Going Deep – Balance and Fun

timthumb.png
I have been thinking a lot recently about balance. Both Magic and Hearthstone have been struggling with some serious challenges with balance, which has gotten me really chewing on the subject. Specifically, I would point to a Podcast with Patrick Chapin and Michael Flores, as well as a rant on SCG Live by Patrick Sullivan that were extremely thought provoking. If you have not seen either of these before, go check them out. They are transcendent. In addition to just being stellar rants, both Patricks work at DWD, so their perspective on the subject of balance will have an impact on in Eternal. This morning I was looking at some comments about the upcoming balance patch, and a few things suddenly snapped together, and I felt the need to share my thoughts.

Before we get into that, we need to talk about fun. Over my entire time as a content creator I have been threatening to write a long sprawling manifesto on the subject of “fun”, but I have suddenly reached some conclusions on the topic that can actually be summarized in a fairly short piece (well, at least short by my standards). So, what makes strategy games fun?

Fun

I want to be very precise about what we are talking about here. I am not talking about all forms of fun, or even all the forms of fun you can get from games. I am not talking about the fun I experience talking with friends, or watching a good movie, or playing with my son. In some respects these are all slightly separate phenomena that seem to tickle a similar part of the brain. I am not even talking about the fun you get playing strategy games with friends, or the rush of doing something spectacular within a strategy game. Those things are clearly fun as well, but I am zooming in entirely on the subject of “what makes strategy fun?” The core of the matter is this:

The fun of strategy games comes from making meaningful decisions, and being rewarded for them.

This statement may seem slightly self-evident, but I think you will see that it is useful. First, this gives us a simple explanation of what makes one game or format more fun than another. A game where you are faced with interesting decisions, and you feel rewarded from making good decisions is going to feel fun. A game where there are few interesting decisions, or when the rewards of good decisions are easily nullified is going to feel less fun. When someone says they don’t think Set 3 draft is fun I think what they are really saying is that they don’t think there are enough meaningful decisions, or that good decisions are not sufficiently rewarded.

Second, I think this gives some insight into the deeper psychological phenomenon at play. It is satisfying to be faced with a tough decision and then being able to “solve” it, making us feel powerful and clever. When faced with a challenge we showed skill and mastery. This satisfies a certain urge within our psyche, and the entire exercise of playing strategy games revolves around scratching this mental itch. You could even make an argument about why this exists from evolutionary psychology; it is a competitive advantage to be attracted to tools and methods that are more effective at achieving our goals. In that sense the urge to compete in strategy games is likely a side effect of the same mental features that leads humans to explore, discover, and build. At some level we don’t do these things because we will derive the direct benefits of these advancements, but rather doing something no one else has been able to makes us feel like a badass. It is that great quirk of human nature that somehow tricks us all into spending thousands of hours playing children’s card games.

With this simplified explanation of what makes strategy interesting let’s slice-and-dice it up further to understand some of the implications of the components. None of these points are totally separate from the others, but rather they are all different aspects of making strategy games fun.

1. There must be some meaningful strategic decisions

The most obvious ingredient of making a strategy game is that there must be some decision points with strategic implications. I would define a strategic decision as one where you must make choice on how to allocate resources to meet different goals or objectives. There are clearly countless examples of this: which 1-drop do I play on turn 1? Do I spend my minerals on building my military or my economy? What technologies should I prioritize? In good strategy games it is not essential that every decision needs to be meaningful, but at least some decisions in the majority of games need to be meaningful. Basically every game of Eternal is going to have some decisions that have no strategic relevance, as well as some decisions that have lots of strategic depth. I want to stress that all the decisions that have an impact on the game can be strategically meaningful. What deck did you choose before the game started? What build of your deck did you use? Those are just as important as in-game play, and can similarly be a source of strategic depth.

