Review: Artifact

Epistemic Sta­tus: Alpha tester

Bot­tom Line: If you are will­ing to de­vote the time and at­ten­tion to a deep strate­gic game, Ar­ti­fact will re­ward you hand­somely. I highly recom­mended those who like such ex­pe­riences to make the time. If you are not will­ing to de­vote the time and at­ten­tion, you will likely be frus­trated and bounce off, and what time and at­ten­tion you do have to game with is bet­ter spent el­se­where.

Ar­ti­fact is an amaz­ing game. Ar­ti­fact is gor­geous, im­mer­sive and fla­vor­ful, hilar­i­ous, in­no­va­tive, ex­cit­ing, sus­pense­ful, skill test­ing, strate­gi­cally com­plex and re­ward­ing. The ex­e­cu­tion is bug-free and flawless. It is the most fun I have had play­ing a game in a long time.

It streams well and is an ex­cel­lent spec­ta­tor sport for those who know the game and cards, and will be sup­ported by an inau­gu­ral tour­na­ment with a one mil­lion dol­lar first prize. It is a Valve game, so you know it will get the level of sup­port and at­ten­tion it de­serves in ev­ery as­pect.

The eco­nomic model is the right one. Rather than ad­dict­ing play­ers to daily re­wards and grinds, Ar­ti­fact charges money for a game worth play­ing. You own your cards and will soon be able to buy and sell them. For your ini­tial $20 you get 20 packs, each con­tain­ing at least one card of the high­est rar­ity and of­ten two or even three. Ad­di­tional packs are $2. Play­ing events costs only a sin­gle event ticket ($1), and you turn a large profit if you can get three wins be­fore your sec­ond loss. I will say more on this later, but the model pre­sented is ex­traor­di­nar­ily gen­er­ous, and those who are com­par­ing it un­fa­vor­ably to Magic On­line should be ashamed of them­selves.

The catch is that Ar­ti­fact is com­plex. Very com­plex. Com­plex enough that I have had mul­ti­ple Magic pro­fes­sion­als try the game only to have them re­port back that they bounced off the game be­cause they did not un­der­stand what was go­ing on. Ar­ti­fact makes the most of its com­plex­ity, and uses its roots in DOTA 2 to jus­tify much of it, but the com­plex­ity is still there and com­plex­ity is bad. Your first hour is likely to be over­whelming and con­fus­ing, as lots of cool things are hap­pen­ing all around you but it’s im­pos­si­ble to fully know why or what they mean, or what is likely to hap­pen next.


Ar­ti­fact matches up two teams of five heroes each, who do bat­tle across three dis­tinct lanes. This mimics the struc­ture of games like DOTA 2, where there are five heroes on each side and three lanes in which to do bat­tle, which we will call left, cen­ter and right.

In ad­di­tion to lanes and teams of five heroes that re­turn when they die, Ar­ti­fact also takes its con­cepts of tow­ers, an­cients, creeps, boun­ties and items di­rectly from DOTA 2. Things that would seem need­lessly com­plex or ar­bi­trary in an­other con­text… still feel some­what that way on oc­ca­sion, but it helps a lot that they’re car­ry­ing the con­cepts over from an­other very com­plex game. This helped me even though I never learned how to play DOTA 2 or any similar game.

The ba­sic jist of Ar­ti­fact is that play­ers have five pow­er­ful heroes that let you do things of their color while they are in the ac­tive lane, and that fight hard, and they also sum­mon a va­ri­ety of other things to fight. If heroes die, they come back with a one turn de­lay. Play­ers start with 3 mana on turn one in each tower, which in­creases by one each turn, and use this to pay for stuff. After play­ers are done do­ing stuff, each unit does dam­age to the unit op­po­site it. If there’s noth­ing there, it dam­ages the en­emy tower or an­cient. Do 40 dam­age each to de­stroy two of the en­e­mies’ three tow­ers to win, or do 40 and then on fu­ture turns 80 dam­age to one of them to kill the an­cient, which also is a win.

At the start of the game, each player sends one of their three start­ing heroes into each of the three lanes at ran­dom, and ran­domly is given three 24 (mean­ing two power and four health) ‘melee creeps.’ Each player starts with a tower in each lane that has 40 health.

The turn or­der is:

  1. Play­ers are shown where new ‘creeps’ will en­ter the bat­tle. By de­fault, each turn each player gets two 24 creeps that are each put in a ran­dom lane.

