Card Balance and Artifact

Previously: Artifact Embraces Card Balance ChangesCard Collection and Ownership

VII. Card Balance

To what extent should cards in a collectible card game be intentionally unbalanced?

Before Artifact’s recent changes, it was clear that Axe was the best red hero, and Drow Ranger was the best green hero. Playing a red deck without Axe, or a green deck without Drow Ranger, was not a strategic choice. It was a sure sign that the player didn’t own the card in question.

Is that… bad?

Queens are better than rooks, which are better than bishops. On a level playing field between players, there’s nothing wrong with that.

If everyone was constructing a chess deck, and it was always a king and queen, two rooks, two bishops, two knights and eight pawns, something would have gone wrong somewhere, even if the game remained chess. But what if everyone who owned them always played a queen and usually two rooks, but disagreed about how many bishops and knights to use? That might be fine. Especially if given a few similar other pieces as options.

Colors are balanced by giving each uniquely powerful cards and abilities. Decks and strategies are created by giving out powerful tools to choose between.

The tension between the card that fits your color requirements, the card that does the thing you want, and the most powerful cards, is central to any collectible card game.

What everyone agrees is bad are oppressive decks. 

To only a lesser extent is an oppressive color or card an issue.

When fields in Magic’s standard format were recently over 50% very similar red/black decks, with nothing even the best players could do about it, that was very bad. In that case, it was too late to pull out the banhammer, so we rode out the damage. But there is a clear issue that Magic’s standard format seems recently to be quite vulnerable to having a Best Deck take over and proving unable to adjust to fix it.

If there were eight tier one decks, each with a different strategy, but there was a card that was in all of them, would that be a problem? It would indicate a likely color imbalance. That would be an issue. But so long as games were not too often determined by who drew this excellent card, it would not be a major concern of mine. The more generic the card, the less concerned I would be.

Games need building blocks. If cards like Shock, Duress or Cancel became automatic includes for a time, as long there’s still room left for customization, it seems mostly fine.

Now return to Axe and Drow Ranger.

From one point of view, pre-change Axe was a key part of What Red Does in Artifact, and Drow Ranger was a key part of What Green Does in Artifact.

If you play red, you play Axe and Legion Commander, then if you have a third hero it was probably but not obviously Bristleback. If you wanted five red heroes, you’d have to accept some benchwarmers like Beastmaster and Ursa, or now Timbersaw. I have heard talk of Pugna or an occasional Sven or Tidehunder, even in decks with a second color.

If you play green, you play Drow Ranger, then you have a second tier group of Omniknight, Magnus, Trent Protector and Lycan, and arguably Rix, Abaddon and Chen, from which you choose your additional heroes according to what you prefer and what your deck is up to. Trent Protector is usually what competitive green wants to be doing, so it usually ends up with the two slot.

What about blue and black?

Blue’s best hero is probably Kanna, but it’s not obvious or universal, followed by Luna, Zeus and Ogre Magi, then likely Skywrath Mage. Specialists can get some work out of Prellex, Venomancer or Crystal Maiden.

Black’s best hero is Phantom Assassin, which should be all but universal. The second tier is Bounty Hunter (which is much stronger now that Axe does not kill him in one blow and can’t ever survive two of his), Sorla Kahn, Tinker, Sniper and Lich. I can see arguments for any of them. If you run mono-black, Storm Spirit becomes playable.

Each color has twelve heroes, one of which is the fallback free basic hero. Of the 48 heroes, I just named 32 of them. That’s not only a lot of heroes, that’s two thirds of all the heroes in Artifact. Each color has a signature hero so that hero quality is stronger in decks with extra colors, but most heroes have a constructed purpose.

What, then, went wrong? Why was this not acceptable? Here are some theories, which likely combined to cause the issue.

  1. The cards in question, Axe and Drow Ranger, are rare and cost dollars.

This is definitely a lot of it. We had headlines like “The most expensive card in Artifact costs more than the game” being thrown around, despite this reflecting that the game is cheap, and only being true for a day or so due to a much inflated price.

If the dominant red hero had been the uncommon Legion Commander instead of Axe, I doubt there would have been half as much complaining. By definition, if something is rare, it is going to be tough for everyone to have enough copies of it. The situation can be seen as a money grab, where cards that are effectively required for play are not sufficiently available.

Ironically, I also believe that if Artifact had contained mythic rares, but Axe and Drow Ranger had remained rares with similar rarity per pack, then there would have been far fewer complaints about Axe and Drow Ranger.

Magic, on the other hand, kind of justifies needing four copies of cards that cost a similar amount and have a similar rarity to Axe or Drow Ranger. Even though you also often need four copies of mythic rares, and you need four copies of each card instead of one for heroes in Artifact.

What is salient, and what is actually going on under the hood, are not as linked as one might hope.

For packs to be worth money, something in them needs to generate that value. If there are no rares (or mythic rares) to do that, packs won’t be valuable. That has some potentially quite bad consequences, but that is another topic.

2. Heroes in Artifact are there every game and don’t stay dead

It is one thing if every deck has Axe.

It is another thing if every game starts with Axe in play.

It is a third thing if killing off Axe means he comes back two turns later. Which he does.

If players understood this as analogous to the queen in a chess game (and chess is an interesting metaphor for Artifact, more so than it is for Magic), then they might be fine with the idea, but most players didn’t see it that way. And yes, it has to be pretty annoying when the card you don’t own is there every game and won’t stay dead.

3. Heroes in Artifact determine the flavor

Not sure this one was important, but I definitely noticed it. Artifact has a wonderful set of voiced lines for its various heroes and creeps, depending on situations and the combinations of cards in play. My favorite moment playing is still when Crystal Maiden shouted out “I finally get to kill someone!” If I never get to play Crystal Maiden, I miss out on that type of discovery. Axe and the other top tier heroes have good lines too, and more of them, but by now I have heard them all.

