The thing is if you try to balance for uncoordinated and unskilled players, you will unbalance for skilled and Coordinated players.
Look at heroes like Io, Oracle or Broodmother. For These heroes to have anywhere close to a 50% win rate in solo queue 2k (around where the majority of players are) then they would be way to strong at the highest levels of play. Do you remember when Io's tether was a nuke + stun, he/she was a free win in pro play, and his/her pub win rate was still just below 45%.
Imagine how much of an instant win last pick brood would be in 5-7K if she was strong enough to have a 50% win rate in 2k.
If you balanced for the majority of players things like viper and riki would be useless for anyone who wasn't trash, and any hero who has a high entry level skill (invoker, meepo, NP, brood, visage) would be overpowered as hell for anyone past that entry level of skill.
http://puu.sh/kRrRo/067da79ac4.png (6.84)
http://puu.sh/kRrSR/847f8cc50c.png (6.85)
The top heroes have barely changed. They basically removed leshrac and added Riki.
these are just random consequences of competitive scene aimed patching. the rate of imbalance in pub games for all the patches, given the central limit theorem, can be assumed to be normally distributed around the average level, so sometimes we get less or more balanced patches for pubs. there are no real hidden tendencies here.
They actually are. If your mid players goes magnus and has a 28% winrate on magnus, you might have lost the game.
There is no logic behind. Icefrog only needs to distribute buffs/nerfs randomly to keep up our interest for this game.
Number 12 and TripleSteal are correct. There is an intrinsic tension between balancing for high skill and low skill that makes it impossible to have both be completely balanced simultaneously.
Those heroes with extreme positioning skills are usually the weakest in pub dota (low skill) and the strongest in pro games because positioning around the map, in the lanes, and in the fights becomes more and more important the more skillful the players.
Items face a similar balancing challenge. Active items are potentially far stronger than equal cost passive items, but pubs get far less than full utility from active items making passive items much more popular.
IS BALANCE IMPROVING FOR PUB GAMES? IN GENERAL DISCUSSION
Relentless
Relentless3 hours ago
One commonly assumed goal of balance changes is to make every hero a reasonable choice some of the time, but for all the heroes to be able to be countered.
One way to measure this is to consider the winrates of the heroes. I generally believe that Ice Frog/Valve attempts to balance the game more for pro play than pub play, but it needs to be balanced for pubs as well. Are they getting closer to 50%?
When we consider the All Time stats we find that the winrate of the average hero is 48.9% (its below 50% because the better heroes are played more so the average hero is a losing hero) and the standard deviation is 3.8%. There are usually 20 heroes significantly below average (more than 1 standard deviation) and 2 heros waaaay below average (more than 2 standard deviations). These two re-pick heroes are the same in every data set (Io and Oracle) which simply require far to much coordination for random pub players to play. In the three previous patches there have been 16 heroes more than 1 standard deviation better than average and 3 or 4 exceptional heroes over 2 standard deviations above the average hero winrate.
In 6.84b the wealth was spread to a larger group of heroes and the standard deviation increased to 4.4% with the average hero's winrate jumping up to 49.2% . This could be regarded as a balance improvement overall.
In 6.84c the statistics are nearly identical to 6.84b.
But so far 6.85 has been a less balanced patch. Although the average hero's winrate increased to 49.4% the standard deviation contracted to 4.3% and the gap between the best and worst heroes increased. So that there are now 24 heroes significantly below average (compared to the usual 20) and only 14 above average (compared to the usual 16) with a record 5 heroes having extremely highwinrate more than 2 standard deviations above average.
These 5 heroes => Omnikinght, Spectre, Necrophos, and Ursa are taking far more than their share of wins and have made 6.85 a patch that is more IMBA than earlier patches. It appears the while Valve has clustered the middle of pack more tightly together, this came at the expense of granting excessive power to a few o f the top heroes in pub play.
mhysa
mhysa3 hours ago
This comment was deleted 2 hours ago
The Number 12
The Number 123 hours ago
The thing is if you try to balance for uncoordinated and unskilled players, you will unbalance for skilled and Coordinated players.
Look at heroes like Io, Oracle or Broodmother. For These heroes to have anywhere close to a 50% win rate in solo queue 2k (around where the majority of players are) then they would be way to strong at the highest levels of play. Do you remember when Io's tether was a nuke + stun, he/she was a free win in pro play, and his/her pub win rate was still just below 45%.
Imagine how much of an instant win last pick brood would be in 5-7K if she was strong enough to have a 50% win rate in 2k.
