Rigged so the ladder LOOKS in your deck to see if you have X cards or abilities and then doesnt match you against a deck you've sideboarded for? I think although you may have statistical backing, it's just bad luck. I really dont think blizzard does ANY looking in your deck to decide matchups at all, it's simply on rating and who's around right now.
'I am a Statistician by trade, with over 15 years of experience crunching p-values and t-tests. So I decided to run a statistics experiment to test my theory.'
I am a Doctor by trade, with over 15 years of experience curing colds and using band aids. So I decided to run a Medicine Experiment to test my theory because yes, I definitely am very much a real doctor
Screenshots? Deck tracker? Sampling error? Z scores, null hypotheses, alpha level, significance test output, etc? 15 years experience lol
I believe him because he simply has no reason or bias to try to slender Blizzard. If anything, it looks like he truly loves hearthstone and only wishes for it to be a fair game. I personally dislike it so much, how judgmental people are in these forums. The sample size and the difference in percentage is DEFINITELY something, at the very least, not normal, and it is simply idiotic, to dismiss it just because of unfounded suspicion.
Thanks, OP, for your hard work, and if I get to rank 5 this season I'll 100% check this out.........
You had me at Slender. C'mon--- OP puts "PROOF" in caps and just vomits out personal experience and expects to be taken seriously. Glad there are "Professional statisticians (by trade) in this world... can you imagine if there was not!!??
Do you honestly think it’s that hard to create code that recognizes tech cards and mana cost? They already tamper with arena draw and matchmaking, yet the standard ladder is such a stretch.
Are you serious bro? You are equating arena card appearance probability, a thing that can be implemented in the background engine with chances of encountering certain opponents? And you are telling me arena matchmaking is rigged too? How so? And if it were rigged, how come streamsnipers are still so successful at facing their targets? It's simple, if you queue at the same time as your opponent, with a comparable number of wins and losses, you get to play against that opponent. Done. No rigging needed, no illuminati to be found.
And to answer: No, it is not hard to build a piece of code that recognizes mana cost and tech cards. But then you'd need a piece of code that equates these tech choices with winrates, and you'd need another piece of code to modulate encounter rates. And for that you'd need an engine where you need to specify what your desired outcome is. Do you want to see more or less tech cards, do you want to see more or fewer druids. You'd have to have some people constantly changing settings for each card that is "techy" or "in too many decks" or whatever. And seeing the rate at which blizz nerfs cards, this is simply on a scale that is impossible to maintain.
Also: the @op's post indicated that he faced 60% of one deck, then changed a card and faced 60% of the other deck. Now this is just terrible news, since now we have confirmed: diddly squat. Why would blizzard make an algorithm that alters the game to such a great degree. Most decks have an average winrate of 55%. By changing the decks you face by 5 or 10%, you could lower that to close to 50% if you keep facing unfavourable matchups. The numbers he has provided are just ridiculously high. So either blizz is stupendously stupid about this, or he pulled those numbers out of his nose. In addition, there are many confounding factors like how long did he need to record this set, what rank were his opponents, at what time of day did he play. All of this can influence what enemy you're facing.
But we all know blizzard, they're always like: 1. Screw the players by making unnecessary secret changes 2. ? ? ? 3. Profit
Despite I'd also like to see some sort of proof for the data provided I'm inclined to believe the OP and the reason for that is my own little survey I did at the time when MSoG just came out. If you remember, people were pretty hyped about Kazakus and were trying out all three versions of Reno Jackson decks (priest, mage and warlock). After a couple of months it became obvious that warlock one is the most successful, but before people were actually trying to make mage and priest work too. At that time I made a meme combo deck which could only win against reno priests. The wincon was to wait till priest plays Raza the Chained, copy it with Faceless Manipulator and bounce it with Youthful Brewmaster. Then I played Raza myself and then, finally, OTKed priest with Coldarra Drake. My plan was to concede each time I face not a reno priest. Before playing that deck I encountered them pretty often, but when I did start playing my meme deck, everything went to utter shit - I couldn't find a single fucking priest in the entire universe, and if I did find a priest, it was not a reno one. I conceded 312 times in tries to find my desired matchup. In 323 games I found only 11 priests, and only two of them were reno priests - and none of those two played Raza in their decks. Thus I wasn't able to make my meme deck work a single fucking time, and after that I never tried to do it again.
I understand it isn't very conviencing or proving anything but the statistics were atrocious without any sane reason. After that case I was pretty sure there's some kind of rig in ladder.