Players who are coming from DOTA 2, and who are more engaged with the world, story and characters, have even more reason to want more variety of heroes to be played.

4. Heroes in DOTA 2 are all playable or close to it, and are constantly rebalanced

DOTA 2 has an insane number of heroes, such that the barrier to full entry is beyond prohibitive. I recently saw in my Google news feed a recommendation that players who want to be good at the game choose 2-3 heroes and stick to them, so you could focus on other aspects of the game, but that it was fine to choose any of the dozens and dozens for your specialization.

Coming from that context, even without the flavor considerations, it’s easy to see why one might have otherwise unrealistic or unwise demands for heroes to be balanced against each other. It’s also easy to see why they think rebalancing heroes isn’t an issue.

To me, having some awful choices (also known as ‘skill testers’) is actively great, and not only for limited play, because (among other reasons, Mark Rosewater has written extensively about this) it means some people can try to make them work for fun, and new players get to learn about what is good by figuring out which cards are bad.

5. Complexity and lack of progression issues were misidentified, and players be whining

The players gonna play, play, play, play, play but they also gonna complain, complain, complain, complain, complain. One of those complaints is always that something in a game is too good, or not good enough.

That doesn’t mean their complaints are invalid, but it does mean that even in the best of times the complaints exist and ‘have to go somewhere.’

In this case, it was not the best of times for other reasons. Players lacked any progression or ranking system (other than the misnamed ‘perfect run’ count that pisses me off every time I go 5-1 and it counts as ‘perfect’). Players were all starting from zero in a super complex and hard to understand game. Players were starting with zero collection. Players were comparing the game to the ‘free-to-play’ model purely on cost and looking for ways to be frustrated by the expense, rather than comparing Artifact to Magic, or to an AAA software title, or thinking of the game as a $20 unlimited drafting experience with an upside option.

So players be whining more than average, especially about various aspects of the economic model. This then spilled over into card balance complaints becoming louder than they would have otherwise been.

6. Deck balance was hurt by player inexperience

I talked a bit about this in previous posts but I’ll reiterate a bit.

In the expert (now ‘prize’) constructed queue, you faced Red/Black aggression a lot, and still do. When one deck dominates, players see the situation as broken and demand action and change.

Those of us who have been around since the Alpha, or play or watch the major tournaments now, know that Red/Black was never a problem for experts. It is a low tier one deck, at best fourth strongest. I am always happy to see my opponent playing it from a win-expectation standpoint. Despite that, I sometimes think ‘again?’ since I am playing to have fun and to learn, not to get easy wins or grind out free packs.

Over time, with or without the changes, players would develop and learn additional strong strategies, and the new hotness would change. The resulting red decks would have still used Axe, and the green decks would have still used Drow Ranger, but the rest of the decks would have been more diverse, and that would have taken a lot of the pressure off.

7. Players never bought into the economic model

If you don’t own Axe, Axe being expensive looks bad. That is money out of your pocket.

If you own Axe, Axe being expensive does not look bad to you. If you own a lot of copies of Axe, it looks mighty fine, thank you very much.

When every player is starting fresh with no collection is exactly the time for players to hate everything that is expensive or necessary, and want entry to be cheap. Only later will they realize the upside of preserving value.

Which is another way of saying, no, players don’t care about card ownership and value. At least, not yet.

8. Equality and card balance is the level zero instinct

People instinctively hate inequality (unless they have the better deal). Because of reasons. Some are even good reasons.

The natural instinct of most players is to want all the cards to be mostly equal. That seems like the most fun and interesting option.

I had that preference in Magic for years, as a professional whose dream was to work in Magic R&D. It took years of conversations with those who make the game, and the game’s top players and writers and thinkers, to understand why this instinct was wrong.

Since then, better explanations have become available, so it is easier to get to where one understands these issues better and embraces card inequality.

Compare the situation now along all these dimensions, with the situation when Artifact was being tested. The players were more heavily invested in time and attention. They played better, and focused on different decks (partly because a few cards were different, but mostly for other reasons). The metagame shifted multiple times. They had richer experiences with collectible card games and their long term needs. They had unlimited card access for testing purposes.

I believe strongly that, not only for limited but also constructed purposes, cards should not be of equal power level. There should be staple cards that are in many or most decks of the appropriate color. There should be very good cards that are difficult to use for various reasons, from requiring other effects to work, to being a myriad of colors. There should be bad cards that are exactly what you need in special situations, and good cards that aren’t what you need as often as you would like or expect. There should be bad cards, and terrible cards, to provide skill testing and fun quirky experiences.

Players should be excited to go out and get better cards that upgrade their options and power level. Not having access to the cards should hurt you, so long as having full access is a reasonable goal for the serious. Demand should be driven. That’s the point.

Wizards does this consciously with Magic: The Gathering. They take each set, and they ‘push’ selected cards to make them the cream of the crop. Planeswalkers frequently get the nod, as do many other rares and mythics. They are not subtle about this. Their preferred cards will frequently hit you over the head with a ‘play this, everyone!’ They do it more and more obviously and dramatically than I would like, but some of it is good to get people excited and shake things up.

One alternative would be dramatically smaller card sets. If every card is good enough, then you would want less of them to get the same level of depth, complexity and choice. You would also want less of them to avoid giving decks with limited colors (or similar factions) similar card quality to multi-color decks, and leave people interesting choices. Rather than think of ‘take these 1000 cards we print each year, and instead of 250 of them being viable and 50 being the top, make it 750 and 250’ and think instead of ‘only print 400 cards.’ Or, having printed a set, rebalance existing cards periodically and make less new expansions over time. Those both seem more like reducing choice and discovery. I don’t think they are better.

This brings us to the related but distinct question of card rebalancing, which I’ll talk about next time. You can embrace a goal of cards being balanced or unbalanced, without implying a stand about when cards should change.




This entry was posted in Uncategorized. Bookmark the permalink.