  2. Play­ers si­mul­ta­neously choose which lane to de­ploy each newly available hero to. Your fourth hero be­comes available on turn two, your fifth on turn three. Any heroes that are kil­led ‘re­turn to the foun­tain’ and be­come available two turns later, and any heroes that re­turn to the foun­tain with­out be­ing kil­led be­come available the next turn.

  3. All new units are de­ployed to each lane. First, the game at­tempts to place new units in empty spaces across from en­emy units. If that can’t be done, then a new pair of open spaces is cre­ated (ran­domly on the left or right side of the ex­ist­ing units) un­til no more such pairs are needed. If that leaves empty spots on one or both sides of the bat­tle, those slots are filled by a straight ar­row (50%), left ar­row (25%) or right ar­row (25%). Then all new units and all ar­rows are shuffled and dis­tributed to the lane at ran­dom. Or, in English, first you fill in any empty space, then you de­ploy to the side, and any empty spaces ran­domly have ar­rows half the time.

  4. Play­ers draw two cards and the mana ca­pac­ity of all tow­ers in­creases by one.

  5. Play pro­ceeds to the left lane. The player with ini­ti­a­tive can ei­ther take an ac­tion or pass. Ac­tions in­clude cast­ing a spell, de­ploy­ing a piece of equip­ment, de­ploy­ing an im­prove­ment, de­ploy­ing a creep card or us­ing an ac­ti­vated abil­ity of a unit or im­prove­ment. If you play an im­prove­ment, you can play it in any lane you like. Any­thing else must go in or im­pact the ac­tive lane un­less it ex­plic­itly says oth­er­wise. A new creep must first go in the empty space of your choice. If there are no empty spaces, you can choose to put it on the left or right side. If a card de­ploys other things rather than it­self, the game will choose for you.

  6. If take an ac­tion rather than pass­ing, and you had ini­ti­a­tive, you lose it.

  7. This con­tinues un­til both play­ers pass.

  8. Each unit does dam­age to the unit across from it. Da­m­age per­sists. If a unit runs out of health it dies. Units that are not op­posed dam­age the en­emy tower di­rectly. Ar­mor re­duces dam­age taken from any source, nega­tive ar­mor in­creases it. There are also a bunch of other abil­ities.

  9. If a tower has taken 40 dam­age, it is de­stroyed and re­placed by the an­cient, which has 80 dam­age. If the an­cient has taken 80, it dies. If you de­stroy two tow­ers or one an­cient, game is over and you win.

  10. Each en­emy you kill gives you one gold, or if it is a hero it gives you five gold. At end of turn, you can shop to buy items, which cost 0 mana and are played on your heroes to make them bet­ter. You come with a deck of items, and can choose be­tween buy­ing those (in a ran­dom or­der each turn), buy­ing a fully ran­dom item, and/​or buy­ing a con­sum­able. Gold can be kept for fu­ture turns.

There are a lot more de­tails, but that’s Ar­ti­fact. There are of­ten com­plex tac­ti­cal ques­tions that can play out over many turns with a lot of bluffing and un­cer­tainty, and you must di­vide your re­sources be­tween the lanes while an­ti­ci­pat­ing how your op­po­nent will do the same. If you in­vest the wrong amount, you can have more than enough to win a tower with­out enough to kill the an­cient in time to mat­ter, and/​or find that you’re a turn be­hind on the tower that ac­tu­ally mat­ters. Often a lot of power effec­tively goes to waste in a side fight un­likely to change the out­come, and one player figures this out well in ad­vance of the other.

Ini­ti­a­tive on most turns is a li­a­bil­ity, as you must com­mit to and show what you are do­ing be­fore your op­po­nent. But on key turns, ini­ti­a­tive is ev­ery­thing, be­cause your first ac­tion kills or stuns their heroes, pre­vent­ing them from do­ing any­thing. Keep­ing the right color hero al­ive in the right lane is a fre­quent fo­cal point. Often games come down to who can have ini­ti­a­tive in the key mo­ment, and how much play­ers can af­ford to give up to make sure they get it.

On the first turn, play­ers face a ran­dom set of matchups be­tween heroes. Cards that turn un­fa­vor­able or neu­tral matchups into more fa­vor­able ones (e.g. they were go­ing to kill you and live, and now at least you both die, or be­fore you both die and now you win) are usu­ally the most valuable things in the first two turns. If you win those early fights, you get five gold, which you can use to buy an item that threat­ens to win you the next fight, and mean­while you’re do­ing early dam­age and can cast spells while your en­emy sits on the sidelines. Some decks try to do se­ri­ous dam­age to tow­ers right at the start, which shifts fo­cus to that.

