Predicting the Cost of Artifact - Neon - October 12, 2018


We have had a lot of hot takes on how expensive Artifact will be. You can find comments ranging from “Artifact will be so expensive it will kill the game!” to “Artifact is going to be one of the cheapest digital card games ever!” and basically everything in between. While I am generally pretty tolerant to people contributing their thoughts to an ongoing discourse, this conversation has been particularly maddening. Everyone just seemed to pull random numbers out of their ass, with no actual thought on the subject, which is frustrating for me, as I have spent a lot of time thinking about digital card game economics. Speculation is fun, but in this case you can do real analysis using the publicly available information to estimate the cost of competitive Artifact decks! I suppose the main reason people don’t come up with mathematically reasoned predictions is because they are smart enough to avoid falling down a well of endless math.


In this article we are going to do all the fucking math in an attempt to predict the cost of competitive Artifact decks. Now, I know there are some of you who are “mathphobic” – never fear! I will protect you! Or at least I will protect you from the worst of it. There are a few sections that I have deliberately written as optional, and can be skipped by anyone who gets overwhelmed by too many numbers. For those who are here to get into the nitty gritty details, I lay out as much as I can to make my calculations obvious. If you do notice an error in my mathematics please let me know so I can fix it (Twitter and Discord [Neon3989] are best)! I should also note that this article is not dealing with a critique of the structure of the economy. I have discussed that elsewhere (part 1, part 2), and will likely return to it in the future, but for now we are completely focused on deck costs. If you are not familiar with my previous articles/videos on card economics, don’t worry about it.

It is worth mentioning that I am going to be using Magic the Gathering as an analog throughout this article, as the Magic economy is very similar at its core to the Artifact economy. There are some important differences, which I address later on, but it is still worthwhile using Magic as a guide to check our work. There are two thing that you really need to know about Magic to follow this conversation. First, they use a 4-rarity system, where “mythic” cards sometimes show up in the “rare” slot in packs. Second, the universal constructed format is called “Standard”, and it only includes cards from recent sets. Outside of that, everything will be pretty self-explanatory.

With that out of the way, let’s start by reviewing our knowns, as well as a couple simple assumptions.

The knowns:

  • There are 4 rarities, the most exclusive being “rare”. “Basic” cards are not included in packs, and are automatically part of your collection.

  • Packs contain 12 cards.

  • Each pack is guaranteed 1 non-hero non-item rare.

  • Each pack is guaranteed 1 hero, which can be any rarity (other than basic).

  • Each pack is guaranteed 2 items, which can be any rarity (other than basic).

  • The rest of the pack (9 cards) are commons/uncommons.

  • The maximum number of copies of non-hero cards you can play in your deck is 3. The maximum for hero cards is 1.

  • Packs cost $2 each.

  • Decks require 5 heroes, at least 40 main deck cards, and at least 9 item cards. Each hero puts 3 signature cards into your deck automatically, and does not require collecting the signature cards.

  • Draft uses packs that are the same as typical packs.

  • ASSUMPTION: the rate of rares showing up in the common/uncommon slots is not high enough to impact the overall cost of decks. There has been some speculation that you can get multiple non-rare, non-hero cards in a pack which is based on some comments from RobAJG from a Twitch stream. The rate of rares showing up in non-hero non-item slots is unclear, so I will assume it is low. Whatever the chances are of getting a rare in the common/uncommon slots (say 5%) you would expect deck prices to fall by about that amount (decks are 5% cheaper).

  • ASSUMPTION: we are going to ignore the market’s transaction costs. There is good reason to believe that the transaction cost will be 15%, like we see in the conventional Steam market, but we do not have full confirmation that the Artifact market follows the same rules. This will certainly add to the cost of building decks.

  • ASSUMPTION: there are no “premium”, “golden” or “foil” cards that are part of the game, or they are not important enough to shift the overall economics.

