|
On April 18 2011 22:09 [F_]aths wrote:
I find it strange that many players like to play vs. AI. I only did so (in AoE for example) because I had no mates to play against. Then I found the online gaming service and played there.
Vs. AI it's mostly to learn how to abuse AI weaknesses. I did like to play the campaign, though.
While I never recommend to practice vs. AI for mulitplayer, if someone likes to do so, he should of course do so.
When I play against AI it is because I either want to warmup after being away from SC for a week or two, to test out a build, but also with the mindset from time to time that "if I can't beat a very hard AI, how can I ever win on ladder then?".
Properly not that smartest way to train I know, but I am trying to change that. 
|
On April 10 2011 05:33 enzym wrote:Show nested quote +On April 10 2011 05:28 [F_]aths wrote:On April 10 2011 05:25 protoss_machine wrote:
I think I have a right to know what my win/lose ratio is whenever I want to. It was fine before the patch... The win/loss ratio is meaningless when you are not playing in Master or higher because the system tries to get you to 50% regardless. If you have deviations from 50%, the only reason is that you did not yet play enough. Win/loss in Diamond and below says nothing about your current skill. The ladder now hides an irrelevant statistics to you to prevent you from falsely putting any meaning in it. There are some statistics one would like to have and we hope that Blizzard will implement. Win ratio is not one of them. You do not get to dictate the meaning of that statistic for other people. User was temp banned for this post.
Let me explain what I meant here and why I expressed myself like I did.
Your post assumes that there can be no other meaning to the w/l statistic for other people than the one which you can see for yourself and have addressed in that post, despite several people in this thread explaining why they would like to see this statistic for different reasons. I am one of these people, but there are more.
But throughout the majority of the discussion, if not all of it, and with the post which I quoted and replied to you managed to ignore such explanations and did not recognize criticism and opinions running counter to your own.
To me my response appeared to make an objectively correct statement with its reasoning having been established more specifically by previous posts in the thread.
Additionally I did not want to express myself as especially respectful towards you because I was under the impression that you were not respecting your own discussion and its participants. I understand that this is unacceptible in a majority of social environments and must admit that I made a mistake by expressing myself in such a harsh and condescending manner.
Nevertheless I hope that my criticism of your statement has now been explained in a way that allows it to be more well understood.
On April 18 2011 22:09 [F_]aths wrote:Show nested quote +On April 18 2011 21:01 Lizarb wrote:
I was one of these old guys that loved the SC universe and wanted to play ladder, but was afraid to loose. "Just need one more practice against the AI".
Good thing that I finally got over this after hearing what Day9 said... "The most important thing is that you play and have fun". I find it strange that many players like to play vs. AI. I only did so (in AoE for example) because I had no mates to play against. Then I found the online gaming service and played there. Vs. AI it's mostly to learn how to abuse AI weaknesses. I did like to play the campaign, though. While I never recommend to practice vs. AI for mulitplayer, if someone likes to do so, he should of course do so.
Here I have to disagree again. Many players like to practice vs AI for different reasons.
LiquidTyler for example likes to practice buildorders vs AI (or rather in absence of an opposing player) in order to make corrections, adjustments, to test and to flesh out very precise timings.
I like to practice vs AI in order to improve my mechanics, because I am out of shape and I see that it greatly impacts my ability to efficiently learn the game.
I have played Broodwar in the past so I know how fast I can be and carry some knowledge of Starcraft basics. I have barely played SC2, so I don't know anything about SC2 matchups at all, but continuously get placed into Diamond regardless of the race I play
SC2 is a strategy game. I want to learn SC2. My knowledge of SC2 matchups doesn't live up to the knowledge of other people in Diamond. Yet I can't focus on learning the game because executing what I know from Broodwar alone still consumes a majority of my focus.
The most easily improved limiting factor for me are mechanics. I can't focus on learning strategies, but at the same time most of my losses will come from more knowledgable people. I can't practice mechanics, because unknown or unprepared for buildorders catch me by surprise and throw me off my game. So I like to practice mechanics vs AI.
Like for LiquidTyler, it allows me to focus on a specific thing while ensuring a distraction free environment. This also works for people wanting to practice one matchup in many consecutive games, which ladder does not allow for.
|
United States12235 Posts
On April 18 2011 21:19 Timm wrote:Show nested quote +On April 18 2011 20:29 Blurb wrote:How many have noticed that win/lose ratio is still partially visible? There's this blue bar that represents total wins and losses, so if there's any consistency to it, we can determine a player's win/lose just by glancing at said bar. Example of 0% win and 100% win. + Show Spoiler +(You can add me on EU server to verify that I my win/lose ratios are in fact 0% and 100%, character code is 938) Hey, I;ve checked a few profiles and i think the bars are actually representing the % of wins per leage.. For example: 5 1v1 wins and 5 2v2 wins means they are both 50%
I can't figure out where it's pulling the data anymore since they removed the "total games" counter, because now you get situations like this: http://us.battle.net/sc2/en/profile/1251071/1/sjArtosis/ That's a perfect 50% ratio but the green bar is the same length as the blue one.
|
On April 07 2011 00:23 [F_]aths wrote:
Now let's say you had a bad day, you are a bit tired and afraid of playing 1v1 since you could lose a lot of games today. You could decide to rather not play to keep your 51% win statistic. Okay, a week or two weeks later you feel great but you are afraid that the strategies developed further and you would get crushed.
