Hearthpwn

Why are people humoring Hooghout on here? He's not here to present anything worth responding to. He's a troll that spouts utter nonsense while attempting to appeal to people who just want to be in an echo chamber and don't care about facts. The fact that he spends so much time writing out so much while not saying anything of value is the only thing to be impressed by. 

He's the poster child for the IYI's (intellectual yet idiots) of the world. https://medium.com/incerto/the-intellectual-yet-idiot-13211e2d0577#.itukmi4hm

Quote from GrandPatzer >>

This is nonsense, you can't imply hearthstone statistics to p-value testing. Or you can but the results don't give any insight.

Quote from GrandPatzer >>

Quote from Lysenko >>

Quote from GrandPatzer >>

This is nonsense, you can't imply hearthstone statistics to p-value testing. Or you can but the results don't give any insight.

 That's not the case.  Let's say pro X is playing deck like freeze mage.  Freeze mage has a 52.96% win rate across a broad population in Legend according to vicioussyndicate.com.  Your null hypothesis is that each game is nearly a coin flip.  You can easily calculate a p-value that tells you the likelihood that that null hypothesis is correct given that you have X number of pros with a 70% win rate with that deck.

Ok, ok, we all know that SOME skill is helpful but you want to test whether being legend rank 1000 is the same skill-wise as being legend rank 1.  So, you take the records of a cluster of players around 1000 (let's say they win 55% of their games with the deck) and figure out the likelihood that you have another cluster of players with the win rates you see among rank 100+ players (who win 70%.)

What you're hanging on to (the idea that above a certain level each game is a random process and skill doesn't matter) is the perfect null hypothesis against which you can use P-value testing to find the likelihood that skill didn't get those players to the top.

You can't calculate probabilities without having a certain value to begin with. Statistics from any site that provides data from games played in HS is not certain in any way. It is always influenced by the rng factors in game. Someone who is sitting at ~legend 3k could have a sick winstreak and end up in legend #1. The guy who is #1 legend could have better luck and que only into favored match-ups.

You can play with the numbers all you like, they still wont give you any insight. Get ur maths str8 boy.

What are you comparing my belief into? Please teach me.

 

Quote from GrandPatzer >>

Quote from MrKruzan >>

Quote from GrandPatzer >>

Quote from Lysenko >>

Quote from GrandPatzer >>

This is nonsense, you can't imply hearthstone statistics to p-value testing. Or you can but the results don't give any insight.

 That's not the case.  Let's say pro X is playing deck like freeze mage.  Freeze mage has a 52.96% win rate across a broad population in Legend according to vicioussyndicate.com.  Your null hypothesis is that each game is nearly a coin flip.  You can easily calculate a p-value that tells you the likelihood that that null hypothesis is correct given that you have X number of pros with a 70% win rate with that deck.

Ok, ok, we all know that SOME skill is helpful but you want to test whether being legend rank 1000 is the same skill-wise as being legend rank 1.  So, you take the records of a cluster of players around 1000 (let's say they win 55% of their games with the deck) and figure out the likelihood that you have another cluster of players with the win rates you see among rank 100+ players (who win 70%.)

What you're hanging on to (the idea that above a certain level each game is a random process and skill doesn't matter) is the perfect null hypothesis against which you can use P-value testing to find the likelihood that skill didn't get those players to the top.

You can't calculate probabilities without having a certain value to begin with. Statistics from any site that provides data from games played in HS is not certain in any way. It is always influenced by the rng factors in game. Someone who is sitting at ~legend 3k could have a sick winstreak and end up in legend #1. The guy who is #1 legend could have better luck and que only into favored match-ups.

You can play with the numbers all you like, they still wont give you any insight. Get ur maths str8 boy.

What are you comparing my belief into? Please teach me.

 

 This is just plain wrong, but fortunaly you made it easy for me to do the maths to prove you wrong.

First of all there is an entire field of statistics that is about calculating probabilities given certain information called Bayesian statistics. So more than that I won't get more  into that.