Cool. From this, we can do a lot of work. To start with, let’s figure out how much it would cost to get a single copy of every non-hero, non-item card in the game. This might seem like a weird place to start, but trust me. As of this writing (according to Artifact Fire) there are 35 rare cards confirmed that fit these criteria, so let’s assume there are a total of 50 in the full set. If this is true, the cost of getting each non-hero non-item card would be about $100, as you can just take $2 times the number of rares (50) to get the expected cost of all the rares. (This implies that a playset of everything will cost about $300, for those who care)

 Wait, what? Really? Why can you just multiply $2 by the number of rares to predict the cost of getting one of each? This is a result of “law of large numbers”, and just basic supply-demand relationships. While it is true that if I spend $100 on packs I am very unlikely to have exactly one of every rare, but if 10,000 people spend $100 each on cards we would expect that the total number of each specific rare would be more-or-less the same (assuming no funny-business on Valve’s side). We can therefore say that the “cost of production” for one-of-each-rare is set at $100, meaning that the demand is forced to meet this price.

For the purpose of this article I should define “cost of production” more formally, as it comes up a lot in the course of this article. The “cost of production” is the amount of money you need to spend to purchase a single copy of all the cards. It would be possible to define it around playsets of everything, but that has some rigidity to it that I don’t like. If you can find the “cost of production” that gets you most of the way to predicting deck prices. Unfortunately, we don’t have enough information to definitively figure out the “cost of production”, so we are going to need to make some assumptions, and estimate based on that. In this initial conversation, I am going to just tie this to 50 rare non-hero non-item cards, but we will return to that later.

I should also note that commons and uncommons will contribute almost nothing to the price of decks. Those who are hunting for competitive rares will totally flood the market with commons and uncommons to the point that they hold almost no monetary value. Maybe there are one or two uncommons that crack $1 each, but this will be extremely… uncommon. Trends in Magic bear this out, where there maybe one or two uncommons worth a dollar or more in every Standard legal set.

Before we go any further, I do need to offer a caveat to this simplified model. If you look at the price of Magic cards, they do not quite follow the pattern where the cost of a set equals the cost of opening all the cards from sealed product. In fact, when you open booster boxes you lose between 15-50% of the value assuming “average” opens. This may seem illogical, but given Magic culture this isn’t a huge shock. Sealed product can be used for drafting, so there is a premium on unopened cards versus opened cards. There is also the “lottery ticket” effect, where a sealed pack has the possibility of holding the most expensive card in the set. Once the pack is opened those two functions are gone. It is difficult to tell if this factor will hold for Artifact, so I am going to ignore it for the sake of this conversation. I just wanted to note that this could mean that all of my estimates are off because of the “sealed bonus” effect.

Nerdy tangent on how I did this math for those who care

There are a few ways to calculate this, and all of them have some problems. After working through a few methods, I decided to use a relatively simple method, using MTG Dawnglare to find the expected value from each Standard-legal booster box, and compare that with the booster box prices listed on TCG Player. For those who don’t know what all that means, I am essentially comparing how much it costs to buy a box with 36 packs, versus the expected value I would get from opening those 36 packs. In both cases I used the “TCG Player Low” price, which could exaggerate this effect, as some stores might be selling chase rares at low prices as loss-leaders, or just charge horrible shipping fees. Dawnglare’s program also ignores cards that cost less than $1, which can add a surprising amount of value to the set. Using this method, it appears Guilds of Ravnica loses 17% of the value upon opening, Rivals of Ixalan loses 45%, and everything else falls in between. I should also note that I am not using Magic the Gathering Online prices, as there is some weird nonsense going on there, which I talk about in a little more detail after the next “nerd blurb”. 

Nerd tangent over

So, we have the cost of the total set, but why does that matter? Some collectors will care about having one-of everything, but how does this relate to the cost of actual decks? Well, now that we know the cost of the entire set, we can begin make some predictions on how value might be allocated. A naïve person might assume that each rare is going to be worth $2 each, as this is the price of a pack, and each pack gets a rare, but anyone with any experience will know card prices tend to be concentrated in a relatively small percentage of the total rares. To put it bluntly, good rares cost more than bad rares, but how can we predict the price difference between good and bad rares? Well, let’s start off by coming up with a simplified model. 

