|
I was having a discussion with my peers about what stats should be tracked and shown to players in a given game, and the effects of those stats on player psychology.
The issue that ignited the biggest debate is whether or not losses should be tracked/shown to players.
For the purpose of the debate, we assume the following:
* The automatch system works properly (or, in other words, it constantly adjusts your rating after each game so that you should inevitably win 50% of your games)
* There is a big enough player pool to feed the automatch
* The automatch system cannot be abused
Given then rules above, the argument is this: Showing Losses only allows players to fixate on an unimportant statistic. If the automatcher rigs everyone’s stats to hover around 50%, it is delusional to judge someone’s skill based on their win/loss ratio. It diverts the viewer away from the true indication of skill: their skill rating.
Having the hardcore background that I have, it’s hard for me to disenchant myself with win percentage. If I was competitively participating in a game that did not track this statistic, I’d be really turned off.
What are your thoughts?
|
wtf of course it should track losses, it only negatively affects the psychology of pussies.
this isn't only true in BW but in anything competitive. imagine if track runners just ignored every race they lost like it didn't happen. thats not how you improve.
saying losses are an unimportant statistic is naive.
|
Its true that the win percentage will not be a good indicator for the middle 80% of players or so, who, as you say, should be near 50%, but at the highest (and lowest) levels it is still a good clue.
It is also useful to see the W/L ratio even for those in the middle, especially in the beginning of the season. for instance:
Player A is 41-43, mid gold league. We can surmise that they are an average gold league player.
Player B, with a similar rating to player A, is 16-7 in the same league. We can surmise that he will soon be moving up leagues, and is probably stronger than player A. With so few games and so little information, we can't know for sure, but our guesses seem reasonable, and we wouldn't have been able to make them without seeing the W-L record.
Something interesting I just noticed: When I was in silver league, I only lost 3-6 points when losing a game, and won 15+ when I won. In Gold, if you disregard Bonus Points, the amounts won or lost are even. So it seems they have taken a page fro iccup in this regard.
EDIT:
Nony is right, the higher the population in the ladder, the larger the percentage of players who should be around 50% win. They are probably already close to 95% - I think some Blizz employee said they felt that the matchmaking system didn't work for the lowest 3% of players.
|
Calgary25986 Posts
Comparing the rating to the win/loss ratio let's you perform a quick mental check about someone's skill. I will play a 15-0 C- differently than a 500-790 C- and you would be irrational to say otherwise.
When the league becomes completely fully developed, then yes, both wins and losses become obsolete statistics to rating.
|
8751 Posts
As a top player, I prefer win percentage because most top players don't hover around 50%. If the guy with the highest rating has 60% win and 2nd place has 70% win, then of course I think the 2nd place guy is performing much better. I know that only applies for a very small percentage of players on the ladder, but a similar concept comes into play when a player destined to have a high rating starts playing late in the season. His opponents should know that the guy they just lost to has a 90% win rate so it wasn't really a fair game, though the matchmaking system is doing the best it can to give fair games.
|
8751 Posts
On July 13 2010 03:20 HCastorp wrote:
Its true that the win percentage will not be a good indicator for the middle 80% of players or so, who, as you say, should be near 50%, but at the highest (and lowest) levels it is still a good clue.
If it's a decently populated ladder, then I think it should really be like the middle 95%+ people have around 50% win.
|
On July 13 2010 03:21 Chill wrote:
Comparing the rating to the win/loss ratio let's you perform a quick mental check about someone's skill. I will play a 15-0 C- differently than a 500-790 C- and you would be irrational to say otherwise.
When the league becomes completely fully developed, then yes, both wins and losses become obsolete statistics to rating.
I agree with this.
However: is the statistic being obsolete relevant? If the majority of players do not believe it's obsolete, then is more harm caused than prevented?
I guess that begs a list of pros. I'll update/edit when i have more than idle program-loading time.
|
Korea (South)17174 Posts
theres always the anomaly unreal gosu that climbs to the top of the ladder without losing, and everybody enjoys seeing this
to not show losses takes out alot of the fun
|
how is rekrul always right
he is that unreal gosu in life
|
United States12237 Posts
Hiding losses is about as relevant as showing wins, in the grand scheme of things. Win ratio is the most important thing, and if you don't have access to both your win and loss information, there's not much of a point. However, hiding losses has some psychological effects:
- No more people accruing losses on "loss accounts" (not really an issue in SC2 anyway)
- No more people starting new accounts when their win ratio dips (again not an issue in SC2)
Chill is right when he says that rating trumps everything in a more saturated ladder, and again I'm speaking of SC2. That's partially because win ratio isn't as heavily emphasized (you have to calculate it yourself), but also because rating is the only relevant metric.
|
I think just giving players the option of playing non-ladder/non-record basically fixes anyone with embarrassment issues. Kinda like how you can play people in UMS (obs maps or otherwise) and it doesn't count on your record.
I think it's a reasonable option, since people don't like to try new things when their all important record is on the line  Although, I think ladder is alienating anyway. I prefer playing friendly matches where people aren't keeping track most of the time, and then playing 2 or 3 important matches against a newcomer or in a show match for pride.
So basically I think as many stats as can be given about a player should be included in competitive scores... What races they use, how often they win with each, their apm, their game length with each matchup... etc etc. That kind of stuff is fun when analysing how good someone is (and not so fun when being analysed if you're a casual player lol)
|
United States12237 Posts
On July 13 2010 03:41 Chef wrote:I think just giving players the option of playing non-ladder/non-record basically fixes anyone with embarrassment issues. Kinda like how you can play people in UMS (obs maps or otherwise) and it doesn't count on your record. I think it's a reasonable option, since people don't like to try new things when their all important record is on the line Although, I think ladder is alienating anyway. I prefer playing friendly matches where people aren't keeping track most of the time, and then playing 2 or 3 important matches against a newcomer or in a show match for pride. So basically I think as many stats as can be given about a player should be included in competitive scores... What races they use, how often they win with each, their apm, their game length with each matchup... etc etc. That kind of stuff is fun when analysing how good someone is (and not so fun when being analysed if you're a casual player lol)
I'm fully expecting those kinds of statistics to be featured on the SC2 Battle.net site, just as many additional stats were shown on the War3 site. I wouldn't expect the in-game reports to be as detailed, though.
|
Ahh, interesting.