27 Responses to Card Balance and Artifact

  1. scmccarthy says:

    I largely agree. I want to respond to your list of potential ways it went wrong.

    Regarding 1, I think it’s actually arguable that Legion Commander WAS the dominant red hero. It seems that every red deck played both, so it’s hard to disentangle. Axe being focused on more then LC looks to me like it’s because of rarity, to me, but personally I thought both of them were a bit of a problem.

    2/3/5/7/8: totally agree.

    4: I don’t play DOTA but this seems quite plausible.

    6 (Deck balance was hurt by player inexperience): I think I have more to add to this point. I first paid attention to the game when they started the public beta, but I was still unable to play. Here are the first impressions I got (accurate or not):

    * The insiders (alpha testers) have already figured out the metagame.
    * The insiders are playing largely the same heroes in every deck.
    * The insiders are largely saying they think constructed is stale and lame, and thus draft is the “real” game.

    So there are several problems here.

    First: Constructed requires money to play, so I’d rather trust other sources about whether buying in is worth it. The sources say draft is better, and that’s free, so why should I bother? Result: I don’t buy in, and so my complaints about what’s wrong with constructed are necessarily ignorant.

    Second: The main thing I like about constructed is exploring the metagame and trying out new ideas. But apparently, that’s all done. Some unknown number of people who supply the vast majority of information about constructed have been playing it for some large, unknown timespan. And apparently, hero lineups are relatively homogeneous. Result: I’m not excited about making discoveries and wait for a shakeup when everyone’s on an even playing field and there are more unknown decks to discover.

    Third: The primary obstacle to me even trying out constructed is that I’d have to buy the expensive cards. Result: I wish for those cards to become less expensive or less necessary, so I can try it out at lesser expense.

    In summary, I think they screwed up by giving the insiders too much and the public too little. A balance patch when moving from beta to release would have been entirely unsurprising, and would not have created the expectation of regular balance patches, but it would have shaken up the metagame in time for everyone to join in the fun. (But, I think this is only one factor among all the others you mentioned.)

    • TheZvi says:

      Insiders thought they’d figured out the metagame and how constructed works. Doesn’t mean they were right. I certainly have figured things out post-Alpha that we didn’t know back then. Also, during the Alpha Cheating Death cost 3, which was a huge deal – I thought it made RG the best deck by a fair bit which was why the UG combo deck was a thing even though it’s objectively kind of janky, since it wins that matchup. After the first nerf to CD, I shifted to liking RU.

      It’s worth noting that moving CD from 3 to 5 *was* a release nerf patch of sorts, and shifted the color balance considerably. It may have been a major reason why RB was so popular with regular players out of the gate and might have made the situation worse (although given how much players hate CD, it’s not like they could have left it like it was!)

      Even pre-nerf, you could certainly play RG, UG, Mono-U, UB or RB in at least one form. And while the main stuff everyone agreed upon, the minor stuff there was still disagreement on – it’s a different type of constructed experience to work on details, but it is still very much a thing.

      I do think that the Alpha people getting out in front hurt people’s perceptions of the metagame and constructed, and we can call that reason #9, and it was a big deal. The experience of exploring the format when it’s wide open was definitely pretty great for a while. Especially when I first got to Valve and learning the game, and a bunch of cards got changed in response to what they saw that first week (e.g. Horn had a 1 cooldown, CD/Oath didn’t need a hero in the lane…)

      A lot of this is a question with heroes is, should we think of heroes like a mana base, where mostly you play these two colors so you use these duallands and then have marginal decisions, or should we think of them like spells/creatures and demand more variety. In constructed I got used to thinking of them more like mana, at least for now.

      On Axe vs. LC, I’m saying Axe was best based on my experiences, I had Axe substantially above LC when it was 7/2/11, although others disagreed. I did play one major tournament with Axe and no LC, because I had 4 green heroes. I learned this was a mistake, because Axe landing in a poor lane early could be very awkward and because opponents could plan against it too well once they saw my list/plan and were good, so next tournament I switched out the 4th green hero for LC.

  2. scmccarthy says:

    Separately, I want to make an argument that Axe/Drow did deserve to be nerfed.

    I grant your argument that not all cards should be the same power level. It sounds to me like you’re more in favor of compressing the band of power levels than, say, WotC, and I think I’m more in favor of compressing it than you, but we are working in the same framework. (I’d be curious if you can point to an example of a CCG compressing the power level band too much.)

    That said, there is a very apparent pattern in the heroes that were played near-universally: most of them are used to stop your opponent from playing cards. Axe and Legion Commander, in combination with their signature spells, are direct hero removal. Drow’s Gust is an uncounterable Abeyance. Phantom Assassin is removal both with Gank and her signature.

    The exception was Kanna. She was played because Prey on the Weak is an undercosted win condition. (Though you could even make an argument that she was a hero who could survive a red removal spell or Gank, and cast Annihilation, and therefore she was also removal.)

    Looking at the differences between constructed and draft, the prevalence of these hero removal abilities and the resulting increased importance of initiative seems like the main distinction. And I think it might be an unfavorable one for constructed. Initiative and denying plays via hero kills are already present to an interesting extent in draft, and in constructed, my limited impression is that the dynamic feels overbearing. I could be wrong – I’ve played next to zero constructed, and of course I’ve literally played zero constructed in the counterfactual world where the hero-killing is weaker. But I think it’s plausible that it makes the game less rich, by overemphasizing these big swings and underemphasizing smaller plays. If the alpha testers are mostly saying that draft is better, it’s certainly worth considering whether this difference in dynamics is a contributing factor.

    The changes to Axe, Drow Ranger and Jasper Daggers are small steps in the direction of less hero denial. I can’t say they’re the changes I would have picked but I do think they’re positive.

