It Failed

Artifact currently hovers at a max concurrent player count in the hundreds, an order of magnitude below Yu-Gi-Oh! Duel Links. Steam reviews are below 50% positive, with recent reviews "mostly negative". As I type this it has 139 Twitch viewers. (Note: These were the metrics when I began writing this blog months ago - they’re worse now)

Unlike games like Fallout 76 and Anthem, which endured rounds of public humiliation, Artifact has been a very quiet failure - but a failure nonetheless. After months of barely updating the game Valve announced that they would stop updating it entirely, instead focusing on a long-term rework. 

My goal is to explain why this happened. Why a game in a popular genre, that appears competently made by a well-respected developer, landed with a quiet thud. The simple version of my answer: it’s just not well-designed.

Judging Games on Apparent Craft

I suspect some will reject this out of hand - much of the analysis I’ve seen online is “the game may have problems but you have to admit that it’s brilliantly designed!” The initial consensus was that design of the game is a strength and that it failed for vague reasons like “poor market fit.” (A trivial observation true for any product that underperforms) But I'd suggest something different: that Artifact looks like a well-designed game but isn't one.

Artifact is clearly a professionally produced product. It feels like a game made by smarties - people who did well on their SATs and went to Carnegie Mellon. It has that good product look - the closest a card game has come to being a prestigious 3rd person sad dad simulator. I think there’s a strong temptation, especially among game developers, to look at the game from afar and declare that it’s a brilliant design because it looks like the kind of game for which that should be true. It fits the profile.

For this blog to have value you must accept that a game can have high apparent production value, a strong pedigree, and superficial indicators of quality while ultimately being no great shakes. And for the remainder of this piece I'll do my best to convince you that that's the case here.

A Monetization Aside

This is a blog about design but it would be almost irresponsible to talk about Artifact’s failure without mentioning the monetization. If you’ve heard any criticism of Artifact it’s probably this one, so I don’t think I need to elaborate much. But to summarize Artifact includes four different monetization models: up front payment, paying for packs, a transaction fee on card sales, and a cost of entry for the big-boy modes. I once heard King of Fighters 95 described as one big fireball/uppercut experiment; Artifact feels like one big monetization experiment, a Frankenstein’s Monster of monetization strategies.

Some have argued that the monetization is fine or even generous. That it’s not hard to “go infinite” and play forever without paying, or that the total cost of ownership is low. Those people are wrong. How they are wrong could be a blog post in itself, but even if they’re right (which, to be clear, they aren’t) the fact that the game has such an apparently punitive monetization model is a huge problem. But not the problem. Plenty of people jumped every monetization hurdle and still quickly stopped playing the game.

Issue 1: It’s Hard to Learn and Conceptually Confusing

A frequent complaint from Artifact players is that they don’t understand why they won or lost and don’t grasp how to improve. What Artifact appears to be about differs dramatically from what it’s actually about, making intuitive strategies incorrect.

The Deceptive Unimportance of Long Term Value

A common theme in European-style board games, like Dominion or Ticket to Ride, is managing potential vs kinetic energy. In these games players choose between actions that build up future potential and actions that expend resources to make immediate progress. The high-level strategy is recognizing when to be the tortoise vs the hare - when to hang back and build up steam and when to beat other players in a sprint to the finish.

At first glance Artifact appears pressed from this mold. Prior to release the developers heavily stressed the “modification” system; modifications are long-term investments that permanently buff cards. The item-buying system in Artifact also looks like a variation on this theme, allowing players to buy cheap low-impact items or save up for higher-ticket more impactful ones. With 3 gold you can immediately invest in a lowly Traveler's Cloak that gives a hero +3 health, or save up 7 gold for a Fur Lined Mantle that gives a hero +8 health instead. The latter represents more power per-card and per-item-slot but is slower - the tortoise approach.

There’s just one problem: Artifact games are too short for the tortoise approach to be viable. Even in draft, which is a relatively slow format, games often end in 5-7 turns.

When a hero dies they sit out a turn before returning to the field. If they die to an upkeep kill (die before they can act) they effectively sit out for 2 turns. Modifications and high-cost equipment persist when a hero returns after dying, but by the time a modified or equipped hero dies and then returns the game is often nearly over.

As a result many modification cards are bad, and the ones that are good would probably still be good even if they granted only a non-permanent buff. It’s telling that one of the best modification cards, Time of Triumph, costs 8 mana and is used as a “finisher” - it can win the game immediately and often represents no long-term value at all.

Midrange items in Artifact are almost entirely useless. Saving up to buy an item on turn 4 that provides good long term value makes little sense when the game can end in 3 more turns. The specifics of item buying also work against midrange items. Items are presented to the player in random order, and to reshuffle your item deck you either wait a turn or buy the topdeck item and draw again. So the dominant strategy is either to load up exclusively on cheap items, or use mostly cheap items plus a few high-cost game-winning items that you can quickly cycle to. 