Let’s continue our assumption above that there are a total of 50 (non-item non-hero) rares in the set. We can break up these rares into 5 equally sized tiers, with the “Tier 1” rares being the most played and most expensive, and the “Tier 5” rares being the least played and the least valuable. Let’s assume that half the value of the whole set falls into tier 1, meaning that the 10 most expensive cards in the game cost $5 each ( $100 / ( 2 x 10 ) = $5 ). Half of the remaining value falls on the tier 2 cards, meaning they cost $2.5 each ( $50 / ( 2 x 10) = $2.5 ). The remainder is then evenly distributed between the tier 3, 4 and 5 rares, leaving them at about $ 0.83 each ($2.5 / 3 = $0.83). In this world let’s construct a variety of possible decks, and estimate their cost. (Remember: we are only looking at non-item non-hero cards, meaning that we are only talking about 25 cards).

Screen Shot 2018-10-12 at 10.11.50 AM.png

So, my assumptions were a little off. The top 20% of cards takes up a bit more than 50% of the total value, and the next 20% of cards take a bit less, but overall, they take almost exactly 75% of the total value. Using these values, we would adjust the model above to price tier 1 rares at $5.50 and tier 2 rares at $2 approximately. This does not have a major impact on the projected deck costs listed above, other than the “deluxe” deck moving up by $3. This is because the tier 1 cards are stealing the value from the tier 2 cards, and assuming you are running them in relatively similar amounts, there shouldn’t be a massive increase in deck price. 

One thing I did notice in my research through MTG prices was that my “quintile” model is likely too broad. The drop off between the first 10% and the next 10% is pretty brutal. Let’s take a look at a revised table, but break the top tier into two. (Numbers are rounded)

Screen Shot 2018-10-12 at 10.12.29 AM.png

This difference will only matter for decks that are particularly top-heavy in cost, as a deck that is an equal mix of the top 10% and the next 10% of cards would have a similar cost to that predicted above. Still, let’s look at a quick example of an “Ultra Deluxe” deck that is the same as the “deluxe” deck described above, but all of the “tier 1” rares are now shifted to the top 10% rather than spread through the top 20%. Using the numbers predicted from Magic, the average cost of the top 10% cards would be $8.40 ( [$100 x 42%] / 5), the next 10% would be $3.20 ( [$100 x 16%] / 5), the next 20% would be $1.60 ( [$100 x 16%] / 10), and the tier 3+ cards would be $0.87 ( [$100 x 26%] / 30). Given these predicted prices the “Ultra Deluxe” deck might cost about $114, which is certainly getting expensive.

While $114 is not crazy relative to what other people spend in other games, it is certainly not cheap, but before moving on from this number I want to stress that this is not particularly realistic. Think about it: this deck would need to be playing playsets of 4 out of the top 5 most expensive cards in the game, which is a bit twisted. Obviously, it is possible that value is really compressed into just very few cards, which is a legitimate concern. I will return to that later in the article.

Overall, I am satisfied with the set of assumptions that I made. There is a lot that can go wrong with them, but they seem to vaguely resemble what we see in Magic the Gathering, which is the closest analog we have for comparison. Before transitioning out of this nerdy math dump, I want to touch on a few of the short-comings of this analysis, in the interest of full disclosure, and so other geeks can improve on my work. First off, when preforming the analysis of MTG card costs I am treating rares and mythic rares as equivalent, when this is not exactly accurate. Let’s take a specific rare as an example – Assassin’s Trophy. While opening a rare when crack a Guild’s of Ravnica pack is 8-times more likely than opening a mythic, your chance of opening an Assassin’s Trophy is about 2.5 times more likely than opening a specific mythic like Doom Whisperer. While Doom Whisperer and Assassin’s Trophy are each worth about $25 each (October 9th prices), they consume unequal amounts of the total set value given that you are going to open a lot more Assassin’s Trophies than Doom Whisperers. I really don’t know how I should factor this into the model, so I am just ignoring it. If someone wants to suggest a good correction factor please contact me - preferably through Twitter or Discord (Neon3989). 

Nerd blurb over!