    • TheZvi says:

      Good question on a game that went too narrow! I’m tempted to call out Hearthstone, at least in the sense that they did this *to the base set* and systematically nerfed every high-quality card in it down to size. It’s also the trap I don’t want to fall into where everything great gets nerfed but nothing is then made great. I certainly think things like taking out Innervate and Wild Growth make me sad. They do of course print stupid new cards in sets periodically, from what I hear – I no longer follow it. ‘

      Eternal also feels like it’s getting this habit of lining up similarly powerful cards that organize into lots of strategies, which gives you a very diverse metagame and has some big advantages, but feels off to me in important ways. It’s like there were guys in a room who took each of 20 things to do, and figured out what their curve looked like and made good cards at each step. Then, if a card did prove good, they deal with it. Rather than, lets intentionally make some tools you want not so great and others very good, and force players to make tough choices. They then make up for the ‘all good tier cards are similarly good’ issue by making their mana so generous that playing two colors is all but free, three is cheap and four is somewhat doable.

      It’s not so much that I think Magic pushes cards so much too much, more that they do it with generically good cards that are often mythics and in ways that are almost locks to happen, as opposed to letting cool emergent things happen more. I especially hate the planeswalker love, of course, and no matter how many things in my game end up similar to Magic I definitely don’t want an analogue to those.

    • TheZvi says:

      Now to talk about the specific card issues.

      I do think that Artifact is largely about fights to kill heroes and stop spells, and that the top heroes being about that isn’t an accident, although it’s also just where they chose to make the best heroes good.

      For red, Axe and LC have great stats plus great cards. Other red heroes can get you one but not the other. There was just a very large gap there, but a lot of this in my mind is that we need to not give mono-red a good hero set. I don’t want their 4th and 5th heroes to be close to Axe/LC, I want there to be a price – again, that’s the mana mindset here. I do think that with a smaller Axe there’s still enough drop-off here to make this OK, but I wouldn’t want to go any further. And Axe going down to 6 makes 7s like Drow a ton better.

      Drow offers TWO unique things in green, to me having Gust available was a key part of what green does, and so was the +1 aura. I like that green has that ability, and playing green decks that cut Drow feels completely different. I’m tempted to say, actually, that the thing to do now is to restore Gust *but remove it from Drow*, if we can do that, and swap it with a rare that becomes Drow’s new signature, and then adjust Drow’s stats based on how good the new card is? Dunno if that is possible for various good reasons, though.

      I think you are wrong about Kanna. Nothing wrong with Prey, but the reason I play Kanna all the time is not that. It’s that PotW is *decent*, directing the guys is net positive, and Kanna’s stats are 2/12. This means Kanna will survive against most attempts at removal, so Kanna is excellent at getting your spell off, doesn’t face the situation where everywhere you deploy risks instant death that I often find on turn 2 with a 3/7, and also the best hero to cast AAC since that doesn’t leave Kanna on death’s door like everyone else.

      If green doesn’t have the true Gust ability, it can’t stop Annihilate and the other mass removal (yes, there’s that stop-cards theme, and yes it’s how constructed works and I embrace that), so its strategy of going wide in various ways gets blown apart. The game is balanced around green having that power. Green needs *some* way to efficiently stop mass removal that works on Annihilation, assuming we’re not targeting that next.

      Similarly, blue is balanced around the fact that its heroes are constantly killed or silenced or stunned. Can’t just take that away…

      Full constructed is about managing/crossing lanes aggressively, having big haymaker plays and answers to them all the time, and knowing when to knock out your opponents’ ability to play spells (and often fighting for initiative to do it). Commons only is a cool constructed variation we tried in the alpha, where crossing lanes well is hard, hero placement is much more long term, and spells are easier to cast. I wish there was a queue for it.

      Draft is a blast in some ways but I’m not entirely happy with the current draft format. But I’ve always been a constructed guy.

  3. Sniffnoy says:

    Huh. This is certainly interesting. My own views lean more towards those of, say, David Sirlin, who would say that
    A. Of course things should be balanced; while not every matchup has to be even, the main factor in who wins should be the choices made during the game, not before it, and, moreover
    B. Having to pay for the particular cards you need, with better cards more expensive, rather than having a pretty-much-flat buy-in from everyone (as you would have in non-CCG games; fighting games are I guess the standard example here) is pay-to-win bullshit.

    Of course, it’s not clear that you can try to do what Sirlin does, and also make a game where you’re building something as large and complex as a Magic deck beforehand. I mean, he tried, didn’t he? Sirlin said that Codex would be like a CCG but addressing (A) and (B), to the point that playing random is viable (which he seems to use as something of a criterion of balance — I mean, it sounds ridiculous in anything where you’re constructing something more complicated than picking a single pre-defined character, but he made it work in Kongai[0]).

    And of course if you take the goal as originally stated, he failed. Picking three characters and an item on each, you can balance to the point that playing random is viable; building a whole Magic-style deck, I don’t think you can. In Kongai or Codex if you go random you lose out on some synergy; in a complex 60-card deck (or 40, or 30) you’d lose out on so much it’s just silly. Of course what he did was go for sucessfully addressing (A) and (B) while abandoning the whole deckbuilding-like-Magic thing, instead having pre-game deckbuilding consist solely of picking three distinct heroes and a matching starter (which you can do randomly!); most of the deckbuilding takes place during the game instead. So he succeeded at making a well-balanced game, just not at what he originally said. Of course this is Sirlin we’re talking about so (being the pompous ass he is :P ) it’s not like he’s going to admit that[1].

    I guess a lot of it comes down to point (A), really. I don’t want to build a deck outside the game. It takes so long and I don’t find it fun, and then when I do play so much of the game is determined by choices that have already happened. And if I am making pre-game decisions, I want them to be discrete (by which I mean, like, *effectively* discrete); I don’t want the opportunity to endlessly tweak small choices, because, well, then I will and that’s not fun. (I guess that’s (C).) Which is why I don’t play Magic, really. :P Someone hands me a deck I’ll play it, but then I’m at the mercy of what deck they’ve given me; this addresses (B) and (C) but not really (A). Or I’ll play in a draft (though I’m awful), because that moves the deckbuilding to be essentially a phase of the game, rather than a separate outside-the-game thing, and it reduces your ability to tweak endlessly (so that’s all of (A), (B), and (C) each addressed to some extent). But I refuse to actually own cards.

