what do you think. real or fake?
for any game
Elo Hell
posted in
Esports
what do you think. real or fake?
for any game
http://www.strawpoll.me/11798268
for any game
http://www.strawpoll.me/11798268
i think theyre elo hell in high ranks but if your a low rank you're just bad
i think theyre elo hell in high ranks but if your a low rank you're just bad
Nub_Danishi think theyre elo hell in high ranks but if your a low rank you're just bad
I agree mostly
some times I find in middle ranked games its just random if you win
[quote=Nub_Danish]i think theyre elo hell in high ranks but if your a low rank you're just bad[/quote]
I agree mostly
some times I find in middle ranked games its just random if you win
I agree mostly
some times I find in middle ranked games its just random if you win
if you are in elo hell, you are at your correct skill rank. every good matchmaking system will attempt to force you into a 50% win rate by pairing you against better players until you cant overcome them, or by pairing you against worse players until you can no longer beat them. people who say that they are at a forced 50% winrate (i see this a lot from dota players) literally just dont understand how matchmaking systems work apparently, and that if you win a bunch of games in a row it logically follows that you will lose a bunch of games in a row afterwards as a result of your increased matchmaking rating pairing you against better players.
if elo hell really existed then for every person experiencing elo hell, another person would be experiencing elo "heaven" where they just get carried well above their skill rank, but strangely enough you never hear about that effect.
if you are in elo hell, you are at your correct skill rank. every good matchmaking system will attempt to force you into a 50% win rate by pairing you against better players until you cant overcome them, or by pairing you against worse players until you can no longer beat them. people who say that they are at a forced 50% winrate (i see this a lot from dota players) literally just dont understand how matchmaking systems work apparently, and that if you win a bunch of games in a row it logically follows that you will lose a bunch of games in a row afterwards as a result of your increased matchmaking rating pairing you against better players.
if elo hell really existed then for every person experiencing elo hell, another person would be experiencing elo "heaven" where they just get carried well above their skill rank, but strangely enough you never hear about that effect.
if elo hell really existed then for every person experiencing elo hell, another person would be experiencing elo "heaven" where they just get carried well above their skill rank, but strangely enough you never hear about that effect.
#5
The problem is that if elo is completely random you will get exactly the same 50% winrate.
Elo is incredibly difficult when you're talking about teams, different classes, different maps, single games and have no real way to gauge how close a match was.
Examples for each:
1. Teams: You have one player with way too high elo on your team. You might lose even though it's not your fault, but if you win their elo gets skewed even more. Unless you take individual performance into account the system can't tell who is dragging the team down until it gets a larger sample size.
2. Classes: Self explanatory, being good at class A does not mean you're good at class B. Class limits / meta can force you to play a class that you're worse at than your elo suggests. Unless you're equally good at all classes that might mean that you end up in the middle in terms of elo where you lose every game on classes you don't like, but win every game when you can play your main.
3. Maps: Self explanatory.
4. Elo relies heavily on larger series of games to judge whether a player is over or underperforming. You're not going to win every single time against someone with only slightly lower elo. If elo predicts a 75% winrate and you win 3 out of 4 rounds your elo is correct and should stay the same. A lot of system will only count the win and increase your elo instead. On the other hand if elo predicts a 40% winrate and you win 2 out of 5 rounds your elo should stay the same, but in most system you will lose elo instead. Then there's also the problem that "almost capping the point 5 times" will not show up in even the most sophisticated systems.
tl;dr
Perfect system gets you 50% winrate.
Completely random system gets you 50% winrate.
The difference is how close the games are.
If you either roll or get rolled the system is bad. If the games are close you have exactly the right elo.
#5
The problem is that if elo is completely random you will get exactly the same 50% winrate.
Elo is incredibly difficult when you're talking about teams, different classes, different maps, single games and have no real way to gauge how close a match was.
Examples for each:
1. Teams: You have one player with way too high elo on your team. You might lose even though it's not your fault, but if you win their elo gets skewed even more. Unless you take individual performance into account the system can't tell who is dragging the team down until it gets a larger sample size.
2. Classes: Self explanatory, being good at class A does not mean you're good at class B. Class limits / meta can force you to play a class that you're worse at than your elo suggests. Unless you're equally good at all classes that might mean that you end up in the middle in terms of elo where you lose every game on classes you don't like, but win every game when you can play your main.
3. Maps: Self explanatory.
4. Elo relies heavily on larger series of games to judge whether a player is over or underperforming. You're not going to win every single time against someone with only slightly lower elo. If elo predicts a 75% winrate and you win 3 out of 4 rounds your elo is correct and should stay the same. A lot of system will only count the win and increase your elo instead. On the other hand if elo predicts a 40% winrate and you win 2 out of 5 rounds your elo should stay the same, but in most system you will lose elo instead. Then there's also the problem that "almost capping the point 5 times" will not show up in even the most sophisticated systems.
tl;dr
Perfect system gets you 50% winrate.
