There is a big problem with deck building on this site. The most obvious is win percentages over small samples. For example, if I make 100 of my own decks in a year I am bound to soon or later get on a win streak with one of them. I then post it saying it has a 60 or 70% win rate and people jump on board. Let's put that to one side. There is another problem.
A significant issue is the guides indicating a matchup is favoured when it really is not. A matchup can be favoured when climbing but not at say high legend. Imagine if, for example, a particular deck was very hard to pilot. It may give the impression that when climbing your deck is favoured vs this hard to pilot deck. However, if that hard to pilot deck was in the hands of a high legend player it may be favoured vs your deck and that comes down to skill. Where this leads into small samples though is that maybe over a small sample you have a good run against DH/bad DH players and then you think your deck is good vs them when it isn't really. Other people then try and get different results. Possibly they pilot the deck differently or are more skilled or less skilled. You will often find comments of people having success with a particular deck but if enough people try it statistically some people are going to go on a winning streak with it. The people who do not may believe it's their fault for not piloting the deck correctly.
All in all, just be careful with bias in your guides and your deck building. I built a deck recently that I thought was good vs DH since I won three initial games with it until then I lost the next 10 vs DH and realized the deck is actually trash against it and so I didn't post it since it would be awful in the meta.
Agreed. In statistics, sample size is hugely important. Until you have a large enough sample the data might be outright lying to you. Save all your conclusions until you collect sufficient data.
Agreed. In statistics, sample size is hugely important. Until you have a large enough sample the data might be outright lying to you. Save all your conclusions until you collect sufficient data.
The trouble of course is that the sample size required is far higher than anyone would expect. I understand there are services which provides large samples if everyone plays that one list but it doesn't assist for a new deck.
There is a big problem with deck building on this site. The most obvious is win percentages over small samples. For example, if I make 100 of my own decks in a year I am bound to soon or later get on a win streak with one of them. I then post it saying it has a 60 or 70% win rate and people jump on board. Let's put that to one side. There is another problem.
A significant issue is the guides indicating a matchup is favoured when it really is not. A matchup can be favoured when climbing but not at say high legend. Imagine if, for example, a particular deck was very hard to pilot. It may give the impression that when climbing your deck is favoured vs this hard to pilot deck. However, if that hard to pilot deck was in the hands of a high legend player it may be favoured vs your deck and that comes down to skill. Where this leads into small samples though is that maybe over a small sample you have a good run against DH/bad DH players and then you think your deck is good vs them when it isn't really. Other people then try and get different results. Possibly they pilot the deck differently or are more skilled or less skilled. You will often find comments of people having success with a particular deck but if enough people try it statistically some people are going to go on a winning streak with it. The people who do not may believe it's their fault for not piloting the deck correctly.
All in all, just be careful with bias in your guides and your deck building. I built a deck recently that I thought was good vs DH since I won three initial games with it until then I lost the next 10 vs DH and realized the deck is actually trash against it and so I didn't post it since it would be awful in the meta.
You're looking way to much into it. I understand what are you trying to say, but in the end it's a card game based on RNG.
Meaning a basic deck in some situation with little to no powerful cards can beat a tier 1 or tier S deck.
DH is the prime example of that. Use only DH cards and you will achieve more or less a decent win-rate. Not great but you will win games. Same goes to any other class, even though more in DH case.
Skill comes from "understanding" the game. That's why I always thought anyway. But RNG is the greatest factor here. You can win 10 in a row but as well loose 20 after that with the same deck.
There's no point really analyzing it more.
This one guy claiming his priest deck has 94% win-rate and my score vs priest as a rogue is 11-1. So all in all I don't even bother with those decks and people claiming their deck is super OP. Just playing my game and keeping my statistics and learning from them
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
There is a big problem with deck building on this site. The most obvious is win percentages over small samples. For example, if I make 100 of my own decks in a year I am bound to soon or later get on a win streak with one of them. I then post it saying it has a 60 or 70% win rate and people jump on board. Let's put that to one side. There is another problem.
A significant issue is the guides indicating a matchup is favoured when it really is not. A matchup can be favoured when climbing but not at say high legend. Imagine if, for example, a particular deck was very hard to pilot. It may give the impression that when climbing your deck is favoured vs this hard to pilot deck. However, if that hard to pilot deck was in the hands of a high legend player it may be favoured vs your deck and that comes down to skill. Where this leads into small samples though is that maybe over a small sample you have a good run against DH/bad DH players and then you think your deck is good vs them when it isn't really. Other people then try and get different results. Possibly they pilot the deck differently or are more skilled or less skilled. You will often find comments of people having success with a particular deck but if enough people try it statistically some people are going to go on a winning streak with it. The people who do not may believe it's their fault for not piloting the deck correctly.
All in all, just be careful with bias in your guides and your deck building. I built a deck recently that I thought was good vs DH since I won three initial games with it until then I lost the next 10 vs DH and realized the deck is actually trash against it and so I didn't post it since it would be awful in the meta.
Agreed. In statistics, sample size is hugely important. Until you have a large enough sample the data might be outright lying to you. Save all your conclusions until you collect sufficient data.
The trouble of course is that the sample size required is far higher than anyone would expect. I understand there are services which provides large samples if everyone plays that one list but it doesn't assist for a new deck.
You're looking way to much into it. I understand what are you trying to say, but in the end it's a card game based on RNG.
Meaning a basic deck in some situation with little to no powerful cards can beat a tier 1 or tier S deck.
DH is the prime example of that. Use only DH cards and you will achieve more or less a decent win-rate. Not great but you will win games. Same goes to any other class, even though more in DH case.
Skill comes from "understanding" the game. That's why I always thought anyway. But RNG is the greatest factor here. You can win 10 in a row but as well loose 20 after that with the same deck.
There's no point really analyzing it more.
This one guy claiming his priest deck has 94% win-rate and my score vs priest as a rogue is 11-1. So all in all I don't even bother with those decks and people claiming their deck is super OP. Just playing my game and keeping my statistics and learning from them