Pre-release the developers stressed playing for value as a core concept, but the way games play out in practice value-oriented play is usually a trap.

Killing Heroes....Is Good?

Heroes are the centerpiece of Artifact, so you’d figure the game is largely about killing enemy heroes while keeping yours alive. But this too is wrong.

One of the quirks of MOBAs is that while the goal is to kill the enemy structures trying to do that as quickly as possible is often counterintuitively a bad strategy. Hard pushing a lane can leave you open to ganks while having minimal upside. While Artifact doesn’t import this exact design quirk it imports that same sort of quirk, where the obvious thing to do is often the wrong one.

In Artifact once a hero enters a lane their ability to move to another is very limited (especially in draft format) - consequently a key strategy is to strand an enemy hero in an irrelevant lane where they contribute nothing to the game. When an opponent's hero is stranded in such manner killing them can do your opponent a tremendous favor, allowing them to redeploy that hero on respawn into hotly contested lanes.

Imagine a Chess where killing the enemy Queen was often wrong - that would be weird. (Yes, I know gambits are a thing)

Because Artifact encourages periods of inaction followed by combo turns (more on that later) killing heroes at the "wrong" time often has little upside. If you kill an enemy Zeus on 6 mana they can't cast their powerful 7-mana spell on the next turn as they'll be dead. But if you kill an enemy Zeus on 5 mana they respawn in time to cast their 7-mana spell on curve, and can redeploy Zeus to a more advantageous lane. As such killing Zeus of turn 5 is often irrelevant or a waste of resources.

Artifact appears to be a clash-of-heroes game but killing enemy heroes is often just plain wrong.

It's Noisy

Artifact has many random factors and involves many micro-decisions, which makes evaluating the impact of those decisions difficult. 

When you lose on turn 10 it could be because you made the wrong choice on turn 9, but it could also be because you made the wrong choice on turn 1. In a game like Hearthstone, where in some matches there are only a few reasonable plays each turn, it's possible to track back to an early turn and understand how the game could have unfolded differently. But in Artifact this sort of analysis is difficult due to the number of possible decisions, which is further compounded by large amounts of randomization obscuring the outcome of those decisions. 

People often compare Artifact to Poker (especially when defending the RNG aspects) but there are two key differences: Poker is constrained enough to be mathematically well-understood, and Poker hands go by fast enough to build an intuitive understanding in a reasonable amount of time.

When you sit down to play Poker for the first time and luck out on an inside straight you might assume this is a winning strategy, but you’ll quickly be disabused of that notion after a few hours of play. It’s also simple enough to calculate or memorize the odds of common scenarios in games like Hold'Em. 

In Artifact you may consistently be doing the wrong thing on turn 1 and as a result losing on turn 10, but you can’t play enough games to spot a pattern and it’s hard to evaluate the impact of plays in a sea of other plays and random events, either intuitively or mathematically.

The end result is, again, people losing without understanding why or how to improve. The game has no real ranking, no ladder, no replays and no meaningful stats, so if you want to track the performance of various cards or strategies you best bet is to fire up Excel.

Issue 2: It’s Not Fun

A common theory of game development is that there are two main ways to make a good game: include great stuff and/or remove bad stuff. If you’re a weirdo who thinks of game design only in liquid analogies this is the juice and oil of “juice it or lose it” and “oil it or spoil it”- “juice adds pleasure, oil removes pain.” (My own catchphrase, “balsamic vinegar it or rejigger it” has yet to catch on)

There are plenty of good games that focus on just one of these two things: games like Killer 7 and No More Heroes include fun inspired ideas with low polish, while Blizzard games remove pain points while having few novel elements.

Unfortunately Artifact fails to execute on either of these. When I say it’s not fun I mean something specific: there’s very little wow factor or inspiration, and rather than avoiding pain points the game often seems centered around them.

Heroes Are Bland

Heroes should be the centerpiece of the game but are often simple stat bundles.They have mostly underwhelming (or no) abilities, and what should be signature abilities are instead moved to separate cards. (When you include a hero in your deck you also include 3 copies of their signature spell)

One of the joys of card games is opening a new pack of cards and considering how to incorporate them into a deck. In Artifact you’d think that heroes serve as the “build around” cards - the linchpins of novel strategies. But hero design in Artifact is very conservative. How do you build around a hero that does 1 point of random damage a turn? Or around a hero that has reasonable stats and no abilities at all?

It’s telling that so many Artifact heroes are generically good or bad, either showing up in a variety of top-level decks or none at all, rather than appearing in strategy-specific decks. If you’re making a red deck there are some heroes you almost certainly want to include, regardless of what your deck is supposed to do, and other heroes you almost certainly want to leave out. Axe, one of the best heroes, has no abilities at all. He has good stats so he’s good at killing guys, and his signature card is also good at killing guys. He’s good in the least interesting way possible.

