|
Now that we finally have unranked games (Which we don't have win/loss games shown), do you think it's ok to have win/loss back into the ranked games?
Personally I had quite a strong detest seeing my own win/loss ratio and didn't start grinding games until it was removed.
but right now, it's back and I still feel kinda uneasy looking at it and this is before the game is public (right now it's beta and so it's kinda ok to just win and lose some games- my mindset)
So how do the rest of you feel?
Poll: Win/Loss is BACK!YAY (370) 92% I am not sure how I feel about this.. (19) 5% NAY (15) 4% 404 total votes Your vote: Win/Loss is BACK! (Vote): YAY (Vote): NAY (Vote): I am not sure how I feel about this..
|
I am not sure how I feel about this. I dont want other people to look at my win/lose ratio. If only I can see my own then YAY
|
Overall win/loss values are absolutely useless; due to matchmaking, your overall win/loss is always going to trend towards 50/50 unless you are at the very extremes of the top or bottom of the ladder. If you want to see how many wins you have right now, just take your total matches played and halve it and you'll be pretty much spot on. Individual race matchup statistics or map winrate statistics hold value, but overall winrate is irrelevant. It's 50%!
To elaborate on this, it doesn't even matter what league you're in. Silver players will have a 50% win ratio. Gold players will have a 50% win ratio. Diamond players will have a 50% win ratio. Unless you're at the very top or bottom of ladder (or have played very few games), your win ratio is ~50%. If you're looking for a metric by which to measure your skill or progress, you would be hard-pushed to find something worse.
EDIT: just glanced at my sc2gears history since my harddrive exploded; Guess what my win ratio is out of 229 games? + Show Spoiler +
|
On November 23 2012 18:13 Hairy wrote: Unless you can see individual race matchup statistics, or map winrate statistics, this data is of no value.
You can. There's a new stats page to my knowledge.
|
OP, you are literally the only person on TL who detests seeing W/L, you are a minority in this forum.
|
On November 23 2012 18:15 DemigodcelpH wrote:Show nested quote +On November 23 2012 18:13 Hairy wrote: Unless you can see individual race matchup statistics, or map winrate statistics, this data is of no value. You can. There's a new stats page to my knowledge.
Yup there is, shows your win % on map, your win % vs a race to. Really glad they added those finally ^_^
|
100-fucking-% YAY.... Finally I don't have to fuck around checking my match history to get an idea of how I'm doing. It's so much better than WoL
|
Might as well, even though there is absolutely no reason to assume that your winrate will be anything but 50% (+-2) unless you're top 100 GM.
The only thing I use it for is to measure progress inbetween seasons.
But it's good to have it back for everyone
|
On November 23 2012 18:19 emc wrote: OP, you are literally the only person on TL who detests seeing W/L, you are a minority in this forum. there were tonnes of people who said the win/loss caused ladder anxiety. And that for sure happens to me when I go on a losing steak that just seems never ending.
|
Silly idea, the vast majority of the playerbase are casual gamers being able to see you win/loss ratio causes ladder anxietly which in return means less players will play Starcraft II -- which is exactly what Blizzard wanted to change with the removal.
Yes competetive players of all leagues wouldn't mind seeing it. But fact is the majority is casuals, which Blizzard obviously will favor in actually staying as active players in return they will cater to balance for the competive players as it'll make them stay too. Win win.
|
United Kingdom14464 Posts
is it an option to show it or remove it? It should be optional, but tbh the arguments about it on both sides are dumb, bar the top and bottom end it trends to 50%.
|
Well the poll says enough ^^ Sure if people want to remove it, it should be an option. I don't get why things arn't more variable. Customizing pleases so many more people.
|
The W/L ratio was brought back because players now have the ability to play unranked. So if you have ladder anxiety, you play unranked instead. No win count, no loss count, only your XP that goes up with every game.
|
On November 23 2012 18:13 Hairy wrote: Overall win/loss values are absolutely useless; It is odd how people keep claiming that win - loss statistics are meaningless. W-L-stats are not meaningless, but cannot directly be used to deduce the player level, but are a progress statistic. In simplest form they provide information how many games have been played, but can also e.g. be used to _roughly_ estimate what kind of changes there have been to your MMR in those games.
After a player reaches his typical MMR range and is stable on his all mu:s, win-loss ratio typically approaches 50% for games played _after_ that moment as Battle.net tries to provide similar level opponents for him. But Win-loss _difference_ (short term) still provides good short term progress information as you can roughly estimate based on it has your MMR dropped or risen lately. For every win your MMR increases and for every loss decreases (E.g. if your W-L difference is -10 for last week and you generalize that you faced always 'similar level' opponents, you deduce that your MMR has roughly dropped about half league MMR range (based on the MMR tool, MMR range for each league other than bronze & master is roughly 18 wins wide (wins against similar MMR level opponents))). Also by looking at other's win-lose differences it quickly tells if the player has risen or dropped (large difference) based on MMR (highlights especially MMR abusers quite nicely).