As the game de­vel­ops, it be­comes more and more about de­cid­ing which tow­ers to fight for, and de­ploy­ing power to those tow­ers, and shifts from build­ing up long term re­sources to win­ning fights on the spot, sav­ing your tower or dam­ag­ing theirs, or un­leash­ing your pow­er­ful high-cost spells and not let­ting them cast or cap­i­tal­ize on their ver­sions. Com­mon endgame sce­nar­ios in­clude each player win­ning one tower and a big fight over the third one, one player go­ing for an an­cient and try­ing to stall in one or both of the other two tow­ers, and a race to take down his an­cient be­fore the en­emy takes down yours. Sacri­fic­ing one tower is of­ten a wise move, ei­ther be­cause you can’t rea­son­ably fight back, or be­cause you can strand a lot of re­sources there, of­ten in­clud­ing mul­ti­ple heroes, with­out a prac­ti­cal path to kil­ling the an­cient be­fore the game ends.

A key ex­pe­rience is that all the things are hap­pen­ing all over three boards. It is up to you to figure out which of those things are worth fight­ing for and spend­ing re­sources to pro­tect or at­tack. Due to the op­tion­al­ity and ran­dom­ness of de­ploy­ment and ar­rows, and not know­ing your op­po­nents’ hand, as­sum­ing the game will go a cer­tain way has a habit of back­firing, but I also of­ten have the bad habit of try­ing to win ev­ery­thing at once when that is not nec­es­sary.

A very im­por­tant rare is An­nihila­tion, which de­stroys ev­ery unit in a lane. When you play against a blue deck, it likely has this card and that forces you to play the en­tire game differ­ently to avoid an over-com­mit­ment, and to not let them have an open­ing to use it later. Another op­tion is At Any Cost, which does six dam­age to ev­ery unit in the lane.

Another im­por­tant rare is Time of Triumph. When red decks reach eight mana, they can give all heroes in a lane a huge boost in effec­tive­ness. As the hero Axe puts it, your an­cients’ days are num­bered.

Green decks have Emis­sary of the Quo­rum. This costs eight mana, and is a creep with high health.

A key item is Blink Dag­ger, which al­lows heroes to shift be­tween lanes.

You choose your five heroes and the other they will en­ter the game. Each also adds three copies of its as­so­ci­ated card to your deck, which ac­counts for 15 of your 45 cards. You also build a nine (or more) card item deck.

As a re­sult of these and a few ad­di­tional similar cards, games where play­ers have ac­cess to rare cards feel very differ­ent than games where they have ac­cess to only com­mon cards. Games of limited, or where play­ers lack key rares, are about figur­ing out where to fight how hard, lin­ing up key re­sources and ac­cu­mu­lat­ing board ad­van­tage. Games with higher level con­structed decks force play­ers to worry much more about over-com­mit­ment and be on the look­out for sweep­ers and pow­er­ful knock­out blows that threaten to ren­der those early bat­tles ir­rele­vant.

Ways to Play

Cur­rent op­tions are to play con­structed, phan­tom draft or keeper draft, and do so in so­cial, ca­sual or ex­pert mode. Ca­sual mode is like ex­pert ex­cept with no en­try fees and no prizes. I gen­er­ally dis­like play­ing on ca­sual mode, since play­ers won’t take it se­ri­ously when there is noth­ing at stake, so I have only ex­per­i­mented with the ex­pert queues. A new player would pre­sum­ably start out in ca­sual for a while, to avoid hem­or­rhag­ing cash and/​or get­ting con­sis­tently crushed while learn­ing the game.

You can also play against the AI. The AI is not good at Ar­ti­fact, and lacks strate­gic plan­ning, but it is good enough to al­low one to get a sense of whether a deck has been built in a rea­son­able fash­ion and get a han­dle on how it plays. It is also very good for play­ers start­ing out.

Ex­pert queues in con­structed and phan­tom draft cost one event ticket ($1) to en­ter. You play un­til you get five wins or two losses. For three wins you get your ticket back, and the fourth and fifth wins each grant a booster pack ($2). If we value boost­ers at their re­tail price, we get an ex­pected value of 0.906 at a 50% win rate, and even a lit­tle bet­ter than that puts you ahead of the game. The catch, of course, is that packs can­not be effi­ciently con­verted into event tick­ets, and there are trans­ac­tion costs to us­ing the Steam Mar­ket­place, so even if packs trade at the full $2 you’re go­ing to take a hair­cut. Keeper drafts re­quire you to bring the packs and al­low you to keep what you draft, and cost two event tick­ets rather than one. I ex­pect the de­fault draft mode to be phan­tom draft.