OK, so for those of you who skipped the math bit, the TLDR was that the assumptions were relatively reasonable, though they might underrate the chance of a few cards holding the majority of the set value. Once issue that is worth noting is that I am using the example of physical cards to predict the behavior of digital cards. Since physical cards need to be stored in warehouses, packaged, and shipped there is a reasonable incentive to allocate some cost even to unpopular cards. In a digital world, we might see value shift more aggressively to the top as the operational cost of managing the least expensive cards drops to zero. In fact, if we look at Magic the Gathering Online we see that the top 10% of cards takes up 68% of the total value, as opposed to the 42% we see in paper Magic. We do not yet know the specifics of the rules around the Artifact market, but it will have a big impact on whether we get hyper-concentration of value on a few key cards. I will touch on what hyper-concentration might look like later in the article, but the online marketplace may increase the chance that this happens. 

Since we are on the topic of MTGO, I should explain why I am using paper data as opposed to MTGO data. Simply put, the MTGO economy has some very weird stuff going on that I can’t fully fold into the model. First off, MTGO has set redemption, which means it interfaces with the paper economy in a strange way. Second, there are treasure chests, which inject more copies of standard legal cards into the system, following a pretty complicated set of mathematical acrobatics. Third, the economy is heavily influenced by the popularity of draft, which injects a ton of cards into the environment, causing something of a collapse of traditional supply-demand economics. Take Core Set 2019 packs for example. They are being sold in the official store for $4, but the going rate through trading bots is $2. One day I might try to tackle this head-on and come up with a better model to explain how all the parts of the economy are connected, but today is not that day. 

Finally, I should once again note that I am still ignoring the “sealed bonus”, which actually super-charged for MTGO. This is probably because of draft as well, but reinforces how strange the MTGO economy is. But with that, there are still a bunch of questions that need to be addressed, the most obvious related to heroes and items. At their core, these questions are much easier than it may seem, but there are a few marginal questions that require cross-examination, so let’s dig into the first part of the question. 

Turkey Talk


Using the model that I have outlined here it might be totally unnecessary to include heroes and items in my calculations. This isn’t because heroes and items are going to be free, but because the driver to the overall economy is the “cost of production”. Up above I proposed that “deluxe” decks would cost about $80 because top tier rares would cost $5 on average and that second-tier rares would cost $2.50, etc. I still think $80 is a reasonable estimate for a competitive deck, but that value will be spread over items, heroes and main deck cards. Why is that? Well, let’s use an example, and since I am Canadian and we just had Thanksgiving, let’s talk turkey.

You are a turkey farmer, and business is fine. Every turkey costs you $20 to raise, and you are able to sell a turkey breast for $11 each, so you make $2 each on each bird, same as all your turkey farming friends. One day one of your buddies comes up to you and says:

“Hey did hear? Those city slickers don’t just want turkey breast anymore. They want legs! Drumsticks! I hear they are buying them at $4 each! I’m going to make $10 on each bird!!” 

Well, that is some very interesting information. $10 a bird sounds like a killing! You want in on that sweet feathery action. But wait a minute… If you start selling your turkey legs at the same price as this guy, who says anyone is going to buy from you? Wouldn’t it be better to sell the legs and breast at a $1 discount so you can pull in all the profit? Sure, it is not $10 profit per bird, but $6 a bird is still way better than what you were getting before, and you will sell so many turkeys!

You are feeling like a genius, selling legs at $3 each and the breasts at $10 each for a while, until suddenly demand dries up. No one is buying! It takes you a little while to figure it out, but it turns out that “buddy” of yours that you backstabbed went and backstabbed you right back! He is selling his turkey legs at $2 each and the breasts at $9 each! He is back to making $2 profit per bird just to screw you! You can’t go any lower than him, since the turkeys still cost $20 each to raise, and you need to make a bit of profit. After a couple of weeks of gobbling up profits, you are right back where you started.

So, what happened here? One-minute turkey was an extremely profitable market, but the next you were back to making $2 a bird. The production cost is set per turkey meaning that the price will ultimately be set the production costs. The production costs do not care one bit what happens to the bird once it is sent to the butcher. Maybe turkeys suddenly become popular pets, and are no longer killed at all! How much does it cost to buy these house turkeys? About $22. In classical economic theory it is assumed that fluctuations in supply and demand will impact the price of the good, but in cases like this we are describing the price is artificially fixed. For digital cards there are also very few of the normal constraints on the system. In the case of turkeys, if you want to expand your production there is a lot of work involved. For Valve, doubling or tripling the number of cards release costs nothing. As a result, you will see the selling-price-per-bird stay tied to the production costs. 