Win - loss difference is a meaningful progress stat, but you cannot deduce a player level based on it. Even in masters the system aims for 50% matchmaking when you have reached your typical level (it is not more meaningful in masters than the lower leagues). Also people rarely have exact 50 % ratio even if the system aims for it. The thin top of the ladder of course will get better than 50% ratios as the population in their level is too thin to get 'equal level' opponents all the time.
To get more accurate idea where the system ranks you with its hidden skill rating (MMR), it is recommended to try out the MMR tool, which provides quite accurate relative ratings.
|
On November 23 2012 18:13 Hairy wrote:Overall win/loss values are absolutely useless; due to matchmaking, your overall win/loss is always going to trend towards 50/50 unless you are at the very extremes of the top or bottom of the ladder. If you want to see how many wins you have right now, just take your total matches played and halve it and you'll be pretty much spot on. Individual race matchup statistics or map winrate statistics hold value, but overall winrate is irrelevant. It's 50%! To elaborate on this, it doesn't even matter what league you're in. Silver players will have a 50% win ratio. Gold players will have a 50% win ratio. Diamond players will have a 50% win ratio. Unless you're at the very top or bottom of ladder (or have played very few games), your win ratio is ~50%. If you're looking for a metric by which to measure your skill or progress, you would be hard-pushed to find something worse. EDIT: just glanced at my sc2gears history since my harddrive exploded; Guess what my win ratio is out of 229 games? + Show Spoiler +
This guy has it right. Seeing overall W/L is nice, and I'm not opposed to it, but in the end it doesn't really matter.
You can go 25-5 in Diamond league and feel like a fucking superstar, but since each win or loss affects your rating differently (beating a gold vs. beating a master), the 25-5 itself is useless as a metric.
Likewise, if you're a masters player who goes 2-10 vs. grandmaster players, your rating still might go up, as each win is worth much more than each loss.
Edit:
On November 24 2012 01:13 korona wrote:Show nested quote +On November 23 2012 18:13 Hairy wrote: Overall win/loss values are absolutely useless; It is odd how people keep claiming that win - loss statistics are meaningless. W-L-stats are not meaningless, but cannot directly be used to deduce the player level, but are a progress statistic. In simplest form they provide information how many games have been played, but can also e.g. be used to _roughly_ estimate what kind of changes there have been to your MMR in those games.
Then why not simply display it in match history rather than on the ladder board? Maybe add a league icon next to points gained and date played. The problem is that people use W/L as a scoring metric when looking at the ladder ("oh, I'm 5th in my division but I'm down 6 games so I must be doing bad"). Like you said, W/L can be used to track short-term progress, but it's useless for long-term.
|
On November 24 2012 01:13 korona wrote:Show nested quote +On November 23 2012 18:13 Hairy wrote: Overall win/loss values are absolutely useless; It is odd how people keep claiming that win - loss statistics are meaningless. W-L-stats are not meaningless, but cannot directly be used to deduce the player level, but are a progress statistic. In simplest form they provide information how many games have been played, but can also e.g. be used to _roughly_ estimate what kind of changes there have been to your MMR in those games. After a player reaches his typical MMR range and is stable on his all mu:s, win-loss ratio typically approaches 50% for games played _after_ that moment as Battle.net tries to provide similar level opponents for him. But Win-loss _difference_ (short term) still provides good short term progress information as you can roughly estimate based on it has your MMR dropped or risen lately. For every win your MMR increases and for every loss decreases (E.g. if your W-L difference is -10 for last week and you generalize that you faced always 'similar level' opponents, you deduce that your MMR has roughly dropped about half league MMR range (based on the MMR tool, MMR range for each league other than bronze & master is roughly 18 wins wide (wins against similar MMR level opponents))). Also by looking at other's win-lose differences it quickly tells if the player has risen or dropped (large difference) based on MMR (highlights especially MMR abusers quite nicely). Win - loss difference is a meaningful progress stat, but you cannot deduce a player level based on it. Even in masters the system aims for 50% matchmaking when you have reached your typical level (it is not more meaningful in masters than the lower leagues). Also people rarely have exact 50 % ratio even if the system aims for it. The thin top of the ladder of course will get better than 50% ratios as the population in their level is too thin to get 'equal level' opponents all the time. To get more accurate idea where the system ranks you with its hidden skill rating (MMR), it is recommended to try out the MMR tool, which provides quite accurate relative ratings.