    It’s interesting, I think, to note the (suggested) counterpicking rules, or character/deck-fixedness rules, for different games. For fighting games (and many other things), you can pick a new character each match, and can counterpick to any character[2]. For Codex, the suggested tournament rules are sticking with the same deck (as in, hero/starter choices, not in-game deck obviously) throughout and having no counterpicking at all; you’re already mostly building your deck during the game, so there’s no need for counterpicking, is the reasoning. Meanwhile Magic, as you certainly know, has its sideboarding rule, which is somewhere inbetween — you have to pick one main deck that you will use in the first game of each match, and in subsequent games you can “counterpick” via sideboarding (done double-blind, so it’s not really “counterpicking” in the usual sense since counterpicking is meant to advantage the loser of the previous game, but I’ll count it).

    Well, it’s somewhere inbetween ruleswise, not in effect — a given Magic deck is obviously not as versatile as a given Codex “deck”. And like, that’s interesting, yeah, but… ick. Not like I play any of these games at a tournament level, so none of this is really relevant to me, but, IDK, I still just have a big problem with (A), I guess. Counterpicking and sideboarding rules reduce the effect of (A) by saying, well, hey, at least it’s only per-game, effectively making the choice of which character or deck to play part of the game rather than the metagame (much as drafting also does as mentioned above) — you’re not making a single choice before the whole tournament based purely on what you expect other people to be playing. But Magic’s sideboarding rules, while obviously they reduce it, are built so as to not entirely give up (A), because there (A) is regarded as a good thing. Blech. I don’t like it.

    I don’t really know where I’m going with this. That’s the contrary point of view, I guess — but, as mentioned above, it’s not just contrary to imbalance but also to complex pre-game deckbuilding; those who want balance but also complex pre-game deckbuilding may not be satisfiable. <shrug>

    [0]I mean, playing random wasn’t great or anything; if you weren’t playing random, item choices were pretty much standardized, and there were generally-recognized best characters; but consensus was that almost every character was nonetheless viable and people even did really play random and do well with it despite the suboptimal items it might give you, so I guess he succeeded there.
    [1]Admittedly, he never explicitly said deckbuilding like Magic, but if you talk about making a CCG and don’t say it won’t be like that, I think it’s implied that it will be.
    [2]Well, traditional fighting games have a rule of “only loser counterpicks”, rather than Smash-style slob picks or Magic-style double-blind (considering sideboarding as counterpicking), but that’s not what I’m concerned with here; the point is, when counterpicking, you can pick anything.

    • Sniffnoy says:

      Although, um, rethinking things there, I guess part of that conclusion is because somewhere along the way I implicitly accepted Sirlin’s extremely strong “you should be able to play random” balance idea. If you throw that out and just want more card-level balance, then, IDK, maybe not so impossible. But it’s not clear to me how much that’s a worthwhile goal, really, if you still have to do complex pre-game deckbuilding.

      I guess also DOTA2 works as an example in place of fighting games above, though I’m much less familiar with that.

      • Sniffnoy says:

        Oh, duh, the reason it might be a worthwhile goal is to address (B), even though it doesn’t do much about (A) or (C). Should’ve thought that through more before posting. (To my mind (A) and (C) are my primary problems with such games, I guess, and (B) would be a problem except it never actually becomes relevant due to (A) and (C) keeping me out in the first place.)

      • TheZvi says:

        I thought Sirlin’s writing about games that weren’t TCGs was mostly excellent. I think his views on TCGs are madness. I did not stick around to check out his game at all, though. Do you think that was a mistake?

        In terms of your A/B/C:
        A: I certainly think that picking random being a viable strategy is both a crazy bar and destroys the whole thing about coming up with a plan and strategy. Random choices should be bad, or choices don’t matter. This idea that choices *during the game* need to trump choices made outside it? Sounds like you want to play a different type of game, sir. Which is fine! But it’s not different from draft, it’s all part of the game if you accept the genre, whether or not cost is an issue. And to make it a level playing field does not require that every component be equal strength, since we’re *not* choosing at random, we would just need all players to have equal access to the same choices. Yes, we also introduce randomness, but random is part of most games. Could have another discussion about matchups and how fair or lopsided they ideally should be, either by default or based on player choices, but I actually am a pretty big proponent of hate cards and think Magic made a AD-style Huge Mistake by abandoning them.

        B: Pay-to-win is, by default, some bullshit. No question. So by default, your game shouldn’t do that. The reason TCGs are so popular and generate so much revenue is some combination of: That players are getting good experiences out of the collectible side of things that compensate of the issues, and because the games are just so good as games that we are willing to pay what they cost. We can think of it as a form of price discrimination that maximizes revenue. Magic in its pure constructed form is pay to win, but more accurately it is simply pay to play, but the amount you pay is too high to just charge outright, so it needs to be obscured. But we pay because it’s worth it, man. The question is, are you providing that level of value, and what is the collectible aspect doing for your players and your game. Otherwise, you can just do a living card game / standard fixed fee.

        But suppose Magic changed (ignore transition costs for now) to a subscription/purchase model. What would you charge? If you charge $50/set you give up most of your revenue. If you charge much more, you lose most of your players. Etc. Hard not to sympathize.

        The other response is, this game incorporates prices and value in its structure. Poker is the obvious existing example, in its own way. That’s the approach I’m exploring and I’m pretty excited by it.

        C: A lot of people including myself do enjoy going that deep. Others don’t bother going deep. It’s that weird space in the middle (e.g. the guy who learned Deus Ex Human Revolution gave a tiny XP bonus for hacking terminals and then couldn’t not do it every time, so the game became terrible for him) that you’re in where small tweak options become bad. One could recommend copying deck lists from others and then not changing them, if you dislike tweaks?