This actually happened to me (or a very similar scenario) and then I got hooked on another game for a while and now it's been like a month or two and it's definitely nerve-wracking to try to come back to playing regularly. What if I can't adjust to the way the game has developed, or the new maps? Overall I think the way the ladder is will probably help me get over this hump.
|
On April 18 2011 23:48 enzym wrote:
Let me explain what I meant here and why I expressed myself like I did.
Your post assumes that there can be no other meaning to the w/l statistic for other people than the one which you can see for yourself and have addressed in that post, despite several people in this thread explaining why they would like to see this statistic for different reasons. I have been ignorant about it because I consider it a matter of facts, not of opinions.
Since Blizzard designed the AMM (automatic match making) to give you a fair match, it tries to give you an opponent you can beat with a 50% chance. Deviations in win ratio from 50% only prove that the AMM cannot always be 100% correct in estimating the player's skill. The more games you play, the more data the AMM has about you and the more correct it can work. As I checked via sc2ranks.com the ratio of players with many games, I found no unusual high deviations, save for very top-players or inactive / extremely poor players.
I admit that my tone is may be a bit more aggressive/ignorant than it should be. However I see no real point to continue the discussion about win ratio meaning until someone proves that the aforementioned argument is wrong.
In this thread we had some discussions about short-term as well as long-term influence of certain actions regarding the win ratio. Overall I never saw a proof that skill can be measured in win ratio because this AMM is designed to get you to 50%.
Some users still think "here, i have a positive win ratio. I crush other people more than I get crushed.", implying "This has to say something". But it don't. If you buy a new account, smurf yourself to bronze league and intentionally lose 5-10 more games to get on a streak where you stomp 20 guys in a row, what does it prove?
If you not smurf, but somehow were considered too week comparing to your true skill and therefore have a positive ratio for some times (it eventually will even out unless you are really pro) what does it say about your skill?
If you were on a perfect 50% ratio but had a skill boost and stomp 10 or 20 guys with few losses inbetween, what does the >50% ratio tell about your absolute skill level?
Nothing. Someone in a higher league which currently is in a slump and has <50% win, will probably still crush you.
On April 18 2011 23:48 enzym wrote:
Here I have to disagree again. Many players like to practice vs AI for different reasons. Yes, you can use the AI mode as sandbox to practice build orders or other mechanical stuff. I meant to actually play against AI to beat it to prepare for multiplayer. Even if someone never played any RTS before, I would advise him to practice that in real games. He will lose a lot in the beginning, but eventually his MMR adapts. He also needs that low MMR to get points and ladder ranking later.
I play a few vs AI games for another reason – achievements. Completely useless for my skill but I still want to have some easy achievement points.
|
i think the league system as it is now, is pretty unnessesary because of that matchmaking rank you cant see, for whatever reason that might be
i, myself, am a bronze league player, atleast from what it shows to me, top 1 there with like 200 points more than the 2nd one, and still i dont get promoted to silver
now if i would have to play against bronze league players i could understand that, but if i check my opponents after the matches, they are atleast top10 silver, mostly gold and sometimes plat
i havent played against a bronze for ages, so what is the use of those leagues? and why cant you see how long it takes you to finally get promoted to the next one
its like, start of season 2, i made my placement stuff and got silver, thought i could go zerg for 5 matches, lost all of them, went down to bronze, went back to terran, won 22 out of 27 matches and have to stay bronze for the rest of my live
a mate of me instead got placed in silver, lost 1 match won 2 got promoted to gold, i think thats pretty random for a "system"
so 1 lose 2 wins is more than 5 lose 22 wins for that league thingy?
one thing i would like to see is having all bronze players in 1 division and stuff, so you can really see how good you are compared to the rest of the league, 100 players each division is kinda poor and doesnt show you anything
but well i guess even tho the ladder system has some problems, its probably a little cooler than just playing random games without any kind of points to get
|
On April 19 2011 02:12 [F_]aths wrote:Show nested quote +On April 18 2011 23:48 enzym wrote:
Let me explain what I meant here and why I expressed myself like I did.