Completely random system gets you 50% winrate.
The difference is how close the games are.
If you either roll or get rolled the system is bad. If the games are close you have exactly the right elo.
The problem is that if elo is completely random you will get exactly the same 50% winrate.
Elo is incredibly difficult when you're talking about teams, different classes, different maps, single games and have no real way to gauge how close a match was.
Examples for each:
1. Teams: You have one player with way too high elo on your team. You might lose even though it's not your fault, but if you win their elo gets skewed even more. Unless you take individual performance into account the system can't tell who is dragging the team down until it gets a larger sample size.
2. Classes: Self explanatory, being good at class A does not mean you're good at class B. Class limits / meta can force you to play a class that you're worse at than your elo suggests. Unless you're equally good at all classes that might mean that you end up in the middle in terms of elo where you lose every game on classes you don't like, but win every game when you can play your main.
3. Maps: Self explanatory.
4. Elo relies heavily on larger series of games to judge whether a player is over or underperforming. You're not going to win every single time against someone with only slightly lower elo. If elo predicts a 75% winrate and you win 3 out of 4 rounds your elo is correct and should stay the same. A lot of system will only count the win and increase your elo instead. On the other hand if elo predicts a 40% winrate and you win 2 out of 5 rounds your elo should stay the same, but in most system you will lose elo instead. Then there's also the problem that "almost capping the point 5 times" will not show up in even the most sophisticated systems.
tl;dr
Perfect system gets you 50% winrate.
Completely random system gets you 50% winrate.
The difference is how close the games are.
If you either roll or get rolled the system is bad. If the games are close you have exactly the right elo.
I don't think being in elo hell means you're bad. I think in some games the only way to get out of it is if you are miles better than the rank you're stuck at. Someone can still be better than their mmr but not good enough to carry 4 or 5 people, however, I do think if you are better then you should be able to slowly climb out.
I don't think being in elo hell means you're bad. I think in some games the only way to get out of it is if you are miles better than the rank you're stuck at. Someone can still be better than their mmr but not good enough to carry 4 or 5 people, however, I do think if you are better then you should be able to slowly climb out.
trizzaaAsk tyler1 @YOya :^)
he got challenger while intentionally losing games lol
[quote=trizzaa]Ask tyler1 @YOya :^)[/quote]
he got challenger while intentionally losing games lol
he got challenger while intentionally losing games lol
Setsul#5
The problem is that if elo is completely random you will get exactly the same 50% winrate.
To not be able to tell the difference you would need to have been placed by accident at exactly your level (assuming the system is not random). If you begin by having a streak of good or poor results then even out to 50/50 then you almost certainly have been placed in an appropriate skill bracket. A player's whole experience over time matters.
Regarding rolls vs close games that depends a lot on the nature of the game and it's propensity for one sided results, and the effect of other factors on that such as (assuming a team game) whether having a very large skill gap between players on one team will render it more likely to lose than the other even though overall team skills may appear equal.
[quote=Setsul]#5
The problem is that if elo is completely random you will get exactly the same 50% winrate.[/quote]
To not be able to tell the difference you would need to have been placed by accident at exactly your level (assuming the system is not random). If you begin by having a streak of good or poor results then even out to 50/50 then you almost certainly have been placed in an appropriate skill bracket. A player's whole experience over time matters.
Regarding rolls vs close games that depends a lot on the nature of the game and it's propensity for one sided results, and the effect of other factors on that such as (assuming a team game) whether having a very large skill gap between players on one team will render it more likely to lose than the other even though overall team skills may appear equal.
The problem is that if elo is completely random you will get exactly the same 50% winrate.[/quote]
To not be able to tell the difference you would need to have been placed by accident at exactly your level (assuming the system is not random). If you begin by having a streak of good or poor results then even out to 50/50 then you almost certainly have been placed in an appropriate skill bracket. A player's whole experience over time matters.
Regarding rolls vs close games that depends a lot on the nature of the game and it's propensity for one sided results, and the effect of other factors on that such as (assuming a team game) whether having a very large skill gap between players on one team will render it more likely to lose than the other even though overall team skills may appear equal.
#9
I'm not sure if you didn't understand what I meant or if I don't understand what you mean.
I'm talking about completely random placement. So you get matched with a mix of far better, far worse and somewhat similarly skilled players against another random mix. The wins and losses will basically be random data. You will end up with ~50% winrate and wildly fluctuating elo.
If the system is good there should not be any long streaks, not if you get another one in the opposite direction or equal length right after that. If you were placed to low then a short streak is normal, maybe 2 or 3 losses after that if you overshoot and then pretty much 50% winrate. But if you were placed correctly, go up e.g. 400 elo and then go down 400 elo something is seriously wrong. Your chances of winning the later games were single digit percentages according to your original elo so if that were correct you would've just had a 1 in 100000 streak. Your elo must've been wrong. But if you then lose all that elo again then either you had a 1 in 100000 negative streak or your original elo was correct. If that happens regularly and/or to a fair share of the playerbase then the system is probably more random that accurate.
