|
The more I think about it, the more strange it seems that in individual competition that even the best players have near 50-60% win rates. Now, I know how the general scheme of the MMR system works and how a player is adjusted to play opponents of the same skill levels. But if I recall correctly, even in BW pro leagues (and sc2 ladder), it seems there is that range of win rates that players fall under and cannot break.
I am a fan of tennis and I like to draw comparisons to the pro players due to the individual nature of the sports. It's common among most people to say that Roger Federer and Rafael Nadal are perhaps the best tennis players in the history of tennis. This is evident in their persistent appearances in Grand Slam finals and championships. In the same way, I consider Flash and Jaedong the best players of the BW era. However, if I recall correctly, I believe Flash has had around 60-70% win rates throughout his career and likewise for Jaedong. I'm curious as hell why there isn't a "Federer" of SC who breaks that threshold of win-loss ratio.
I made the conclusion that there's some erratic element in SC (call it some arbitrary factor-- luck, skill cap due to parameters of the game, w/e). I'm assuming I'm wrong though but I was hoping someone had an understanding or an inclination of why the win rates or so similar in top players.
Sorry if this doesn't seem like a discussion worthy topic. I tried looking up the forums and I don't have IRC cuz I have a mac 
|
What about Select's ratio? Although that's his NA account,
|
The game's young. There hasn't been enough time for someone to become that consistent, confident, and talented.
|
Because there are so many variables in this game compared to any other sport like Tennis
User was warned for this post
|
One major factor is that luck plays a huge role in Starcraft I & II. Being good can limit the damage to how unlucky you can get, but it can only do so much.
|
Or it could just be that there is no player who is objectively 30% better than every other player, whatever the objective parameters may be.
|
1) Intense multitasking will inevitably lead to mistakes and large mental strain - you don't need to multitask when playing tennis
2) Luck plays a large role (mostly build order and scouting luck)
3) Players going for risky builds and/or trying to play mind games (boxer comes to mind)
4) Top players are not skilled enough to compensate for above reasons to win consistently over the next tier players
|
I have a 75% win rate on ladder, and I suck balls. The trick of the game is to stay ahead of the metagame. If someone dooes the exact same thing everyone else does then it's a cointoss, and most people will copy the top players- so people will know how to deal with them when the time comes.
|
On March 15 2011 14:13 Zeke50100 wrote:
Or it could just be that there is no player who is objectively 30% better than every other player, whatever the objective parameters may be.
If you were 30% better than another player you would probably beat them close to 100% of the time. Being only slightly better will yield a much larger win%.
|
It is because there are many unknowns in an individual game of Starcraft. To use your tennis analogy, it would be like if there was a mirror (that magically let balls through) in the middle of the court. Yes good "tennis" players would distinguish themselves, because they could predict their opponent's shots based on what they are doing and what they have seen, but you cannot know for sure.
So Flash can do his double armory build THINKING it is correct, but if he missed something in the scout he could find JD in his base right after he "wasted" gas and minerals on upgrades, and just get rolled.
|
I think it's the "limited information" aspect of the game... similar to blind chess. Imagine playing tennis with limited information. By the way, for anyone who's curious, a dominant player like Federer in tennis seems to have around an 80-90% win rate.
Edit: cLutZ puts it better than me.
|
I would say that ladder ranks aren't a good way to look at skill in SC2. Tournament stats are definitely better. In ladder you are largely facing people you don't know how to counter. It's totally random, people all have different styles so you have to adapt to the situation at hand.
|
Well if you're talking about ladder it's because the system is designed to do that. If you're winning all the time it's going to put you against higher skill people.
|
@kazzabiss, and those factors would be...?
@complete, that's what I thought initially but some might say it could be lack of scouting or map vision.
(EDIT)
@exwOn+ Show Spoiler + sorry, my bad! like I said though, for people in MOST cases in top tier play, they seem to fall under
EDIT
@munchmunch and clutz: nicely put. makes sense.
If it is true that there is a factor of "luck" involved, doesn't it seem like the game should try and fix that? or is something like that irreparable since it's sorta incorporated into the dynamic of the game? (i.e. fog of war in clutz's/munchmunch's explanation)
I'm sure people who believe in competitive sports would agree luck has no bearing in competition but is this factor of ambiguity just a part of the game I should just adhere to then?
|
At all but the highest levels of tennis, comparable people will have ~50% win rates against each other.