Your post assumes that there can be no other meaning to the w/l statistic for other people than the one which you can see for yourself and have addressed in that post, despite several people in this thread explaining why they would like to see this statistic for different reasons. I have been ignorant about it because I consider it a matter of facts, not of opinions. Since Blizzard designed the AMM (automatic match making) to give you a fair match, it tries to give you an opponent you can beat with a 50% chance. Deviations in win ratio from 50% only prove that the AMM cannot always be 100% correct in estimating the player's skill. The more games you play, the more data the AMM has about you and the more correct it can work. As I checked via sc2ranks.com the ratio of players with many games, I found no unusual high deviations, save for very top-players or inactive / extremely poor players. I admit that my tone is may be a bit more aggressive/ignorant than it should be. However I see no real point to continue the discussion about win ratio meaning until someone proves that the aforementioned argument is wrong. In this thread we had some discussions about short-term as well as long-term influence of certain actions regarding the win ratio. Overall I never saw a proof that skill can be measured in win ratio because this AMM is designed to get you to 50%. Some users still think "here, i have a positive win ratio. I crush other people more than I get crushed.", implying "This has to say something". But it don't. If you buy a new account, smurf yourself to bronze league and intentionally lose 5-10 more games to get on a streak where you stomp 20 guys in a row, what does it prove? If you not smurf, but somehow were considered too week comparing to your true skill and therefore have a positive ratio for some times (it eventually will even out unless you are really pro) what does it say about your skill? If you were on a perfect 50% ratio but had a skill boost and stomp 10 or 20 guys with few losses inbetween, what does the >50% ratio tell about your absolute skill level? Nothing. Someone in a higher league which currently is in a slump and has <50% win, will probably still crush you.
I like to base arguments around facts rather than opinions as well. Opinions are then built upon these arguments.
I will try to explain how I see the situation according to the above mentioned principle.
Regarding the issue at hand there are at least two camps.
Camp 1 is of the opinion that the only meaning of the w/l statistic lies with its statistical meaning, with its statistical significance. Its statistical significance is a fact which is, as I see it, uncontested. It is a fact that w/l ratio loses significance the less close you are to either the very top or the bottom of the ladder and the more games you play. You are part of that camp.
Camp 2 is of the opinion that there can be other meaningul ways to interpret w/l ratio, that there are valid reasons to value the visibility of it other than for its statistical significance. Reasons by which this statistic has meaning for people were given in this thread:
-as an indicator of proximity to a stabilized MMR (how far are you from approaching your true, current skill and is it going up or down)
-as an indicator of the perceived Blizzard opinion about the use/function of the ladder and its representation of certain ideological values (competitiveness, honesty/transparency, completeness, others)
I am part of that camp.
There are indisputable facts which are made into arguments and around which opinions are based.
You are of the opinion that the meaning of w/l lies with its statistical significance. That is a valid opinion.
I, and others, are of the opinion that there can be and in fact are other reasons for people to value the visibility of that statistic. This is a valid opinion.
I cannot force you to find meaning in it for yourself in the same way we do. I acknowledge that your opinion is different from my own but is valid.
You cannot force others to find meaning in it for themselves only in the way you do. You cannot argue that there is only one meaningful way to interpret w/l, because evidently as by this thread there are people who do find it meaningful in other ways. This is a fact. You should acknowledge that our opinion is different from your own but is valid.
On April 19 2011 02:12 [F_]aths wrote:Show nested quote +On April 18 2011 23:48 enzym wrote:
Here I have to disagree again. Many players like to practice vs AI for different reasons. Yes, you can use the AI mode as sandbox to practice build orders or other mechanical stuff. I meant to actually play against AI to beat it to prepare for multiplayer. Even if someone never played any RTS before, I would advise him to practice that in real games. He will lose a lot in the beginning, but eventually his MMR adapts. He also needs that low MMR to get points and ladder ranking later.
I understand what you mean, but doubt that the differentiation between "practicing against the challenge which the AI poses" and "practicing AI unrelated things in a game with an AI" is useful in this context. Both of them have the AI in the game and both of them prepare you for multiplayer play.
Of course you can't practice matchup strategies against the AI other than the way LiquidTyler does it. It is possible that I am missing the forest through the trees here though. Right now it seems strange to me.
|
On April 18 2011 21:11 Timm wrote:
IMO this is the flaw with the current division system: The points do not actually reflect how high up you are in your division (in other words your hidden MMR). In platinum and lower it essentially means how well you where able to deplete your bonus pool which is a matter of playing enough games.
If you start winning more you'll eventually move up, so you cant really focus on becomming first in your division (< masters).. Yes the current system has several flaws. But one can approximate the current rating with the average of the opponents (regardless if one won or lost those games.)
If one often gets opponents from a higher league, it is probable that the MMR is comparable to the MMR of a higher division tier within that league. I see how I am improve with the opponents I get. Even with the tier obfuscation it is possible to roughly guess where I stand.
You need just to win a game per day (roughly 7 per week or just play 14 games per week if we assume 50% win ratio) to use up the bonus pool. The bonus pool advantage awards an active player a free win every day for the ladder ranking (of course not for MMR skill ranking.) This is not too much I think.
