No matches

There are two things we need to make clear right at the start of this article.

First, all card games need certain randomness. This is just the nature of the genre. The RNG effects can vary from simple “what do you draw from the top of your deck” to “this card can win the game if the coin flips well”, but they need to be there. Players need to solve curveballs, even if it’s just a problem of what do they do without having all their resources in hand.

Without randomness, card game matches become streamlined. One optimal play follows the other. Even if this doesn’t mean the game gets solved, it certainly means the game gets boring to watch.

This leads to point #2: professional card game players will always hate the existence of randomness. When the success of your career ties to how many games you can win on a day, you don’t want randomness getting in the way of your skill. You want to control as many factors as possible, preferably all.

“Good RNG” is almost oxymoronic.

With these two axioms in mind, card game developers then need to solve a challenging task: how do they implement RNG effects, while making them good for the game, the players and the viewers. Some get it right and some get it wrong, but none get it perfect. “Good RNG”, after all, is almost oxymoronic.

To exemplify different developers’ solutions to RNG, let’s take a look at two of the largest card games on the market, before we move to our main topic.

The good and bad RNG of Magic

In Magic: The Gathering, RNG on cards is nonexistent. Every card does the same thing every time, so their effects are predictable. The game remains complex, however, because a lot of cards allow you to perform a variety of options. One player then has to choose what they want to do out of several options and their opponent then have to predict what the other person could do and try to play around it. So far, so good — it sounds like pro player’s dream.

Where MTG sins, however, is its resources system. As in most card games, you cast spells using mana. Unlike most card games, you need to draw the land cards that give you that mana. What’s more, you have to draw the exact mana color you need, otherwise you don’t get to play the spells. If you draw fewer mana than you need, you are essentially X turns behind. If you draw more mana than you need, then you don’t have enough threats or answers to play, and are also behind.

This leads to a situation where a significant percentage of games end because one player gets defeated by his own mana system. Even the best player in the world can’t win a game of Magic if they don’t get to play any cards.

Therefore, Magic has turned a good, or at least neutral, RNG effect — that of the top deck card draw — to a very bad one. Nobody wants to watch or play a game where one player gets to play stuff, and the other doesn’t.

It’s turn 6. I’ve not drawn lands in two turns, and my Ral Izzet and Cleansing Novas are dead in hand all because I’ve missed a mana drop.

The good and bad RNG of Hearthstone

When Hearthstone launched, it knew that it shouldn’t follow Magic’s bad example of mana draw. Thus, it made sure that players get the same mana every turn. They wouldn’t have to worry whether they could play their 5-mana spell on Turn 5, which is often a concern in Magic. If you want to, then you can and both players can approach the match with this information in mind.

But if Hearthstone had left it at that, it would’ve been a boring game to watch and not at all have the stream success it has enjoyed. And while the devs did realize that, it was long years before they identified which RNG effects are good or bad, enjoyable or frustrating. It was a journey to find where the graphs of “most exciting” and “least bad” met to produce memorable results.

The early days of Hearthstone carried terrible RNG decisions with them. Cards like Ragnaros were so game-swinging that they made for polarizing outcomes. Hit right, and you win the game. Flop, and you just invested significant resources in a card that did nothing.

Hearthstone’s was a journey to find where the graphs of “most exciting” and “least bad” met to produce memorable results.

An even worse offender perhaps was a card called Tinkmaster Overspark, which needed a correction to balance the game. The vanilla version would target a select minion and either transform it into a 1/1 or a 5/5 body — RNG effects way too far apart to make for a balanced outcome.

Eventually, Hearthstone card design arrived to a point where some RNG effects were actually not only exciting for their unpredictability, but also sometimes skill-testing. Primordial Glyph allowed players to discover new lines of play outside their deck to adapt to various board situations. The entire Discover mechanic, in fact, channeled a form of pseudo-randomness so well, that it became one of the most appreciated keywords in the game.

Furthermore, there’s Lyra the Sunshard which also generated random spell cards, but is well-designed enough to be part of a whole new archetype of Priest decks. And, for all the memes and salt surrounding it, Pavel’s incredible Babbling Book into Polymorph to kill Amnesiac’s Malygos remains etched as one of the most memorable, jaw-dropping, crowd-roaring single moments in Hearthstone.

And since we’ve already established that RNG has to exist and pros will always hate it, the excitement RNG generates is the metric that ultimately lasts.

Artifact’s boring RNG

When we get to Artifact, it’s only fair to assess it’s RNG the way we do for other card games. The resource RNG is gone and the draw RNG is significantly mitigated by the power of drawing two cards each turn. Two red points for Valve’s game right then and there. Who plays first — which is a huge deal in both Hearthstone and MTG — has an elegant solution through the Initiative mechanic. Another red point for Artifact.

But then, we run into problems, weaved into Artifact’s core mechanics and card design alike. Let’s start with the former.

There are a number of RNG effects associated with how each turn begins:

  1. Where do neutral creeps spawn
  2. Which direction does each unit attack
  3. What items do you get from the shop

Of the three, the item shop scenario is the least offensive, because the RNG there is both controlled and exciting. The unpredictability of Secret Shop items is a fun moment for both players and viewers, but it’s also not the single one case of RNG that defines the shopping phase. Players have some control over what Item Deck cards they get and the list of Consumables is small enough to produce somewhat consistent results. With that in mind, let’s turn to #1 and #2.