There’s an argument to be made that separating out heroes and their signature cards, like I'm doing here, is unfair since they’re a package deal. But signature cards are often also underwhelming - in some cases being forced to include them in your deck feels like punishment. And casting a spell from your deck feels (and mechanically is) different from using a hero ability.

Another dull aspect of heroes is how similar they are. Green has a hero that gives adjacent allies +1 defense, and a hero that gives adjacent allies +2 defense, and a hero that gives adjacent allies +2 regeneration, and a hero that gives adjacent allies +2 attack. That’s 4 heroes that give adjacent allies a basic stat buff, all in the same color, and one of those buffs is a strictly better version of another. 

The hero Debbie does extra damage to heroes and structures. Phantom Assassin does extra damage to heroes. Bounty Hunter does extra damage to units and structures but only half of the time. These are cards in the same color and all variations on a theme. Crystal Maiden and Outworld Devourer both restore mana; Kanna, Prelix and Venomancer all summon extra units - all blue heroes.

Card games often use cycles of similar cards across colors to emphasize color differences. In Magic Healing Salve is a white spell that heals for 3 and Ancestral Recall is a blue spell that draws 3 cards, hinting to players at the different philosophies of these two colors. In Artifact similar cards exist in the same color - rather than underscoring the differences between colors these cards accentuate the bland uniformity of a color. Green has four heroes that buff stats - I suppose this tells players that green is the stat-buffing color with the subtlety of a hammer, but the overlap is unexciting and the best stat-buffing spell is red anyway. 

The philosophy here is like from a fighting game full of palette swaps. But palette swaps serve a utilitarian purpose: they save development work. In a card game if you have a hero that gives allies +2 defense why would you create a another one, in the same color, that gives allies +1 defense? You still have to draw new card art. The work saved is the work of coming up with a unique idea.

Perhaps the logic is that exploring minor variations on a theme is a way to manage complexity and understandability. But this seems like a severe miscalculation: of all the ways to simplify Artifact, which can be complex and impenetrable, simplifying it by making heroes dull is one of the worst ways possible.

Even hero art is dull. Each piece of art is fine by itself but it’s as if the artists were directed to work from a few standard reference poses and camera angles. Many pieces of art are similar and demonstrate nothing of that hero's abilities or personality. In Dota 2 Axe is a bad ass and Ogre Magi is comic relief - in Artifact art they look like the same stoic imposing figure.

Am I racist or do these all look the same?

It may seem petty to complain about art in a design review but the point is that Artifact displays a level of conservatism that extends even beyond game-design.

Way Too Much RNG

Implosion was a notoriously hated Hearthstone card that did 2-4 damage and summoned that number of 1/1 Imps. The problem with Implosion is that, if you count a 1/1 creature as roughly equivalent to a point of damage, it could do between 4 and 8 damage, which is just too high a range. A highroll on Implosion could win a player the game almost single-handedly.

Artifact is full of Implosion-style RNG.

Bounty Hunter sometimes does 4 extra damage for a turn. Sometimes he doesn’t. There’s no tradeoff - he just flips a coin, heads he does extra damage and tails he doesn’t. In a MOBA a character may attack hundreds of times over the course of a game so crit chance is subject to the law of large numbers. In Artifact Bounty Hunter might attack only 8 times in one game, which isn’t enough for the law of large numbers to apply. If you roll heads on the first 3 turns of the game you're way ahead, and if you roll tails Bounty Hunter has just been an inert pile of stats.

Ogre Magi has a 25% chance of making a copy of any spell cast. 75% of the time his ability does nothing, which feels bad if the Ogre Magi is yours, and 25% of the time he duplicates a spell, which feels terrible when the Ogre Magi isn't yours. Over the course of the entire game you might cast 6 spells that Ogre Magic has a chance of duplicating so again there’s no law of large numbers in play.

Sometimes you line up against the enemy hero at the start (purely based on luck), roll heads on Bounty Hunter for extra damage, cast Track and Payday (spells that generate gold), use that gold to buy a Helm of the Dominator (an instant-win item in draft, more or less) that happens to be at the top of your item deck and win the game in the first couple turns.

That was how my last game of Artifact went before I uninstalled it.

This is the part where someone says “I bet the best player in the world could win a game where their opponent bought a Helm on turn 2.” Maybe - but matchmaking is supposed to match you against evenly-skilled players, not match the best player in the world against beginners. Another defense I’ve seen is that yes these sorts of bad beat stories do happen but the majority of the time they don’t. Nobody would accept a fighting game where a low short could instantly KO someone from full health 1% of the time. People hated tripping in Smash Brothers. It doesn't matter that these sorts of effects are rare, or than both players are subject to them equally and thus "balanced" - they're more frustrating than fun. Losing to them doesn't mean you lack skill and winning via them doesn't take any.