Excalibur somewhere mentioned the league mindset. To somewhat combat this, Blizzard introduced the division rankings with top-8 and so on but of course this is still inaccurate because we don't know the division modifier. But it does give anyone something to work for. Even if you suck, Top-50 can be achieved with activity and you can tell yourself that just 50% of the players will be ranked Top-50. It still does not really describes the skill but it is a goal one can work for.
|
On April 18 2011 23:48 enzym wrote:
I understand what you mean, but doubt that the differentiation between "practicing against the challenge which the AI poses" and "practicing AI unrelated things in a game with an AI" is useful in this context. Both of them have the AI in the game and both of them prepare you for multiplayer play.
Of course you can't practice matchup strategies against the AI other than the way LiquidTyler does it. It is possible that I am missing the forest through the trees here though. Right now it seems strange to me. Sandbox play like versus AI is of course good to practice build orders since you can setup a sandbox game fast and with no need of having another player available.
While I don't think one should play too much versus AI in the sense of playing full games versus AI and not just for mechanical practice, I LIKE that Blizzard gives the option even for teamplay versus AI. That is great for casual players who don't like the pressure of facing real humans as opponent. While those players will not improve very much, they still have fun. Blizzard even awards this mode with some achievements, making it one of the officially approved ways to enjoy starcraft. Elitist gamers may be look down to vs AI players but in my opinion, it's upon them to decide how to experience the game.
On April 19 2011 03:04 enzym wrote:
Camp 2 is of the opinion that there can be other meaningul ways to interpret w/l ratio, that there are valid reasons to value the visibility of it other than for its statistical significance. Reasons by which this statistic has meaning for people were given in this thread:
-as an indicator of proximity to a stabilized MMR (how far are you from approaching your true, current skill and is it going up or down) To see if your MMR stabilized, you can check your match history and check win/loss ratio for the last 50 or so games. You don't need the total win ratio. The total win ratio will get you in the wrong mindset. Lets say you have 49% and are afraid of getting even lower. So you play less ladder overall. This is not how ladder should work.
There is also more statistical noise than one initially assumes. The average distance of a 4-streak series (wins or losses) is just 30 games. You will have winning or losing streaks even with no skill improvement or degradation. How can you know that you lost 5 or 6 games in a row is due to being in a slump or due to random noise? You cannot know, you can just assume. You can still have your assumptions without an overall loss counter.
But you can compare the opponent's level you get. This gives you an idea about your true standing. It also gives you an idea about your skill plateauing or not. This is meaningful information. The win ratio is not.
There are even more factors. Assume you are good versus terran but bad versus protoss. That means you lose MMR rating because of your losses versus protoss and therefore get terrans below your overall skill. If you by coincidence gets 6 terrans in 10 games, you will probably win a lot. But this says nothing about your overall skill.
On April 19 2011 03:04 enzym wrote:
-as an indicator of the perceived Blizzard opinion about the use/function of the ladder and its representation of certain ideological values (competitiveness, honesty/transparency, completeness, others)
Originally I was against win ratio removal, too; even though I already knew that it is meaningless. I felt that Blizzard takes something away from me. Even though it is not useful, I don't like a patch where I have less afterward than I had before. omgwtfbbq, just let it how it was! That were my thoughts.
But are those toughts right? A win ratio display does not do anything good for Diamond or below. While some people would consider it more competitive, there would be less ladder games overall. Ladder match search therefore would be longer and less accurate. This should not be a design goal.
It is no sign of honesty to show the win ratio when whe win ratio is meaningless by the very design of the AMM. It is also not more complete when irrelevant data is shown. Having data visible which leads you into the trap that it would mean something while in fact it does not, leads to a wrong mindset.
Therefore it is acceptable to remove irrelevant data.
With few games played, the expected deviation of the data is high of course. With more games played, you get closer to 50% regardless no matter what (unless you are really pro.) But the absolute distance from exact 50% will still rise. Since few people have the statistical education to understand those data, it is good to hide it. Otherwise they will put meaning in what is nothing than random noise. Showing statistical noise has nothing to do with ladder transparency. It distracts from real skill improvement since you put your thoughts on the ratio while you should not.
|
On April 19 2011 02:50 xRivoNx wrote:
i think the league system as it is now, is pretty unnessesary because of that matchmaking rank you cant see, for whatever reason that might be i, myself, am a bronze league player, atleast from what it shows to me, top 1 there with like 200 points more than the 2nd one, and still i dont get promoted to silver Because you are not winning versus silver players consistent enough. Either you don't get much silver opponents in the first place, or you are still too much losing versus bronze players. Because you are active, you gets the full bonus pool advantage and can be top in your division.
This is one of the things which confuses people, that especially in bronze with a big lead in points, you are not necessarily eligible for league promotion. Blizzard should explain the ladder a bit better. Excalibur's ladder guide explains it very good. It's a shame that the community must do the work for Blizzard.
You say that you sometimes play versus "top" silver players. Those players could from the lowest silver tier (there are 3 tiers in silver which one cannot see however) and have the issue that they just win versus bronze players like you. Since they are active, they are ranked quite high.