All RNG is to some extent bad, so might as well make it exciting.

The first occurrence dictates how each lane shapes up each turn and for novice players it’s very unintuitive. Creeps don’t spawn in any sort of pattern and you can easily get creeps spawns that benefit or wreck your game plans every single turn. Imagine you have a Sorla Khan pushing a lane against the clock, but every turn, a pesky 2/4 spawns to block her 12 damage. Your entire strategy in that lane is now falling apart.

The issue with creep spawns is that its effect is a product of double-fold RNG. First, the game has to decide which lanes the creeps spawn into. If that was the only checkmark, that wouldn’t have been that big of a problem. There are only six available options: 1&1, 2&2, 3&3, 1&2, 1&3 and 2&3. If the players knew that, for example, creeps always spawned in the leftmost free combat slot, they could play around the six possible outcomes and now this RNG becomes skill-testing, i.e. the good kind of RNG.

As it stands, after creeps “choose” their lane at random, they also choose their combat slot at random. The wider the board, the more possibilities there are and the more difficult it becomes to predict and play around outcomes. Instead of good and skill-testing RNG, we get one that is bad, especially if it results in the aforementioned Sorla Khan-esque scenarios.

The direction of attack can be another irritating outcome that once again has limited counter-play options. 50% of the time, units will attack straight ahead. 25% of the time they will attack right and 25% they will attack left. And having a 2:1:1 options ratio is not good odds to have, because you can’t reliably predict what will happen in the most important phase of Artifact: the combat. High percentage chances of diagonal attack mean that you’ll often see a situation where a single neutral creep is absorbing three attacks worth five, six, 10 times its health value. A lot of times, this can be an improbably outcome that decides a game all by itself.

18 damage on the board likely means a dead tower next turn… unless this exact thing happens.

Here, an argument can be made that the Babbling Book into Polymorph example from before was just as improbable (the math is not as important here, so let’s just say it is) as a creep spawning at the perfect position in the perfect lane to absorb all three of the important attacks. So why do we list the first as “good” and the second one as “bad”?

The inherent problem with these two Artifact RNG effects is not just that they are bad in terms of probability. Their sin is that they are boring and leave nothing to remember them by.

As we established with the Hearthstone example, Blizzard were right to identify that all RNG is to some extent bad, so they might as well make it exciting. Valve did include their own RNG effects, but made them bland, unappealing and dull. There will never be a viewer who gets amped up by how an attack arrow curves. Neutral creep spawns also just happen and carry no intrinsic “wow” effect. Good chance is that you’ll also miss both of those RNGs until the game zooms on the lane in question, at which point the only reaction they’ll invoke would be, “Oh, that’s how it is? OK, well, too bad.”

There will never be a viewer who gets amped up by how an attack arrow curves.

The gutting RNG combination of “bad and boring” is not only present in Artifact’s engine fundamentals, but some of its cards too. Bounty Hunter is by far and away the biggest offender with his Jinada ability.

50% of the time, Bounty Hunter is an 11-attack hero, the highest in the entire game. The other 50%, he’s a dull vanilla 7/7 unit. This is the very definition of a coin flip and coin flips are never exciting. Ogre Magi’s 1-in-4 Multicast and Tidehunter’s 50% Ravage aren’t much better either: they are a combination of bad RNG leading to swinging effects which, worst of all, is not even cool to watch.


These cards’ effects are neither good, nor exciting, taking the worst of both worlds.

What’s worse is that the few instances of YouTube-worthy RNG effects in Artifact are so polarizing in respect to player’s odds at winning that they are bad for the game. The old Cheating Death could produce memorable moments, but nobody would like their entire spell, turn and game plan foiled by a few lucky rolls.

The Golden Ticket item is a similar offender. Yes, it could pull off a historic Apotheosis Blade and send the crowd in ecstasy, but it’s also an effect that breaks the rules of the game because you don’t even have to pay for that Apotheosis Blade. A player just tripled one of their resources by doing nothing.

The easy fix

The good news for Artifact is that it’s closer than Hearthstone and Magic ever were to solving its RNG. After all, Magic won’t likely eliminate how mana lands works, and Hearthstone will always push for fun first and care about game’s integrity second.

Artifact is closer than Hearthstone and Magic ever were to solving its RNG.

Artifact, on the other hand, is in the golden middle. The game is still in its first month and its competitors have given it enough examples of possible RNG routes to go. Then, what Artifact needs to decide is whether to go the “boring but good” or “exciting but questionable” RNG route.

The first one might actually be the way to go for Valve’s game, as it will fit the world of Dota 2 more. Hearthstone is a gimmicky twist on World of Warcraft, so cards doing stupid shit makes sense. Artifact has a more serious twist on both the lore and the gameplay, and if it fixes its bad RNG it will become an even more solid game, engine-wise.

The solutions aren’t that complicated either. A tweak on the attack arrow percentages and/or changing how creeps spawn, for example, can make combat more predictable and liken Artifact to best parts of Magic: The Gathering strategy. Card changes like the one we saw with Cheating Death already can bring more positives to how turns play out — and there aren’t even that many cards that need addressing either.

Artifact is very close to being the perfect card game when it comes to its fundamentals, but solving its RNG problems is a step it must take. Randomness can be bad or it can be boring, but it must never be both at the same time.

Share on FacebookShare on TwitterCopy hyperlink