I'd guess that ~40 million people in the world play tennis, and today's greats have been practicing for 25+ years. SC and SC2 would need a long time to find their "greats" comparable to today's tennis stars.
Maybe a kid like Leenock will become that great, with another 10 years of practice. Then again, he probably didn't start young enough.
|
Its worth pointing out that Federer plays the lowest seeds in every tourney for the first 3 or 4 rounds, while on the ladder the top 20 all play just each other for the majority of their matches. Also in SC2 one mistake at a certain point can put the game in an unwinnable position, while in a game like tennis each event is independent. I think mainly it is that one game can be completely lost in a moment.
|
On March 15 2011 14:20 kamicom wrote:
@exwOn, no offense but your MMR is probably weak/you haven't been playing enough if you have a 75% (unless you're grand-masters level)
I do take offense. I was just above points and w:l of someone who made top 200 before I stopped laddering. I've played SeleCT, Masq, Destiny, and HuK on the ladders. I can guarantee you the reason they win or lose is because of the metagame and how the average gamer plays.
This is a game of limitations. The metagame will often be based off the top ladder players. People lower down will learn how to counter the builds, and the top players begin to lose because their opponents have been trained how to deal with it.
|
win/loss ratio should be win over loss, not win divided by total game
|
On March 15 2011 14:16 Gnial wrote:Show nested quote +On March 15 2011 14:13 Zeke50100 wrote:
Or it could just be that there is no player who is objectively 30% better than every other player, whatever the objective parameters may be. If you were 30% better than another player you would probably beat them close to 100% of the time. Being only slightly better will yield a much larger win%.
It was a random number >.<
Also, saying that you would win 100% of the time is a pretty big generalization, since you don't have any information about the sample at all.
|
Top tennis players ALWAYS play much weaker opponents the first few rounds of a tournament.
SC tournaments weed out a lot of the weaker players in the qualifying tournaments (like MST in BW), and the top players are just seeded into the final tournament, so the top players face other strong players at a higher % of their total games played than the top tennis player counterparts, who maybe don't face a legitimate threat until the quarter-finals.
At least, maybe that is a factor.
EDIT: Should have refreshed before I posted, this was already mentioned a few posts up.
|
Well SC2 is still early and being figured out, but their are a lot of people with high win rates above 60%. Naniwa had a very high one, Select dominates the NA server, MC has a 60+ winrate along with MVP, its not uncommon to see above 60 but its just still a lot of random losses going on
|
Mostly because Blizzard's matchmaker is really, really good, I'd say. Also, because we haven't really reached the point where the game is understood enough that the very best can reliably beat the merely very good.
|
Because in tennis you apply your skill advantage over the opponent several times per point, several of which happen per game, several of which happen per set, 5 of which happen per match.
No single point in tennis can cost you the match, while a single moment in SC2 can.
|
I would say that a lot of the top players are fairly close in terms of skill level
The fact that the system is constantly trying to get you down to 50%, only the top player, in theory, would have a higher win rate than 50%. Obviously this isn't perfect though...
|
Now mathematically there are 2 choices a win and a lose. Now the percent of a random draw(no skill just lottery) according to the numbers it would be 50/50. Now what does this say. Because te only 2 possible outcomes will ne'er change Ina head tail situation the percentage will always always be 50/50. Now over to sc2. We all know that you have a 50/50 chance but cause there is skill it is different. If a master player versed bronze every game he would win almost every one of it, breaking our 50/50. However the match making system will match you with evenly ranked players so you will never e able to win as many and lose. It just comes down to the match making system. And mainly cause pros have hit a dead end, they seem to be at a consistent level with each other mainly since they can not play better players because try aren't so they are forced to play the same rank of player meaning someday the slightly lower player will catch and go find his dead end, no one better. If there is no one better to improve to you can not get better.
Also because pros play so many games a 10 game win streak comes out as a 0.5% increase which is nothing
|
@exon sorry my apologies. Regardless, I meant MOST players in top tier playing at that certain range of win-loss ratio.