#9
I'm not sure if you didn't understand what I meant or if I don't understand what you mean.
I'm talking about completely random placement. So you get matched with a mix of far better, far worse and somewhat similarly skilled players against another random mix. The wins and losses will basically be random data. You will end up with ~50% winrate and wildly fluctuating elo.
If the system is good there should not be any long streaks, not if you get another one in the opposite direction or equal length right after that. If you were placed to low then a short streak is normal, maybe 2 or 3 losses after that if you overshoot and then pretty much 50% winrate. But if you were placed correctly, go up e.g. 400 elo and then go down 400 elo something is seriously wrong. Your chances of winning the later games were single digit percentages according to your original elo so if that were correct you would've just had a 1 in 100000 streak. Your elo must've been wrong. But if you then lose all that elo again then either you had a 1 in 100000 negative streak or your original elo was correct. If that happens regularly and/or to a fair share of the playerbase then the system is probably more random that accurate.
I'm not sure if you didn't understand what I meant or if I don't understand what you mean.
I'm talking about completely random placement. So you get matched with a mix of far better, far worse and somewhat similarly skilled players against another random mix. The wins and losses will basically be random data. You will end up with ~50% winrate and wildly fluctuating elo.
If the system is good there should not be any long streaks, not if you get another one in the opposite direction or equal length right after that. If you were placed to low then a short streak is normal, maybe 2 or 3 losses after that if you overshoot and then pretty much 50% winrate. But if you were placed correctly, go up e.g. 400 elo and then go down 400 elo something is seriously wrong. Your chances of winning the later games were single digit percentages according to your original elo so if that were correct you would've just had a 1 in 100000 streak. Your elo must've been wrong. But if you then lose all that elo again then either you had a 1 in 100000 negative streak or your original elo was correct. If that happens regularly and/or to a fair share of the playerbase then the system is probably more random that accurate.
SetsulCompletely random system gets you 50% winrate.
Correct me if I'm mistaken, but is this really true? Take this example:
you are put in a match with 11 other players of truly random skill, and you are the worst player to ever play the game. Now, there are 6 players on the other team who could be any skill level and there are 5 players who on your team who could be any skill level, plus one who is guaranteed to be terrible.
With a big enough sample size, aren't you at a statistical disadvantage? Every time you go into a game, there is a higher likelihood of being the worse team than there is of being the better team, due to your own individual skill level. This would (again with a large enough sample size) mean you're likely to win fewer than 50% of your games.
And the opposite would be true for the best player in the history of the game.
[quote=Setsul]Completely random system gets you 50% winrate.[/quote]
Correct me if I'm mistaken, but is this really true? Take this example:
you are put in a match with 11 other players of truly random skill, and you are the worst player to ever play the game. Now, there are 6 players on the other team who could be any skill level and there are 5 players who on your team who could be any skill level, plus one who is guaranteed to be terrible.
With a big enough sample size, aren't you at a statistical disadvantage? Every time you go into a game, there is a higher likelihood of being the worse team than there is of being the better team, due to your own individual skill level. This would (again with a large enough sample size) mean you're likely to win fewer than 50% of your games.
And the opposite would be true for the best player in the history of the game.
Correct me if I'm mistaken, but is this really true? Take this example:
you are put in a match with 11 other players of truly random skill, and you are the worst player to ever play the game. Now, there are 6 players on the other team who could be any skill level and there are 5 players who on your team who could be any skill level, plus one who is guaranteed to be terrible.
With a big enough sample size, aren't you at a statistical disadvantage? Every time you go into a game, there is a higher likelihood of being the worse team than there is of being the better team, due to your own individual skill level. This would (again with a large enough sample size) mean you're likely to win fewer than 50% of your games.
And the opposite would be true for the best player in the history of the game.
CitricSetsulCompletely random system gets you 50% winrate.Correct me if I'm mistaken, but is this really true? Take this example:
you are put in a match with 11 other players of truly random skill, and you are the worst player to ever play the game. Now, there are 6 players on the other team who could be any skill level and there are 5 players who on your team who could be any skill level, plus one who is guaranteed to be terrible.
With a big enough sample size, aren't you at a statistical disadvantage? Every time you go into a game, there is a higher likelihood of being the worse team than there is of being the better team, due to your own individual skill level. This would (again with a large enough sample size) mean you're likely to win fewer than 50% of your games.
And the opposite would be true for the best player in the history of the game.
I think he means to say that the expected winrate is 50%. You wouldn't believe that to actually be the true winrate for every player, especially over a "small" number of games.
[quote=Citric][quote=Setsul]Completely random system gets you 50% winrate.[/quote]
Correct me if I'm mistaken, but is this really true? Take this example:
you are put in a match with 11 other players of truly random skill, and you are the worst player to ever play the game. Now, there are 6 players on the other team who could be any skill level and there are 5 players who on your team who could be any skill level, plus one who is guaranteed to be terrible.