If you balanced for the majority of players things like viper and riki would be useless for anyone who wasn't trash, and any hero who has a high entry level skill (invoker, meepo, NP, brood, visage) would be overpowered as hell for anyone past that entry level of skill.
Mokujin
Mokujin2 hours ago
fuck yoursekf
Mokujin
Mokujin2 hours ago
Aternatively: stop posting shithead
DelCredere
DelCredere2 hours ago
Yeah, very simple logic behind 40% winrate heroes: wanna play those cancers? Learn to play
Pomi
Pomi2 hours ago
http://puu.sh/kRrRo/067da79ac4.png (6.84)
http://puu.sh/kRrSR/847f8cc50c.png (6.85)
The top heroes have barely changed. They basically removed leshrac and added Riki.
Reese
Reese2 hours ago
when i see omni in the enemy team i assume i just lost
ILoveUndertale
ILoveUndertalean hour ago
no
TripleSteal-
TripleSteal-an hour ago
these are just random consequences of competitive scene aimed patching. the rate of imbalance in pub games for all the patches, given the central limit theorem, can be assumed to be normally distributed around the average level, so sometimes we get less or more balanced patches for pubs. there are no real hidden tendencies here.
Kitrak
Kitrakan hour ago
honestly why do people even give a shit about statistics theyre not indicative of anything useful
This comment was edited an hour ago
Anatoli
Anatolian hour ago
They actually are. If your mid players goes magnus and has a 28% winrate on magnus, you might have lost the game.
Soultrap
Soultrap44 minutes ago
There is no logic behind. Icefrog only needs to distribute buffs/nerfs randomly to keep up our interest for this game.
Relentless
Relentless2 minutes ago
Number 12 and TripleSteal are correct. There is an intrinsic tension between balancing for high skill and low skill that makes it impossible to have both be completely balanced simultaneously.
Those heroes with extreme positioning skills are usually the weakest in pub dota (low skill) and the strongest in pro games because positioning around the map, in the lanes, and in the fights becomes more and more important the more skillful the players.
Items face a similar balancing challenge. Active items are potentially far stronger than equal cost passive items, but pubs get far less than full utility from active items making passive items much more popular.IS BALANCE IMPROVING FOR PUB GAMES? IN GENERAL DISCUSSION
Relentless
Relentless3 hours ago
One commonly assumed goal of balance changes is to make every hero a reasonable choice some of the time, but for all the heroes to be able to be countered.
One way to measure this is to consider the winrates of the heroes. I generally believe that Ice Frog/Valve attempts to balance the game more for pro play than pub play, but it needs to be balanced for pubs as well. Are they getting closer to 50%?
When we consider the All Time stats we find that the winrate of the average hero is 48.9% (its below 50% because the better heroes are played more so the average hero is a losing hero) and the standard deviation is 3.8%. There are usually 20 heroes significantly below average (more than 1 standard deviation) and 2 heros waaaay below average (more than 2 standard deviations). These two re-pick heroes are the same in every data set (Io and Oracle) which simply require far to much coordination for random pub players to play. In the three previous patches there have been 16 heroes more than 1 standard deviation better than average and 3 or 4 exceptional heroes over 2 standard deviations above the average hero winrate.
In 6.84b the wealth was spread to a larger group of heroes and the standard deviation increased to 4.4% with the average hero's winrate jumping up to 49.2% . This could be regarded as a balance improvement overall.
In 6.84c the statistics are nearly identical to 6.84b.
But so far 6.85 has been a less balanced patch. Although the average hero's winrate increased to 49.4% the standard deviation contracted to 4.3% and the gap between the best and worst heroes increased. So that there are now 24 heroes significantly below average (compared to the usual 20) and only 14 above average (compared to the usual 16) with a record 5 heroes having extremely highwinrate more than 2 standard deviations above average.
These 5 heroes => Omnikinght, Spectre, Necrophos, and Ursa are taking far more than their share of wins and have made 6.85 a patch that is more IMBA than earlier patches. It appears the while Valve has clustered the middle of pack more tightly together, this came at the expense of granting excessive power to a few o f the top heroes in pub play.
mhysa
mhysa3 hours ago
This comment was deleted 2 hours ago
The Number 12
The Number 123 hours ago
The thing is if you try to balance for uncoordinated and unskilled players, you will unbalance for skilled and Coordinated players.
Look at heroes like Io, Oracle or Broodmother. For These heroes to have anywhere close to a 50% win rate in solo queue 2k (around where the majority of players are) then they would be way to strong at the highest levels of play. Do you remember when Io's tether was a nuke + stun, he/she was a free win in pro play, and his/her pub win rate was still just below 45%.
