Developer Insights: Boomsday Arena Update with Kris Zierhut
The Arena is receiving updates with the next Hearthstone patch. Read on for more information!
Quote from BlizzardWelcome back for another inside look at updates coming to the Arena! Your feedback helps shape our approach to this game mode, so we’ve been following the Arena-related conversations closely in the community.
Card Variety
Just like how we want to keep win rates per class within a percentage range, we also want to maintain variety in Arena card selection. Each class had a group of cards that appeared almost twice on average per Arena run, and we focused on reducing this appearance rate in the coming 12.0 update.
We also looked at the appearance rate of class cards, spells, and weapons. We think it’s valuable for these cards to show up often for several reasons: they make playing each Class feel unique, they change the pace of a match, they help create more skill-testing situations, and finally, many of these cards offer opportunities for a player to catch up from behind, which results in more interesting games.
That, the appearance bonus for these cards is currently too high, causing them to appear so frequently that they can make Arena matches feel repetitive. To solve that, we’re reducing the appearance rates of these cards as follows in the upcoming update:
- Class Cards: +100% -> +50%
- Spells and Weapons: +75% -> +50%
Note that these bonuses multiply, which means that spells and weapons will have a net 125% increase in appearance rate, instead of their old 250% net increase.
Rarity
In our last blog, we also took a close look at card appearance by rarity. Each class had a group of Rare-quality cards that showed up more often than desired. Reducing the appearance bonuses for class, spell, and weapon cards helped that problem somewhat, but it didn’t fully resolve it.
Additionally, when Rare cards are offered side by side with Basic and Common cards, they get picked more often, which can result in them being selected and played more frequently than their appearance rate would suggest.
For those reasons, we’re lowering the appearance rate of Rare and Epic cards. Picks 1, 10, 20, and 30 always offer Rare or better cards and will continue to do so. For all the other picks, Rare cards will appear about 11% of the time, and Epic cards about 5% of the time.
Appearance Rates
As part of our effort to better explain how Arena works, we’d also like to share with you the current appearance rate for all cards. For each class, we ran a simulation of 100,000 Arena drafts. The spreadsheet linked below reveals how often each card showed up for each class. We hope this gives some insight into the chance of seeing various cards. These rates will change over time, but this one-time look should give you a clue as to how the system works.
What’s in the Bucket?
Since our last blog, players have also been requesting a glimpse into the contents of each bucket in the Arena draft system. Today, we’ll give you a one-time look at that as well. Remember that card promotion or demotion is almost entirely due to a very high or very low pick rate (how many times picked divided by how many times offered). One notable exception is Mind Control Tech. That card is being picked a reasonable rate for the typical card, but still often felt oppressive, so in the 12.0 update, it will be promoted anyway.
Thanks for tuning in, and we hope this look behind the curtain was illuminating. The Arena system will continue to evolve, so keep your feedback coming. We’ll see you in the Tavern!
A 1 hour video of Kripp explaining the spreadsheet is coming PogChamp
i want a 2 hours version of this with Kripp face all'over the screen
This system had a few strange quirks. In one of my recent Paladin drafts I started out with two identical sets to choose from and than very similar iterations happened two or three more times.
As for the spreadsheets... Well, I will wait for Kripp to do analysis and try to learn something useful ;)
Except public stats from Blizzard routinely place him in the top 100 Arena players in the world so... maybe your analysis is total bullshit? Just a thought.
Kripp lets Heartharena sway him a bit too much, but 99% of the time he makes absolutely correct picks, and he often realizes when HA is dead wrong.
You said "Not sure how useful anything he says is going to be" - I'll take the advice of a top 100 Arena player any day of the week. Just because HearthArena offers similar advice, that doesn't make his advice bad. I've watched him draft dozens of decks, and when he disagrees with HearthArena he often gives good reasons why.
"Just like how we want to keep win rates per class within a percentage range, we also want to maintain variety in Arena card selection."
This is the proof directly by Blizzard that Hearthstone is rigged
Rigged tombe balanced? What a crime!
I was joking lol
So what are you waiting for starting a campaign promoting to drain the swamp and selling Make Hearthstone great again hats?
Looking at the bucket and I can't believe Omega Agent is in bucket 5 and 5.5 of Warlock. Is it just me overrating it?
T10, which makes it fairly slow.
Kripp would love that spreadsheet!
That spreadsheet is Kripparrian's wet dream
Ok, so the devs increase and decrease the % of appearance of certain cards in the Arena. Cool. They say it themselves. Very cool. So, why do people refuse to believe, that such hidden system exists in ranked play mode? That some cards are more likely to appear in the mulligan, to be the first and last card in the deck.
?
There are many sites that collect Hearthstone data from thousands of players over thousands of games. If there were those kind of systems in ranked the community would hear about it. Blizzard knows this and has literally no reason to secretly modify statistics in ranked. Why the fuck would blizzard ever implement that? You have no evidence whatsoever that Blizzard modifies such things.
If you're the one that claims the system is rigged, the burden of proof rests on your shoulders. Find information, gather statistical data, run tests by yourself with help of people, try to get conclusive results.
It'd be interesting to figure out if that kinda thing actually happens in the game. It's a lot of hard work to do, and it'd certainly give us better insight on how the game actually works, but so far nobody cares enough to test it out or give solid arguments other than that one article that talks about some algorithm some company made to affect match making.
Until then, it just seems like confirmation bias and people not understanding how statistics work overlap.
You're just out trolling everywhere today aren't you?
To adjust winrates of certain decks and archetypes. To "nerf" certain broken cards without changing their stats, text or manacost. For example, there is a card Master Oakheart, that allows you to develop 40+ mana on board in one turn. It is broken. If they nerf it directly, they will have to refund it's cost to everybody, who crafted it and still they will face a ton of hate from the community, cause the taunt druid deck would be much less powerful and competitive and in order to play that deck people has crafted not only Master Oakheart, but also Hadronox, Dragonhatcher and the girl, who doubles the battlecry of minions.
There are meta defining (or deck defining) cards every expansion, like Master Oakheart, Malygos, Bloodreaver Gul'dan, Twig of the World Tree, Carnivorous Cube, Prince Keleseth, Patches the Pirate and others. They are not that many in numbers so it is quite easy to adjust their appearance percentages manually.
Why should they do that? Because they can. It's a very effective way to control the meta without frequent nerf of cards.
I mean, I'm happy they share those internal spreadsheets but they at the very least could've cleaned up this mess a bit. It gives some insights, sure, but good lord is that a hard read.