With that, let’s return to the question of heroes and items. Artifact packs are kinda like a turkey with three cuts of meat; the rare slot, the item slot and the hero slot, which are all worth money. The exact distribution of money in each of those slots is going to be determined by the relative demand/scarcity of each type of card, but the overall cost is going to be determined by the overall “cost of production”. It should actually be noted that the consumers in our story see a price cut in their meat, but that is not exactly how things will pan out in our game. Applying this model to Artifact we most resemble people who are buying whole turkeys, as our decks will be made up of all three “cuts of meat”, meaning our savings in one area are balanced out by costs in another.

Before we move on to explain this further I want to make a bold prediction: the “cost of production” will be very close to the cost of the number of rares times the value of a pack. This means that the “cost of production” for heroes and items will either be less than or similar to the cost of production for non-hero non-item cards. With that said, let’s talk about heroes, and where they might fit in. 

We Can Be Heroes (just for one slot)


 So, we know there is one slot per pack that will be a hero, but we don’t know the drop rate of common, uncommon, and rare heroes. As of this writing (October 9, 2018) there are 12 heroes that have been confirmed to be rare. There are a few heroes that have not yet revealed, and there is a possibility that some of the hero rarities are mistaken, which suggests we can probably expect between 12-16 rare heroes. With those number in mind, we can actually predict the number of packs you would need to open to receive one of each rare hero.

Screen Shot 2018-10-12 at 10.14.20 AM.png

With these numbers it may seem like heroes are going to be hella expensive. Even the best-case scenario would lead to a “cost of production” that is higher $120, which is a good deal higher than what I predicted above. It should be stressed that in this case we are discussing heroes, which means they are a playset of 1. Let’s take a somewhat radical example. First, let’s assume that all the value of a deck is in the heroes. Value of heroes is evenly distributed between 16 rare heroes. The cost of production would be $320, meaning all heroes would be $20 each. A deck that played all rare heroes would cost $100 in that scenario, and that requires some pretty bold assumptions. Even a deck that only played 4 rare heroes would drop the cost to $80 in that case, which is quite reasonable in my opinion.

These numbers are important because they set the cost of production. If the hero-centric cost of production is less than the main-deck cost of production then the cost of heroes will not be a major contributor to deck costs. Let’s go back to the turkey example. The hero cards resemble the legs. There is some demand for turkey legs, but it is much less than breasts. There is still some demand for heroes, so they will take up some price, but I expect it will be a fairly small percent of total deck prices (usually). Valve would need to use a very low rare drop rate to actually push up the cost of heroes. That 10% line seems like an important value to watch. If the drop rate is below 10% you might expect competitive rare heroes to be some of the most expensive cards in the game. If the drop rate is at 10% or above then expect hero prices to be quite reasonable.

Tyranny of the Vesture of the Tyrant


 Much of that concepts described above about heroes also applies to items, but we have a lot less information to work off of unfortunately. For heroes we are pretty sure there are 44 heroes that appear in packs, and we know the rarity for most of them, so it is pretty easy to make a couple simple assumptions and then make predictions. For items, it is a lot harder to guess how many remain to be previewed and how many rares there will be in total. Most importantly, we don’t know how rarity works for the two slots dedicated to items. Do each have a chance to be a rare? Or only one? What are the chances that it is a rare? This kind of information is important enough that I am not going to put any effort into solving this problem. Ultimately, if there are a lot of rares and the drop rate is low this will lead to a high production cost, which in turn will dictate the price of decks. Pretty broad statement, I know, but I can’t give much more right now. 