Imagine how much of an instant win last pick brood would be in 5-7K if she was strong enough to have a 50% win rate in 2k.
If you balanced for the majority of players things like viper and riki would be useless for anyone who wasn't trash, and any hero who has a high entry level skill (invoker, meepo, NP, brood, visage) would be overpowered as hell for anyone past that entry level of skill.
Mokujin
Mokujin2 hours ago
fuck yoursekf
Mokujin
Mokujin2 hours ago
Aternatively: stop posting shithead
DelCredere
DelCredere2 hours ago
Yeah, very simple logic behind 40% winrate heroes: wanna play those cancers? Learn to play
Pomi
Pomi2 hours ago
http://puu.sh/kRrRo/067da79ac4.png (6.84)
http://puu.sh/kRrSR/847f8cc50c.png (6.85)
The top heroes have barely changed. They basically removed leshrac and added Riki.
Reese
Reese2 hours ago
when i see omni in the enemy team i assume i just lost
ILoveUndertale
ILoveUndertalean hour ago
no
TripleSteal-
TripleSteal-an hour ago
these are just random consequences of competitive scene aimed patching. the rate of imbalance in pub games for all the patches, given the central limit theorem, can be assumed to be normally distributed around the average level, so sometimes we get less or more balanced patches for pubs. there are no real hidden tendencies here.
Kitrak
Kitrakan hour ago
honestly why do people even give a shit about statistics theyre not indicative of anything useful
This comment was edited an hour ago
Anatoli
Anatolian hour ago
They actually are. If your mid players goes magnus and has a 28% winrate on magnus, you might have lost the game.
Soultrap
Soultrap44 minutes ago
There is no logic behind. Icefrog only needs to distribute buffs/nerfs randomly to keep up our interest for this game.
Relentless
Relentless2 minutes ago
Number 12 and TripleSteal are correct. There is an intrinsic tension between balancing for high skill and low skill that makes it impossible to have both be completely balanced simultaneously.
Those heroes with extreme positioning skills are usually the weakest in pub dota (low skill) and the strongest in pro games because positioning around the map, in the lanes, and in the fights becomes more and more important the more skillful the players.
Items face a similar balancing challenge. Active items are potentially far stronger than equal cost passive items, but pubs get far less than full utility from active items making passive items much more popular.IS BALANCE IMPROVING FOR PUB GAMES? IN GENERAL DISCUSSION
Relentless
Relentless3 hours ago
One commonly assumed goal of balance changes is to make every hero a reasonable choice some of the time, but for all the heroes to be able to be countered.
One way to measure this is to consider the winrates of the heroes. I generally believe that Ice Frog/Valve attempts to balance the game more for pro play than pub play, but it needs to be balanced for pubs as well. Are they getting closer to 50%?
When we consider the All Time stats we find that the winrate of the average hero is 48.9% (its below 50% because the better heroes are played more so the average hero is a losing hero) and the standard deviation is 3.8%. There are usually 20 heroes significantly below average (more than 1 standard deviation) and 2 heros waaaay below average (more than 2 standard deviations). These two re-pick heroes are the same in every data set (Io and Oracle) which simply require far to much coordination for random pub players to play. In the three previous patches there have been 16 heroes more than 1 standard deviation better than average and 3 or 4 exceptional heroes over 2 standard deviations above the average hero winrate.
In 6.84b the wealth was spread to a larger group of heroes and the standard deviation increased to 4.4% with the average hero's winrate jumping up to 49.2% . This could be regarded as a balance improvement overall.
In 6.84c the statistics are nearly identical to 6.84b.
But so far 6.85 has been a less balanced patch. Although the average hero's winrate increased to 49.4% the standard deviation contracted to 4.3% and the gap between the best and worst heroes increased. So that there are now 24 heroes significantly below average (compared to the usual 20) and only 14 above average (compared to the usual 16) with a record 5 heroes having extremely highwinrate more than 2 standard deviations above average.
These 5 heroes => Omnikinght, Spectre, Necrophos, and Ursa are taking far more than their share of wins and have made 6.85 a patch that is more IMBA than earlier patches. It appears the while Valve has clustered the middle of pack more tightly together, this came at the expense of granting excessive power to a few o f the top heroes in pub play.
mhysa
mhysa3 hours ago
This comment was deleted 2 hours ago
The Number 12
The Number 123 hours ago
The thing is if you try to balance for uncoordinated and unskilled players, you will unbalance for skilled and Coordinated players.