So basically what you want is a graph showing your MMR over time? That I can understand, hell, it would even make sense. However, since W/L is a metric derived from a system that strives to keep the ratio as close to 50% as possible, you only really need to know games played per season. In fact, the more you play, the more accurate this system will be at pinning you on 50%. We dont want to know a single number, we want to know how that number relates to other numbers.
So in response to the OP: I don't care either way.
|
On November 24 2012 01:33 Fenris420 wrote:Show nested quote +On November 24 2012 01:13 korona wrote:On November 23 2012 18:13 Hairy wrote: Overall win/loss values are absolutely useless; It is odd how people keep claiming that win - loss statistics are meaningless. W-L-stats are not meaningless, but cannot directly be used to deduce the player level, but are a progress statistic. In simplest form they provide information how many games have been played, but can also e.g. be used to _roughly_ estimate what kind of changes there have been to your MMR in those games. After a player reaches his typical MMR range and is stable on his all mu:s, win-loss ratio typically approaches 50% for games played _after_ that moment as Battle.net tries to provide similar level opponents for him. But Win-loss _difference_ (short term) still provides good short term progress information as you can roughly estimate based on it has your MMR dropped or risen lately. For every win your MMR increases and for every loss decreases (E.g. if your W-L difference is -10 for last week and you generalize that you faced always 'similar level' opponents, you deduce that your MMR has roughly dropped about half league MMR range (based on the MMR tool, MMR range for each league other than bronze & master is roughly 18 wins wide (wins against similar MMR level opponents))). Also by looking at other's win-lose differences it quickly tells if the player has risen or dropped (large difference) based on MMR (highlights especially MMR abusers quite nicely). Win - loss difference is a meaningful progress stat, but you cannot deduce a player level based on it. Even in masters the system aims for 50% matchmaking when you have reached your typical level (it is not more meaningful in masters than the lower leagues). Also people rarely have exact 50 % ratio even if the system aims for it. The thin top of the ladder of course will get better than 50% ratios as the population in their level is too thin to get 'equal level' opponents all the time. To get more accurate idea where the system ranks you with its hidden skill rating (MMR), it is recommended to try out the MMR tool, which provides quite accurate relative ratings. So basically what you want is a graph showing your MMR over time? That I can understand, hell, it would even make sense. However, since W/L is a metric derived from a system that strives to keep the ratio as close to 50% as possible, you only really need to know games played per season. In fact, the more you play, the more accurate this system will be at pinning you on 50%. We dont want to know a single number, we want to know how that number relates to other numbers. So in response to the OP: I don't care either way. You look at only from one viewpoint like most who claim that win-loss stats are meaningless. Win-loss stats are not for deducing player skill level due to how the ladder system works (for more info check Excalibur_Z's guide). They are progress statistics that tell how many games you have played and distribution of the end results of those games. These numbers can also be used e.g. for the purpose I mentioned. The main problem for these numbers is that most seem to try deducing player skill based on them, which cannot be done only based on the raw numbers. Also many do not understand how the ladder system & matchmaking works. People also look at the percent ratio too much - The short term difference is much more interesting.
|
On November 24 2012 01:29 sevia wrote:Show nested quote +On November 23 2012 18:13 Hairy wrote:Overall win/loss values are absolutely useless; due to matchmaking, your overall win/loss is always going to trend towards 50/50 unless you are at the very extremes of the top or bottom of the ladder. If you want to see how many wins you have right now, just take your total matches played and halve it and you'll be pretty much spot on. Individual race matchup statistics or map winrate statistics hold value, but overall winrate is irrelevant. It's 50%! To elaborate on this, it doesn't even matter what league you're in. Silver players will have a 50% win ratio. Gold players will have a 50% win ratio. Diamond players will have a 50% win ratio. Unless you're at the very top or bottom of ladder (or have played very few games), your win ratio is ~50%. If you're looking for a metric by which to measure your skill or progress, you would be hard-pushed to find something worse. EDIT: just glanced at my sc2gears history since my harddrive exploded; Guess what my win ratio is out of 229 games? + Show Spoiler + This guy has it right. Seeing overall W/L is nice, and I'm not opposed to it, but in the end it doesn't really matter. You can go 25-5 in Diamond league and feel like a fucking superstar, but since each win or loss affects your rating differently (beating a gold vs. beating a master), the 25-5 itself is useless as a metric. His W-L difference is +20. We do not know his relative starting MMR level, but we can deduce that his MMR has risen most likely considerably during these games. Based on these numbers we do not know if he for example started in platinum league and was then promoted to diamond. Also if his MMR was already in the diamond range, now his MMR may already be in the master range (diamond MMR range is about 18 wins against similar MMR level opponents according to the MMR tool). Also we do not know if he has reached his 'typical' range & maintains it from where his win ratio usually approaches 50%. To make more accurate deductions you would have to examine other data such as his match history, his opponent's profiles & opponent's opponents. Or he could be e.g. using the MMR tool, from where he gets fairly accurate relative ratings and shares this data with you.