The point I will make here that is worth mentioning is the potential for items to do some weird things for the market. From what I understand, Vesture of the Tyrant is one of the meta-defining cards in the beta. There are a lot of decks that run some copies, and some decks that run a lot of copies. It is hard to know if this will hold when we move to post-NDA beta or full release, but I have no reason to believe it will change. In that case, Vesture has a chance to warp the economy, as everyone will be fixated on getting their hands on one specific card. This is one of the dangers of powerful items, since they can go in any deck, causing the demand to be turbo-charged. In a worst-case-scenario I can imagine a production cost around $100, of which $30-$50 is allocated to Vesture, meaning decks playing 3x Vesture cost over $150. I am a little concerned about this, but it is a pretty easy problem to avoid, as some simple tweaks to the rare drop rate will avoid this problem. For example, if there is a 50% chance of opening a rare item in every pack and there are 12-16 rare items, Vesture would only be able to reach about $25, assuming it consumed a full 50% of the set’s total value. In that case the price is unlikely to be sustainable, as it will be washed out by the value of non-hero non-item cards. While I have no special information that verifies things one-way-or-another, it should just be noted that Vesture might end up being the chase rare if the drop rate is relatively low. 

50 rares assumption

At the top I assumed that there would be 50 non-item, non-hero rares in the game, and have built all my predictions on that figure. What if that is wrong? Does that totally throw off these predictions? Not really, it just requires a simple correction factor. If there are 40 rares this brings down the total cost of production down to by 20%, meaning all the other projected costs will drop by 20% as well. If there are 60 rares then the cost of production has increased by 20%, meaning all my projected costs just increase by 20%. This is pretty easy to correct once we see the whole card file.

Hyper-concentration/Meta Diversity

One of the potential problems that could end up being a big deal is if the cost of cards ends up being distilled down to a couple key pillars of the format. To show this, let’s imagine that all the value of the format is distilled down to 4 cards – one of each color. These pillars of the format hold all the value, and every deck of that specific color must play 3 copies of these cards. Using the 50 rare assumption spelled out above, this suggests that each of the money rares will be $25 each. A 2-color deck would runs 6 money rares, meaning it would cost about $150, which is not a great price.

While this is a vaguely interesting thought exercise, this is a pretty exotic set of conditions, so why does it matter? Well, this is what happens when the value of the set gets concentrated in just a small number of decks. Imagine instead of 4 cards holding all the value you had two decks that held all the value. In that case the value of each deck is likely to push up to $150 (assuming no overlapping cards, and both decks are equally represented). Compare this to a world where there are just 6 competitive decks (with no overlapping cards, equal representation), and the cost will fall to $50. A well-balanced set is not only important for a healthy metagame, but it will help distribute the cost more evenly.

This leads to a couple of pet peeves that I want to rant about. The first is complaints about decks that are “all rares”, or Valve putting powerful cards at rare, saying this will increase the cost of the game. This isn’t quite true. First off, let’s pretend that Valve assigned card rarities totally at random. About a third of the most powerful cards would be rares because there are only 3 bloody rarities! Unless you expect them to deliberately put the most powerful cards at low rarities (which could ruin draft) there are going to be powerful rares, so calm yourself. Second, the number of rares in a deck is much less important than the percentage of value that is concentrated within specific rares. Let’s look back at my hyper-concentration example, where 4 cards consume all the value of the set. Every other card in your deck could be common or uncommon, and your deck would still be $150. Let’s take another example. Imagine that all rares were equally valued, making them $2 each. Even if you played a deck of 100% rares your deck cost would be only $50. In fact, bad rares are a bigger problem than good rares, as it dilutes the chance of getting competitive cards, and increases the cost of production. Don’t bitch when they put Annihilation at rare, you should bitch when they put Wrath of Gold at rare.

In some ways, this is counter-intuitive, and it especially hard to wrap your mind around if you are coming from the world of Hearthstone or Eternal where rarity is heavily tied to the value of a deck. This is just a fundamentally different economy, and works on totally different rules. If you still don’t quite believe me, let’s take a look at a couple specific decks from Magic the Gathering to emphasize this point (prices from October 9th).

Boros Angels               Esper Control              Boros Midrange

# of Rares: 24              # of Rares: 25              # of Rares: 28

# of Mythics: 16          # of Mythics: 7            # of Mythics: 9

$430                            $420                            $266

Looking at Boros Angels on paper, one would think it would be the most expensive deck of the bunch by a good deal, followed by Boros Midrange, and Esper would be the cheapest. In reality, Angels and Control are basically tied, and Boros Midrange is much cheaper than the other two. What is disguised in these numbers is that Esper plays Teferi, the most expensive Mythic in standard and Boros Midrange uses a handful of less popular rares and mythics. I should also note that finding decks that fit this narrative was not hard. Steel Leaf Stompy uses 7 more rares and 1 more mythic than Esper control, and costs $70 less (and almost $100 more than Boros midrange). If I went to look at historic data I expect the price to see just as much variation as I am describing here. It is far more important what rares and mythics you are playing rather than how many rares and mythics you are playing.