Despite I'd also like to see some sort of proof for the data provided I'm inclined to believe the OP and the reason for that is my own little survey I did at the time when MSoG just came out. If you remember, people were pretty hyped about Kazakus and were trying out all three versions of Reno Jackson decks (priest, mage and warlock). After a couple of months it became obvious that warlock one is the most successful, but before people were actually trying to make mage and priest work too. At that time I made a meme combo deck which could only win against reno priests. The wincon was to wait till priest plays Raza the Chained, copy it with Faceless Manipulator and bounce it with Youthful Brewmaster. Then I played Raza myself and then, finally, OTKed priest with Coldarra Drake. My plan was to concede each time I face not a reno priest. Before playing that deck I encountered them pretty often, but when I did start playing my meme deck, everything went to utter shit - I couldn't find a single fucking priest in the entire universe, and if I did find a priest, it was not a reno one. I conceded 312 times in tries to find my desired matchup. In 312 games I found only 11 priests, and only two of them were reno priests - and none of those two played Raza in their decks. Thus I wasn't able to make my meme deck work a single fucking time, and after that I never tried to do it again.
I understand it isn't very conviencing of proving anything but the statistics were atrocious without any sane reason. After that case I was pretty sure there's some kind of rig in ladder.
I do not doubt that this is what it felt like to you at the time. But just let me ask you this to see if I understand you correctly: Are you saying, that the algorithm blizzard has is so great, that based on the deck you built, a deck noone else was playing at the time, the algorithm was able to anticipate the combo you were going to play (namely steal your opponents Raza) and decided with no prior knowledge that your deck was too strong against that particular deck therefore lowering your chances of encountering that particular deck?
How about people here explain why does blizzard have a patent for matchmaking to encourage people to puchase packs, where is the proof that they are not using it? i have not seen the proof, blizzard simply saying they are not using it is not enough prove, its not illegal for them to lie about it and anyone with a half working brain can understand that they wouldn't admit to it, i want real proof.
As it happens, I am a patent attorney. I'm frankly surprised Blizzard got that patent through. Alice vs. CLS Bank and its progeny have largely killed off this kind of patent. In any case, as the claims are written, I don't think they could implement this in Hearthstone. They could do something similar, but the existence of this patent alone doesn't mean much. Companies routinely pursue patents that never see commercial embodiment, especially big companies.
I think it's totally reasonable to ask questions like this, do some testing, and share the results with your views.
I've posted before on this topic - my view is that it seems unlikely. Primarily because if proven it's damaging to Blizzard/Team 5 to be seen to artificially working against players in this way. And even if they were, they are basically hampering you whilst helping your opponent. Why?
Blizzard wants tech cards to exist, and ideally for you to craft that epic giest, crab, rat, cleaner etc or rare tech card. Tech cards are a design mechanism to help, in some degree, to stop specific decks getting too out of control. It's works in their favor on two fronts but people are suggesting they are working to undermine it. It's not in their interest to do that.
Large statistic collection sites like hsreplay - or someone with access to their data could choose to look for statistically significant patterns with 100,000s/millions of games. Such an algorithm would be detectable with such a large sample size. It's not like Blizzard could get away with this, isn't in their interest anyway, and would be damaging to them upon discovery.
And I wouldn't call this "proof" of rigged matchmaking. I'd call it "evidence" of a functional matchmaking algorithm. Matchmaking, by design, should make it harder and harder to keep a streak going. The point is to level you out at an overall 50/50 win rate. It is likely that Blizzard's matchmaking algorithm is designed match players running a deck with a high win rate against decks that are hard for that to beat to push that win rate to 50/50. Thus, when you add a third silence, they see that the deck composition crushes Warlocks and they avoid pairing you with them. That's not "rigging" the ladder. I'd love to see you run the same experience at, say, rank 25, and see if you keep getting paired up with decks that are easy for your deck to beat to push you UP to 50/50.
Do you honestly think it’s that hard to create code that recognizes tech cards and mana cost? They already tamper with arena draw and matchmaking, yet the standard ladder is such a stretch.
No, I don't think it, I know it, not all code is equal no matter how similar you think they should be. Adjusting arena draw is likely modifying a field in a database and I assume when you talk about tampering with matchmaking you're referring to MMR. Assessing and matching MMR is based on a fairly simple formula so in very basic terms not that difficult to code, to introduce some kind of deck assessment factor into this matching adds hundreds of different criteria into the assessment, so for me yes it is actually a stretch that they'd waste time and resource on maintaining this.