        Let us return to the fighting games, though, because I think that’s instructive. Each fighter has its strengths and weaknesses. She’s good at jumps and kicks, he throws fireballs, he teleports, her moves are super fast, etc. As individual components we want them to be balanced. But do we want them actually all balanced? If there are 50+ smash brothers heroes, all equal, then I can’t learn what’s good and bad or be surprised by how to use what’s not as good, or handicap better players, or have a good single player campaign progression, etc etc. Plus, 50+ good heroes means I have to learn 1000+ matchups to know all the heroes, or 50 just with my one (and because of countering I need to know a bunch of ’em). So I would argue, no, I do *not* want all 50 to be good choices at random. I want 10+ of them to be good choices at random, and I want each other fighter to *serve a purpose*. Maybe Zelda is mostly terrible but she utterly crushes the top-tier Luigi. Maybe Sonic is really easy to play and wins if you’re new, but is bad at the higher levels when people can react fast enough. Maybe Toad is what Axe wants, a real challenge (to get to do anything). And so on.

        You could also look at the fighters as colors, and your kicks, punches, jumps and so on as cards. In that case, Chun Li has a great kick card, and will tend to play a lot of kicks. You don’t then say, her kick is broken and her punch sucks, we need to balance this. That would be silly.

        Anyway, I should get back to the next section, but good thoughts here from both of you, so thanks.

      • Sniffnoy says:

        (Going out of order here for simplicity)

        I did not stick around to check out his game at all, though. Do you think that was a mistake?

        Which game, Codex? Um, dunno. Seems like the sort of thing you’d like, if I had to guess? I like it but am probably not the best to evaluate it (I am awful at it). Dunno if I like it so much it was worth the price, but oh well. The biggest problem I have with it is that it’s easy to fall behind and then just never be able to recover, and this can take a while to play out. (Like, you can fall into a loop of “OK, I’ll put some stuff out, and if I can just keep it alive till next turn it’ll help me recover” — and then you can’t keep it alive till next turn.) But it’s also possible that this is just me/people I play with not yet understanding how to recover better. Also, I would definitely prefer if it were more old-Magic than modern-Magic (in the sense that you’ve talked about — more weird global effects and less everything-is-creatures), which seems like a relevant warning to you, but, oh well. Anyway certainly it’s fun, but I’d probably say you should read other people’s reviews rather than ask me TBH. Like do other people get into tech 2 fights as rarely as I have when playing with my friends? (As in, both sides make it to tech 2 and fight it out with tech 2 stuff, not one side makes it to tech 2 well before the other and then steamrolls.) Because tech 3 units were intended to be the game-enders, and sometimes they are, but it sure seems like tech 2 is a lot of the time. But I get the impression maybe that’s just me/who I play with? (Most of my games have been against one particular friend of mine who is way better than me at this sort of game.)

        Regarding the fighting game example at the end: Yeah I agree completely with this. This is roughtly the sort of thing I had in mind when I said above that I didn’t see the point of card-by-card balancing, but thanks for making that more explicit. Especially since really it’s less “I don’t see the point” and more “That would actually be a bad thing in a number of ways”.

        Basically yeah I think I’m concluding that these two styles don’t mesh well, and likely Magic is right to do things its way, even as that way is not for me. Definitely interesting that you’re trying to see if you can go even further in that direction, even though I’m sure the result will be something I’ll stay well away from.

        I guess what is likely worth mentioning here — and this touches on (A) and your response to it as well, though more on that later — is that I’m big on the idea of, like, the magic circle, that games should be self-contained. Obviously, this is going in the opposite direction of that! But then I guess one of the motivating ideas of Magic was that it should be “bigger than the box”, so…

        Btw regarding (C), it might be more accurate, rather than saying “and then I’ll do that and it won’t be fun”, to also include the alternative “and then I won’t do it, and then when I lose I won’t know how much of my loss to attribute to skipping this detail, and it’ll be frustrating”. Or something like that.

        As for your response to (A), quickly:

        But it’s not different from draft, it’s all part of the game if you accept the genre, whether or not cost is an issue.

        Right, I mentioned above that draft resolves this problem by bringing the choice inside the game. Or did you mean something different?

        And to make it a level playing field does not require that every component be equal strength, since we’re *not* choosing at random, we would just need all players to have equal access to the same choices.


        Could have another discussion about matchups and how fair or lopsided they ideally should be, either by default or based on player choices, but I actually am a pretty big proponent of hate cards and think Magic made a AD-style Huge Mistake by abandoning them.

        I am definitely more on the “matchups shouldn’t be too extreme” side, but once again it’s worth noting (as I’ve already said) that IMO there’s less of a problem here if you just bring the choice inside the game (which, as already noted, can be done by expanding what counts as inside the game, as happens e.g. in draft, or in say Smash tournaments with their free choice of characters but rules about the order of picking. (Constructed Magic tournaments don’t fully bring the choice inside the game seems to me, but I guess they do so to a greater exent than informal Magic just by being tournaments in the first place; we can imagine that the choice of deck is a multi-way blind choice beforehand.))

        I certainly think that picking random being a viable strategy is both a crazy bar and destroys the whole thing about coming up with a plan and strategy.

        So, there are several things here I want to respond to.

        First, I agree that “random is viable” is probably too high a bar; but obviously it’s not impossible on account of it’s been done. But Sirlin used it as a bar so I thought it was a useful point of comparison — it’s something that’s possible for things more towards the self-contained, pick-a-character end (but only for those who want to obsessively balance their game (or not introduce any substantial differences in the first place but we’re not talking about that)) and impossible for something more towards the Magic end. There’s probably a region in the middle, I guess, where (unlike in Magic) balance is actually a reasonable goal, but “random is viable” isn’t just a high bar but actually an impossible bar. IDK. Anyway this paragraph isn’t really saying much interesting so I’ll move on.