A final note: concentration really does matter. One of the most expensive Standard environments of all time was Battle for Zendikar Standard. Decks we packed with rares and mythics, but that doesn’t fully explain what was going on. The format allowed you to play almost any card you might want in your deck, meaning demand was super focused on a couple key cards since everyone was playing them. There was also some crazy hype around certain cards, driving up their costs in a way that wasn’t sustainable (you can learn more about the topic of “irrational exuberance” here).

The Competition


Through all of this analysis I feel that the $50-$80 estimate for competitive decks is fairly solid. It is possible that some decks edge up to the $100 mark if they become overly reliant on the most popular cards, but assuming the meta is reasonably well balanced, this is unlikely to be common. I should also note that his assumes no “sealed bonus”, which could push the cost of competitive decks as low as $30-40. Throughout the piece I have been saying things like “this seems reasonable”, but it is worth taking a moment to reflect on this.  

I have actually written extensively on the economy of other games, so we can use those numbers for this conversation (if you are interested here are two relevant links). I am not going to delve into my methodology here, but let’s just talk purely about how much money you would expect to spend to build a top tier deck starting with absolutely no collection, and no play rewards. I also want to stress that the only comparison that I can effectively offer here is the relative cost of different games, not so much on more broad questions like “is it worth it?”. Every time I write a piece like this I get a response like “You said $50 was a reasonable price of a deck, but that is WAY TOO EXPENSIVE FOR ME!” I mean…. sure. I’m not telling you how to live your life. If you can’t or won’t spend money on card games that is fine, but what are we even doing here? This is most useful for people who want to determine which game offers the best “bang for your buck” in terms of deck costs, so with that in mind let’s take a look at some other digital card games:

Shadowverse: $70
Eternal: $75
Hearthstone: $105
Magic Arena: $120 

In the interest of full disclosure, some of these numbers are a touch out of date, but are likely accurate within +/-20%. In this world, we can see that Artifact matches up quite well, as the upper end of the $50-$80 range is roughly in the zone of the cheapest card games available on the market, with the lower end of Artifact’s range being very affordable. Obviously, these games offer some free-to-play rewards, and Artifact does not, which explains why Artifact might be aiming for a low price point. Comparing free-to-play and pay-to-play experiences of games is very much like comparing apples and oranges, so I am not going to try to account for this in any way. I will note that both Hearthstone and Magic Arena have a pretty rough free-to-play experience, as it is very difficult to maintain a competitive collection without regularly spending money on the game. Given that, I would predict that Artifact will be noticeable cheaper to maintain a competitive collection than Magic Arena and Hearthstone for the average player. 

I want to take a moment here to dispel a misconception about trade-based economies. Some people talk about trading as if it is a magical tool where the cost of cards evaporates into thin air. “I buy cards cheap, and then trade them back before the price falls! I basically play for free! This means that trading-based economies are better for everyone.” While this is a cute narrative, someone is paying for these cards. Strategic trading is a method to transfer some of the metagame’s cost burden from more attentive players to less attentive players. Trading does not somehow miraculously erase the cost of buying decks for an “average” player. In one thread comparing the costs of different games a commenter said that Magic the Gathering Online was cheaper than Eternal (which I consider best-in-class) because they could rent whatever deck they want for $20 a month. Obviously, it is true that $20 is cheaper than $75, so if you are playing the game for exactly one month you are coming out ahead, but this means you are spending $240 a year on MTGO, which is way more than most Eternal players spend. I have seen a shocking number of comments that employ bogus reasoning to erase the cost of decks through the power of trading. This is not how the world works.

