Talk:Arena

Table Question
On the table, what does the "2:", "3:", or "3+:" mean to the left of the rewards? (Unsigned comment from Anonymous user 72.129.254.134)
 * That number appears to refer to the number of different possible rewards. E.g., for Novice it says 2 and the reward possibilities are either (1) A pack and gold or (2) dust. I agree its not necessarily clear though. Also, please try to sign your comments using ~ so people know who left the comment! CrsBenjamin (talk) 18:05, 20 January 2014 (UTC)
 * Actually, the number describes the number of individual rewards granted by each Key. For example, a Journeyman Key rewards 2 rewards in total: a card pack, plus either a small amount of gold or dust, or an individual expert card. I've added a column and a little explanation to make this clearer. -- Taohinton (talk) 12:31, 24 January 2014 (UTC)

Cards
- I just got a golden Ysera (legendary) card after only 1 win in arena, so it is definitely possible, if someone would like to update that. In fact, in that same pack (from 1 win), I got 1 legendary, 1 epic, 1 rare, and 2 commons.
 * "Higher rarities/golden versions of cards appear to be more common at higher levels, and may be entirely unavailable at lower levels, or simply less common."
 * It sounds like you're talking about opening a card pack awarded from the Arena? The sentence is referring to individual cards granted as rewards; card packs all have equal chances of rarities/goldens regardless of where you get them. I've added the word 'individual' to that sentence to try to make that a little clearer. -- Taohinton (talk) 00:01, 10 June 2014 (UTC)

Number of wins table is incorrect
The number of wins table seems to have statistics based solely on mathematical modelling,with no real world data taken in. Players do win 3 matches on average, but that doesn't mean that 50% of players make it to 3 wins. For example, consider 5 players: A, B, C, D, and E. Player A being much stronger than the rest, he wins all his matches against the other players. So all other players go 0-3 and player A goes 12-0, and retires. In this scenario, only 20% of players make it to 3 wins. Of course this is an extreme scenario.

Another scenario would be the following. All players start with an 0-0 record. Then B wins 3 games against A. Now C wins 3 games against B. D wins 3 games against C. Etc, etc. In this scenario, all players except A make it to 3 wins.

---


 * Ended up spending a small amount of time reviewing the situation, and produced a program that would simulate 'fights' based on a gaussian distribution of skill level. Over 1 million simulations, the numbers come out very close to what is displayed via simple mathematical modeling.


 * Also, for the record, the law of large numbers comes into play here making your example false, though I wasn't certain the premise was until I ran these simulations.


 * Anyway, results:

Wins	-	Players 0	-	12992 1	-	18564 2	-	18765 3	-	15388 4	-	11466 5	-	8174 6	-	5464 7	-	3528 8	-	2236 9	-	1392 10	-	837 11	-	504 12	-	690


 * As for the graph: http://i1336.photobucket.com/albums/o646/caladin2/HearthstoneWinDistribution2_zpsff21nt28.jpg


 * When these are converted to a percentage and compared to the figures already in use, the differences are relatively minor & small enough to be rather irrelevant.


 * ValAichi (talk) 05:48, 24 October 2015 (UTC)


 * Thanks for the investigation! The fact that it's very close to the bare mathematical model is interesting. -- Taohinton (talk) 01:30, 6 November 2015 (UTC)

Capital 'A' or not?
Should we write 'Arena'? -- Karol007 (talk) 10:45, 15 August 2015 (UTC)
 * Thanks for prompting me to answer this. I've been correcting usage to 'Arena' for some time, but hadn't gotten round to really verifying it. The capital 'A' is used consistently in the game's string files, which means yes we should use it. The files have 39 instances of 'The Arena', 20 if which use a capital 'T', 19 of which don't, so I don't think we should bother with that; lower case 'T' is more natural and less fiddly anyway. Feel free to correct all non-capital 'Arena's, though, I know there are quite a few. -- Taohinton (talk) 20:26, 15 August 2015 (UTC)

Videos
I'm interested in opinions on whether we want a Videos section on this page, and if so what should be in it. At present we have 3 videos of a guy mostly talking about his Arena run, and a Trump video with him choosing a draft.

I don't feel like we really want just random videos of people playing Arena on this page. For a start, there are thousands of such videos - how do we decide which should be featured here? We also don't do this for other pages, and the vast majority of Arena play is no different from any other type of play, only the deck selection and reward processes. If there's a place for any videos on this page, they should be short illustrative ones, showing the basic features of the Arena so new readers can see how it looks.