With a big enough sample size, aren't you at a statistical disadvantage? Every time you go into a game, there is a higher likelihood of being the worse team than there is of being the better team, due to your own individual skill level. This would (again with a large enough sample size) mean you're likely to win fewer than 50% of your games.
And the opposite would be true for the best player in the history of the game.[/quote]
I think he means to say that the expected winrate is 50%. You wouldn't believe that to actually be the true winrate for every player, especially over a "small" number of games.
Correct me if I'm mistaken, but is this really true? Take this example:
you are put in a match with 11 other players of truly random skill, and you are the worst player to ever play the game. Now, there are 6 players on the other team who could be any skill level and there are 5 players who on your team who could be any skill level, plus one who is guaranteed to be terrible.
With a big enough sample size, aren't you at a statistical disadvantage? Every time you go into a game, there is a higher likelihood of being the worse team than there is of being the better team, due to your own individual skill level. This would (again with a large enough sample size) mean you're likely to win fewer than 50% of your games.
And the opposite would be true for the best player in the history of the game.[/quote]
I think he means to say that the expected winrate is 50%. You wouldn't believe that to actually be the true winrate for every player, especially over a "small" number of games.
#12
Yeah, sample size is the problem though. Given enough time you'll get the usual bell curve and the better players will probably be higher up. But if you need 10000 games per player to get to that point because even the best player ever only gets a 51% winrate then it's rather irrelevant.
And then you have the fact that your playerbase is not static. New players messing things up etc.
Also improving a lot but only changing your winrate from 50.1% to 50.2% would be rather frustrating and unnoticable.
#12
Yeah, sample size is the problem though. Given enough time you'll get the usual bell curve and the better players will probably be higher up. But if you need 10000 games per player to get to that point because even the best player ever only gets a 51% winrate then it's rather irrelevant.
And then you have the fact that your playerbase is not static. New players messing things up etc.
Also improving a lot but only changing your winrate from 50.1% to 50.2% would be rather frustrating and unnoticable.
Yeah, sample size is the problem though. Given enough time you'll get the usual bell curve and the better players will probably be higher up. But if you need 10000 games per player to get to that point because even the best player ever only gets a 51% winrate then it's rather irrelevant.
And then you have the fact that your playerbase is not static. New players messing things up etc.
Also improving a lot but only changing your winrate from 50.1% to 50.2% would be rather frustrating and unnoticable.
Setsul#9
I'm not sure if you didn't understand what I meant or if I don't understand what you mean.
I'm talking about completely random placement. So you get matched with a mix of far better, far worse and somewhat similarly skilled players against another random mix. The wins and losses will basically be random data. You will end up with ~50% winrate and wildly fluctuating elo.
If the system is good there should not be any long streaks, not if you get another one in the opposite direction or equal length right after that. If you were placed to low then a short streak is normal, maybe 2 or 3 losses after that if you overshoot and then pretty much 50% winrate. But if you were placed correctly, go up e.g. 400 elo and then go down 400 elo something is seriously wrong. Your chances of winning the later games were single digit percentages according to your original elo so if that were correct you would've just had a 1 in 100000 streak. Your elo must've been wrong. But if you then lose all that elo again then either you had a 1 in 100000 negative streak or your original elo was correct. If that happens regularly and/or to a fair share of the playerbase then the system is probably more random that accurate.
Assuming no random placement matches occur, if we make the assumption that a new player is low skill and the system adds them at medium skill then they will lose their first game. The system will reduce it's opinion of their skill and try again, they will still lose. This will continue until they also start to win games at which point the system will stabilise it's opinion of their skill and their results should start to even out.
This assumes that the system is not tuned to overreact to their performance and actually reaches a stable position. When they first enter the system they should definitely experience a set of similar results based on the differential between the system's assumption about their starting skill and their actual skill unless those things happen to match which won't be the case for most players.
There is also the matter of population movement. In a system where all the players start at the bottom (a season for example) different skill level players will experience change differently. The best or most dedicated (activity is always a factor in this kind of situation) should very quickly form a vanguard that separates itself from the main bulk of players, but a medium skill player will spend more time sat in the general body of players including the weakest because they are all dominated by the better players.
A regularly resetting skill ladder is an example of this kind of situation. A mature rating ladder that doesn't reset doesn't really see this problem. This is why some games that have seasons use a placement match system to separate their players so they can more quickly arrive at a skill appropriate level without this experience. In TF2 if you look beyond match results it's fairly easy to bracket players quite quickly off the back of 10 games or so.
So depending on how the placement matches system work, a game with seasons that uses them could well produce the 50/50 experience you discuss. A system that relies purely on it's rating system and has a mature ladder must make an assumption about new player skill at the start which will very likely vary from their objective skill that should produce an initial adjustment that creates a streak of results that reflect that.