On November 24 2012 01:29 sevia wrote: Likewise, if you're a masters player who goes 2-10 vs. grandmaster players, your rating still might go up, as each win is worth much more than each loss. GM is a subset of master league. It is very hard to get dropped from GM, if somebody reaches it and it has limited amount of spots. Often there are many GM players who actually have lower MMR than many top master players as their MMR may have decreased during their time in GM. The matchmaker usually pairs people quite nicely (not too large MMR difference). Or course at the top of ladder where population is thin, it may have to pair people with large MMR difference much more often than e.g. in lower leagues, where there are lots of similar level people to choose from. Thus if you are at the top of the ladder, deductions based on the win-loss difference are less accurate (master league is huge regarding MMR range).
On November 24 2012 01:29 sevia wrote:Edit: Show nested quote +On November 24 2012 01:13 korona wrote:On November 23 2012 18:13 Hairy wrote: Overall win/loss values are absolutely useless; It is odd how people keep claiming that win - loss statistics are meaningless. W-L-stats are not meaningless, but cannot directly be used to deduce the player level, but are a progress statistic. In simplest form they provide information how many games have been played, but can also e.g. be used to _roughly_ estimate what kind of changes there have been to your MMR in those games. Then why not simply display it in match history rather than on the ladder board? Maybe add a league icon next to points gained and date played. The problem is that people use W/L as a scoring metric when looking at the ladder ("oh, I'm 5th in my division but I'm down 6 games so I must be doing bad"). Like you said, W/L can be used to track short-term progress, but it's useless for long-term. The current league does not tell too much about where a person is regarding his MMR (especially after the changes to the promotion system Blizzard made in the beginning of S9. The league icon only tells inside which league his MMR was when he was placed there). And if people try to use W-L numbers for wrong purposes, like directly deducing the skill, it does not make the numbers themselves useless. The numbers are still a progress statistics that can be used for other purposes. E.g. for understanding how many games you have played and how many you won. And by personal opinion these kind of basic statistics should always be available regardless how different people use them. If they are often used for wrong purposes, more advanced statistics should be provided and/or players be educated regarding the possible uses.
|
They are progress statistics that tell how many games you have played and distribution of the end results of those games.
What use would such a statistic be? I am assuming you are talking about the remaining games after you take (wins - losses). If the number is positive, the player has improved faster than the average player, and if negative the opposite. The bigger the number, the bigger the change since last season. However, since it does not show your MMR, you still don't know where the person was last season. They may have improved faster than their "peers" in a league below them, but it doesn't mean that their positive win ratio is of any interest to you.
[...] but can also e.g. be used to _roughly_ estimate what kind of changes there have been to your MMR in those games. and then
Win-loss stats are not for deducing player skill level due to how the ladder system works
Then what value is there in knowing you had a sudden win streak or an improving MMR from the last 10 games? If you are winning more, you are doing better. If you are doing better, you are a better player. The whole reason we have leagues is because we consider them divided by player skill. League placement is based on MMR and MMR is based on how many games you win and against whom you win.
The reason I don't like win rates as a statistic is that it only shows how much you win, but is blind to whom you win against. There is no context in which each win can be valued.
Also by looking at other's win-lose differences it quickly tells if the player has risen or dropped (large difference) based on MMR (highlights especially MMR abusers quite nicely).
Looking at players best league placement also highlights MMR abusers. Someone dropping 10 or even 15 games over a season is not necessarily significant if they played 1500 games. Especially not if 10 of those drops occured because they were playing on tilt last night. 15 negative w/l over an entire season is a lot more significant than recent fluctuations, as those will normalise. Not to mention the overall win/loss ratio for several seasons.
Then again, I never did care about win/loss ratios since I will need to become amuch better player in order to play against someone that I would value keeping statistics of. Either way, my point is that I think a graph and the actual MMR would be a much better tool for evaluating performance than simply adding losses, which is most likely not going to be all that surprising anyway.
|
To be honest, a simple feature like that is KEY to have. With all the new analysis tools as well, SC2 is really starting to turn out more awesome than it already was!
|
|
|
|