There are a couple of reasons why showing wins and losses could be a bad or good thing. I am not sure on percentages, but the majority of gamers are casual, if that, while a smaller majority of the community are active "hardcore" gamers.
From a hard core's perspective I would want to see my ratio to constantly improve on. Win loss ratio is a good display of a players consistency. You can have a person with 100 wins - 10 losses with 1000 Rating, vs. someone with the same rating with 300 more games played. Who is better, more refined, and more consistent? It can really distinguishes skill gaps.
I can see how not displays losses would please a majority of a lot of the community.
"Oh this player has 2000 wins, he must be very experienced", but maybe this player also has 2000 losses or more, so he is breaking even, but they dont see that. OR "oh this guy is 2000-2500, wow, baddie" Sure the auto-match tries to aim for a 50/50 average, but as Rekrul said, there are always going to be that phenomenal player that simply sets the pace for the rest of the ladder. Sure, its pretty demoralizing for some people that care, but thats how the cookie crumbles, there are the elite, and then there's the rest.
Blizzard has done stat tracking with wins/losses and percentages on ladder, but their PROFILE, displays wins, so Ladder Wins, and All-time wins can be two different things. Some players can accept the fact that their lower tier, for whatever reason that is their own. Whether they enjoy vs. AI or whatnot doesn't matter.
So when a ladder is reset wins are still accounted for, and eventually reset at the end of the ladder season, and over the long haul, you still have an experienced account with many games played.
Wins/Losses are a pretty vital stat for players looking to ladder. Perhaps you can track ladder records according to seasons to also track progression.
I think there can be tweaks to be done to please both the competitive and casual players.
Profiles = Wins and other stats to account for time playing the game.
Ladders/Scoreboards = Win / Losses displayed..
It wouldn't make much sense to see a 100 Win ranked #1 while 350 wins is running 2nd place.
|
i think it's critical for a ladder to show win/lose ratio. otherwise it loses a lot of it's appeal.
in the OP the hypothetical can never really be more than just that though. the vast majority of people wont ever have a 50/50 ratio.
Also I think this forum will show a huge favoring of lose/win ratio being shown, i mean remember how much rage there was/is about the division system in bnet 2.0? lol (ITS STILL SO STUPID! >.< )
|
I think all three (win, loss, level) are important. Because the system autoadjusts you to be 50%, win/loss mean nothing, but if someone has only played 5 games, then the level becomes very important. if you take away any one of these elements, then it is impossible to understand whether or not you've improved, or whether you're ranked properly (i.e. whether you have a 50% w/l).
of course, this is purely from a competative perspective. it does not address the psychological aspect.
|
Calgary25986 Posts
Which perspective are we looking at this from? The competitive player's? The most ideally fair situation? A spectator's?
|
On July 13 2010 05:42 Chill wrote:
Which perspective are we looking at this from? The competitive player's? The most ideally fair situation? A spectator's?
I posited my OP from the point of a developer: "Is it good for the game to display losses to its players?"
good for the game as in: attracting and retaining the most amount of people.
|
Calgary25986 Posts
So for a casual base? I would imagine they would prefer you hide it since I think the average player is embarrassed about their record.
|
United States47024 Posts
On July 13 2010 03:14 Mora wrote:
Given then rules above, the argument is this: Showing Losses only allows players to fixate on an unimportant statistic. If the automatcher rigs everyone’s stats to hover around 50%, it is delusional to judge someone’s skill based on their win/loss ratio. It diverts the viewer away from the true indication of skill: their skill rating.
It's worth noting that if the automatcher works the way you say it does (assuming of course, that the automatcher cannot be perfect), then showing losses is not an unimportant statistic. The further it is from 50%, the less likely it is that a player has evened out at their appropriate skill level. This is where the difference between a 15-0 C- player and a 500-790 C- player comes in. One player has played enough games to accurately determine his skill level, while the other clearly hasn't yet. In this sense, win ratio acts as a measure of variance to show how close or how far the displayed rating is from the player's actual skill level.
|
Korea (South)17174 Posts
On July 13 2010 06:12 Mora wrote:Show nested quote +On July 13 2010 05:42 Chill wrote:
Which perspective are we looking at this from? The competitive player's? The most ideally fair situation? A spectator's? I posited my OP from the point of a developer: "Is it good for the game to display losses to its players?" good for the game as in: attracting and retaining the most amount of people.
well this will not effect how many copies are sold in any way imo, but as far as retaining casual players on a subscription model...you can't just look at it as simply as the question you are asking
there are other things involved, like can you reset your stats, can you have multiple ids, possibly having an option where you can choose to publically display your stats or not, etc etc all effect this question. u also have to ask the question do they want to retain casual players or push it as an e-sport etc etc. having a 'what is better a. or b.' when there are possible c. d. and e. options etc is really pointless, especially in business
not a fan of debating theoretically useless things lol, but in the scenario you described i think the benefits of showing stats motivation/competition/blah far outweigh the downsides embarassment etc even for casual players. psychologically people like to know ALL their details. this is why ppl take IQ or personality tests etc and shit online even though its pretty useless. a system that just shows ranks and no stats will be very bland and won't promote real gaming action
|
|
|
|
|
|
EPT