Rating systems have moved on from Elo of course, it's not really used any more and dominates things like Chess for historical reasons rather than quality.
Streakiness is another consideration, and is related to random elements of the game being measured. Hearthstone is a notorious example where the best players who operate at a very high level can still lose many games in a row due purely to random events inherent in the game design, and the way that those random events can snowball into a huge advantage. Competitively configured TF2 is probably pretty stable in this regard, the vanilla version has a lot more luck involved.
[quote=Setsul]#9
I'm not sure if you didn't understand what I meant or if I don't understand what you mean.
I'm talking about completely random placement. So you get matched with a mix of far better, far worse and somewhat similarly skilled players against another random mix. The wins and losses will basically be random data. You will end up with ~50% winrate and wildly fluctuating elo.
If the system is good there should not be any long streaks, not if you get another one in the opposite direction or equal length right after that. If you were placed to low then a short streak is normal, maybe 2 or 3 losses after that if you overshoot and then pretty much 50% winrate. But if you were placed correctly, go up e.g. 400 elo and then go down 400 elo something is seriously wrong. Your chances of winning the later games were single digit percentages according to your original elo so if that were correct you would've just had a 1 in 100000 streak. Your elo must've been wrong. But if you then lose all that elo again then either you had a 1 in 100000 negative streak or your original elo was correct. If that happens regularly and/or to a fair share of the playerbase then the system is probably more random that accurate.[/quote]
Assuming no random placement matches occur, if we make the assumption that a new player is low skill and the system adds them at medium skill then they will lose their first game. The system will reduce it's opinion of their skill and try again, they will still lose. This will continue until they also start to win games at which point the system will stabilise it's opinion of their skill and their results should start to even out.
This assumes that the system is not tuned to overreact to their performance and actually reaches a stable position. When they first enter the system they should definitely experience a set of similar results based on the differential between the system's assumption about their starting skill and their actual skill unless those things happen to match which won't be the case for most players.
There is also the matter of population movement. In a system where all the players start at the bottom (a season for example) different skill level players will experience change differently. The best or most dedicated (activity is always a factor in this kind of situation) should very quickly form a vanguard that separates itself from the main bulk of players, but a medium skill player will spend more time sat in the general body of players including the weakest because they are all dominated by the better players.
A regularly resetting skill ladder is an example of this kind of situation. A mature rating ladder that doesn't reset doesn't really see this problem. This is why some games that have seasons use a placement match system to separate their players so they can more quickly arrive at a skill appropriate level without this experience. In TF2 if you look beyond match results it's fairly easy to bracket players quite quickly off the back of 10 games or so.
So depending on how the placement matches system work, a game with seasons that uses them could well produce the 50/50 experience you discuss. A system that relies purely on it's rating system and has a mature ladder must make an assumption about new player skill at the start which will very likely vary from their objective skill that should produce an initial adjustment that creates a streak of results that reflect that.
Rating systems have moved on from Elo of course, it's not really used any more and dominates things like Chess for historical reasons rather than quality.
Streakiness is another consideration, and is related to random elements of the game being measured. Hearthstone is a notorious example where the best players who operate at a very high level can still lose many games in a row due purely to random events inherent in the game design, and the way that those random events can snowball into a huge advantage. Competitively configured TF2 is probably pretty stable in this regard, the vanilla version has a lot more luck involved.
I'm not sure if you didn't understand what I meant or if I don't understand what you mean.
I'm talking about completely random placement. So you get matched with a mix of far better, far worse and somewhat similarly skilled players against another random mix. The wins and losses will basically be random data. You will end up with ~50% winrate and wildly fluctuating elo.
If the system is good there should not be any long streaks, not if you get another one in the opposite direction or equal length right after that. If you were placed to low then a short streak is normal, maybe 2 or 3 losses after that if you overshoot and then pretty much 50% winrate. But if you were placed correctly, go up e.g. 400 elo and then go down 400 elo something is seriously wrong. Your chances of winning the later games were single digit percentages according to your original elo so if that were correct you would've just had a 1 in 100000 streak. Your elo must've been wrong. But if you then lose all that elo again then either you had a 1 in 100000 negative streak or your original elo was correct. If that happens regularly and/or to a fair share of the playerbase then the system is probably more random that accurate.[/quote]
Assuming no random placement matches occur, if we make the assumption that a new player is low skill and the system adds them at medium skill then they will lose their first game. The system will reduce it's opinion of their skill and try again, they will still lose. This will continue until they also start to win games at which point the system will stabilise it's opinion of their skill and their results should start to even out.
This assumes that the system is not tuned to overreact to their performance and actually reaches a stable position. When they first enter the system they should definitely experience a set of similar results based on the differential between the system's assumption about their starting skill and their actual skill unless those things happen to match which won't be the case for most players.