Artifact is stuffed with all sorts of randomness. What position heroes deploy to, what position and lane creeps appear in, which direction they attack in, which cards you draw from your deck, which cards you draw from 3 separate item decks, two of which you don't author. The spell Eclipse targets randomly, Chain Frost bounces randomly, Roar moves units to random other lanes, sometimes an ability like Ravage stuns a bunch of units and other times it doesn’t. And these are, almost without exception, “bad” RNG like Implosion: instead of getting two different but roughly equivalent outcomes you either get lucky or you don't.

The strangest thing about the RNG is that Artifact is pitched as a serious card game for high-IQ players - it’s complex and mentally taxing. But it includes a dozen different Mario Party mechanics.

The Spreadsheet is Too Apparent

Here's a scandalous confession: I don’t love European board games. Too many of them are transparent optimization problems with a weak theme bolted on - an Excel spreadsheet with graphics. Artifact suffers from this in that combat, which in many card games is a significant phase, is a single instantaneous resolution of a math problem. Your numbers crash into their numbers in way that look suspiciously like the subtraction of two linear equations, and many cards and abilities merely adjust the coefficients of those equations.

In Hearthstone it’s “my poisonous snake bites your taunt guy killing him, then my 3 wolves attack your bear.” In Artifact it’s “my integer tuple subtracts from your integer tuple, but first I play a card that increases the second element of that tuple by 2."

Not giving players direct control over combat can work. It's a feature of the Autochess / Underlords / Teamfight Tactics games. But in those games the combat is visually engaging with some degree of unpredictability. Observant players can watch how a battle plays out then adjust their composition or placement accordingly. In Artifact the combat isn't visually engaging and before it resolves the game informs you via UI what the result will be. There's no point in even watching the screen during combat, it's just a formality. 

Elaine: Yeah. I met this lawyer, we went out to dinner, I had the lobster bisque, we went back to my place, yada yada yada, I never heard from him again.
Jerry: But you yada yada'd over the best part.
Elaine: No, I mentioned the bisque.

Combat in Artifact is the yada yada yada in this exchange. It should be a main attraction but instead it's an afterthought.

It Encourages Inaction

Far too often the best (or only) strategy in Artifact is to prevent your opponent from acting or to do nothing yourself.

Preserving initiative in Artifact is extremely important. (Though the concept is almost invisible at first glance - another thing that makes the game unintuitive) If you have initiative and pass you get to act first in the next lane and passing again will carry that initiative to the next lane, ad infinitum. The best play is often to pass, sometimes in two or three lanes in a row, to preserve initiative in a critical lane. 

The flip side of doing nothing being a good strategy is that, when it does come time to do something, that something often begins by disabling your opponent. There are many cards and abilities in Artifact that prevent your opponent from acting. Artifact is most enjoyable when you and your opponent have a very involved back and forth - you play out a minion, they play out a minion to block it, you play a low-impact item hoping to get them to commit to some mana, they also play a low-impact item, you use a spell to kill that blocking minion, they replace it with another cheap minion, etc etc. But while that’s the most fun type of turn the best strategy is to prevent that sort of turn from happening. Instead of outwitting your opponent in a complex back-and-forth the better strategy is to go first, entirely disable them, then do whatever you want without opposition. It’s like the design goal of Artifact is to make every deck play like a Magic: The Gathering blue control deck. (The most complained-about type of deck)

In most card games you draw one card per turn and start with a sizable hand. In Artifact you begin with a small hand size and draw 2 cards per turn despite sharing that hand across 3 lanes. (Essentially 3 different game boards) So not only is doing nothing often the best strategy, it can be a mathematical inevitability. If you use a card in every lane you quickly run out.

"What if we could prevent players from acting while also indulging in more frustrating RNG?" you ask. No worries, Artifact has you covered. It includes a mechanic that allows you to "lock" random cards for a duration, during which they can't be used at all. It's like a random discard effect from Magic, but Magic has all sorts of mechanics that leverage discard and the graveyard for interesting effects while Artifact has no equivalent.

Sometimes you lock 3 key cards from your opponent and win the game, sometimes you lock 3 irrelevant cards and it has no effect. The locked cards are never revealed so it's not only frustrating for the player being locked out, it also denies the caster the satisfaction of knowing the impact of their play. 

The Initial (and only) Set of Cards is Uninspired

It’s not uncommon for the first set of a card game to have major problems. Often the developers don’t quite understand their own game yet and the initial set has bad balance and wonky mechanics. Artifact has those issues in spades. Symmetrical effects tend to be terrible or great with little in-between. Balancing and costing is way off. I believe the first Artifact card set is particularly bad, even as far as first sets go, but bad first sets are pretty common in card games and I wrote about this previously so I’m not going to dwell on it.