Secondly since you argue that all people at top 1000 has equal skill, and since picking decks is part of that skill we get that every player has an equal chance of winning a game apriori, that is 50%. Now lets asume that it would only take you 20 straight victories to go from 1000 to legend. The probability of that happening is (1/2)^20=9.5*10^(-7) which is almost litterally 1 in million. 
Now you say he queues only into favored matchups, however that doesn't even matter for the probabilities, since the chance of queueing into these better matchups given our assumptions exactly cancles out with increased winrate% since every person has chosen a deck that gives them the same average probability to win given all the matchups avilable.

Now I won't say that this kinda thing never happens. But really I think I made my point. 

I have never said that every player has equal chance of winning?? I just said that after certain level of experience, people know the game so well, all that matters is RNG. This does not mean 50-50% games. Get a grip.

To reject this null hypothesis, you'd have to calculate millions and millions of them since the issue is so wide. The numbers behind the calculations are also rigged since the game is rng-based and you cant base your calculations on constantly changing random numbers, therefor this whole process is nonsense and provides us with no information. Obviously you can use this test to get some answer but it's not reliable in any means. Ask your maths professor for further guidance, you're way off my friend. 

You could not be more wrong, randomized effects don't follow the concept of logic.

I DON'T BELIEVE IN MATH! CLIMATE CHANGE IS FAKE NEWS!

Quote from banstylejbo >>

 

In general, humans have a hard time understanding the effect randomness plays into their lives. When things go good we naturally attribute it to some kind of skill we must possess. When things go bad we look to blame the factors we can't control rather than ourselves. It's human nature. In an environment such as Hearthstone it leads to a culture of negativity when the easy way to make ourselves feel better is to assume we just always have bad beats instead of looking into why things didn't go our way. Sure, not every scenario is winnable, but many could be if we had done something differently.

 

Also please stop the 50/50 argument, all 2 player games over all players have an average 50/50 win rate, it has nothing to do with the skill of the game. The coinflip guy is correct in a certain respect, the top 1% probably are pretty close in skill. it's no surprise, but that has nothing to do with a "low skill cap" seeing as out of Xmillion players the top 1% or less is a significant number.  Yes, it  might be that the game as a competetive scene would be better if the skill cap was even higher, but that doesn't mean it's a "low" skill cap.

The best that can be said is that it may be too low a skill cap for a game with this many players.

If we put 100 pro players to play the same mulligan game (with no rgn cards) agains the same machine oppenent, we could see and evaluate all discrepancies and the most concurrent sequence choice. 

Doing that, we can demonstrate if there no difference  on top ceilling rank.

Then if we add random cards in deck, we could see the impact between skill and luck. 

Quote from Nicholasjh >>

Quote from banstylejbo >>

 

In general, humans have a hard time understanding the effect randomness plays into their lives. When things go good we naturally attribute it to some kind of skill we must possess. When things go bad we look to blame the factors we can't control rather than ourselves. It's human nature. In an environment such as Hearthstone it leads to a culture of negativity when the easy way to make ourselves feel better is to assume we just always have bad beats instead of looking into why things didn't go our way. Sure, not every scenario is winnable, but many could be if we had done something differently.

 

 Good point, and I think it illustrates part of what is happening. When you think... I made a good play and won, I lost and it was random, you have a confirmation bias, because of your thoughts your brain automatically looks for patterns based on those thoughts, so any loss that was randomly influenced becomes "hearthstone is random, that's why I lost". And any win based on your winning plays makes you think" it was actually a good play" (whether or not you actually played optimally) If you want to become a better player you have to look for how you could have played differently and won(including times when you have already won), eventually you will be able to make more optimal plays, but that never happens if going in what you are seeking is what are the random plays that I can excuse my loss for, and even what are the plays that won me the game? If you do that you are wasting your brain power.

 

Quote from CyberWarrior1337 >>

TL,DR: Skill matters, but only with a significant number of games and therefore not in tournaments and for the normal hearthstone player.

Quote from GrandPatzer >>

 Or the one with the more luck.