On April 19 2011 02:50 xRivoNx wrote:
its like, start of season 2, i made my placement stuff and got silver, thought i could go zerg for 5 matches, lost all of them, went down to bronze, went back to terran, won 22 out of 27 matches and have to stay bronze for the rest of my live No-one has to, because there are always at least 20% more lazy than you. (So by definition of bronze, 20% have to stay in there, but I guess that more than 20% of the active players don't really want to improve.)
Coming myself from gold, you can believe me that I know from first-hand experience how it is to feel to be a bronze league hero. But it is a lesson to learn. You are free of any angst that you could get demoted further. You suddenly can focus on the game without worrying to lose because you have to experiment a lot.
On April 19 2011 02:50 xRivoNx wrote:
a mate of me instead got placed in silver, lost 1 match won 2 got promoted to gold, i think thats pretty random for a "system"
so 1 lose 2 wins is more than 5 lose 22 wins for that league thingy? It is not how much you win (hence the meaningless of win ratio.) It's about against who you win. If your friend lost 1 to gold but also won 2 to gold, he belongs in gold league.
On April 19 2011 02:50 xRivoNx wrote:
one thing i would like to see is having all bronze players in 1 division and stuff, so you can really see how good you are compared to the rest of the league, 100 players each division is kinda poor and doesnt show you anything It is even worse because Excalibur found out that bronze consists of 7 skill tiers. So you hardly can know your real standing within bronze. But you should check the opponents you get (regardless if you won or lost.) When you get only bronze players, you have to improve a lot to finally get promoted. If you win versus Silver and you don't get promoted, you win only versus sucking silver players which are close to be demoted to bronze or you are not playing consistent enough yet (meaning that you still lose too much to bronze players.)
|
On April 19 2011 04:04 [F_]aths wrote:Show nested quote +On April 19 2011 03:04 enzym wrote:
Camp 2 is of the opinion that there can be other meaningul ways to interpret w/l ratio, that there are valid reasons to value the visibility of it other than for its statistical significance. Reasons by which this statistic has meaning for people were given in this thread:
-as an indicator of proximity to a stabilized MMR (how far are you from approaching your true, current skill and is it going up or down) To see if your MMR stabilized, you can check your match history and check win/loss ratio for the last 50 or so games. You don't need the total win ratio. The total win ratio will get you in the wrong mindset. Lets say you have 49% and are afraid of getting even lower. So you play less ladder overall. This is not how ladder should work. There is also more statistical noise than one initially assumes. The average distance of a 4-streak series (wins or losses) is just 30 games. You will have winning or losing streaks even with no skill improvement or degradation. How can you know that you lost 5 or 6 games in a row is due to being in a slump or due to random noise? You cannot know, you can just assume. You can still have your assumptions without an overall loss counter. But you can compare the opponent's level you get. This gives you an idea about your true standing. It also gives you an idea about your skill plateauing or not. This is meaningful information. The win ratio is not. There are even more factors. Assume you are good versus terran but bad versus protoss. That means you lose MMR rating because of your losses versus protoss and therefore get terrans below your overall skill. If you by coincidence gets 6 terrans in 10 games, you will probably win a lot. But this says nothing about your overall skill.
None of this invalidates the opinion against which you are arguing. This is the case because you base your counter-argument on very specific scenarios instead of basing it on all of them.
You speak of streaks of 5 or 6 games, but Camp 2 never defined any streak length to base their argument on. Statistical expressiveness of streaks increases with growing number of wins/losses making up that streak, and the validity of Camp 2's argument increases with it. Larger streaks can occur when players return to the ladder after a break in which their skill either deteriorated or increased, while their MMR did not get any of that information.
Yes, displaying streaks would be more meaningful in that regard than displaying w/l, but gauging it by w/l is still better than not being able to gauge it at all, at least according to Camp 2.
Yes, you can check your match history for that. Or you can express a desire to have a dedicated stat displaying it better (w/l, streak).
Yes, differing proficiency in different matchups reduces the meaningfulness of overall MMR for gauging a trend. But gauging a trend by overall MMR is preferable to not being able to gauge it at all, according to Camp 2. Furthermore the matchup breakdown argument calls for the inclusion of more/better statistics (matchup specific ones), like some people in this thread have asked for, instead of the removal of information, which you are defending.
Lastly, the claim that w/l will get you into the wrong mindset is purely anecdotal and is not applicable to everyone. People will stop playing for fear of losing w/l, and people will play more because of a desire to restore or achieve a better w/l, just like you pointed out in a different thread: "Yesterday I needed 8 games to get mere six points to get back my old division placement. If I would have fail at that I just would play more games today. I think I will play today anyway even if I risk my rank."
