but on the flip side you also need to play with more new cards to learn about them. From there I can move into drafts with a more reasonable vacuum view of new cards and learn what happens with relative color preferences and pairings.

This gives you an idea of card power levels pretty quickly, lets you test color pairings, and enables you to note key power and toughness jumps through creature interactions. This also puts the focus on in-game performance of cards instead of the draft portion. This is often ignored since the draft process has so many obvious decision points, but the rounds are where card values are actually determined.

Knowing this, you can then focus on drafting specific things. For example, are there narrow cards you had underrepresented in your test decks that swing around late and make an archetype? If given the choice between two equivalently powered cards, which one is more replaceable? How do overall color power levels bias the pick order? How does pack order change how people pick colors?

I’m unsure if there will be time to implement this for Pro Tour Born of the Gods, but for the next Pro Tour I want to try to get two team members solely assigned to these tasks. Build approximate decks with cards and battle with all ten color combos plus a niche deck or two. Then start drafts, ideally multi-queuing online, with the two people reconvening every few drafts to update on everything they did.

SolForge Drafting Or Cards Versus Decks

I’ve been playing a ton of SolForge lately. Last week’s announcement of StarCityGames.com teaming with Stoneblade Entertainment was awesome to hear.

The best part? I’ve been learning about Magic Draft theory from playing the game. Specifically, I picked up something that I think in the long run will help me grasp new Limited formats much better.

I’m going to cut to the chase here.

SolForge, like Magic, has two major resources beyond life. One is deeply rooted in manipulating probabilities, while the other is a set counter that certain cards allow you to accelerate. In Magic, you have mana as the first, and physical card count (+1 per turn barring cards that draw more cards) as the second. In SolForge, the first is the leveling system, and the second is your two plays per turn. I’m going to focus on the latter.

In the end, the way to win the war on plays per turn is the same as the way to win the war of card count: trade two for one. The way to do this in SolForge is just through Alpha Tyrannaxes instead of Divinations (the former is referencing Scars of Mirrodin Limited where green decks generated card advantage by forcing group blocks on their larger creatures).

As a result, Draft is mostly a game of "I pick this card because it has the most raw power" (also a result of almost every creature being on curve, but that’s a different story). In Magic, you can have similar things occur. For example, Magic 2014 had very few cards that were reasonably capable of winning a game and not just trading for an irrelevant card on your opponent’s side or being bricked by a good one. As a result, you had a format where everyone focused on getting those cards and filling in the rest with whatever interacted with them in any way.

Theros is similar, though the filler versus filler battles tend to end with trades and games that can end instead of stalls. The good cards (Wingsteed Rider, Nimbus Naiad, Nessian Asp) are just miles better than the other cards, and drafting is about finding the way to maximize your count in that category and the number of cards you have that interact with those cards.

As an aside, note that this can manifest as "counting cards is all that matters" (M14 to an extent) or as "having the one card that kills them is all that matters" (Theros). These may actually be the same thing, only in the first case the card that ends the game is the Opportunity you cast three turns before the Air Servant.

On the flip side, you can have a format where you are building a deck of interchangeable cards. Think Zendikar and Magic 2012, where there were so many ways to apply pressure and so few ways to stop it. Or Magic 2013, where there were a large number of ways to interact with cards and a balanced number of card advantage sources and the most important part of the format was just making sure your deck had the resources to let you play around what your opponent drew each game.

There is some middle ground. Formats like Rise of the Eldrazi exist on occasion, where you find the powerful enabler that is open and then fill in a known shell for that enabler.

How does this distinction help us? By making parallels to a known format, we can enter the early drafts of a format knowing certain key factors. For example, a format with a very high power level discrepancy between the best and worst cards is likely to be a format based on individual card power. Curve might matter a little less, and finding the open color will matter a little more. If the power level is flat and the format is about synergy, you can focus on refining your generalized deck outline each draft.

That brings you to the next layer faster. How much do the mediocre cards matter? Are they virtual unplayables (M14 again)? Based on that you can figure out when you need to commit to colors (as discussed a couple of weeks back).

Obviously there are unique twists to each format, but having a framework to follow helps a lot. When it comes to playtesting, this kind of focus is what changes your progress from throwing darts blindfolded to something that actually finds an optimal strategy.

Making A Better Game

I’ll admit this last part has little to do with drafting.

For every bit I find SolForge awesome and interesting, I hate Hearthstone.

First of all, the World of Warcraft card game existed. In making Hearthstone, they fixed two of the large issues with the game: the weapon system (Cursed Scroll that cost life to activate if they targeted a creature were a base game mechanic; they now effectively have depletion counters) and draft (it was basically "choose one of ten colors, your spells are only that color"; now your color is chosen before, and you don’t see uncastables). They didn’t fix the big one: the combat system.

For those of you who haven’t played either of these games, there are an infinite number of combat + post-combat main phases in a turn, though things still "tap" to attack. Only one creature attacks per combat. Instead of attacking players, creatures attack creatures or players, and only certain creatures have the special ability of blocking. In the World of Warcraft card game, creatures with the ability tapped to block. In Hearthstone, the creatures must be the defender on all attacks (think Gideon Jura, only they have the option of no attacks or attack Gideon).

This makes having the initiative on attacks way more important. If you attack first, you get to choose how to trade profitably. This tends to lead to cascading board advantages and games that are impossible to stabilize.

Wait, didn’t I say SolForge is similarly focused on creating cascading board states? Why is that system more interesting? Simply put, in the Hearthstone/World of Warcraft combat system, the aggressor has too much of an advantage, preventing a back and forth. Instead of a 4/4 being good enough to stop some early beats, you need something on par with a Baneslayer Angel.