What I will dwell on is that not only is the first set bad in the ways first sets typically are, it's also just plain dull. Most first sets in card games need “bread and butter” cards - simple cards that serve a basic function or introduce a new concept. But Artifact has very few interesting build-around cards to supplement the more basic ones. The Artifact card set reads like a Magic starter set (already the beginner set) with the most interesting 10% of cards removed. It’s telling that some of the best creatures in Artifact are generic big green guys with good stats - in Magic that class of creature is a noob trap.

There’s maybe one item in the game with a splashy fun power. Everything else is straightforward stat adjustment. Most of the top end creatures either have good stats or boost the stats of others rather than having flashy powers - their strength is in the ability to inflate the coefficients of the system of linear equations that governs combat.  

Lack of Direct Control is Frustrating with an Inappropriate Metaphor

What I'm calling the "metaphor" is the contextualization of mechanics. In autobattler games you're implicitly a commander making high-level decisions. Many sports games feature coaching modes where you call plays without executing them, or modes where you play as a specific player running plays called by an AI coach. The lack of total control in these modes makes sense because it's contextualized properly: a real-life coach doesn't also catch the ball and a wide receiver doesn't call plays.

In Artifact what you can and can't control feels arbitrary, with no logical explanation either offered or theoretically existing. Why do my creeps, soldiers I employ as part of my army, pick their lanes randomly? If I'm the commander shouldn't I be able to tell them where to go? I can make a hero drink a potion, use an ability or equip an item -  micro-level tactics that should be up to the individual - so it seems as the player I embody the heroes rather than a coach. So why can't I decide which direction to attack in, which should also be a micro-level decision? I choose which lane they deploy to, a high level strategy decision, then they decide which slot to a deploy to and which slot to attack towards, mid-level tactical decisions, and then I can tell them to drink a potion, a low-level tactical decision.

This may seem like a minor, heady objection, but often times systems that are impenetrable or frustrating are so because they lack the proper framing. Players are a lot more willing to accept limitations, or consider limitations features, when they follow real or video game logic. 

Issue 3: Weak Core Pillars

Now that Valve has announced that they're overhauling Artifact rather than doing small updates I keep coming back to two questions: what are the core design principles and are they good?

My answers: “I’m not sure” and “not really.” 

What does a fixed Artifact look like? What's worth keeping? These are tough questions because the core pillars of Artifact are weak - a "fixed" Artifact might just be an entirely different game.

The Three Lane Structure is Unique - and Awkward

Probably most unique element of Artifact is that, taking a page from the structure of DOTA,  it plays out across three “lanes” that share the same hand of cards but are otherwise impendent. Unfortunately it's not one of the best elements. If anything it's a major detriment.

At any given time in Artifact you're either looking at one of three lanes or you’re looking at a zoomed-out view that makes it hard to see anything. As a player this means that much information is hidden - not “hidden information” in a game-theoretic sense but hidden offscreen. Playing Artifact involves a lot of paging around to remind yourself of the total state of the board. Maybe you have an item that you're thinking of playing in lane 1, but you need to double check that it wouldn't better be suited for lane 2 or 3. It's like playing a board game by peering through a small porthole that only allows you to see a third of the board at a time. It’s not uncommon to see even the best players make mistakes because they simply forgot about the state of other lanes, and the constant checking and rechecking of game state draws out game length.

As a viewer it’s even worse. While a player can switch between lanes whenever they need to refresh their memory as a viewer you’re at the mercy of the player controlling the action. If you forget the exact state of lane three you’re in the dark until the player decides to pan back over to it. One of the goals outlined by Gabe Newell was to leverage Valve's knowledge of DOTA esports, but Artifact is poorly suited for esports as the viewing experience is awful. (The million-dollar tournament date came and went without any tournament occurring)

When a computer formally enforces the rules of Magic: The Gathering the game is slow-paced and fussy, something Hearthstone took pains to avoid. A great strength of Hearthstone is that it feels designed for digital. Artifact feels like a game that was designed to be played on a large table and was transferred to digital with an unwillingness to compromise.

The argument in favor of the three-board structure is that it mimics the three-lane structure of DOTA. But Artifact is a card game with a DOTA skin, not a DOTA card game. In DOTA all lanes lead to a central Ancient that must be destroyed to win the game. In Artifact the lanes never converge and the win condition is different. So the three-lane structure has a number of downsides and the potential upside, that it ties into DOTA, is weakly realized.