There is also the matter of population movement. In a system where all the players start at the bottom (a season for example) different skill level players will experience change differently. The best or most dedicated (activity is always a factor in this kind of situation) should very quickly form a vanguard that separates itself from the main bulk of players, but a medium skill player will spend more time sat in the general body of players including the weakest because they are all dominated by the better players.
A regularly resetting skill ladder is an example of this kind of situation. A mature rating ladder that doesn't reset doesn't really see this problem. This is why some games that have seasons use a placement match system to separate their players so they can more quickly arrive at a skill appropriate level without this experience. In TF2 if you look beyond match results it's fairly easy to bracket players quite quickly off the back of 10 games or so.
So depending on how the placement matches system work, a game with seasons that uses them could well produce the 50/50 experience you discuss. A system that relies purely on it's rating system and has a mature ladder must make an assumption about new player skill at the start which will very likely vary from their objective skill that should produce an initial adjustment that creates a streak of results that reflect that.
Rating systems have moved on from Elo of course, it's not really used any more and dominates things like Chess for historical reasons rather than quality.
Streakiness is another consideration, and is related to random elements of the game being measured. Hearthstone is a notorious example where the best players who operate at a very high level can still lose many games in a row due purely to random events inherent in the game design, and the way that those random events can snowball into a huge advantage. Competitively configured TF2 is probably pretty stable in this regard, the vanilla version has a lot more luck involved.
GentlemanJonAssuming no random placement matches occur, if we make the assumption that a new player is low skill and the system adds them at medium skill then they will lose their first game. The system will reduce it's opinion of their skill and try again, they will still lose. This will continue until they also start to win games at which point the system will stabilise it's opinion of their skill and their results should start to even out.
Yeah, you didn't understand what I meant.
You're still in the mindset "most ratings are correct, false ratings will quickly be corrected".
My point was if all ratings are random (which means most are completely wrong) then it'll not sort itself out quickly. It's just self perpetuating randomness.
Going back to my original points:
1. Rating systems are rather difficult because in most games results aren't 100% comparable and there's no real benchmark that you can use to get a few initial guaranteed correct ratings that you need for the system to work.
2. The global average winrate must be 50%. You will not deviate far from that in a perfect system and you won't either in a random system. So unless it's a system that's somehow far worse than random (or random with huge overreactions) your elo will only change very slowly if at all (within a certain range obviously) unless you actually improve. Which is one of the 2 main points of the elo/rating system.
"I have 50% winrate / my elo is not changing -> I must be stuck in elo hell" is simply not a logical conclusion. It could be a terrible system (or a localized area of randomness, e.g. all players start with the same elo in the "middle" and it needs some time to sort that out) or just as well Dunning-Kruger.
Any rating system that's not completely broken will get everyone close to 50% winrate. That's the easy part. The difficult part is to produce close games. If that happens then you can reasonably assume that the original goal, accurately rating the players' skill, was met.
So that's what I mean. If there's completely shut outs (both wins and losses for you) then the rating system is not working properly, if you have mostly close games and even very close games then it probably is. Your elo won't change much either way, but one means you actually are in elo hell and the other means you're exactly where you should be and should stop bitching and start improving if you want to get higher elo.
[quote=GentlemanJon]
Assuming no random placement matches occur, if we make the assumption that a new player is low skill and the system adds them at medium skill then they will lose their first game. The system will reduce it's opinion of their skill and try again, they will still lose. This will continue until they also start to win games at which point the system will stabilise it's opinion of their skill and their results should start to even out.[/quote]
Yeah, you didn't understand what I meant.
You're still in the mindset "most ratings are correct, false ratings will quickly be corrected".
My point was if all ratings are random (which means most are completely wrong) then it'll not sort itself out quickly. It's just self perpetuating randomness.
Going back to my original points:
1. Rating systems are rather difficult because in most games results aren't 100% comparable and there's no real benchmark that you can use to get a few initial guaranteed correct ratings that you need for the system to work.
2. The global average winrate must be 50%. You will not deviate far from that in a perfect system and you won't either in a random system. So unless it's a system that's somehow far worse than random (or random with huge overreactions) your elo will only change very slowly if at all (within a certain range obviously) unless you actually improve. Which is one of the 2 main points of the elo/rating system.
"I have 50% winrate / my elo is not changing -> I must be stuck in elo hell" is simply not a logical conclusion. It could be a terrible system (or a localized area of randomness, e.g. all players start with the same elo in the "middle" and it needs some time to sort that out) or just as well Dunning-Kruger.
Any rating system that's not completely broken will get everyone close to 50% winrate. That's the easy part. The difficult part is to produce close games. If that happens then you can reasonably assume that the original goal, accurately rating the players' skill, was met.
So that's what I mean. If there's completely shut outs (both wins and losses for you) then the rating system is not working properly, if you have mostly close games and even very close games then it probably is. Your elo won't change much either way, but one means you actually are in elo hell and the other means you're exactly where you should be and should stop bitching and start improving if you want to get higher elo.