Not every decision in a game is going to be meaningful. Rock-Paper-Scissors is an obvious example. This is clearly a game all about making decisions, but it is hard to argue that those decisions are particularly meaningful. Though it might feel good to win a round of Rock-Paper-Scissors, you will not find it particularly satisfying if you were playing it for hours. There is just no strategic depth to the decisions you are making. On the flip side, it is actually possible for games to have too much “strategic depth”. I haven’t played the most recent iteration of the Civilization franchise, but one of the defining features of the series is the ability to micromanage every facet of your nation. While I found this very interesting, I know a lot of other people would just glaze over rather than trying to squeeze out the last bit of efficiency from their cities. Not only were these micromanaged decisions exceptionally fiddly, they barely had any impact on the game, making the “correct” decision barely an improvement over making an “incorrect” decision. As such, it is easy to feel like these decisions would begin to lose strategic meaning, leading to frustration for a lot of players.

It should also be mentioned very briefly that one of the most important questions in strategy game design is whether execution should be an important element of the game. The entire genre of strategy games is divided between turn-based and real-time strategy. Card games as a sub-genre obviously fall on the turn-based side of this arrangement, but real-time games place weight on both technical ability as well as strategic choice. Though this has a profound impact on game design, the source of the fun in both cases is pretty similar. The satisfaction you get from perfectly sequencing your cards to win a game is very similar to the satisfaction you get from perfectly microing your units to win a key battle. Games that stress execution like Fighting games and the First Person Shooters do have strategic depth to them, but are not really strategy games. The relationship in these genres between technical proficiency and strategic thinking is complicated, but it is clear skilled execution and timing is an essential ingredient that differentiates these types of games from strategy games.

2. Good decisions should be rewarded (appropriately)

Games should be structured such that the player who makes the most correct decisions is going to be more likely to win, or at least enter a game state from which winning is more likely. In some respects this is trying to link the strategic importance of individual game decisions to the outcome of the game. One of the great challenges of designing strategy games is tuning the rewards you get from good decisions. It is pretty easy to design games where a small early game advantage can snowball into an insurmountable lead, or where single game pieces are so powerful that they wipe out any advantages that came before. Hearthstone has – somehow – managed to be guilty of both of these sins. Board-centric match ups are often decided in the opening turns of the game, while decks like Priest and Warlock offer late-game cards so absurdly powerful that they invalidate everything that happened up to that point. So, on the one hand, good strategy should be rewarded, but not too rewarding where individual cards are too swingy for a given match up.

In some respects I think the desire to get rewards from good play explains certain types of player behavior that might seem irrational at the surface. Take for example players who like prison style decks, or hard control decks with no real win condition other than decking your opponent. Once the game is firmly under their control the strategic meaning of their opponent’s decisions is totally irrelevant. Does that stop this prison-enthusiast from enjoying the game? Nope! They just relish the opportunity to develop a larger and larger advantage, basking in all the rewards they can accumulate. I am sure we have all heard someone say something like “Wait, no! Come back! I wasn’t finished winning yet!” when their opponent is totally locked out of the game. A purely rational player would be happy for the game to end so that they can move onto another match, but given that some players derive their fun from revelling in their own rewards, they are still able to enjoy a game long past when it is functionally “over”.

3. Minimizing the amount of “non-game” moments

“Non-game” moments are times in a game where there are no relevant game actions that can be made by at least one player. If we look back to our definition of what makes strategy games fun, it seems pretty obvious that game-states with no relevant decisions to be made would not be enjoyable. It is also interesting how non-game moments feel worse than just sitting around doing nothing. By playing a strategy game you were promised an experience of making interesting decisions which could lead you to compelling rewards, but instead I am stuck on two power with all 5-drops in hand. This essentially violates the whole logic around why we play strategy games, right?! As such, game designers should try to minimize the experience of non-game moments. This might seem trivial, but as you will see, it is a lot more difficult than it looks, given how it ties in with balance.