@Diader, hmm I have to disagree on that. I'm not too familiar with the ATP (men's tennis league) format but you need to have a repertoire of accolades from ATP-certified tournaments before trying to win qualifiers to get a spot on the Grandslams. Relative to the competitive scene in tennis, the players in a grandslam are a small fraction. Also, the ATP rank system is much like MMR and ELO rating for chess as winning many matches against lower players doesn't mean as much.
@lolsixtynine starcraft has the similar increments as tennis. Each battle is a rally and while there might not be as many rallies as battles, there are also sets (Bo3, Bo5, Bo7) to eliminate a lucky game or point.
EDIT
Again, to those who say SC2 is new and match making system is too good-- even with those assumptions, there should be a few players who (hypothetically being the best) should just dominate their way up through ladder. It's at that skill level that top-level players are at that I didn't really understand why Player #1 out of top 200 did not have significantly more wins than player #112.
|
I don't think the matchmaker is as great as people say -- at least so far for me. I guess I need to play a lot more so it can bump me up, as I have a 75% win rate right now and I am completely dominating my opponents (I did get smurfed once though...but was BO loss -- I really hate TvT).
|
Your win-loss ratio isnt decided by how good you are, or the game you play. That has nothing to do with it.
The win-loss ratio is only decided by who you play against.
Imagine for example, that there are 2 players that are insanely good compared to everyone else, but about as good as each other.
What will their win-loss ratio be?
Well you dont know.
If the tournament is a 2 player tournament, these 2 guys that are insanely better than everyone else, will still have a 50% win-loss ratio.
If its a 4 player tournament, then they will have 75% win-loss ration on average, and so on.
Make a million player tournament, where the pros have to play against thousands of bronze level people before getting to the top, and suddenly, the pros will have a 99% win-loss ratio.
|
Tennis matches contain a ton of repetition; each match is decided by the total of hundreds of points, each of which is a small self-contained battle of wits / skill. In Starcraft, on the other hand, matches are often decided by a single good or bad decision. Statistically speaking, the more repetition that occurs per match, the higher the chance of a marginally better player winning over a marginally worse player. So, a player who's "10%" better than everyone else at tennis can dominate the sport, but a player who's "10%" better than everyone else at Starcraft might only win 55% of their matches.
|
@morimacil, but relative to all the other players, he would still gain more wins from them (you can't assume they only play each other because that obviously isnt the case in the ladder system)
but I guess that does make sense that if 2 hypothetically good players had high MMR's, they'd only be playing each other on the ladder. ... but then, if one player just dominated the other which is plausible, he should be at a higher win loss and push the 2nd player down to win against the rest of the players.
I think I made sense there... :x
EDIT
Also, general consensus sorta tells me there is a factor of ambiguity and lack of variance/repetition which can be considered luck in SC2.
I'm a bit of a cynic when watching sports in that I pick things out which I consider flaws. Every sport in my opinion has flaws, many of them that cannot be changed due to the nature of the game. I'm not necessarily saying it's a problem that needs to be fixed however. Labeling this factor of ambiguity as a "flaw", I'm guessing it's just something people have to abide by?
If so, it sorta saddens me that there won't be a "Federer" of SC2 in the near future (unless I made some wrong judgement somewhere in the assumptions).
|
Being a former tennis instructor, and playing alot of SC2...I would say the reason a good tennis player can be more dominent than a good SC2 player are how advantages play out...
Say you have a better builder order than the guy...what's that going to do for you? It'll give you an edge, if it's TvT, your better builder order maybe give you an extra marine or two, that might come into play later on. Or maybe you scouted better than him...it gives you an advantage, but it's a small one...
In tennis; if you're both equally skilled, but you have a better first serve..you're dominating him. If you have a better slice, well that comes into play often. If you have a better topspin forehand than him, the game's over...in tennis, if the other player is better, however minor it may seem, gives that person a huge advanage.
Say you're better in micro than your opponent in SC2...how often does that play out? 1, 2 maybe 3 battles in a game? But if you have a better topspin forehand than the other guy, you're using that forehand every point to set up, to finish, to tire your opponent out because he has to put even more topspin to counter your topspin...
Just my 2cents...
|
On March 15 2011 14:15 eXwOn wrote:
I have a 75% win rate on ladder, and I suck balls. The trick of the game is to stay ahead of the metagame. If someone dooes the exact same thing everyone else does then it's a cointoss, and most people will copy the top players- so people will know how to deal with them when the time comes.
but how many games have you played? under 100 i bet. either that or you are REALLY good, which you claim not to be.