Unlimited Hand and Board Size Sounds Cool - but Isn't

One of the features the dev team emphasized is that both your hand size and number of units on the board are unlimited. Like the three lane structure this core feature sounds neat on paper but has significant downsides. The UI doesn’t do a great job of handling large boards or hands - in particular large boards force players to scroll to see units that are off the screen. The three lane structure already has the downside that you can only see one-third of the action at any time, and if you have a large board you can’t see even a single full lane at once.

Even ignoring the awkward UI it's not clear that unlimited board size is a net positive. The wider a board is the less predictable unit placement is, which moves the game away from calculated strategy and more towards pure chance. It’s also unfortunate that the color that’s best at producing wide boards, blue, is the color best at clearing them. The other colors often have no meaningful way to interact with a large board.

In this game state there are 6 enemy units off screen, and this is only viewing one of the three lanes.

Unlimited hand size allows for a strategy where you draw a huge number of cards and save them up for one mega-turn. In most card games discard mechanics and being punished for tempo loss keep players from hoarding cards, and in both Magic and Hearthstone there are penalties for drawing too much of your deck. In Artifact it’s very rare for players to draw their entire deck (I don't even know what happens - I've never seen it), there are no discard mechanics and doing nothing is often a correct strategy, all of which make card hoarding too viable. A limit on hand size would force players to take action more often and make tough decisions about which cards to keep and which cards to play or discard.

Overly Cute Creativity

One last core philosophy I’ll mention is the emphasis on “clever” card interactions. In a game like Magic: The Gathering a clever card interaction, one the player can feel smart for "discovering", is something like playing a spell that makes you discard your own creature, then using another spell to revive that expensive creature for low cost. 

In Artifact many of the clever card interactions are based on math and wording technicalities rather than on the emergent interactions of systems. Take Rend Armor, which reads “modify a unit with -X Armor where X is its armor.” Intuitively Rend Armor destroys armor; that’s what "rend" means. But here's the catch - if you negate an already negative number you get a positive number, and thus you can use Rend Armor to gain armor.

As another there's Shop Deed, which reads “Each item in your Secret Shop costs X less gold, where X is equal to its base cost.” This seems like a complicated and formal way of saying “items in your secret shop are free.” But what if you have TWO Shop Deeds? Then the cost of an item is (x - 2x), or -x, which means “buying” an item from the Secret Shop gives you money.

While it makes mathematical sense based on the exact wording of these cards it doesn’t make intuitive sense - owning one shop deed might make the stuff in your shop free (assuming the items have no wholesale purchase cost...) but why would owning two deeds to the same shop give you money? What does it even mean to own two deeds to the same shop? If you own a Subway franchise can you somehow buy the same Subway again and get rich by scarfing down meatball subs? 

This sort of math-based "creativity" is a minor aspect of Artifact but I call it out because the developers have done so - it's a feature they're clearly proud of. I don't understand why. This sort of "clever" design reminds me of those SAT questions like "1/X is less than X - always, never or sometimes true?" Once you learn to carefully parse the text of each card to spot unintuitive math weirdness the design philosophy offers little else. 

Least Bad Game Design

Much of the design of Artifact feels motivated by “what can we do to keep the game from falling apart?”

Why is hero placement in lanes random? Probably because choosing exactly where they go would break the game by allowing you to easily snipe out weak heroes by lining up across from them.

How come on the first turn 3 creeps are randomly deployed to the lanes, but then on subsequent turns it's 2 creeps? I assume because this is what made the game work, and other variations broke it.

Why are "arrows" (that determine whether units attack straight ahead or diagonally) random? If units always attacked straight ahead games would be too fast-paced and non-interactive; if they always curved to hit enemy units the game would be too grindy. Allowing players to choose would slow the pace of the game down to a crawl while adding fussy UI. So to make the game work arrows are assigned randomly with probability derived from data. But while data can tell you what works best there's no guarantee that the optimal percentage makes for a fun mechanic - and in fact it does not.

These sorts of decisions are local problem solving - local maxima solutions that are better than immediate alternatives but don't make for a fun game. They're limited by core design principles the team presumably wasn't able to get away from.

Issue 4: A Misguided Approach to Feedback

Three Donkeys, the team that developed the Artifact design and worked with Valve to turn it into a game, has done a number of interviews and podcast appearances, both pre and post-release. I try to avoid too much speculation about process and motives, but there's a lot I can say about the development of Artifact, and in particular the attitudes of the developers, that isn't speculative.

One thing I want be very clear about up front is that many people peg Three Donkeys as "the problem" with the game and portray Valve as hapless victims. Artifact is a Valve game and Three Donkeys were contractors; any substantial decisions either originated at or were approved by Valve. The idea that independent contractors should bear the brunt of the blame for failure is nonsense. This section is going to be critical of Three Donkeys - they've done a number of problem interviews while Valve has remained tight-lipped - but that criticism should be understood in the context that Artifact is ultimately a Valve game, not a Three Donkeys one.