In SolForge, you can trade life (letting them attack) for setting up good blocks. Life totals are higher relative to early cards, and midgame cards help rectify early advantages as the game proceeds in phases, so it’s more difficult to fall so far behind early that you can’t recover. The late game power level differences between "low-drops" (relatively better at lower levels) and "high-drops" (relatively more powerful at level three) are marginal compared to the difference between a one-drop and a three-drop. All of this dampens the movement of board advantages.

Hearthstone games are also complex in a bad way. There are just too many possible configurations of actions in most turns with fewer ways to distinguish them then in Magic.

It turns out that lands are secretly one of the most important aspects of Magic here. Your opening hand in Hearthstone is three cards on the play, four plus a draw step on the draw (one is a Lotus Petal). In Magic, you have seven cards either way plus one on the draw. This may seem like given the similar systems that Magic would create more decision trees, except lands don’t really do things. 40% of the cards you draw don’t interact with your opponent’s cards.

Turns out that the average Magic opener is about +1 card on the average Hearthstone opener, and that is erased quickly. Not only do you draw 40% less actual cards a turn in Magic, but in Hearthstone you have a once per turn two-cost Vanguard ability. While that ability is definitely worth less than a full card of difficulty a turn since it doesn’t accumulate if unused, you’re still drawing nearly a full extra card a turn in Hearthstone.

Because of the importance of curving out early in Hearthstone so you have attack initiative, your deck is relatively low curve. Cards trade up well, and high-drops are slow to stabilize a game, so playing multiple two-drops over a four-drop is a much closer decision than in Magic. I’m reminded of counting points ala cribbage and how adding a fifth card can cascade into so many more counts of fifteen. Cards tend to accumulate in your hand, and at some point you’re faced with multiple turns with a huge number of play/sequencing options on top of combat options.

Basically, imagine everyone has to learn Magic by playing Legacy Delver. It might work, but there are way too many decision layers to start off with.

SolForge combats this with the steady five-card hand size. No matter what point of the game it is, you only have five cards to play. While this branches into a lot of possibilities based on board position and timing relative to attacks, a lot of those tend to be eliminated by simple heuristics. Play the properly leveled guy into the empty lane, not the level one guy.

In Magic and SolForge, this lower number of tactical/sequencing decisions allows more processing power to be used on strategy. A good example in Magic is "how do I order my guys to play around removal (also made more relevant because attacking creatures aren’t removal)?" In SolForge, this manifests as "what do I want my deck to look like on the next shuffle and how much tempo can I sacrifice to achieve that without dying?"

Aside: I assume this is Hearthstone trying to fix the "quest issue" of its predecessor. Imagine any card can be played as a basic land, only some cards draw cards if they are lands and do nothing otherwise. Hand sizes whittle down, and decisions are easier. (Do I have a good play? If no, try to draw one). But this really punishes control decks by making the board state where the aggro deck is out of things harder to accomplish. As a result, the control decks almost all looked like Prison decks in larger formats. There were specific cards that made this a bigger issue (for example, Energy Field that does one wither damage to your opponent and all of their creatures each turn), but without cards on that level, control would be fundamentally unplayable.

Really, this pulls into my big issue. Hearthstone is just Magic trying to "fix" mana, something that isn’t really broken in Magic.

Simply put, if you are trying to out-Magic the people at Wizards of the Coast, the odds are not in your favor. You are behind twenty years of development if you change any base game mechanics. Magic does introduce new things, but most of them are relatively compartmentalized. If you make a new game, the things you introduce can’t really be dropped after a year and explained away as "we tried it, and it didn’t work" unless you reboot the whole concept.

Want an idea of how hard it is to properly balance a new fundamental game idea? Planeswalkers were first printed seven years ago and have only really hit their sweet spot in the last two years. Want to go back further? Counterspell was still in almost every set seven years after Alpha, and saying you would print Rakdos Cackler was an automatic disqualification from R&D eligibility. That’s actually a world where barring really bad opposing cards aggro is mathematically unplayable. You actually can’t pressure your opponent enough to kill them before they play something you can’t beat.

Are people really going to stick around that long through poorer game play experiences if you don’t offer something unique?

The one thing I dislike about Hearthstone I’ve actually been able to excuse is the user interface. I’m not going to laud Magic Online’s greatness here, but it plays like a card game. Animations don’t blur into subsequent actions. There aren’t really superfluous spell casting animations to begin with that take up time for no in-game effect. There’s no 3D battlefield image with moving parts. Hearthstone feels more like a game that isn’t turn based because of these things. I hate this because I just want to play cards, but I can understand there is an audience without the same background who is more familiar with this style.

That said, there is honestly no reason a player who disconnects from a turn-based game with a turn clock should auto lose. In a RTS game the reconnect time basically assures victory for the connected party, but as I’ve experienced many times you can have your Internet drop out in the middle of a Daily Event and reconnect six minutes later without losing any real advantage.

Unfortunately, my ability to extrapolate ends there due to my limited breadth of trading card game knowledge.

The big question on my mind is now whether the "resource focused on creating controllable variance" is something that has to exist in a card game. Is there a point in using cards if shuffling isn’t really a thing? Or is there a game where variance simply exists for certain sub actions that don’t bubble up nearly as far? My gaming experience is shockingly narrow for someone who has played for so long, but I would love to look deeper into this.

Actually, the big question on my mind is still "when is Chris Fennell telling me how to draft this format?" But let’s take things one step at a time.