Draft Could Totally Ruin This (In a Good Way) 

As a final point, I want to talk about the role draft may play in the economy. While we don’t know how much it will cost, beta testers have been hyping draft constantly. I am fairly certain that we were told that drafts would use a “keep-what-you-draft” model, which means draft would be a supply for cards, as people are constantly opening up new packs. It is difficult to anticipate the relative popularity of draft and constructed, but if draft ends up being way more popular than constructed, then the card prices will basically collapse. Maybe premium rares like Annihilation, Axe, or Vesture are the most in-demand constructed cards, but the constant flood of product from draft will push down the price over time no matter how much they are played in constructed.

It should be noted that this phenomenon will take a little while to come into effect simply as a function of how long it will take people to draft. I anticipate the first month or so the markets will be really wild as people get accustomed to the system, but after the initial chaos things will smooth out, and we will begin to see the price of staple cards begin to fall. The rate of that decline will be tied to the relative popularity of draft. Obviously, some specific cards will spike and fall based on tournament results or high-profile articles, but the overall metagame cost will decline as a result of drafters. If draft ends up being much more popular than constructed then all my math is going to go to shit as a result, but as an upside it means that constructed will be extremely affordable. If this phenomenon holds for future releases you should expect a pattern where constructed deck costs are initially very high, but they drop off rapidly over the first month, similar to what we see in Magic the Gathering Online.

The relationship between drafters and constructed players is actually more interesting than this, as constructed players help off-set the cost of draft. Draft uses 5 packs of cards, meaning that it should cost $10 if there are no prizes associated. Let’s say that draft doesn’t end up being very popular, and the vast majority of players spend most of their time playing constructed. In this world we would expect the overall prices to follow the model I have laid out very closely. If this is the case, then drafters would be able to draft almost for free as they can sell the rares they open in draft and get most of their money back. If draft becomes more popular, then the prices of constructed cards will fall as everyone opens a ton of product. In this case, draft functionally becomes more expensive because less money is reimbursed when drafters try to flip their rares. Most likely we will see an equilibrium, where both drafters and constructed players share the cost burden, partly as a function of players switching between the two as the functional prices change. It is worth quickly noting the potential impact of draft prizing on the overall prices of the metagame: very little. And prizing that is offered is likely going to be related to an increase in entry fees. If it isn’t then the whole pricing scheme needs to be shifted to reflect the “actual” cost of packs. Similar to having fewer than 50 non-hero non-item rares, this would be a simple adjustment, though it might be slow-acting in practice.


Hopefully that was a helpful run down of potential scenarios! This was a lot of fun to write up, and gave me an excuse to geek out on a topic I am very passionate about. Let’s quickly summarize what we learned here.

  • Assuming 50 non-hero non-item rares, and that non-hero non-rare cards set the cost of production, competitive deck costs are likely to fall between $50-$80.

  • If there are more than 50 non-hero non-item rare cards than you will expect deck prices to be higher, and if there are fewer non-hero non-item rares than the price will be lower.

  • There is a chance that the price of decks is set by the items or the hero cards, depending on their relative drop rate. I personal predict the cost of decks is going to be closely tied to the “cost of production” for the non-hero non-item rares.

  • This model assumes no “sealed bonus” for packs, and that draft is not popular enough to influence the prices in constructed. If draft ends up being very popular you will see prices of metagame staples fall over the course of time. If draft does end up being very popular then the cost of competitive decks could fall as low as $30-$40.

  • Given this information, Artifact is very likely to be less expensive than games like Magic Arena and Hearthstone, and could be cost-competitive with games like Eternal and Shadowverse.

  • The main danger for Artifact is an unbalanced metagame where the majority of the set’s value is concentrated in a few specific cards. Vesture of the Tyrant is a lead contender for this, as it is very likely to see extensive play between several decks.

  • We have not included the transaction costs in this analysis. If Artifact uses the same model as the Steam market, it will be a pretty big drag on the economy, causing prices to be noticeable higher.

Thank you all for joining me today! If you liked this piece, you can find some of my other articles on card game economics here and here. I also want everyone know that A+Space is now partnered with Inked Gaming, where you can use the code “ASPACE12” to get 12% off your orders of all sorts of sweet nerdy gear. Be sure to share your thoughts on Twitter and on Reddit! I love to discuss these questions as you can see! You should also check out my podcast, if you haven’t already.

Thanks again for joining me today! Talk to you again soon.