One doesnt simply know how to deal with the strat a pro does, if you 3gate sentry expand, which is what 95% of toss platyers do now days on the ladder, its not like. "oh i do X and insta win" the reason they are the strats the top players do, is because they are so solid against everything.
@OP its because the MMR matches you against players at the same skill level as you, if you win a lot it matches you against better players untill you lose, and vice versa until you lose. Therefore unless you are one of the best on the ladder, or one of the worst, youll average at to about 50%
|
Sorry if I sounded harsh. I guess what I'm trying to say is if some is facing SeleCT, a well know player for his use of early bio, they'll have a pretty good grasp on what to do because they've faced opponents who've mimicked him a dozen times over. It makes it harder on the top players to get a high win ratio, but being able to pull off such a high win rate amongst such opponents is even more of a testimony to how good they actually are.
under 100 i bet. either that or you are REALLY good, which you claim not to be. I was at 160ish before I quit I think. But really, just stay one step of the metagame and put your opponents into new situations they've never been in.
|
that is because of the matchmaking of course. You cannot compare tournament records with ladder records with properly functioning matchmaking system.
Federer has 70 percent win rate because he plays against all sorts of opponents. From rank 500 plus to Nadal.
If Federer only plays Nadal, Djokovic, Murray everyday, then his ratio will be closer to 50 percent than 70 percent given sufficient sample match sizes. This is what is happening in the ladder. Player rank 1 does not have statistically more wins than player 112 because player 1 plays against 2-10 everyday while player ranks 112 plays against 100-150
In BW, the nature of winner's league and proleague means Flash can play against anyone who is not close to his skill level. Proleagues makes the bulk of your appearance compared to individual leagues and thus you can get up to 60-70 percent win ratio.
|
at the current state of the game, there's a lot of luck involved that skill can or cannot make up for. for example there are games where i (an above average sc2 player) could've beaten gsl pros. however in many other games, such as fighting games, there is no way i (if i was above average) would even touch the pros. they would beat me 99.9999 percent of the time if they were trying. some games just have less luck involved.
i think when sc2 gets to the point where players can 100 percent survive against 1 and maybe 2 base play, and transition into a big macro game, that's where skill will trump luck a lot more than this current metagame.
|
On March 15 2011 14:50 eXwOn wrote:
I was at 160ish before I quit I think. But really, just stay one step of the metagame and put your opponents into new situations they've never been in.
Trust me, ive you play another 200 games, youll be sitting just above 50%
|
I think a point that should be made is that for example tennis, 1 mistake costs u '1' point (15 w/e you want to say) While in SC1/BW/2 1 mistake can cost you the game. Losing one game in a best of 3 is quite serious. While in a tennis game, after your mistake you get on ur feet, take a deep breath, relax for a second or two until you/your opponet serves again. 1 'leg' (is it called/written like that) basically require you to score '4points', 4*6 = 24 points for a set, and ~2/3/4/5/6 sets for a game (win). (Hope i use the terms correctly). So a single mistake, how huge it may be, can be restored quite easily without depending on your opponent making a similary huge mistake. (perfect games don't exist in real sports)
In football (soccer) 1 goal can be huge, and cost you the game, yet you see very high win/loss ratios there, because you have an 11 man team. If 1 fcks up u sometimes (depending on who fcks up) have 10 people to restore your mistake.
I think in SC1/BW/SC2 mistakes can be punished much more severly because you can't make that 'many' mistakes.
The fact that tenis players play weaker opponents in early rounds kinda makes 'sense'... but they are still pro's, they also train 8~10hours a day similar to Federer/Nadal any other tennis legend. It's kinda a silly arguement, same things happen in early stages of tournaments, at least, that's the general idea. Any tournament that has seeds works like that. Also GSL/MST/TSL kinda work like that.
It's hard to say what the actual reason in, but I feel it has something to do with how severly 1 notable mistake can be punished compared to other (real life) sports
|
Because ladder is a constant stream of BO1, and anyone can lose those pretty consistently.
|
when you're matched against people of your skill level the win rate will ofc even out to 50%
the federer is probably not playing all people of his equal skill level even if they are better than everyone else
|
A big issue here too is to consider how linear Starcraft 2 is, and how dependent every second of the game is on the outcome. In Basketball/Football/Soccer/Whatever, there isn't much stopping a person from coming back from a HUGE deficit of points. The only variable is fatigue and morale (which is consistent between both sides).