It should actually be mentioned that card games are fairly unique in that you can have entire games be “non-games”. Take some other common styles of strategy games, like RTSs or MOBAs. There are no competitive games of Starcraft II or League of Legends where one side is just dead from the start. Yes, I know there it would be possible to create SC2 maps or LoL lineups where one side is essentially doomed from the beginning, but that isn’t something that realistically happens in a competitive setting. In both games if you matched up an inexperienced player or team on one side versus a veteran player or team on the other side the more skilled side is almost guaranteed to win. There is no amount of bad luck that will cause SKT to lose to a band of scrubs. In Eternal, Magic, or Hearthstone it is possible for the most accomplished player to lose purely to variance. I have discussed the importance of variance in other places, so I won’t dwell on it any further, but it should be noted that “non-games” and “non-game moments” are almost unique to card games. The fact that John Finkel can lose to a nobody is a feature rather than a bug. If you want to avoid non-games altogether, you should probably just take up some other kind of turn based strategy game.

4. Limiting repetitiveness

When every game plays out the same things get pretty boring. Perfect information strategy games like chess bump up against this constantly. There are obviously tons of people who have played tens of thousands of hours playing chess, but most people don’t have the appetite to stare at those exact same pieces for all time. One of the most important issues with repetitive games is that they can eventually be solved. Tic-tac-toe is one of the most obvious examples as many people are able solve this game entirely on their own given enough time. Though chess hasn’t been solved yet, much of the game has been distilled into memorizing sequences of moves, and following strategies based on the massive catalogue of all the games of chess that have ever been played. This is only possible because of how repetitive Chess is as a game. At this point, could you actually say you are making meaningful strategic decisions? “I followed this pre-determined series of steps that I memorized and eventually won as a result” is pretty distant from “meaningful strategy”. I know things aren’t actually that simple, but there is a lot of fun drained out of a game when it has been at least partially solved.


I promise we will get to talking about balance really soon, but I just wanted to touch on some of the implication of my view on fun. First off, different players have different appetites for the different aspects I have pointed out. Earlier I mentioned that some players like games to involve an element of coordination and technical fluency in addition to strategic depth, and different games cater to players with different levels of interest in these two sides of gameplay. Someone with a strong passion for strategic decision-making might enjoy the inexhaustible depth of games like Civilization, or durdly control decks like Chalice. Low tolerance for repetitive games? You probably prefer draft to constructed. Low tolerance for non-games? You probably prefer highly consistent decks like Stonescar Burn or Xenan Midrange, to fancier fare like combo decks. I ultimately think strategy gamers are not that much different from one another in what they are looking to get from their games, but the major difference is their tolerance for different aspects of the game. There are a lot of people who speak like the prison aficionado who lives to lock people out of the game is just an entirely different beast than the player who is jams a bunch of 1-drops and burn spells. I don’t really think that is the case. They are getting their pleasure from the same place, but just have slightly different tastes for the game-states they prefer.

Those of you who are familiar with the writing of Mark Rosewater (as well as my own writing) might object to all this. “What about the player psychographics? Don’t they matter? Do you still believe in Johnny/Jenny and Timmy/Tammy?” Yes, I do, but as I emphasized at the start I am only focused on the strategic component of gaming. There is a certain fun that comes from self-expression and exploration, and Johnny/Jenny players are looking for that kind of fun. Given the model that I am proposing these players are likely drawn to non-repetitive games, as well as navigating meaningful decision trees, but the particular fixation on playing a unique deck means we are tapping the fun the fun that comes from creativity. Timmy/Tammy players are clearly drawn to spectacle as well as strategy. These players will likely be drawn specifically to decks that give higher impact rewards, but they are clearly drawing some satisfaction from the game beyond just the strategic merits of the deck. I don’t think there is anything wrong with using strategy games as a means to exercise your creative expression, or a method to create an exhilarating spectacle, but we should be clear that this satisfaction is slightly adjacent to the fun we get from strategic decision making itself.

Finally, I am sure some of you likely picked up on some of the complexity that I am hiding in these different principles. Some of these aims may be at odds with each other at some level. Avoiding repetitive games might increase the chance of non-games occurring. By reducing the number of non-games, you might decrease the rewards of good decision-making. If progress in a game is a net-zero proposition, how can the rewards for good decision-making be justly distributed if both players are making equally good decisions? Even seemingly simple questions like where to set the bar for repetitiveness is incredibly hard. Reaching good solutions to these problems, especially through the indirect tool of card design, is extremely difficult.