Rollback Post to RevisionRollBack
Arguing on the internet is like playing chess with a pigeon. You may be good at chess, but the pigeon is just going to knock all the pieces down, take a shit on the table, and strut around like its victorious.
I don't think they could implement this in Hearthstone.
1. Player X spends $$$ to create an expensive meta deck (cubelock for example)
2. Player X pushes "Play" button and the matchmaking system searches for an opponent
3. The matchmaking system protects player X by not matching him against an opponent who teched with specific cards Z against player X deck
4. Player X is stimulated to spend more $$$ on new meta decks in the future by not having a bitter experience of seeing cards Z while playing the cubelock
How is this not worth it to be implemented when their whole patent is EXACTLY for this?
I know this really is an irrelevant topic......but I still think we have a right to know. I am a Statistician by trade, with over 15 years of experience crunching p-values and t-tests. So I decided to run a statistics experiment to test my theory. I would encourage anyone to repeat my experiment and see if you get similar result.
Procedure: I climbed to Rank 5 using a Pally Murloc Deck. I rapidly ascended to this rank, and then, of course, ran into the Void Lord and Cubelocks. Earlier than rank 5, I got a fairly even mix of opponents, but because the Cubelock is dominating right now, I expected to see more of them at the higher ranks. So......I changed one card in my deck. ONE CARD. I swapped out a knife juggler for a Defias Cleaner to get an extra silence tech. I was already running 2 spellbreakers, so this card had no purpose other than to silence a voidlord.
I immediately noticed the ladder was not facing me against cubelocks anymore. Since I couldn't lose stars at that rank, I started a game, and if it wasn't a warlock, I just conceded. It literally took me 17 games to even face a warlock.......and the deck was a ZOO build....NOT Cubelock. It took me 26 concessions to find a cubelock to go against.....because I changed one card to counter that deck,
I repeated that experiment every 3 hours.....just to allow some new players to come on and off.
I used 26 matches as my standard candle. (No pun unintended)
Trial 1 : 3.8% chance of facing cubelock
Trial 2: 4.0% chance of facing cubelock
Trial 3: 4.2% chance of facing cubelock
Average: With deck holding 3 silences, instead of 2.....the chance of facing cubelock was 4.0%
OPPOSITION EXPERIMENT:
I took the Defias Cleaner out, and put my knife juggler back in.
Again standard candle of 26 games at Rank 5:
Trial 1: 73.2 % chance of facing cubelock
Trial 2: 57.6% chance of facing cubelock
Trial 3: 69.2% chance of facing cubelock
Average 66.6% chance of facing cubelock.
Remember, this is a sampling of over 150 games, 75 WITH a 3rd Silence (No Cubelocks), and 75 with only 2 Silences (Many Cubelocks)
By changing ONE card in my deck, the likelihood that I would face the deck it was DESIGNED to beat went down by a factor of 4.5. When I removed that one card, the likelihood I would face that same deck went UP by a factor of 4.
As a further test, I took out a card and added a Ravenholdt Assassin, which has the same stats as a Defias Cleaner, but no silence. I saw ZERO change in the amount of cubelocks I was facing with the card, but did only run one trial.
So right now, Pally, Priest, and Warlock, and hunter are the classes to beat. You would think you would see them on average every four games, with a couple randoms thrown in.
But based on switching out ONE card.....and running hundreds of trials I found my chance of facing a cubelock DROPPED 62% when I put in a tech card to counter it. One Single Card.....the Defias Cleaner.
I implore someone to recreate my experiment. If Blizzard is fudging the latter, we need to know about it.
"Paranoid Player"
What method did you use to account for the fact that the amount of conceding you did would have crashed your MMR thus meaning that groups of people you are being matched against would have varied drastically between tests?
Rollback Post to RevisionRollBack
Arguing on the internet is like playing chess with a pigeon. You may be good at chess, but the pigeon is just going to knock all the pieces down, take a shit on the table, and strut around like its victorious.
I'd love some proof to, but just to make a point, let's assume you're right and your idea holds weight.
I have the following questions;
What is your hypothesis? Would it be correct to say that it is: "Hearthstone has a matchmaking system that tries to lower the chance that decks including specific tech-cards matches up against the decks that they are intended to counter."?
If so, what would you say motivates them to arrange such a system? I argue that it would just lower the incentive to craft tech cards if it's not working out. You may say it's because they want the expensive tier 1 decks to be mandatory to craft, but I'd maintain that people would still craft these decks as long as they're posted on Hearthpwn, reddit or twitter or whatever. Wouldn't blizzard make more money on tech cards actually working? That way the meta changes more and people will need to try out more decks. Blizzard has nothing to win on a stale meta as people would only build one or two decks.