Look at heroes like Io, Oracle or Broodmother. For These heroes to have anywhere close to a 50% win rate in solo queue 2k (around where the majority of players are) then they would be way to strong at the highest levels of play. Do you remember when Io's tether was a nuke + stun, he/she was a free win in pro play, and his/her pub win rate was still just below 45%.
Imagine how much of an instant win last pick brood would be in 5-7K if she was strong enough to have a 50% win rate in 2k.
If you balanced for the majority of players things like viper and riki would be useless for anyone who wasn't trash, and any hero who has a high entry level skill (invoker, meepo, NP, brood, visage) would be overpowered as hell for anyone past that entry level of skill.
Mokujin
Mokujin2 hours ago
fuck yoursekf
Mokujin
Mokujin2 hours ago
Aternatively: stop posting shithead
DelCredere
DelCredere2 hours ago
Yeah, very simple logic behind 40% winrate heroes: wanna play those cancers? Learn to play
Pomi
Pomi2 hours ago
http://puu.sh/kRrRo/067da79ac4.png (6.84)
http://puu.sh/kRrSR/847f8cc50c.png (6.85)
The top heroes have barely changed. They basically removed leshrac and added Riki.
Reese
Reese2 hours ago
when i see omni in the enemy team i assume i just lost
ILoveUndertale
ILoveUndertalean hour ago
no
TripleSteal-
TripleSteal-an hour ago
these are just random consequences of competitive scene aimed patching. the rate of imbalance in pub games for all the patches, given the central limit theorem, can be assumed to be normally distributed around the average level, so sometimes we get less or more balanced patches for pubs. there are no real hidden tendencies here.
Kitrak
Kitrakan hour ago
honestly why do people even give a shit about statistics theyre not indicative of anything useful
This comment was edited an hour ago
Anatoli
Anatolian hour ago
They actually are. If your mid players goes magnus and has a 28% winrate on magnus, you might have lost the game.
Soultrap
Soultrap44 minutes ago
There is no logic behind. Icefrog only needs to distribute buffs/nerfs randomly to keep up our interest for this game.
Relentless
Relentless2 minutes ago
Number 12 and TripleSteal are correct. There is an intrinsic tension between balancing for high skill and low skill that makes it impossible to have both be completely balanced simultaneously.
Those heroes with extreme positioning skills are usually the weakest in pub dota (low skill) and the strongest in pro games because positioning around the map, in the lanes, and in the fights becomes more and more important the more skillful the players.
Items face a similar balancing challenge. Active items are potentially far stronger than equal cost passive items, but pubs get far less than full utility from active items making passive items much more popular.IS BALANCE IMPROVING FOR PUB GAMES? IN GENERAL DISCUSSION
Relentless
Relentless3 hours ago
One commonly assumed goal of balance changes is to make every hero a reasonable choice some of the time, but for all the heroes to be able to be countered.
One way to measure this is to consider the winrates of the heroes. I generally believe that Ice Frog/Valve attempts to balance the game more for pro play than pub play, but it needs to be balanced for pubs as well. Are they getting closer to 50%?
When we consider the All Time stats we find that the winrate of the average hero is 48.9% (its below 50% because the better heroes are played more so the average hero is a losing hero) and the standard deviation is 3.8%. There are usually 20 heroes significantly below average (more than 1 standard deviation) and 2 heros waaaay below average (more than 2 standard deviations). These two re-pick heroes are the same in every data set (Io and Oracle) which simply require far to much coordination for random pub players to play. In the three previous patches there have been 16 heroes more than 1 standard deviation better than average and 3 or 4 exceptional heroes over 2 standard deviations above the average hero winrate.
In 6.84b the wealth was spread to a larger group of heroes and the standard deviation increased to 4.4% with the average hero's winrate jumping up to 49.2% . This could be regarded as a balance improvement overall.
In 6.84c the statistics are nearly identical to 6.84b.
But so far 6.85 has been a less balanced patch. Although the average hero's winrate increased to 49.4% the standard deviation contracted to 4.3% and the gap between the best and worst heroes increased. So that there are now 24 heroes significantly below average (compared to the usual 20) and only 14 above average (compared to the usual 16) with a record 5 heroes having extremely highwinrate more than 2 standard deviations above average.
These 5 heroes => Omnikinght, Spectre, Necrophos, and Ursa are taking far more than their share of wins and have made 6.85 a patch that is more IMBA than earlier patches. It appears the while Valve has clustered the middle of pack more tightly together, this came at the expense of granting excessive power to a few o f the top heroes in pub play.
mhysa
mhysa3 hours ago
This comment was deleted 2 hours ago
The Number 12
The Number 123 hours ago
The thing is if you try to balance for uncoordinated and unskilled players, you will unbalance for skilled and Coordinated players.