In Starcraft, with VERY few exceptions, an advantage is inevitably a win. Going 30 food up on your opponent 10 minutes in will usually signify a victory within the next 10 minutes, whereas a 30 point lead in Basketball CAN be made up without consideration of this prior disadvantage.
tl;dr the point I am making is that you are punished for smaller mistakes that extend across the length of a game. Dropping your racket in Tennis is a variable isolated to that one set. Forgetting to wall in and losing 5 probes will lose you the game.
|
On March 15 2011 14:55 L3g3nd_ wrote:Show nested quote +On March 15 2011 14:50 eXwOn wrote:
I was at 160ish before I quit I think. But really, just stay one step of the metagame and put your opponents into new situations they've never been in. Trust me, ive you play another 200 games, youll be sitting just above 50%
Maybe, but I'd rather train in BoX and the physiology of the game. I'm done with trying to get to #1 on ladder. BoX really shows the better players, which may also be why the W:L are close with Bo1.
|
They're close to 50% because skill will only get you so far in SC. Just like in poker, there is some luck involved.
|
I like L3g3nd's and Oxb's interpretations of it. I could see how variance is a significant factor (i.e. in poker, 40% odds to 60% of opponent could be ran twice for a more solid result)
Again, to the ppl who said it's due to MMR, you have to realize a really good player should still rip through the matches made by the system and eventually make it out on top. It's just the fact that noone does that puzzled me.
|
|
|
Because balance is really overrated. So much more of this game is in the mechanics and the strategies than the balance of the game.
People went for years without a patch in Brood War and the balance QQ was way less than what we have now.
|
Don't use ladder stats, use TLPD! Only counts official tournament games, much like tennis rankings or whatever, you wouldn't count Federer's practice games, so why are people acting like ladder stats matter? :p
http://www.teamliquid.net/tlpd/sc2-international
Top player Kas has a 66% winrate.
|
It has to do with how the matchmaking works. MMR is a value that represents your raw skill that works with a rolling average. The matching system tries to match you to the closest rated MMR first, with a second priority of not making you wait hours on end in the event nobody with very close to MMR as you enters the matchmaking pool. The difference between the two MMR scores determines how many points you will win or lose from the match. So, you'll always have a near 50% chance of winning or losing with the way the system works. The only exception is when a single player hits the near 70% mark. The highest you are supposed to get ratio wise in a system like this is 60-65% win rate. Although, if a single player is better than everyone else it isn't too improbable that his ratio will be strangely high.
|
On March 15 2011 14:57 Subversion wrote:
Because ladder is a constant stream of BO1, and anyone can lose those pretty consistently.
I don't know a lot about tennis but this seems like the obvious answer to me. A tennis match seems much more comparable to BO5 matches as opposed to individual games in which case someone like Flash could easily have an 80%+ winning record.
|
That's because the game forces a 50% win loss ratio. The few exceptions are when you are so good, or so crap, there is Boone better or worse for you to play.
|
Actually there is a VERY simple answer if you are looking to compare SC to Tennis.
If you think of 1 game in SC as a set in Tennis then an actual match in Tennis is actually a BO5 in SC.
When you do this you get a much higher win percentage for better players because you mitigate the extraneous variables that you are bound to come by in a short SC match.
I would bet that Federer's win percentage is much less if you use sets (or even games) instead of full matches.
|
On March 15 2011 14:14 lbmaian wrote:
1) Intense multitasking will inevitably lead to mistakes and large mental strain - you don't need to multitask when playing tennis
I'm all for the proliferation of e-sports but this needs to be clarified. I'm not going to make any judgments but perhaps some people on the forum aren't familiar with what runs through your head as an athlete. I say this because anyone who has played any sports would balk at this statement.