Assuming no random placement matches occur, if we make the assumption that a new player is low skill and the system adds them at medium skill then they will lose their first game. The system will reduce it's opinion of their skill and try again, they will still lose. This will continue until they also start to win games at which point the system will stabilise it's opinion of their skill and their results should start to even out.[/quote]
Yeah, you didn't understand what I meant.
You're still in the mindset "most ratings are correct, false ratings will quickly be corrected".
My point was if all ratings are random (which means most are completely wrong) then it'll not sort itself out quickly. It's just self perpetuating randomness.
Going back to my original points:
1. Rating systems are rather difficult because in most games results aren't 100% comparable and there's no real benchmark that you can use to get a few initial guaranteed correct ratings that you need for the system to work.
2. The global average winrate must be 50%. You will not deviate far from that in a perfect system and you won't either in a random system. So unless it's a system that's somehow far worse than random (or random with huge overreactions) your elo will only change very slowly if at all (within a certain range obviously) unless you actually improve. Which is one of the 2 main points of the elo/rating system.
"I have 50% winrate / my elo is not changing -> I must be stuck in elo hell" is simply not a logical conclusion. It could be a terrible system (or a localized area of randomness, e.g. all players start with the same elo in the "middle" and it needs some time to sort that out) or just as well Dunning-Kruger.
Any rating system that's not completely broken will get everyone close to 50% winrate. That's the easy part. The difficult part is to produce close games. If that happens then you can reasonably assume that the original goal, accurately rating the players' skill, was met.
So that's what I mean. If there's completely shut outs (both wins and losses for you) then the rating system is not working properly, if you have mostly close games and even very close games then it probably is. Your elo won't change much either way, but one means you actually are in elo hell and the other means you're exactly where you should be and should stop bitching and start improving if you want to get higher elo.
SetsulYeah, you didn't understand what I meant.
You're still in the mindset "most ratings are correct, false ratings will quickly be corrected".
My point was if all ratings are random (which means most are completely wrong) then it'll not sort itself out quickly. It's just self perpetuating randomness.
Going back to my original points:
1. Rating systems are rather difficult because in most games results aren't 100% comparable and there's no real benchmark that you can use to get a few initial guaranteed correct ratings that you need for the system to work.
2. The global average winrate must be 50%. You will not deviate far from that in a perfect system and you won't either in a random system. So unless it's a system that's somehow far worse than random (or random with huge overreactions) your elo will only change very slowly if at all (within a certain range obviously) unless you actually improve. Which is one of the 2 main points of the elo/rating system.
"I have 50% winrate / my elo is not changing -> I must be stuck in elo hell" is simply not a logical conclusion. It could be a terrible system (or a localized area of randomness, e.g. all players start with the same elo in the "middle" and it needs some time to sort that out) or just as well Dunning-Kruger.
Any rating system that's not completely broken will get everyone close to 50% winrate. That's the easy part. The difficult part is to produce close games. If that happens then you can reasonably assume that the original goal, accurately rating the players' skill, was met.
So that's what I mean. If there's completely shut outs (both wins and losses for you) then the rating system is not working properly, if you have mostly close games and even very close games then it probably is. Your elo won't change much either way, but one means you actually are in elo hell and the other means you're exactly where you should be and should stop bitching and start improving if you want to get higher elo.
I don't think I've said anything that contradicts anything you've said. I just wanted to point out that at the start of a player's career in a rating system they should have a clear period of adjustment which should be evidence that they are either being moved up or moved down the rating system to their appropriate level, and it will probably be characterised by a consistency of results that don't appear thereafter.
The only bone of "contention" if you want to put it like that is that some games will have more random effects or meta games that don't produce close games no matter how close in skill players really are. These would probably be fairly low skill ceiling games anyway.
[quote=Setsul]Yeah, you didn't understand what I meant.
You're still in the mindset "most ratings are correct, false ratings will quickly be corrected".
My point was if all ratings are random (which means most are completely wrong) then it'll not sort itself out quickly. It's just self perpetuating randomness.
Going back to my original points:
1. Rating systems are rather difficult because in most games results aren't 100% comparable and there's no real benchmark that you can use to get a few initial guaranteed correct ratings that you need for the system to work.
2. The global average winrate must be 50%. You will not deviate far from that in a perfect system and you won't either in a random system. So unless it's a system that's somehow far worse than random (or random with huge overreactions) your elo will only change very slowly if at all (within a certain range obviously) unless you actually improve. Which is one of the 2 main points of the elo/rating system.
"I have 50% winrate / my elo is not changing -> I must be stuck in elo hell" is simply not a logical conclusion. It could be a terrible system (or a localized area of randomness, e.g. all players start with the same elo in the "middle" and it needs some time to sort that out) or just as well Dunning-Kruger.
Any rating system that's not completely broken will get everyone close to 50% winrate. That's the easy part. The difficult part is to produce close games. If that happens then you can reasonably assume that the original goal, accurately rating the players' skill, was met.