Ar­ti­fact is a com­pli­cated struc­ture as it is, so it is un­der­stand­able that play­ers are limited to a few proven play modes and a sin­gle con­structed for­mat. Hope­fully in the long term this will ex­pand as we get more play­ers, bet­ter col­lec­tions and more sets.

In the fu­ture, we have been promised a $1 mil­lion dol­lar first prize tour­na­ment, doubtless with more to fol­low if the game does well. Play­ers who are good enough to con­sider com­pet­ing for that should not be dis­tracted by small ex­penses or grind­ing op­por­tu­ni­ties, and fo­cus on im­prov­ing their play as rapidly as pos­si­ble.

Graph­ics, Sound and Un­der Interface

This is a beau­tiful and supremely well-pol­ished game. I have not yet en­coun­tered a bug or glitch of any kind in the beta, and there were al­most none even in the alpha. Game play is as smooth as can be and ev­ery­thing just works. Given the com­plex­ity of the un­der­ly­ing game, it does a great job show­ing you what is hap­pen­ing and why.

While the card art is only about on par with other games, the rest of the in­ter­face is full of great touches. Each player gets a highly emo­tive imp that illus­trates what is go­ing on, which is mostly great fun, even if it is kind of an­noy­ing when it is em­phat­i­cally point­ing out that your tower is about to die (since I as­sure you I am already aware of that). There is a ton of in­for­ma­tion on the screen, or on other screens one can scroll to, and it is all easy to nav­i­gate and view once you are used to it. The only thing I dis­like in this area is that it is cur­rently a bit an­noy­ing to see what the as­so­ci­ated hero cards are for the en­emy heroes, but that will doubtless be fixed soon.

Rather than the quick repet­i­tive one lin­ers of Hearth­stone and Eter­nal, the heroes and many of the crepes have rich per­son­al­ities and lots to say. There are many lines that de­pend on the in­ter­ac­tion of mul­ti­ple spe­cific cards, or cards get­ting to do a par­tic­u­lar thing. It ac­tively makes me want to play with a va­ri­ety of cards and heroes, and un­like ev­ery other card game, I never play with the sound off.

Eco­nomic Analysis

Play­ers have come to ex­pect a free to play ex­pe­rience from many of their games. Not only do they ex­pect to not pay, they ex­pect to be paid for play­ing, in the sense of hav­ing a bet­ter and more ex­pen­sive col­lec­tion af­ter play­ing than they did be­fore play­ing.

This is a toxic, no good, very bad thing. I shared some of my thoughts in my write-ups of Eter­nal. I fully echo and en­dorse what Richard Garfield said in A Game Player’s Man­i­festo. In­stead of play­ing games, play­ers are trapped in Sk­in­ner boxes, or play­ing in a way that is effec­tively work­ing for a tiny effec­tive wage. Huge swaths of games are never seen and ex­pe­rienced, while the most effi­cient are used over and over, and I be­lieve the eco­nomic model of free to play and card cre­ation is to blame.

Ar­ti­fact re­turns us to the model of buy­ing packs and then buy­ing, sel­l­ing and trad­ing cards to get what we want. The game costs $20, packs cost $2, en­ter­ing ex­pert-level events costs $1. Play­ing ca­su­ally is free once you’ve paid the $20, in­clud­ing ca­sual draft­ing.

So if all you want to do is draft, you can do so end­lessly for free, for­ever, and never have to worry about build­ing a col­lec­tion.

If you want to build a col­lec­tion, sev­eral things work in your fa­vor that are easy to not fully ap­pre­ci­ate.

Packs always con­tain at least one rare card, but of­ten con­tain two or even three. This adds up to a sub­stan­tial dis­count over time.

Much more im­por­tantly, Ar­ti­fact doesn’t have a fourth rar­ity. There are no ‘Le­gendary’ or ‘Mythic’ cards, at all. If you are get­ting your cards from packs, this cuts the num­ber of packs you need to get a full col­lec­tion by a fac­tor of three or more. If you are buy­ing sin­gles to get what you need, it cuts the cost of the best cards by that amount or more. We don’t know what the price of top cards is go­ing to be, but there isn’t that much room be­fore open­ing packs be­comes a bet­ter solu­tion, and traders start bust­ing open packs to find cards to sell.