More specifically, in tennis you need to be taking in consideration your opponent's positioning, be aware of where you are on the court, constantly on the move, etc. Of course these things become second nature with tons of experience, as is the case in SC. Sure Idra multi tasks really well but I'd imagine he's rarely in a novel position where his actions aren't like clockwork.
|
One thing to remember is that in SC each game is included into your win/loss ratio, not just the series. So you can have a person that have never lost a BO5, but if he goes 3-2 in all of them he'll have a 60% win rate. If he never lost a BO3 but went 2-1 he'll still have 67% win rate in Starcraft.
|
There's also the factor of how you're calculating win/loss ratio.
In pro SC win/loss ratios take into account every game played. In a Bo5 all 5 games are taken into account.
Correct me if I'm wrong, but in a sport like Tennis, Federer's ~81% ratio is for matches, and not sets. I would be willing to bet that if win/loss ratio was worked out on the basis of sets, rather than matches, then Federer's ratio would be a significant amount lower than 81% --- but I can't find any stats set win/loss ratio.
Edit: hehe, poster above me talked about the same thing and posted on the same minute.
|
Starcraft and Starcraft 2, because of their resource system, tend to force players into paper rock scissor scenarios. Even if you're really good and you play a paper to someone else's scissors, you're going to get behind, if not lose outright.
"But good players always scout, so their choices aren't blind!" Good players also deny scouting, and if you deny an opponents scouting you have costed them resources. To do a guaranteed scout the resource cost, at certain points in the game, can be very high. There are quite a few scenarios, where if you scout, you've spent quite a few resources to basically find out your opponent is doing something you didn't need to scout at all, and thus you basically played a small scale scissors to someone else's rock. Depending how far ahead or behind you are, that can make a big difference.
If there were no fog of war skill could be more absolute and you'd see better players win ratio raise, but the game would be so much less interesting then, and the skill ceiling would be much lower.
|
First off:
I tried looking up the forums and I don't have IRC cuz I have a mac
Colloquy is a good IRC client for Mac.
To the topic at hand:
It is easier to dominate in Tennis if you are the better athlete. Tennis takes a drastically longer amount of time to master than SC2 and nobody is created equal due to their natural athleticism or lack there of. While some natural talents come into play regarding SC2 they are drastically less pronounced than in any professional athletic sport. Also, a tennis match is a long drawn out affair. If, in SC2 we played best of 11 for each ladder match it would be much harder for the lesser player to win.
|
One factor is definitely the imperfect information from fog of war. I think top players' win rates would be much higher if both sides had complete vision. With imperfect information you can be outguessed, and it can lead to some silly results like gambling no robo in PvP only to get instantly killed by DTs a minute later. RTS games tend to boil down to a somewhat counter system, and the effects range all the way from instant death to small but noticeable disadvantage. Even the best players can't avoid being out build-ordered in the early game and this certainly makes it harder to achieve 80%+ winrates vs. top competition over a long period of time.
|
Also think about who uses the ladder system, while at lower levels you might play a guy and not see him again for a long time (or probably never), top players are generally playing other top players. Maintaining a very high win rate in this environment would require you to be significantly better than those that are already at the top. And of course the mass gamers near the top of the ladder can take some games off pros, due to imperfect scouting, BO1s, plain bad luck, and other things already discussed.
|
Doesn't seem strange to me considering what you are comparing.
It seems more reasonable to compare one point or game in tennis to one game of starcraft. In these situations, like in a starcraft game 1 mistake or miss calculation will lose you the point or put you at a big disadvantage for the game. The best tennis players don't win 80 percent of their points/games though they might win that % of matches. If you played starcraft tournaments like tennis where first to 6 games wins a set in a best of 5 sets you would get higher win percentages since the better player would win more overall.
Also there are other factors like lack of scouting information, early game build orders, increased number of variables in starcraft etc etc.
|
If you look at higher tier players, playing higher tier players, the win loss ratio will adjust accordingly because they're more likely to lose. Also, through trial and error. How do you think these people got so good? Realistically, a lot of players don't play as well in practice games like they do when real numbers and rankings are on the line, people like myself. I can't play well if I'm doing practice because subconsciously I don't care as much, so I don't really give it a good honest try.
|
top players do have high win percentages. demuslim, idra, select all near 75% since the match making ladder places people even with you then obviously the majority will be near 50%
|
I think the answer to this question is blantly obvious.