With this approach to strategic fun in mind, let’s start talking about balance.

Balance

Balance is actually a problematic word. For a scale to be balanced you need equal weight on both sides. If you have identical weights on each side of your scale you will be able to achieve perfect balance. There are a lot of games that are balanced around this principle, such as chess, checkers, and Go. This concept of balance works for simple games like these, but as soon as each side has non-identical pieces all hell breaks loose. The analogy of the scale thoroughly breaks down when you consider that different strategies will be better or worse as particular aspects of the game, or that strategy A might have an advantage over strategy B, but a weakness to strategy C. In fact, it becomes totally hopeless to meaningfully assess the overall balance of a given format past a certain level of complexity. All the different pieces are so wound up with one another that it becomes impossible to tug at one thread without pulling the entire rat’s nest of relationships along with it.

I will admit that that this description of balance is slightly overcomplicated. It is clearly true that all the different cards in Eternal are tied up with one another, but in reality only a very small percentage of the card pool is actually relevant for balance. In other articles (such as this one and this one) I have touch on the idea that formats are balanced around certain subset of cards. These are some of the most common and efficient cards in the game, and they each offer unique barriers that players need to address in order for their deck to be successful. Surely if we distil down to the list of cards Eternal is balanced around we will see things are not quite as complicated? Well, I’m not so sure. Let’s make a list of the cards Eternal is balanced around (this is in no particular order).

Torch, Sabotage, Valkyrie Enforcer, Auric Runehammer, Tavrod, Sandstorm Titan, Dawnwalker, Heart of the Vault, Slay, Vanquish, Lightning Storm, Annihilate, Steward of the Past, Finest Hour, Rapid Shot, Gearcruncher, Champion of Fury, Oni Ronin, Vara, Vara’s Favor, Argenport Instigator, Champion of Glory, Unseen Commando, Icaria, Predatory Carnosaur, Permafrost, Black Sky Harbinger, Protect, Vadius, Worldbearer Behemoth, Mystic Ascendant, Xenan Obelisk, Bandit Queen, Channel the Tempest, Banish, Obliterate, Grenadin Drone, and Harsh Rule

This isn’t a complete list, but it hits most of the important cards. Any change to the balance of the game or a content release will get players thinking through this list trying to figure out which cards got better and which cards got worse. Each player has a mental model of which components of this complex balance puzzle are most important, and how they are all connected to one another, and chooses decks based on that model. Now imagine you are a game designer trying to rebalance the environment. Depending on which cards from this list you buff or nerf you can cause incredible ripple effects, disrupting the entire balance equilibrium of the system. How would it ever be possible to create a truly level and balanced playing field given all this complexity?

This may sound like I think it is impossible to truly balance things. You know what? Yeah, that is exactly what I am saying. I think true balance is impossible, and is not even worth attempting. For example, let’s just say you took the cards currently in Eternal. I don’t think there is any combination of stats and costs that will lead to a perfectly balanced format. I don’t even really know what a perfectly balanced format would mean. Is it possible to make the format more balanced than it currently is? Probably, but even then there is a question of how you pursue this.

Let’s take for example the Argenport Midrange archetype. Most people agree it is (at least) slightly over the line in terms of power, with many pointing the finger directly at Tavrod, but that is not a full solution to the problem. It is not like you can just shave off 10% of his power level in the abstract, you need to make specific changes to the text of the card. There are like 10 different ways to nerf Tavrod, each having a different potential impact on the format. Also, wouldn’t it be possible to nerf Argenport Instigator and Runehammer to have a similar effect on Argenport’s metagame share? What if Slay was moved to 4JS or 3JJJSSS? I am not saying any of these changes are good, but I am just pointing out that the philosophy with which you direct you balancing decisions is very important. It is specifically important to remember that balance is about both decks and cards. This is best shown through the history of bannings in the Modern format of Magic the Gathering. Deathrite Shaman, Gitaxian Probe, and Treasure Cruise are cards that are each too powerful for the format, while Splinter Twin, Eye of Ugin, and Golgari Grave Troll are each more about the associated deck rather than the card itself. Obviously if a specific card is too good you need to hit the card in question, but if a specific deck is more of an issue the approach can be more broad. Once you break into the digital space where any card can be tweak in any way at any time things just become more and more complicated.