Look at heroes like Io, Oracle or Broodmother. For These heroes to have anywhere close to a 50% win rate in solo queue 2k (around where the majority of players are) then they would be way to strong at the highest levels of play. Do you remember when Io's tether was a nuke + stun, he/she was a free win in pro play, and his/her pub win rate was still just below 45%.
Imagine how much of an instant win last pick brood would be in 5-7K if she was strong enough to have a 50% win rate in 2k.
If you balanced for the majority of players things like viper and riki would be useless for anyone who wasn't trash, and any hero who has a high entry level skill (invoker, meepo, NP, brood, visage) would be overpowered as hell for anyone past that entry level of skill.
Mokujin
Mokujin2 hours ago
fuck yoursekf
Mokujin
Mokujin2 hours ago
Aternatively: stop posting shithead
DelCredere
DelCredere2 hours ago
Yeah, very simple logic behind 40% winrate heroes: wanna play those cancers? Learn to play
Pomi
Pomi2 hours ago
http://puu.sh/kRrRo/067da79ac4.png (6.84)
http://puu.sh/kRrSR/847f8cc50c.png (6.85)
The top heroes have barely changed. They basically removed leshrac and added Riki.
Reese
Reese2 hours ago
when i see omni in the enemy team i assume i just lost
ILoveUndertale
ILoveUndertalean hour ago
no
TripleSteal-
TripleSteal-an hour ago
these are just random consequences of competitive scene aimed patching. the rate of imbalance in pub games for all the patches, given the central limit theorem, can be assumed to be normally distributed around the average level, so sometimes we get less or more balanced patches for pubs. there are no real hidden tendencies here.
Kitrak
Kitrakan hour ago
honestly why do people even give a shit about statistics theyre not indicative of anything useful
This comment was edited an hour ago
Anatoli
Anatolian hour ago
They actually are. If your mid players goes magnus and has a 28% winrate on magnus, you might have lost the game.
Soultrap
Soultrap44 minutes ago
There is no logic behind. Icefrog only needs to distribute buffs/nerfs randomly to keep up our interest for this game.
Relentless
Relentless2 minutes ago
Number 12 and TripleSteal are correct. There is an intrinsic tension between balancing for high skill and low skill that makes it impossible to have both be completely balanced simultaneously.
Those heroes with extreme positioning skills are usually the weakest in pub dota (low skill) and the strongest in pro games because positioning around the map, in the lanes, and in the fights becomes more and more important the more skillful the players.
Items face a similar balancing challenge. Active items are potentially far stronger than equal cost passive items, but pubs get far less than full utility from active items making passive items much more popular.
pls stop posting, we already had enough threads where you made yourself look like an asshat.
Per favore Accedi per inviare commenti.
One commonly assumed goal of balance changes is to make every hero a reasonable choice some of the time, but for all the heroes to be able to be countered.
One way to measure this is to consider the winrates of the heroes. I generally believe that Ice Frog/Valve attempts to balance the game more for pro play than pub play, but it needs to be balanced for pubs as well. Are they getting closer to 50%?
When we consider the All Time stats we find that the winrate of the average hero is 48.9% (its below 50% because the better heroes are played more so the average hero is a losing hero) and the standard deviation is 3.8%. There are usually 20 heroes significantly below average (more than 1 standard deviation) and 2 heros waaaay below average (more than 2 standard deviations). These two re-pick heroes are the same in every data set (Io and Oracle) which simply require far to much coordination for random pub players to play. In the three previous patches there have been 16 heroes more than 1 standard deviation better than average and 3 or 4 exceptional heroes over 2 standard deviations above the average hero winrate.
In 6.84b the wealth was spread to a larger group of heroes and the standard deviation increased to 4.4% with the average hero's winrate jumping up to 49.2% . This could be regarded as a balance improvement overall.
In 6.84c the statistics are nearly identical to 6.84b.
But so far 6.85 has been a less balanced patch. Although the average hero's winrate increased to 49.4% the standard deviation contracted to 4.3% and the gap between the best and worst heroes increased. So that there are now 24 heroes significantly below average (compared to the usual 20) and only 14 above average (compared to the usual 16) with a record 5 heroes having extremely highwinrate more than 2 standard deviations above average.
These 5 heroes => Omnikinght, Spectre, Necrophos, and Ursa are taking far more than their share of wins and have made 6.85 a patch that is more IMBA than earlier patches. It appears the while Valve has clustered the middle of pack more tightly together, this came at the expense of granting excessive power to a few o f the top heroes in pub play.