On April 19 2011 04:04 [F_]aths wrote:Show nested quote +On April 19 2011 03:04 enzym wrote:
-as an indicator of the perceived Blizzard opinion about the use/function of the ladder and its representation of certain ideological values (competitiveness, honesty/transparency, completeness, others) Originally I was against win ratio removal, too; even though I already knew that it is meaningless. A win ratio display does not do anything good for Diamond or below. While some people would consider it more competitive, there would be less ladder games overall. Ladder match search therefore would be longer and less accurate. This should not be a design goal. It is no sign of honesty to show the win ratio when whe win ratio is meaningless by the very design of the AMM. It is also not more complete when irrelevant data is shown. Having data visible which leads you into the trap that it would mean something while in fact it does not, leads to a wrong mindset. Therefore it is acceptable to remove irrelevant data. With few games played, the expected deviation of the data is high of course. With more games played, you get closer to 50% regardless no matter what (unless you are really pro.) But the absolute distance from exact 50% will still rise. Since few people have the statistical education to understand those data, it is good to hide it. Otherwise they will put meaning in what is nothing than random noise. Showing statistical noise has nothing to do with ladder transparency. It distracts from real skill improvement since you put your thoughts on the ratio while you should not.
How can it be more honest to provide less information?
How can the removal of a statistic lead to increased completeness of the statistical framework?
Lastly, the claim that there would be less ladder games overall is purely anecdotal again.
Even if it was true it does not invalidate the argument presented by Camp 2.
On top of that plenty of ladders in plenty of games have seen great activity despite having w/l and other statistics displayed.
To complete the round of doubt: What if the psychological effect of w/l on ladder activity is true, but only acts on ladder activity and anxiety in the short term while shifting that fear level and orientating it towards a different reason in the long term? I.e., assuming that this psychological effect is real and significant, what if that fear shifts from an orientation to w/l to an orientation to a different stat which will then replace it?
I don't see how you are going to resolve this issue to beyond it being an issue of opinion vs opinion, which means that it will be hard to defend the removal of that statistic over an option which would allow people to choose for themselves.
|
On April 19 2011 01:44 Excalibur_Z wrote:Show nested quote +On April 18 2011 21:19 Timm wrote:On April 18 2011 20:29 Blurb wrote:How many have noticed that win/lose ratio is still partially visible? There's this blue bar that represents total wins and losses, so if there's any consistency to it, we can determine a player's win/lose just by glancing at said bar. Example of 0% win and 100% win. + Show Spoiler +(You can add me on EU server to verify that I my win/lose ratios are in fact 0% and 100%, character code is 938) Hey, I;ve checked a few profiles and i think the bars are actually representing the % of wins per leage.. For example: 5 1v1 wins and 5 2v2 wins means they are both 50% I can't figure out where it's pulling the data anymore since they removed the "total games" counter, because now you get situations like this: http://us.battle.net/sc2/en/profile/1251071/1/sjArtosis/ That's a perfect 50% ratio but the green bar is the same length as the blue one.
That does look strange indeed. However, when you look here: http://us.battle.net/sc2/en/profile/1251071/1/sjArtosis/ladder/ it turns out he does in fact have equal wins an losses.
So, on the profile overview page the bar represents wins, but the label next to it represents total games. Could be a bug for the masters league, since in all other leagues the bar and the label both represent wins..
|
On April 19 2011 03:31 [F_]aths wrote:Show nested quote +On April 18 2011 21:11 Timm wrote:
IMO this is the flaw with the current division system: The points do not actually reflect how high up you are in your division (in other words your hidden MMR). In platinum and lower it essentially means how well you where able to deplete your bonus pool which is a matter of playing enough games.
If you start winning more you'll eventually move up, so you cant really focus on becomming first in your division (< masters).. Yes the current system has several flaws. But one can approximate the current rating with the average of the opponents (regardless if one won or lost those games.) If one often gets opponents from a higher league, it is probable that the MMR is comparable to the MMR of a higher division tier within that league. I see how I am improve with the opponents I get. Even with the tier obfuscation it is possible to roughly guess where I stand. You need just to win a game per day (roughly 7 per week or just play 14 games per week if we assume 50% win ratio) to use up the bonus pool. The bonus pool advantage awards an active player a free win every day for the ladder ranking (of course not for MMR skill ranking.) This is not too much I think. Excalibur somewhere mentioned the league mindset. To somewhat combat this, Blizzard introduced the division rankings with top-8 and so on but of course this is still inaccurate because we don't know the division modifier. But it does give anyone something to work for. Even if you suck, Top-50 can be achieved with activity and you can tell yourself that just 50% of the players will be ranked Top-50. It still does not really describes the skill but it is a goal one can work for.
Yeah, I try to find my real skill level as well by looking at opponent levels, and try to end as high as possible in my league.
The funny thing is though that the people that are at the top of for example platinum with a big amounth of games are just not better than people in platinum with less games.The people with fewer games maybe just didnt reach their true skill level yet.