Another bonus is that you can liqui­date any 20 cards to get an event ticket. Packs con­tain 12 cards, so if you liqui­date the bulk of your sur­plus com­mons, you can get back a ticket about once ev­ery three packs, cut­ting the cost by a full sixth. This also should help hold up the mar­ket value of com­mons, pre­vent­ing the mar­ket from be­com­ing flooded.

The mar­ket it­self is not yet op­er­a­tional, as the game re­mains in beta. I worry that the lack of a work­ing mar­ket is giv­ing play­ers the wrong idea of the game’s eco­nomics, but this will be reme­died soon. If the mar­ket runs well and Valve takes only a rea­son­able cut, it will con­sid­er­ably re­duce the cost of get­ting the cards you need and al­low the bulk of the value to be re­claimed from packs.

Magic On­line has long had the is­sue that packs trade for well be­low the cost to buy packs from the store. Either a lot of play­ers who don’t know any bet­ter must be buy­ing packs from the store any­way, tour­na­ments must be giv­ing out packs faster than the play­ers need cards, or some com­bi­na­tion of the two, for this to hap­pen. Ar­ti­fact also has a tour­na­ment sys­tem that eats event tick­ets and spits out packs, which in turn only spit out a frac­tion of their value in event tick­ets. So there is definitely at least some risk of there be­ing a flood of packs caus­ing them to drop be­low store value.

If that does hap­pen, packs get cheaper and so do sin­gles, mak­ing the game cheaper to buy, but mak­ing event pay­outs worse. Magic On­line charged a much higher amount of money to en­ter events, with drafts usu­ally at least $8 or so and con­structed events at least $5, and con­structed leagues now over $10 for five matches, so in­creas­ing the rake was a big deal there. At $1 for an av­er­age of about four games, even a large rake means that the hourly cost re­mains low, with ca­sual be­ing backed by Elo-style match­mak­ing. That makes me far less con­cerned about the drop in value, pro­vided the mar­ket­place fees re­main rea­son­able. It would still need to be fixed, of course, if it be­came a large enough effect, but that is also easy enough – just re­place the sec­ond pack prize with event tick­ets.

Add all these effects to­gether, and the effec­tive cost of play for Ar­ti­fact ends up be­ing on the ex­treme low end of col­lectible card game costs that aren’t free to play. Free ca­sual phan­tom drafts are an ex­treme con­trast with Magic On­line, as is the far less greedy pack struc­ture.

The catch is that those free to play Sk­in­ner boxes are pow­er­ful stuff. When one plays Arena, Eter­nal or Hearth­stone one feels like one is mak­ing money, be­ing paid to play. You’re not. You can’t sell the cards, so you’re be­ing given a dis­count on buy­ing the game, and a pa­thetic hourly rate of dis­count at that even when you are max­i­miz­ing your grind. But that is not how it feels.


Ar­ti­fact is a unique and amaz­ing game, with tons of in­ter­est­ing de­ci­sions and ex­cit­ing games and su­per rich in fla­vor, but it asks a lot of its play­ers. One must pay at­ten­tion and be com­fortable with a lot of com­plex­ity. When play­ers of differ­ent skill lev­els play, if the decks are similar in power level, the suffi­ciently su­pe­rior player will win al­most all the time.

The ex­pe­rience is not for ev­ery­one. You need to know what ex­pe­rience you are aiming for. Are you here to be com­pet­i­tive and try out for the mil­lion bucks against the likes of pro­hibitive early fa­vorite Stanis­lav Cifka? Are you here to do a bunch of fun drafts? Are you here for the sto­ries, fla­vor and lore? All are valid choices.

I would have loves to be an Ar­ti­fact streamer and com­peti­tor, but be­tween my fam­ily and var­i­ous job op­por­tu­ni­ties, I do not have the time to treat that with the se­ri­ous­ness it de­serves, and I can’t pre­tend that at my age I haven’t lost a step. I likely will stream from time to time when I have the chance, and I will of course throw my hat in the ring when the time comes, but I am un­der no illu­sions that I am tak­ing home the mil­lion dol­lars.

Two choices that are not available are the abil­ity to prof­itably grind out a dol­lar or two an hour in game as­sets, or to fully un­der­stand the game and/​or do well with­out giv­ing the game the at­ten­tion it de­serves. I en­courage play­ers to try out the game, but do so with your eyes open.

At a min­i­mum, if you have the gump­tion to get through the first hour or two of learn­ing, the ba­sic $20 product, with its free ca­sual phan­tom drafts, is an amaz­ing bar­gain. I am a big fan of Ar­ti­fact.