I feel we should remove at least the first two videos, and possibly all of them. Thoughts? -- Taohinton (talk) 15:19, 2 September 2015 (UTC)
 * I'm OK with the cleanup. Other pages, e.g. Anub'Rekhan have a bunch of videos too. You can remove some of them, if you want. -- Karol007 (talk) 19:05, 2 September 2015 (UTC)

Without any further input, I've gone ahead and removed the video section. None of them were very good overviews/depictions of the mode, and since we're no longer in closed beta (as we were when the first videos were added) there's no longer any great need to have video of the arena full stop. We also have the gallery and various screenshots throughout the page to visually exhibit how it looks. -- Taohinton (talk) 21:41, 26 September 2015 (UTC)
 * Action

Rewards
Reward pool section is quite the mess. The ordering seems messed up now. 0 wins shows 25-40 gold which I believe is wrong.
 * I'm not sure quite what you're referring to. 25-40 gold for 0 wins has been stated for years, and roughly matches average data from third-party sites. Note that individual gold and dust rewards do shrink when the other rewards or the number of rewards increase, which can look a little odd if you read the table the wrong way. -- Taohinton (talk) 19:16, 10 October 2015 (UTC)

Noticed something about the table that caused me some confusion. I always assumed the random rewards are unique, as in you can only get one of any entry. The wiki doesn't state anywhere if this is the case or not so maybe my assumption is just wrong, but I just received one regular and one golden common from 8 wins. Just saying, either it should be stated that the same entry can be rewarded more than once (if this indeed is the case), or the table should have common and golden common listed as separate entries. Not sure if the same applies to cards of other rarities.

Matched twice against the same player
Arena: "It is possible for a player to be matched twice against the same player over the course of a single Arena run." Around Christmas I've played the same opponents in constructed twice in a row. That wasn't a duel, just a regular play mode match. One of them joked that we were the only ones playing Hearthstone atm ;-) -- Karol007 (talk) 21:02, 5 January 2016 (UTC)

"Upcoming Changes" section needs to be combined with rest of article
I'm not sure of how the integration should proceed. Should the whole "upcoming changes" section be renamed? Hoping for advice from the original author of the section. Link to section: Arena. Aegonostic (talk) 14:28, 2 March 2017 (UTC)
 * Looks like you posted this just as I was finishing off the new version. The changes have been made. -- Taohinton (talk) 15:41, 2 March 2017 (UTC)

100 in 10 challenge
This page briefly mentions the 100 in 10 challenge and TwoBiers as the first (and as far as I know only) winner. However, there is no mention whatsoever of the restrictions on class selection.

I am in doubt as to the best way to fix this. I can just add "(while following certains restrictions)" on this page, add a full section on the challenge and expand on the rules there, or create a new page for it. What is the best way to handle this? (Especially given that, after TwoBiers completed the challenge, global interest has waned completely)

-- BigHugger (talk) 13:27, 26 April 2017 (UTC)


 * Any of those approaches would be fine. New page would be fine, even if it's only linked from here, as long as it's got a few paragraphs of information it seems worthy to me. In a way, the "100 in 10" challenge forms a sort of "Player vs Community" counterpart to the more direct "Player vs Player" tournaments - its rules would seem right at home next to tournament "match formats" like Conquest and Last Hero Standing. Perhaps if we document more challenges like this we could even make a portal page similar to tournaments.


 * That's a lot though. Also makes sense to just summarize with something like "while choosing a new hero whenever possible" and add a reference and/or external link to the reddit/other post where the comprehensive rules were laid out.&#32;- jerodast (talk) 14:49, 26 April 2017 (UTC)

Changed reward structure: Bugged or misunderstood?
When I interpreted "In particular, reward bags will no longer contain single, non-Golden Common cards." (quote from https://us.battle.net/hearthstone/en/blog/20944006/hearthstone-update-august-8-8-8-2017) as "non-golden common cards are no longer offered as a reward for any key", I felt confident of my interpretation. Not anymore.

Yesterday I finished an Arena run at 2-3, and my rewards were a KotFT pack and a non-golden Bloodhoof Brave. I thought that was not expected to happen anymore. Does this mean my interpretation is incorrect, or is this simply a bug in the reward generation logic?