        Second, I’m not sure the inference expressed by this sentence is really correct. I mean, I had some sort of strategy in mind every Kongai deck I came up with, you know? (Though they didn’t necessarily play out that way when I actually tried them.) But also you could just play a random deck and figure out what sort of strategy it required on the fly. But basically I guess this may once again come down to:

        This idea that choices *during the game* need to trump choices made outside it? Sounds like you want to play a different type of game, sir.

        Or in other words, the question is, which part of the game do you find to be the fun part — coming up with a plan, or adapting to circumstance? And I guess I lean more towards the latter side! In Kongai for instance I always thought the most fun parts were having to play against strategy, being in a bad position and having to play defense with an aggressive deck (or what I thought of as one, anyway; maybe it was less so than I thought), because one bad decision there could really sink you.

        I guess it’s also a question of how much you want to invent a strategy. In a game with just fixed characters, random can be a viable choice — you just have to know the (usual?) strategies for every character. You’re not inventing them. (I mean, characters could perhaps have other, less-intended strategies, but ignoring that…) Or if (to use the example of Kongai again) you’re picking 3 characters at random, well, then you have to be able to figure out really quickly how those 3 strategies can be meshed together into a larger one.

        Anyway I don’t know where I’m going with this. Sounds like we don’t really disagree on a lot, just have different tastes in games. :P

      • Sniffnoy says:

        Oh wait I missed the obvious counterexample which is e.g. chess or something similar. :P Barely any asymmetry, yet the idea of strategy doesn’t vanish…

      • TheZvi says:

        Chess is an interesting case, here and otherwise, because it’s so clean. Chess starts out fully symmetrical, but white going first gives incentive to break that, and once the pieces are in subtly different locations, interesting things start to happen. And that’s enough.

        Chess also shows the big plus and minus of a stagnant set of rules and initial conditions, in this case fully stagnant since no luck at all. Knowledge is permanent and becomes oppressive, but it also means that players can develop and learn and be rewarded over lifetimes, which is very cool. If we wanted to change chess to “mix it up” at higher levels, we would want to avoid changing what the pieces do, but perhaps it would be good to force players into selected openings, or change initial piece locations sometimes (e.g. Fisher chess) or similar…

      • Sniffnoy says:

        Yeah that’s worth noting as a disadvantage of keeping things static. In any case, it remains a clean illustration of my point that a kit is not a strategy. :)

        But I want to take a moment to consider your point, because it’s interesting. This ties in with your most recent post as well. While I’m generally in favor of balancing — although, after this discussion, maybe not for games where you build a full deck beforehand! — I’m wary of rebalancing. At least, ongoing rebalancing. I like the idea that there is a point at which the game is done and it’s up to the players to advance the metagame. I guess the question becomes, how long can a game last under those conditions before it ends up like Chess with so much memorization? Obviously randomization helps as you noted. Note that from this point of view, Magic’s addition of new cards (and removal, in the rotating formats[0]) is a form of rebalancing; the game is never “done”.

        Unfortunately I don’t know how one would go about getting data on this. Because how long a game lasts as something that people play has to do not only with the qualities of the game itself, but people’s attention span, to what extent there are other similar games, etc. I don’t know how one could go about making comparisons here.

        [0]I’m excluding bans, because bans can happen in any game when things are sufficiently extreme, and not really what I want to consider here. And there are only so many things you can ban before you have nothing left! :P

      • TheZvi says:

        Even in Artifact I’m developing an ‘opening book’ over time, as patterns repeat and I develop rules for what to do in various situations, despite massive amounts of random elements everywhere – you just learn that this is the right play in each branch or type of branch, and use that to narrow other choices, and so on. How long things last before breaking down depends as you note on a lot of factors. But even with zero random element and tons of time and work, chess is *more popular than it has ever been* which is pretty great, and bodes well for letting good things become static.

        And I couldn’t agree more with, you create a thing, and then it’s done. If you then create another different thing that’s mostly when you fix things. Then you let things break, and do what you need to do, and let interesting things emerge.

        It’s likely not a coincidence here that I’m a fan of speedrunning – note to everyone that AGDQ19 is happening now.

    • deluks917 says:

      I was very active in the sirlin games community for many years (sadly the community is mostly dead now, but it had a good run). I actually have the highest tournament game/match winrate of anyone in Yomi who has a non-trivial number of tournament games. I also had the highest average ELO over my career. Yomi was his most popular game. So I think have substantial experience with the results of Sirlin’s philosophy.

      Yomi had 20 characters and the balance was quite good. But eventually people realized that you mostly only wanted to play 4-5 regularly in tournaments (Troq, Geiger, Zane, Degrey and maybe Grave). Some other characters were occasionally useful but lots of yournaments got stuck in the same counterpicking loop: Troq > Zane > Geiger > Troq. If someone was down enough in the set it made sense for them to switch to someone like Grave/Degrey who were hard to counterpick. This eventually all got very stale. I wish sirlin had reblanced or otherwise shook things up.

      I am discussing yomi because Sirlin arguably succeeded in his aims. The balance is about as good as you can hope for given that he never changed the cards after ‘release’. Despite the balance there were still problems. Here is a mu chart I made years back.

      In Sirlin’s other games he mostly failed to achieve the levels of balance he wanted. All versions of Flash Duel before the current one (which I cannot comment on) have had big balance problems. In one version of Flash Duel he had to nerf a truly massively overpowered character (Quince). Even after the nerf the second best character (Setsuki) was clearly S tier. Tournaments had to run a house rule nerf to Setsuki. Even with a Setsuki nerf the game had maybe six playable characters and they were not even balanced within the top tier. Puzzle Strike and Codex have not had the same problems as Flash Duel but the balance isn’t great. Sirlin and friends did their best but balancing things is very difficult.