The Customer is Always Wrong

I suspect most developers are familiar with a designer who, when confronted with confused and dissatisfied players or testers, blames them for not getting it rather than acknowledging that their design has shortcomings. 

As someone with contrarian leanings I don't think that bending over backwards to satisfy play testers is correct. Tomonobu Itagaki once said about Ninja Gaiden "At first it was easier, but when the testers said 'this is too difficult', I made it even more difficult" and I have to admit I smile at this. But if a majority of players consider your game confusing or off-putting you have to concede that something may be wrong.

Players are often terrible at stating complaints. As a developer you can choose to pick holes in poor phrasing and reject the complaints, or you can assume the role of detective and get to the bottom of them.

I used to read the Star Wars Galaxies message boards, not because I was interested in the game but because it was fun to watch the developers interact with fans. They seemed intent not on understanding player complaints but on proving them incorrect. A player would complain that a weapon was weak, and the response from the devs was "according to our spreadsheet this weapon has above-average DPS - you're wrong." But the more productive approach is to try to determine the root cause of the complaint. Perhaps the weapon really is weak due to a bug. Maybe the spreadsheet has the wrong formula. Maybe the weapon's effective DPS is very different from theoretical - it has a low rate of fire and high overkill, or works well only at specific ranges. Maybe it has very little knockback or hitstun, so while it has high DPS that may not be the best metric, and a metric like "how much damage it does before the enemy is in melee range" is more relevant. Maybe players don't understand how to use it correctly, maybe it's weak against a common enemy type or for a certain group composition. Maybe it just has unsatisfying sound effects. All of those are possible explanations that the typical player is not sophisticated or invested enough to articulate.

Unfortunately interviews with Three Donkeys reveals a dismissive rather than investigative attitude. For every complaint they have a practiced explanation for its rejection out of hand.

The monetization is bad? But owning every card in Artifact is way cheaper than owning every Hearthstone or Magic card - case dismissed! Of course Magic has been around for decades, has orders of magnitude more cards, and has physical cards with real scarcity, so this is a meaningless comparison. When you play Magic Arena, the newest online version of Magic, you get more free cards than you know what do to with. It inundates you with new decks before you have time to understand the last one you got. And Arena is free. Artifact has an up-front cost, shipped with no quest or reward system, and each session of the relatively unsophisticated big-boy mode (prize play) costs money to enter. Yes, Artifact is cheaper - when considering a narrow and contrived play pattern.

The game has too much RNG? But the ELO spread between players is large, which means that the game is heavily skill-based - case dismissed! But the game has no visible ELO or real ranking system, so this assertion is unverifiable. The best players win much more often than the worst players - this one is verifiable but also fairly irrelevant. The game doesn't appear to have skill-based matchmaking (or if it does it doesn't work due to low player counts), which means beginners are matched up against experts. When the game launched players who had been playingin beta for a year were matched against players like me playing their first game - they better have a high win rate! 

While data can prove that a game is balanced (by some metric) or not overly RNG-reliant no amount of  gesturing to a spreadsheet can prove that a game is fun. That the game was fairly unsuccessful and got poor reviews is data that says the game is not fun. Frustratingly in interviews the devs attribute the poor reviews to "review bombing." But review bombing happens when people are angry over the inclusion of "SJW politics" or at a game going Epic Store exclusive; a game getting poor reviews isn't "review bombing." Deuce Bigalow: European Gigolo wasn't review bombed by critics, it's just a bad movie.

The difference between a review bomb and bad reviews

According to the devs virtually every complaint about Artifact is unjustified and a result of unappreciative, unsophisticated players. "You need a 200 IQ to appreciate Artifact" was already a meme on the subreddit, so to see the developers repeat that argument in interviews is dispiriting. I don't know if I have a 200 IQ but I have a graduate degree in CS from an Ivy, was a National Merit Semifinalist, blah blah...I have many flaws but being a dumb-dumb isn't one of them. If the reason I don't enjoy Artifact is that I'm too basic then I have no idea who the target audience is or how it's larger than a dozen people. Which is more plausible: that the game is too sophisticated for anyone other than Isaac Newton, Albert Einstein, Stephen Hawking and Commander Data, or that maybe it's just not that fun due to a variety of design missteps?

The devs have no explanation for why the game failed beyond that players don't appreciate its near-perfection. Just to be clear "near-perfection" is not me being snide - the developers claim it's the best card game they've created. So why did it fail? Even when they identify a potential problem like monetization they clarify that while players see the monetization as a problem that those players are merely confused.

