Hearthpwn

We understand completely what you're saying, but it's just not correct.

This post I made a few minutes ago points to a public data set that makes absolutely clear that across thousands of players the matchmaking is random with respect to what is in your deck.

In my post above, I offered you a source of a huge amount of publicly available data that says you're not right about this.  Have you looked at it yet?

- Rank (shold be around the same rank as you)
- Experience (number of wins in ranked)
- NEW: Deck type (dont know how it works, but if you switch decks you will see that also your opponents change, and will pilot a deck that has a high chance to defeat you)

They have stated that they try to match newer players against each other, but other than that the third on your list is certainly not the case.  How do I know?  Easy.  If you visit HSReplay's Meta Matchups tab, you can mouse over each archetype matchup in their entire database and see both popularity percentages and raw numbers of times those matchups have occurred in their entire data set.

I'll cut to the chase so you don't have to:  The percentage matchups for each archetype closely track the popularity of each opponent archetype.  Or, to put it another way, they are not manipulating your opponents based on what is in your deck.

It's unlikely there's a more comprehensive data set out there for looking at these numbers.  Fortunately, HSReplay.net has put it all out there for anyone to see who cares enough to look.

Regarding whether matchmaking is biased toward certain match-ups based on a player's selected archetype:

It dawned on me just down that hsreplay.net displays raw counts of recorded games for each archetype matchup in their database.  (Go to the Meta tab, click on Matchups, and mouse over each win percentage.)  Their total sample size is millions of games in the past 7 days, and some of the more common matchups have 60-70 thousand games recorded.

Just inspecting the numbers casually, there doesn't seem to be any indication that certain archetypes are particularly over- or under-represented in matches for other archetypes.  So, it doesn't appear that they're doing the kind of thing I was hypothesizing about.

If someone wants to collect those numbers and do a more rigorous analysis, I'd be interested to read it, but I didn't see much interesting to convince me that it would be worth doing so myself.

Quote from Xynot >>

 Currently, game makers deny implementing these patents, but their existence has fueled this debate.

Quote from iandakar >>

You know one thing that does bug me though, everyone bringing up info about it being rigged keeps proving that they are at the 'negative' side of it: the game purposely matching opponents against them.  Given this particular hypothesis, this means we have folks who are on the other end, getting great and easy matches against opponents.

• Ideally, ranked ladder should measure your skill at playing the game to a greater degree than your skill at picking a deck archetype.
• Matchmaking that tries to even out differences by class/archetype also makes the experience better for people without access to or awareness of community data like Vicious Syndicate or HSReplay.net.

I collected some of my own data back in December.  Unfortunately, my sample size was way too small.  It's really hard to grind enough games fast enough, because one needs to compare opponent distribution to a known, measured distribution across the whole community like Vicious Syndicate's, and those can change pretty fast.  So, I got tired before I had a large enough sample size (although I felt my results were interesting enough to warrant a closer look.)

Aside from that, I also posted about the comments of an engineer working on Hearthstone matchmaking from Blizzcon 2017 that seemed to indicate that they were using business intelligence data to tweak matchmaking.  Note that archetype classification and performance are well within what a game company's business intelligence group would be routinely collecting and feeding back into development.  Unfortunately, since the engineer just casually threw the term "meta" around as the subject of their data input, it's not entirely clear what precisely he meant by that.

Finally, some thoughts on reasons that such an approach would be good for meeting game design goals and improving player engagement.

I keep posting links back to some of this stuff because there's so much noise in these threads that the people who want a serious conversation and/or serious data collection and analysis end up talking past each other.

If the different methodology is to exclude mobile revenue, sure.

Regardless, Hearthstone is a tremendous moneymaker for Blizzard at any of these revenue levels.  The team is small, the game is technically simple, and the art costs are relatively quite low compared to any other Blizzard game.

However, I do know a lot of people who used to play on PC who have migrated to playing entirely on mobile because their mobile implementation is so good and the short game length works well with how people play mobile games.

Quote from Marakius >>

I don't exactly know how the vS Deck Power is calculated, as Razakus Priest haven't dropped below 52% winrate since the start of January according to HSreplay, but somehow it doesn't make the cut.

People just complain that it's "autopilot" because it can be very frustrating to lose to a well-played combo deck.  They don't get to see the hours of awful games that led up to that point for their priest opponent. :)

This is the deck I have been running with.  I have not been feeling weak against aggro in general, though there are some games that I just straight-up lose.  Trick is to mulligan heavily for early AOE and card draw vs. paladin, rogue, druid (just in case), and, to a lesser extent, warlock.  Key cards for this are Spirit Lash, Duskbreaker, Dragonfire Potion, Kazakus (which can give you both AOE and card draw), Netherspite Historian (to fish for Duskbreaker), Primordial Drake, and Psychic Scream.

Sometimes aggro decks just walk all over me, but just as often, it feels like I can come up with an answer for everything they do and they just get nowhere while I chip away at them.

Edit:  Pint-Size Potion and Shadow Word: Horror are a great combo, but I find it's easier just to run some heavier AOE that doesn't require having both cards, particularly since Pint-Size Potion doesn't seem that handy on its own.  But, I haven't spent a lot of time working out how best to use them.

Quote from robert_fanr >>

The "skill" required by this deck is to draw RAZA and ANDUIN on time .

Quote from scorpyon >>

Raza Priest isn't very strong any more.

Looking at the stats on HSReplay.net, the difference between average, population-wide murloc and aggro paladin win rates is less than half a percent.  To put this in perspective, you would need to play at least 40,000 games to be able to see a clear statistical difference between these win rates.

I think definitelynotsalty's "rank 10 or better" qualification was to ensure that your opponents are playing at least three class cards in their decks, so you're very unlikely to see neutral cards in the pick.

Fair enough, I wasn't arguing that it shouldn't be nerfed, just that doing so would require a different argument than to appeal to Blizzard's dislike of certain kinds of combo decks.

However, it's not a great candidate to nerf.  Skilled players can do quite well with it, but the population-wide win rate for it is hanging right around 50% (49.8% from Vicious Syndicate's numbers, 51.5% from HSReplay's statistics.)  If anything, a deck that's sensitive to skill like that is the kind of deck I'd imagine the designers would like to encourage.