      I think there are two lessons here. One is that trying to create a balanced game is very difficult. If you don’t rebalance or ban things you will probably fail. If you play them long enough many magic the gathering formats settle on a single best deck. Secondly even if you succeed in balancing a static game it might still get boring. Games like Chess and Go are the exception not the rule. Your game’s core mechanics are unlikely to be as compelling as the mechanics in Chess or Go.

      • TheZvi says:

        Good data!

        I see the lesson here somewhat differently. The lesson is, you need a lot of different meaningful decision points to create reliable and lasting balance, in the good sense.

        The problem with Sirlin’s games, based on what I remember from reading him (but not playing) and your description, is that players picked characters and then got a fixed set of abilities based on what they picked. This in Magic would be if each color combination was given a fixed deck of cards to play with. Then, over time, it becomes clear through proper play and discovery what the equilibrium is, because the decks/cards/fighters can’t adjust.

        DOTA 2 makes lots of heroes relevant, with lots of matchup and team-based dynamics, but if it didn’t do microadjustments that would be hopeless. I think without component knobs for players to turn out-of-game, you have very little margin for error. Magic made those knobs matter less, which is a large part of why recent Standard formats have run into so much trouble. There’s no counterplay response worth a damn.

        Magic formats do not need to degenerate into one-or-two deck formats with extensive play, even without extensive tweaking. We have lots of historical examples of formats that didn’t. Even if there is a zero-level ‘best deck’ that doesn’t mean you should play it any given tournament in the stable state.

      • Eric Fletcher says:

        Another facet of Sirlin games is that they have very little luck in them, which means that the better player wins 80 – 90% of the time in a mirror match. Contrast that with Chess (99%+ win rate for the better player) or Magic (65-70% win rates for top professionals in a Grand Prix).
        Combined with a low number of entries (20 for Yomi/Flash Duel/Puzzle Strike; [20*19*18] for Codex), this leads to matchup charts being actually calculable, and the “best decks” being discover-able. Again, compare to Magic, which has on the order of [100*99*40*39*38*37*36*10*9*8] possible decks, even if you limit yourself to Standard legal, 2 color, only play 9 cards as 4-ofs + 24 lands, ignore sideboards, and only look at the 100 “constructed” cards (so, about 9% of the total cards printed – or about half the total rares/mythics). And Magic doesn’t even try to have their balance last more than 3 months before changing the set of possible cards.
        In terms of total player-hours, though, I suspect Yomi et. al. devolved in about 36 months with about 100 (tournament) players? whereas Magic averages 2 months, and has at least 10,000 Grand Prix level tournament players.

      • sniffnoy says:

        (Finally getting back to this…)

        Huh, that’s quite interesting, as someone who was at best on the periphery on this stuff. (Personally I found Flash Duel pretty mediocre and I never tried Puzzle Strike or Yomi. Kongai was always Sirlin’s great game for me, but that’s really dead…) Particular notes:

        Some other characters were occasionally useful but lots of yournaments got stuck in the same counterpicking loop: Troq > Zane > Geiger > Troq. If someone was down enough in the set it made sense for them to switch to someone like Grave/Degrey who were hard to counterpick.

        Man, counterpicking loops. As someone who doesn’t play/watch traditional fighting games, that’s a phenomenon I’m not familiar with. But I guess with slob picks instead people would just always switch to someone like Grave or DeGrey instead.

        This does all make me wonder if Sirlin’s whole goal of the balanced asymmetric game is maybe just a worse idea than trying to make a great symmetric game…

  4. Pingback: Artifact Embraces Card Balance Changes | Don't Worry About the Vase

  5. Pingback: Card Collection and Ownership | Don't Worry About the Vase

  6. Pingback: Card Rebalancing and Economic Considerations in Digital Card Games | Don't Worry About the Vase

  7. Pingback: Advantages of Card Rebalancing | Don't Worry About the Vase

  8. Pingback: Disadvantages of Card Rebalancing | Don't Worry About the Vase

  9. Jon Ross says:

    I’ve been thinking a lot about the concept of axes as described by JoINrbs in the context of Slay the Spire. (That’s axes as in the mathematical concept, not Axes as in the red Artifact guy.)

    The gist is that often it’s not correct to think of a decks “goodness” as a single scalar. You can have a Slay the Spire deck that’s good on most axes, but has no AOE and struggles with the Gremlin Leader. You can have a deck that’s very good at defending, but has trouble dealing the sustained damage to beat the Giant Head. And so on.

    I think part of the reason Dota succeeds in being well balanced despite having a huge variety of things you can do is that multiple axes matter and heroes have a lot of degrees of freedom to be good on different axes. What emerges from this is that heroes that are not necessarily top tier have drafts where they are situationally very good. This is one of the best things about Dota as a spectator experience in my opinion.

    Magic is also pretty good at this, you can have a deck that has a lot of tiny meta advantages that add up to make it the best deck for a tournament, even if it’s not necessarily powerful in a vacuum. This presents opportunities for deckbuilders who really understand the game to press their edge, and it also adds variety, which is awesome.

    From my limited experience with Artifact I don’t think it’s as good at this. It feels like there are fewer axes to play with in deckbuilding. I’d go into more detail on this but my epistemic confidence is low.

    This is the video where JoINrbs introduces the concept:

    • TheZvi says:

      Great points, definitely going to check out that talk as it seems relevant to several of my interests.

      Artifact definitely has several axes, in this sense, that matter – attachments, AoE, tower damage, single damage, killing creeps, boosting or multiplying creeps, killing heroes, crossing lanes, sustainable hero boosting, speed versus power, armor and armor piercing or big damage chunks are what I can come up with off top of head. That’s not the depth of Magic, but it’s a damn good start. I do think that in limited play, there’s less of these dynamics, but that is true in Magic as well.

  10. Pingback: Game Analysis Index | Don't Worry About the Vase

  11. Pingback: Artifact: What Went Wrong? | Don't Worry About the Vase

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s