Most explanations I see from sophisticated people - other developers, games media, etc - are a variation on "bad market fit." But again, saying that a game failed due to "bad market fit" is just a way of re-stating that the game failed. It's not analysis and has no explanatory power. Many of the people making these claims haven't played the game much or at all. Clearly there's a market for computer card games, so if the market is rejecting this particular game it's because of specific design and implementation issues.

These sections reads harsh but here's the rub: as a developer understanding criticism can only make your games better. You don't "win" by finding ways to dismiss complaints. Someone with a complaint is not an enemy or a challenger in a contest of wills. If people aren't enjoying your game you can't prove that they are or should be - that's not a thing. 

What is the Future of Artifact?

It's instructive to look at a game that successfully turned around: Final Fantasy XIV and the "2.0" version, A Realm Reborn. (Full disclosure: I worked at Square-Enix during that turnaround, but had nothing to do with XIV beyond providing some informal verbal feedback on a handful of occasions)

One of the first steps to turning around Final Fantasy XIV was installing Naoki Yoshida as the new producer. If a certain prop has bad UV seams you can blame it on a modeler; if a certain region has repetitive features you can blame it on a level designer. When a game is full of all sorts of technical and design issues spanning every area either every person working on it was incompetent or you have a top-down problem. 

The development team suffered from collectively poor judgment, and a new producer brought with him a change in priorities.

When Yoshida took over he produced blog posts like this one

I sometimes do consulting work and when I do typically the first thing I do is play the game and write up impressions. They can range from extremely high-level things like "the two main gameplay modes feel disconnected" to nit-picky things like "this graphic effect has some artifacting on the edges." The idea is you identify all the issues, you fix them, and then you have a good game. 

The blog post above is a perfect example of that approach. Some of the issues, like "tutorials for new players", are fundamental. Others, like cooler "fanfare" when you level up, are very low-level. 

I'd contrast this to modern "road maps", which exist mostly to placate players with vague promises of future content and the illusion of responsiveness. Many of the items in the Yoshi-P blogs are simply acknowledgement - they're going to examine a system and figure out if and how they can improve it. In other cases the dev team has specific fixes in mind and lays them out. As a player its a refreshingly honest and ego-less approach. It inspires confidence - when you read his blogs you can't help but think "this guy gets it." 

Square made it clear that relaunching FFXIV was a serious effort they were pot-committed to. It was an all-hands-on-deck situation, and if anything the commitment to A Realm Reborn seemed more serious than the commitment to the original. 

My worry with Artifact 2.0 is that it seems to be taking the opposite approach. Much of the team seems to have transferred at least partially to other projects (notably Underlords) - it's not clear if this is a personnel shakeup or because Artifact is lower priority. There's no outward indication that the team leadership or philosophy has changed. 

Valve leans heavily on data collection, player testing and other "objective" processes but those apparently came up short. Have they been revamped? 

Valve has invited friendly streamers to HQ to look at Artifact changes. But these streamers were notorious for hyping up Artifact pre-release, and frequently claim that it's a great game with only minor issues. These are the same people who beta tested the game for months and gave positive feedback completely out of step with how the game was eventually received.

Most worrying it's not clear if Valve understands what the problems with Artifact are. The only communication Valve has put out is a short blog post acknowledging that issues exist. Do they view the three-lane structure as a non-negotiable "USP" or as an albatross? Is the combat system considered integral to the game or something that requires rethinking?

I get the "speak softly and carry a big stick" approach - instead of putting out a demonstrative roadmap go quiet, fix the game, and release an update that blows everyone away. But one of the advantages of the FFXIV approach is that it serves as a sanity check. When the team blogged about major changes there was a chance for players to say "this sounds way off." Working in a hermetically sealed chamber then presenting the resulting game to the world fully-formed seems dangerous, especially when that same process just came up short. (There was an Artifact closed beta but it wasn't terribly useful - that's a whole 'nother blog post)

So the answer to "what is the future of Artifact?" is "who knows." Maybe an Artifact 2.0 will blow skeptics away, or maybe Artifact will be quietly dropped. Speaking as a player there are two things I'd love to see from the dev team: more specifics on what the dev team considers the current strengths and weaknesses, and a willingness to question even basic assumptions about how the game should play.

The Conclusion

If you aren't familiar with the rules of Artifact some of this may have been hard to follow, but I hope you get the spirit of the criticism even if you don't grasp the details.

In my blogs I try to examine specific issues in the context of a larger point. The larger point here is simple: to critique games you have to be willing to engage with them non-superficially.

To demand that every game review or analysis approach this level of detail is silly; criticism value isn't proportional to length. (Sorry FilmCritHulk) But there's little value in observing the apparent craft of a game from afar and declaring that it must be good.

If you want to know how the pudding tastes you have to eat it. Artifact is like a desert crafted by a Michelin Star chef, expertly prepared and presented. There's every reason to assume it's great. But ultimately you have to taste it, and in the tasting it comes up short.