My second point is the risk involved from such a move from Blizzard. It's pretty easy to test using a quantitative method as you did. Requires fairly low effort, especially if many are doing it. Even more considering various meta reports and third party stat trackers. If Blizzard would be found out that they're rigging the ladder, that would be HUGE and may just destroy a huge part of the community, especially the competitive scene (which I would say is what keep Hearthstone alive). It seems like a huge risk for a low benefit and I see this being sort of a defeater for any conspiracy theory. It just makes no sense, at least for me. But someone can provide more proof and data about this, on a much larger scale as well, then I'd definitely be willing to trust that data and wonder if there's something weird going on from Blizzards side.
Also, data like that wouldn't necessarily conclude that Blizzard made this intentionally, we'd need more proof for that. It might as well originate from an unintentional glitch in the game, or whatever.
i can say what motivates Blizzard to arrange A System like that. If meta is stable and you can climb ladder with any cards or any decks you dont need any New cards to change meta anymore. Because it is stable and every deck is equal. Also you will know that the first deck you have created can climb ladder even after 2 years and you motivate yourself like " ı really dont need New cards to climb so i can spend my money for more challenging stuff ". After that only fun players spend money for try New cards. Competitive players will keep climbing ladder with older decks. Guess what? Blizzard's main money source is competitive players, not fun players. So it seems fair for A company which gains money from these.
Yes, but I argue that rigging it like this would make it more stable than if people needed to change techs and decks to react to tough match ups instead of the game actively trying to nullify tech cards. Nullifying tech cards means that strong decks change less, because they are less challenged.
I made similar experiment, but with two decks - mill kingsbane rogue, and exodia questless mage. Playing mage got me mostly mill decks and the rogue mostly stood versus aggro ones. Never otherwise, but at the time I thought it's just long term big ass coincidence XD lasting two months XD I didn't give a shit, coz I was at 52% winrate
I don't think they could implement this in Hearthstone.
1. Player X spends $$$ to create an expensive meta deck (cubelock for example)
2. Player X pushes "Play" button and the matchmaking system searches for an opponent
3. The matchmaking system protects player X by not matching him against an opponent who teched with specific cards Z against player X deck
4. Player X is stimulated to spend more $$$ on new meta decks in the future by not having a bitter experience of seeing cards Z while playing the cubelock
How is this not worth it to be implemented when their whole patent is EXACTLY for this?
So how much do I have to spend to get this protection as I've spent over a £1,000 on this game and have seen no indication of this. In fact it's the opposite, the number of times my opponent has had the silence, eater of secrets, flare, geist they've needed is far more memorable than the number of times I've not faced a deck I was trying to tech against.
Rollback Post to RevisionRollBack
Arguing on the internet is like playing chess with a pigeon. You may be good at chess, but the pigeon is just going to knock all the pieces down, take a shit on the table, and strut around like its victorious.
Do you honestly think it’s that hard to create code that recognizes tech cards and mana cost? They already tamper with arena draw and matchmaking, yet the standard ladder is such a stretch.
No, I don't think it, I know it, not all code is equal no matter how similar you think they should be. Adjusting arena draw is likely modifying a field in a database and I assume when you talk about tampering with matchmaking you're referring to MMR. Assessing and matching MMR is based on a fairly simple formula so in very basic terms not that difficult to code, to introduce some kind of deck assessment factor into this matching adds hundreds of different criteria into the assessment, so for me yes it is actually a stretch that they'd waste time and resource on maintaining this.
If you think it would be wasted time you obviously don’t know much about business. Do you not grasp the importance of a fair and even matchup? Do you think people want it skewed one way or the other? No. Blizzard is Vegas...they do better with the win rate staying around 50%. Not to mention it’s the easiest way to balance the game without nerfing cards into oblivion. I don’t think it’s as rigged as if x meets this condition, do y. But the probability of x happening if y is present does increase or decrease. I view it like a video game slider. If this kind of matchmaking didn’t exist, new players would get obliterated
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
Rigged so the ladder LOOKS in your deck to see if you have X cards or abilities and then doesnt match you against a deck you've sideboarded for? I think although you may have statistical backing, it's just bad luck. I really dont think blizzard does ANY looking in your deck to decide matchups at all, it's simply on rating and who's around right now.