BigHugger (talk) 07:22, 14 August 2017 (UTC)


 * For reference: "Arena Rewards: Reduced variance of rewards in Arena. In particular, reward bags will no longer contain single, non-Golden Common cards." Maybe see if it happens a second time around in Arena. If it does, then it's probably a bug/misunderstood. Could be that the beginning of your Arena run was before the implementation of the feature. Or the feature has not been implemented yet (doubtful).


 * Also, as an aside, I really dislike how vague the description "Reduced variance of rewards in Arena" is. It could mean anything could have changed with the Arena rewards table. Aegonostic (talk) 17:26, 14 August 2017 (UTC)


 * Definitely not a run I started before the change (which was announced as coinciding with the KFT release). There were one and a half runs in between.


 * I do fully agree with all the rest of your reply. The "reduced variance" is just a mouthful of gibberish. And the only thing I thought I was able to translate into normal English ... well, I'll wait and see if the pattern repeats before jumping to conclusions.


 * (Sometimes I wish I could get access to the data collected by sites such as heartharea.com or hsdecktracker.net and do some number crunching on their data. That would give me a great opportunity to get a lot of data in this article up to date) -- BigHugger (talk) 17:59, 14 August 2017 (UTC)


 * Mystery solved. The patch notes were patchy (no pun intended). The non-golden commons HAVE been removed, but only for runs with three or more wins.
 * I have already updated the page with this info. -- BigHugger (talk) 22:48, 14 August 2017 (UTC)

How Arena actually works right now
With all the changes they publish about Arena, I'm not even sure if some of the past changes the Hearthstone team made to Arena are still being applied anymore. The latest news about the definition of what Arena is is written in this blog post: Blog post here (November 29, 2018). Really though, it was a somewhat confusing read. The main points on how Arena works can be best described by the blog post in May 2018: Developer Insights: Arena with Kris Zierhut (May 31, 2018).

The three Arena cards that appear in each selection screen during the draft come from buckets, which are card lists. Arena uses two bucket systems, one for non-Legendaries and one for Legendaries. Buckets in the middle of the spectrum of a bucket system contain cards from adjacent buckets (e.g. bucket #1.5 contains the weakest half of the cards from bucket #1 and the strongest half of the cards from bucket #2, etc.). This means that cards in the first and last buckets appear half as often as the buckets in the middle of the spectrum.
 * Not sure if they still publicly release the buckets, but last time in August they released the buckets list and also the appearance rates list.

The win-rate and pick-rate of cards are used to decide which cards go into which bucket. Cards with high win rates are in the top buckets (e.g. bucket #1, bucket #1.5, bucket #2, etc.). Cards with very high pick rates are promoted a bucket. Cards with low pick rate are demoted a bucket.

Arena cards have a "card weight", which means they have an increased or decreased chance to appear depending on the card's type (e.g. spell, weapon, Class card, etc.).
 * For example, a card with a 50% increased chance of appearance would have a weight of 1.5.
 * Common and Basic cards are weighted to appear more often than Rare cards, which are weighted to appear more often than Epic cards.

Arena buckets have a "bucket weight", which is the chance for each bucket to be picked within a draft. This is proportional to the total weight of all the cards in that bucket.
 * For example, a bucket with a total weight of 200 would have twice the chance of appearing than a bucket with a total weight of 100.

Then there are micro-adjustment weights, which are made using a holistic algorithm that evaluates the win rates of all cards that can be drafted in Arena. This process is applied a couple weeks after each expansion, after the meta is a bit clearer. The algorithm gives each card a micro-adjustment weighting value between 0.7 and 1.3, and this is multiplied by the card’s other weights. Then, the win rates across the nine classes are "balanced" by tuning these micro-adjustment weights. Ideally, the final win rates for each class come as close to 50% as possible.

Next, a machine learning model is used to answer some questions such as "the win probability given X cards drawn". This is probably a part of the micro-adjustment process, but not too sure. The blog post was pretty confusing. In summary, the process of tuning card weights should equilibrate the overall win-rate across all nine classes to be as close as possible to 50%.

The definition of what Arena is now can be pretty confusing, but I think I've summarized the main points. Some past changes to Arena stated in the past may no longer apply, so it may be difficult to root those out of the article. I think though that the card restrictions and the 50% reductions on Flamestrike and Abyssal Enforcer still apply. The validity of other past Arena changes that have been stated to occur will have to be checked somehow.