Meanwhile everybody is talking about being high platinum or whatever like it means something.
|
United States22883 Posts
Honestly, I agree with the OP. The ladder does exactly what it should do, from Blizzard's end. There are some faults, but they're not particularly relevant to the vast majority of people using the ladder, and those are the people Blizzard really needs to sell the multiplayer aspects to. Pro gamers don't need a good ladder from Blizzard in order to find games.
That said, I don't think the ladder is a very good practice tool and it's unfortunate that it carries all the weight when Blizzard picks candidates for their tournaments. The random match up and map selection means that you really can't refine your play particularly well. You can't keep track of subtle adjustments to a build or aggression, because similar matches are often so far apart.
It can be useful for basics or as a warm up, but with hardcore improvement as the goal, rather than winning as the goal, Custom games are a much better system, and it's a shame we don't have iCCup style channels to find games.
|
I hate that losses were removed. I now can't calculate my win/loss ratio. If they wanted to appeal to the noobs, they should leave an option to display losses or not, and be selective of who to display losses to. Such as, Bob is ashamed of his losses, so he decides to opt into not showing them to others, but Rick over here isn't entirely proud of his losses but would like to see his progression, so he opts to see his losses but not share them. Jeff over here wants people to see his win/loss ratio, and have his friends check out his progression through his win/loss ratio changing. That is how I'd like it, I personally wouldn't show my losses to others until I'm above 50% which I am sure I am at this point, because in season 1 I was at 60% or higher by the time it was done. But still trapped in bronze, facing silvers and golds without a huge loss if I do lose...this is where the Blizzard matchmaking system fails hard. You can beat people way better than you, and still not progress because you may have done something stupid (like me) and skipped practice league out of boredom, considering that it's played at "normal" speed. I didn't get into WC3 because of how slowly it was played at.
|
On April 19 2011 03:04 enzym wrote:
-as an indicator of proximity to a stabilized MMR (how far are you from approaching your true, current skill and is it going up or down)
This isn't an accurate interpretation of win/loss ratios, because the matchmaking system doesn't try to stabilize a long-term win/loss ratio at 50% -- it's looking to find individual player matchups with a 50/50 likelihood of each player winning. A high or low win/loss ratio can simply memorialize some MMR excursion from the past when current games are 50/50. (Eventually, if the total number of games greatly exceeds the length of such an excursion, the long-term w/l will approach 50%, but the system isn't trying to make that happen faster than it would by optimizing the short term likelihood of a win.)
-as an indicator of the perceived Blizzard opinion about the use/function of the ladder and its representation of certain ideological values (competitiveness, honesty/transparency, completeness, others)
The only "ideology" involved in a system that strives for a 50/50 likelihood of a win in matching is that individual games shouldn't be one-sided if it can be avoided. Beyond that (honesty? really?), you're talking nonsense, sorry.
|
On April 19 2011 17:04 Lysenko wrote:Show nested quote +On April 19 2011 03:04 enzym wrote:
-as an indicator of proximity to a stabilized MMR (how far are you from approaching your true, current skill and is it going up or down) This isn't an accurate interpretation of win/loss ratios, because the matchmaking system doesn't try to stabilize a long-term win/loss ratio at 50% -- it's looking to find individual player matchups with a 50/50 likelihood of each player winning. A high or low win/loss ratio can simply memorialize some MMR excursion from the past when current games are 50/50. (Eventually, if the total number of games greatly exceeds the length of such an excursion, the long-term w/l will approach 50%, but the system isn't trying to make that happen faster than it would by optimizing the short term likelihood of a win.) Show nested quote +-as an indicator of the perceived Blizzard opinion about the use/function of the ladder and its representation of certain ideological values (competitiveness, honesty/transparency, completeness, others) The only "ideology" involved in a system that strives for a 50/50 likelihood of a win in matching is that individual games shouldn't be one-sided if it can be avoided. Beyond that (honesty? really?), you're talking nonsense, sorry.
How am I talking nonsense exactly? I wasn't sure how else to call it. But people perceive the removal of that statistic as being part of Blizzard's catering to less competitive, more "fragile" and more casual players by obfuscating an accurate display of your match history and derived stats. That counts as dishonesty to me. It counts as promoting feel good aspects over the competitive ones. And removing the display of a stat makes the framework of available information less complete.
Ideological is probably not the right word, but I don't know how else to call it and honesty, completeness and competition do represent virtues/values to some people.
Regarding the first part - you already acknowledged that it will get closer to 50/50 the more games you play, and I already addressed the issue of its inaccuracy in a post right befor yours. So I don't know why you would mention it again.
|
On April 19 2011 05:00 enzym wrote:
You speak of streaks of 5 or 6 games, but Camp 2 never defined any streak length to base their argument on. Statistical expressiveness of streaks increases with growing number of wins/losses making up that streak, and the validity of Camp 2's argument increases with it. Larger streaks can occur when players return to the ladder after a break in which their skill either deteriorated or increased, while their MMR did not get any of that information. Some weeks ago I had a streak of 9 or 10 games won (1v1 ladder) in a row. This can happen. Day[9] once told he had lost 14 games in a row, and he is a pro. There is less meaning in quite large streaks than one would normally assume.