So that's what I mean. If there's completely shut outs (both wins and losses for you) then the rating system is not working properly, if you have mostly close games and even very close games then it probably is. Your elo won't change much either way, but one means you actually are in elo hell and the other means you're exactly where you should be and should stop bitching and start improving if you want to get higher elo.[/quote]
I don't think I've said anything that contradicts anything you've said. I just wanted to point out that at the start of a player's career in a rating system they should have a clear period of adjustment which should be evidence that they are either being moved up or moved down the rating system to their appropriate level, and it will probably be characterised by a consistency of results that don't appear thereafter.
The only bone of "contention" if you want to put it like that is that some games will have more random effects or meta games that don't produce close games no matter how close in skill players really are. These would probably be fairly low skill ceiling games anyway.
You're still in the mindset "most ratings are correct, false ratings will quickly be corrected".
My point was if all ratings are random (which means most are completely wrong) then it'll not sort itself out quickly. It's just self perpetuating randomness.
Going back to my original points:
1. Rating systems are rather difficult because in most games results aren't 100% comparable and there's no real benchmark that you can use to get a few initial guaranteed correct ratings that you need for the system to work.
2. The global average winrate must be 50%. You will not deviate far from that in a perfect system and you won't either in a random system. So unless it's a system that's somehow far worse than random (or random with huge overreactions) your elo will only change very slowly if at all (within a certain range obviously) unless you actually improve. Which is one of the 2 main points of the elo/rating system.
"I have 50% winrate / my elo is not changing -> I must be stuck in elo hell" is simply not a logical conclusion. It could be a terrible system (or a localized area of randomness, e.g. all players start with the same elo in the "middle" and it needs some time to sort that out) or just as well Dunning-Kruger.
Any rating system that's not completely broken will get everyone close to 50% winrate. That's the easy part. The difficult part is to produce close games. If that happens then you can reasonably assume that the original goal, accurately rating the players' skill, was met.
So that's what I mean. If there's completely shut outs (both wins and losses for you) then the rating system is not working properly, if you have mostly close games and even very close games then it probably is. Your elo won't change much either way, but one means you actually are in elo hell and the other means you're exactly where you should be and should stop bitching and start improving if you want to get higher elo.[/quote]
I don't think I've said anything that contradicts anything you've said. I just wanted to point out that at the start of a player's career in a rating system they should have a clear period of adjustment which should be evidence that they are either being moved up or moved down the rating system to their appropriate level, and it will probably be characterised by a consistency of results that don't appear thereafter.
The only bone of "contention" if you want to put it like that is that some games will have more random effects or meta games that don't produce close games no matter how close in skill players really are. These would probably be fairly low skill ceiling games anyway.
Daggeri have 800 elo in eu pugchamp :^]
get on my level #500ELOLIFE
[quote=Dagger]i have 800 elo in eu pugchamp :^][/quote]
get on my level #500ELOLIFE
get on my level #500ELOLIFE
so is elo hell a bellcurve of rank where >50% winrate still puts you in the lower portion of the graph?
so is elo hell a bellcurve of rank where >50% winrate still puts you in the lower portion of the graph?
eeeso is elo hell a bellcurve of rank where >50% winrate still puts you in the lower portion of the graph?
I think it's more a situation where your improvement doesn't get rewarded by a >50% win rate in a timely fashion
[quote=eee]so is elo hell a bellcurve of rank where >50% winrate still puts you in the lower portion of the graph?[/quote]
I think it's more a situation where your improvement doesn't get rewarded by a >50% win rate in a timely fashion
I think it's more a situation where your improvement doesn't get rewarded by a >50% win rate in a timely fashion
I don't believe that in any multiplayer game you can actually have an elo hell
People are so toxic in general that just by virtue of your team having 5 people who may just go afk, vs the enemy team having 6, and this ending the game immediately - that by just being average statistically you should win more.
I think that anyone who is genuinely stuck in elo hell is probably the kind of person who does go afk etc, the kind of egotistical mind which would tell himself how much better than everyone else in his skill bracket he is, is bound to be that guy.
I don't believe that in any multiplayer game you can actually have an elo hell
People are so toxic in general that just by virtue of your team having 5 people who may just go afk, vs the enemy team having 6, and this ending the game immediately - that by just being average statistically you should win more.
I think that anyone who is genuinely stuck in elo hell is probably the kind of person who does go afk etc, the kind of egotistical mind which would tell himself how much better than everyone else in his skill bracket he is, is bound to be that guy.
People are so toxic in general that just by virtue of your team having 5 people who may just go afk, vs the enemy team having 6, and this ending the game immediately - that by just being average statistically you should win more.
I think that anyone who is genuinely stuck in elo hell is probably the kind of person who does go afk etc, the kind of egotistical mind which would tell himself how much better than everyone else in his skill bracket he is, is bound to be that guy.
Please
sign in through STEAM
to post a comment.