Now that we have my somewhat nihilistic view of balance plainly stated, as well as the issue of how hard it is to balance in a digital space given the virtually limitless freedom, it is time to bring fun back into the mix. The purpose of balance is to make a game as fun as possible. Balance should always acts in service to fun, and all decisions around balance should be made to maximize the fun of the game. Game designers should not be focused on reaching some impossible Zen harmony where all the various elements of the game are perfectly in balance. Instead, as long as the game is fun, things can be slightly imbalanced. This may look more like a wild jumble of disparate forces, but as long as the whole system is balanced enough that it doesn’t topple over in one direction, that is OK, as long as the format is fun.

In some respects this view of game balance is quite freeing. Take for example the math problem that makes balance so hard: the designers of a game will only ever play a small fraction of the amount that the community does. Let’s say Eternal has roughly a thousand players on at all times. There is a lot more than this, especially when you consider mobile users, but let’s just say 1000 at all times for the purpose of this exercise. That means that the Eternal community is logging 24000 hours a day, or 168000 hours a week into this game. Let’s say that DWD has 10 people whose only job it is to playtest Eternal, and that they are all working 40 hours a week, and they are so good at this job that every hour they log is like 10 hours of a random player. That is still only up to 4000 hours a week, or about 2.4% of the hours logged by the community. Testing and optimizing decks just takes incredible amounts of time, so these guys will never be able to replicate the man-hours put in by the community. This is part of the reason why I chaff at comments like “DWD should just get the balance right the first time!” Is that realistic? Is it even possible? The game becomes very easy to balance if everything were to become simplified, but that would also be boring. As long as this math issue exists, I have no idea how designers can ever develop a format that is truly balanced. This is why I would rather designers focus on creating a game that is fun. As long as the balance is not totally egregious, let’s just see how things shake out!

This approach to balance may come off as a little reckless, but let me explain. First, a truly unbalanced game will not be fun. Let’s take a look at this card:

eternal_oops_all_spells.png

Pretty broken right? Imagine playing against a Temporal Control player who jams this on turn 5 while you are on some normal unit heavy deck (let’s say Praxis). The game is completely over, and not only is it over, but nothing that happened in the game up to this point really matters either. Unless you were able to get your opponent into burn range you have basically no route to victory. As such, the entire game was functionally a “non-game”, therefore violating rule number 3 from above. This is clearly an extreme example using a card that would never see print,  so instead let’s talk about actual changes that have happened over the history of Eternal. Flame Blast was one of the balance changes I was most passionate about. Stonescar Burn decks were defined by their ability to consistently burn people out from 15-18 health, leading to games where your opponent would start fireballing your face, and there was just nothing you could do about it. To me, this felt like it violated the first of the points spelled out above, since when you were getting burned out there was just no opportunity for counterplay, and it felt pretty lousy. This was also a violation of 4 since the games that ended in this manner felt pretty homogenous. The old Champion of Cunning fell into the category of being too rewarding for the investment. Inspire encouraged repetitive games states. Excavate lead to both extended non-games as well as repetitive game states. Going down the line I could probably explain how most cards were nerfed not just as a response to some imbalance in the metagame, but rather to emphasize the importance of creating fun games.

I don’t mean to pretend that this standard of focusing on fun is so much easier than focusing on balance. Where do you set the line for each of these factors? What is an acceptable level of “non-game” moments? What is an appropriate level of counterplay for each type of threat? What is an acceptable reward for each of the given scenarios? When do games cross over from fun to repetitive? These are really hard questions, but I think they are slightly easier to answer when the objective is maximizing fun rather than balance. True balance requires a near infinite amount of testing to perfect, while fun is much easier to find.