All this paranoia.. jeez. :)
I enjoy.
1. Screw the players by making unnecessary secret changes
2. ? ? ?
3. Profit
Despite I'd also like to see some sort of proof for the data provided I'm inclined to believe the OP and the reason for that is my own little survey I did at the time when MSoG just came out. If you remember, people were pretty hyped about Kazakus and were trying out all three versions of Reno Jackson decks (priest, mage and warlock). After a couple of months it became obvious that warlock one is the most successful, but before people were actually trying to make mage and priest work too. At that time I made a meme combo deck which could only win against reno priests. The wincon was to wait till priest plays Raza the Chained, copy it with Faceless Manipulator and bounce it with Youthful Brewmaster. Then I played Raza myself and then, finally, OTKed priest with Coldarra Drake. My plan was to concede each time I face not a reno priest. Before playing that deck I encountered them pretty often, but when I did start playing my meme deck, everything went to utter shit - I couldn't find a single fucking priest in the entire universe, and if I did find a priest, it was not a reno one. I conceded 312 times in tries to find my desired matchup. In 323 games I found only 11 priests, and only two of them were reno priests - and none of those two played Raza in their decks. Thus I wasn't able to make my meme deck work a single fucking time, and after that I never tried to do it again.
I understand it isn't very conviencing or proving anything but the statistics were atrocious without any sane reason. After that case I was pretty sure there's some kind of rig in ladder.
"The Naaru have not forgotten us!"
Added third silence .. 4 matches - 3x cubelock in wild rank 10
Are you saying, that the algorithm blizzard has is so great, that based on the deck you built, a deck noone else was playing at the time, the algorithm was able to anticipate the combo you were going to play (namely steal your opponents Raza) and decided with no prior knowledge that your deck was too strong against that particular deck therefore lowering your chances of encountering that particular deck?
Are people still responding to this troll post? LOL
μολὼν λαβέ
I think it's totally reasonable to ask questions like this, do some testing, and share the results with your views.
I've posted before on this topic - my view is that it seems unlikely. Primarily because if proven it's damaging to Blizzard/Team 5 to be seen to artificially working against players in this way. And even if they were, they are basically hampering you whilst helping your opponent. Why?
Blizzard wants tech cards to exist, and ideally for you to craft that epic giest, crab, rat, cleaner etc or rare tech card. Tech cards are a design mechanism to help, in some degree, to stop specific decks getting too out of control. It's works in their favor on two fronts but people are suggesting they are working to undermine it. It's not in their interest to do that.
Large statistic collection sites like hsreplay - or someone with access to their data could choose to look for statistically significant patterns with 100,000s/millions of games. Such an algorithm would be detectable with such a large sample size. It's not like Blizzard could get away with this, isn't in their interest anyway, and would be damaging to them upon discovery.
.
And I wouldn't call this "proof" of rigged matchmaking. I'd call it "evidence" of a functional matchmaking algorithm. Matchmaking, by design, should make it harder and harder to keep a streak going. The point is to level you out at an overall 50/50 win rate. It is likely that Blizzard's matchmaking algorithm is designed match players running a deck with a high win rate against decks that are hard for that to beat to push that win rate to 50/50. Thus, when you add a third silence, they see that the deck composition crushes Warlocks and they avoid pairing you with them. That's not "rigging" the ladder. I'd love to see you run the same experience at, say, rank 25, and see if you keep getting paired up with decks that are easy for your deck to beat to push you UP to 50/50.
μολὼν λαβέ
Arguing on the internet is like playing chess with a pigeon. You may be good at chess, but the pigeon is just going to knock all the pieces down, take a shit on the table, and strut around like its victorious.
If it is a proof of sth is a proof that warlock is a high win rate deck and it would not be matched with a deck with 16 straights ''concede''
Arguing on the internet is like playing chess with a pigeon. You may be good at chess, but the pigeon is just going to knock all the pieces down, take a shit on the table, and strut around like its victorious.
I made similar experiment, but with two decks - mill kingsbane rogue, and exodia questless mage. Playing mage got me mostly mill decks and the rogue mostly stood versus aggro ones. Never otherwise, but at the time I thought it's just long term big ass coincidence XD lasting two months XD I didn't give a shit, coz I was at 52% winrate
i think punintended.
Rejoice, for even in death, you have become children of Thanos.
Arguing on the internet is like playing chess with a pigeon. You may be good at chess, but the pigeon is just going to knock all the pieces down, take a shit on the table, and strut around like its victorious.