To combat the uncertainty of a player's skill when he was inactive for a while, it is believed that the confidence interval for the MMR is automatically widened. This means you get opponents from a greater skill range to allow the MMR to adapt faster. Once you get back from a larger break, the win or loss streak should not be too large.
On April 19 2011 05:00 enzym wrote:
Yes, displaying streaks would be more meaningful in that regard than displaying w/l, but gauging it by w/l is still better than not being able to gauge it at all, at least according to Camp 2. At this point, this camp is wrong. W/L counter should not be compared to estimate improvement. The quality of the opponents you get provides a real hint about the current MMR ranking.
If we assume a theoretically perfectly working AMM (which of course can never be made) your win chance will be 50% every game. The AMM is not perfect of course. This means, deviations from 50% winratio are signs of an imperfect AMM.
The trick is to provide 50% for any current match. If you would get a pro when you are >50% and a noob when you are <50%, the ratio would be very close to 50% but the games would not be much fun because you either get stomped or you get a victim instead of a worthy opponent. Since the AMM wants to provide 50% chance for any new game, there are some side effects which can lead to deviations from a 50% win ratio. But those side effect don't inform you about your skill. It only informs you how good the AMM works.
Getting more wins than losses can even delay your league promotion. Since you win so much, you get opponents which are not strong enough. What you want is to get about 50% because then you get opponents of your true skill. If you want to get up to gold, you should not be happy to bash many silver players in a row. What you want is that the AMM throws the gold players at you since only this qualifies you for the promotion.
Removing information which causes more trouble than insight is a good thing I guess. While there are some players who still don't like the change, overall I see more players voicing that they play more once the loss counter was removed. This alone justifies it.
One cannot expect from everyone to get a competitive mindset just because he wants to play Starcraft multiplayer matches. There is a thread on TL about being nervous when searching a ladder match. Many players admit that they still are nervous. With the removal of the loss counter, the ladder takes some pressure away.
In this thread, I only heard people who stated that they play more now. No-one said that he play less since he cannot see his loss counter.
|
On April 19 2011 16:42 Creegz wrote:Such as, Bob is ashamed of his losses, so he decides to opt into not showing them to others, but Rick over here isn't entirely proud of his losses but would like to see his progression, so he opts
To bash bob for not man-ing up and showing losses. This is the life of the internet.
In a sense, i like what alot of what OP wrote because it applies to me, a guy who bought sc2 day one, threw 20 matches in, then havent touched 1v1 since.
I, myself, honestly want all leagues besides diamond+ to be toned down a bit. Not that placing gold was bad or anything, i hate having to enter a "tryhard" mode to win games for "fun"
Little irrelevant steps in the way dont bother me, ill be unphased by the removal of losses, but i enjoy the lightheartedness that comes from it.
When i go ladder, i find myself to 4gates, timed zerged rushes, proxy terran buildings with abusive pronged attacks at the 8 min mark.
Not that its hard to deal with, and a bit of effort i could leave this, I just want to play to enjoy.
I wanna play long macro games, test unit comps, enjoy myself for 15-30 min a game but the people around me look for the fastest win to the next match.
Hopefully everyone lightens up so i can get some care-free matches in.
|
On April 19 2011 18:02 [F_]aths wrote:
Removing information which causes more trouble than insight is a good thing I guess. While there are some players who still don't like the change, overall I see more players voicing that they play more once the loss counter was removed. This alone justifies it.
[...]
In this thread, I only heard people who stated that they play more now. No-one said that he play less since he cannot see his loss counter.
After all your knowledge about the AMM and about statistics, one would expect that you are aware of the ambiguity of such a display. Seeing more people stating that they play more because of this change can have several reasons, only one of which is an actual overall increase in ladder activity.
It could also appear this way because of a perceptual bias (you pay more attention to them), it could be because these people are somehow more likely to express their opinion or because they are overrepresented on this website, forum or thread in general.
But most of all, even if a majority of people was happy about that change, it still doesn't mean that removing w/l is a better choice than providing an option, because without the option you are screwing other players over, as evidenced by this thread.
I know that I am showing negligible activity on the SC2 ladder and that the meaninglessness of that ladder with all its shenanigans is one reason for it. I did not state that earlier though, because it does not contribute to the discussion and because it has one additional reason (albeit one that is linked to it) as well.
On April 19 2011 18:02 [F_]aths wrote:
One cannot expect from everyone to get a competitive mindset just because he wants to play Starcraft multiplayer matches. There is a thread on TL about being nervous when searching a ladder match. Many players admit that they still are nervous. With the removal of the loss counter, the ladder takes some pressure away.
Nobody is expecting a competitive mindset from everyone. That's why most people who are opposing the removal of w/l are advocating an option to hide/show it.
|
|
|
|
|
|
EPT![[image loading]](http://imgur.com/kPKLu.png)