Argenport and Tavrod

With this philosophy in mind, I actually think it might be helpful to identify some of the potential balance changes that might be most in line with this way of thinking. I personally think the constructed format is fairly healthy, with the exception that Argenport is slightly over the line, specifically in ways that make the game less fun. Let’s talk about this.

Many people are pointing directly at Tavrod as the most likely candidate to be hit with the nerf bat. DWD might still be shy about changing cards that are part of a bonus campaign, but we can point to a few aspects of the card that seem to be violations of the principles of fun. First is the limited number of efficient counters. Most of the popular threats in the 5 power range are vulnerable to a range of the most common conditional removal spells (Annihilate, Permafrost, Obliterate, Vanquish, Predatory Carnasaur). Tavrod is essentially unique in that he is only vulnerable to Vanquish. In Partick Sullivan’s rant about Ravenous Chupacabra he refers to the tension around playing down a premium threat, and being faced with a real risk/reward problem. Part of what makes big midrangy threats interesting is the drama around who gets the reward. Do I get the reward when my opponent doesn’t have the answer, or does my opponent get the reward when they punish me with a removal spell? That drama is interesting, and connects with meaningful decisions both earlier in the game and later in the game. In the case of Tavrod I think there are just too few ways to punish someone for playing him. There is not quite enough meaningful counterplay, which leads me to believe one of the most likely changes would be to give him an additional vulnerability, either by reducing his health to 6 or removing endurance.

One could argue that the rewards you get from playing Tavrod might be too large. Drawing as many as 2 cards from the attack seems like a lot, especially since it is not hard to clear the way for an attack given the quality of removal Argenport has access to. Though this has some merit, I am not really convinced that this will move the game to become more fun. The issue with Tavrod is that opponents feel helpless, and that their game decisions don’t matter because they just can’t beat a 5/7 endurance. Changing Tavrod to a 6-drop might be a viable option as well, and although I like that better than tweaking his attack ability, I think I would prefer the changes to increase his vulnerability to removal spells like Obliterate or Permafrost.

Now it is possible to solve this problem other ways. It is possible to change other cards to create new vulnerabilities for Tavrod. Here is an example: Yeti Furflinger becomes a 2/5 for 3PP with an ultimate that costs 4 instead of 6. Tavrod is now very vulnerable to this revised card, potentially pushing him out of the format. Even more simply, Polymorph could just become a fast spell. I don’t really think either of these changes are particularly likely, but they are possible, and could offer a solution to the Tavrod problem.

I am putting a lot of emphasis on counterplay for a number of reasons. In my opinion it is extremely important for decks to have good counters, and one of the issues with Argenport midrange is that the potential counters are not quite good enough. I am not saying the deck has no bad match ups, but more that too few of the match ups quite bad enough. In fact, one of the cards that we can point to as a key cause for Argenport Midrange’s recent performance is Unseen Commando. Before this card came out, Argenport was much weaker to aggro than it is now. He fills the same role that Barthollo once did, where it offered a potential counter to decks it was potentially weak against.

The other card that I would point to that specifically creates un-fun game moments is Auric Runehammer. We have all been in the position where our hand is a bunch of 3 and 4 drops that all die to Runehammer. We are faced with the choice to either allow our opponent to develop their board for another turn hoping to draw a higher health unit, or we just need to play into the relic weapon, and hope they don’t have it. Runehammer is just so punishing to wide range of decks, and can often generate a massive advantage in the developing stages of the game. As a result, I think Auric Runehammer is at least slightly in violation of the second principle of fun strategy that I laid out above.


It feels quite satisfying to get all that out of my head. I have been mulling over these topics for a long time, and I am almost surprised that I was finally able to collect my thoughts enough that I could get them on paper. Hopefully you found this interesting! Be sure to check out the Reddit thread to share your thoughts!

Love,

Neon