Ok, in summary, these three blog posts should overall describe what Arena is:
 * Developer Insights: Arena with Kris Zierhut (May 31, 2018)
 * The best and clearest blog post in understanding what Arena is so far.
 * Developer Insights: 12.0 Arena Update with Kris Zierhut (August 1, 2018)
 * Developer Insights: Arena Balance Through Science (November 29, 2018)

Aegonostic (talk) 20:41, 30 November 2018 (UTC)

Reward Table incorrect information
The arena reward table does not have up to date information and needs updating. I have collected 2.5k runs, and posted on reddit. com/r/ArenaHS/comments/a2q25u/updated_arena_rewards_table/ Also, the guaranteed pack is always from the most recent expansion, while the second pack at 12 wins will be a standard pack but not from the most recent expansion. It appears there is no second pack in reward possibilities below 12 wins (maybe removed with August 2017 update), or the probability is extremely low. Most importantly the possible values need updating, and random rewards at 8+ wins give a second box with a different range of possible values than the first box, the possible values of both I have detailed in my reddit post. There is also this post from over 2 years ago reddit. com/r/ArenaHS/comments/4m37oq/12_win_arena_rewards_droprates_updated_for_old/ that supports there being two distinct reward boxes with differing ranges of possible values. Gwasp (talk) 06:06, 6 December 2018 (UTC)


 * Alright I will take a look at your post, wow ok it seems really good! Thank you for doing all the hard work and data collecting on the rewards and numbers, I will try to add this to the wiki and reference the two posts! Thank you! It might take me a while though to assimilate the information. I feel like Blizzard should have released the rewards structure information a long time ago since it's similar to the loot box situation in China where they mandated that game companies revealed the chances to receive items from loot boxes or loot in general. Aegonostic (talk) 08:23, 6 December 2018 (UTC)


 * Table updated! Let me know if it works out, thank you again for your hard work and reference links! :D Aegonostic (talk) 10:59, 6 December 2018 (UTC)


 * Thanks for the quick update to the page, looks good. I agree Blizzard should have released the reward structure, but they responded to the China loot box law by selling in game currency with loot boxes being 'free' extras. I don't expect them to do anything unless players raise their voice. They did release the drafting buckets spreadsheet eventually. Think because of the table on this page people haven't really bothered to ask Blizzard for it, as it is just good enough. Gwasp (talk) 15:35, 6 December 2018 (UTC)

1st 10th 20th 30th picks are now implemented differently
Start from Patch 14.0.0.29933, the 1st, 10th, 20th, and 30th picks are no longer guaranteed to offer Rare, Epic or Legendary cards, so there might be decks with solely Free and Common cards.

reward for first arena match with class
I don't see this mentioned in the article, but did anyone else notice that? When I played my 1st match as mage (won) got a bunch of mage cards (forgot to write them down) and when I played my 1st match (loss) as Shaman, got a bunch of Shaman cards (eleven total: two flametonge totems, two totemic mights, two windspeakers, two fire elementals, 1 ancestral healing, 1 ancestral healing, 1 windfury)

Unfortunately I had most of them already I think from cardpacks...

If I'm correct in guessing that Arena is giving out predetermined basic class cards, it seems like it's advantageous to play Arena early on and get them, so that you won't get dupes later when you get card packs later since randomized card packs seem to avoid giving you unusable surplus (beyond 2) dupes.

Of course ironically it's bad to play Arena early if you lack the skill to get very far in it, so that's a bit of a tradeoff. Probably skill could be gained by playing the single-player campaigns like Kobols and Catacombs since you can experiment with a wider variety of cards than you start off with in some classes.

It might be first victory, I lost 1st match and got nothing with Priest but won 2nd match and got 2 x Mind Vision, 2 x Holy Nova, 2 x Shadow Word Death, 2 x Mind Control, 1 x Radiance, 1 x Radiance, 1 x Holy Smite, 1 x Holy Smite.

What's confusing is how sometimes you get a batched 2x a card, and other times consective 1x. Any idea why? Tycio (talk) 01:08, 28 March 2021 (UTC)

Visual Timeline
Here's the removed chart for the visual timeline. I personally don't think it adds anything to the page that the other chart doesn't, but it apparently has its fans. If you want to update it and add it back to the page, feel free. I'm not going to anymore. TheMurlocAggroB (talk) 22:45, 1 June 2021 (UTC)

====Visual Timeline====