The match making system keep searching for a player at your level. This means winning around 50% of your games and losing 50% of your games. ofc if you are so so good, match making system can no longer find a suitable player for you, thus you will suddenly have surplus of winning ratio (or surpuls of losing ratio if you are terrible terrible).
(I know some people who are way above 50% even though they are not even masters, this is because shit happens. Match making system is not perfect, even it was chance of finding a wrong matching player is still there since not everyone at your level can be online at any given time. That is to say for most it will be around 50%, unless you are at god tier, bottom feeder or you just happened to have high win rate by pure chance).
|
Having higher than 50% is not necessarily the fault of the matchmaking system. For example the players you're matched with could have played worse than usual. You will also accumulate some wins at the beggining, if you're better than average, since the system doesn't have enough info to know your skill level.
If you keep getting fair matches after accumulating more wins than losses, your extra wins will remain, since the system does not try to bring your ratio down to 50%, all it tries is to match you with someone you have a 50% chance of winning.
|
I read somewhere that a player that has a 55% chance of winning a game has a 95% chance of winng the match. A tennis match typically takes a couple of hours, while a typical Starcraft2 match takes 30 minutes. (BO3). If all Starcraft2 matches were BO9, we would probably have about as long matches and see the same kind of trend. (The same player winning more consistently).
In addition a really good SC2 player get studied a lot before a match, while the better player don't really have the same opportunity, especially in the current very fluid metagame. A BO9 would probably give the better player a chance to study the other player while in the match, and adjust his gameplay.
|
the matchmaking system is designed to give you a 50% winrate
|
If you are talking about 50% win rate on ladder, thats because the match making system is designed to make you go 50/50. The old saying "there is always someone better than you" applies, no matter how high you are on the ladder, there will always be players as good if not better than you, and on any given day you can beat them or they can beat you.
Even at the top the system is still trying to make you go 50/50 so unless you are insanely better than your potential opponents or abuse the fotm cheese you will rarely find anyone with over 60% winrate.
In tourneys, well anything can happen, nerves can get the better of you, the room might be too cold or hot for you to be confortable, blah blah blah.
|
Because the more you win the higher ur ELO gets and the higher your ELO gets the harder your opponents become.
Also, percent win ratio doesn't matter, its (wins - loses) that really determine skill
|
On March 15 2011 18:18 blackbrrd wrote:
I read somewhere that a player that has a 55% chance of winning a game has a 95% chance of winng the match. A tennis match typically takes a couple of hours, while a typical Starcraft2 match takes 30 minutes. (BO3). If all Starcraft2 matches were BO9, we would probably have about as long matches and see the same kind of trend. (The same player winning more consistently).
In addition a really good SC2 player get studied a lot before a match, while the better player don't really have the same opportunity, especially in the current very fluid metagame. A BO9 would probably give the better player a chance to study the other player while in the match, and adjust his gameplay.
You are bang on here, the longer the series the more chance for the better player to win. You see this in Tennis alot.
When men play 1set or 3 set matches there are often upsets, as you can play a blinding match for an hour before the better player can use their stamina advantage or skill advantage to wear you down. Thats why in smaller tounaments players like Federer sometimes go out in the early rounds but hardly ever go out in Early rounds of things like the US Open or Wimbledon where the matches are all 5 sets.
To beat Federer in a 5 set match you have to play at your best for 2-3 hours usually, he doesn't go down without a fight. On the other hand when Federer is playing a much weaker player he can finish them off in under 1hour 10 minutes (seen him do it at wimbledon, he lost about 3 games total), but when one of those weaker players has given him a good match they usually win the first set or two and then get stomped more and more as the game goes on.
|
I think I have a good explanation for you. I'm also a tennis player and I like to draw comparisons between tennis and starcraft. In tennis, the most similar thing to one game of starcraft is actually ONE point, and NOT the whole match. If you think about it, although one point is very short, nearly of the quick decision-making and skill happens for every individual point, and not as much throughout the match.
So with that in mind, if you take a look at the point split between pro tennis players in a match, you would actually find that the ratio is extremely similar to win/loss in sc2, it's only about 60/40 for a comfortable lead, and 55/45 for a good lead in a single set. So in actuality because the skills are so close (even with the greatest players) players are going to be trading points (or games) fairly evenly for the most part. The difference in tennis is that you need to string a lot more together than in sc2, which favors the player that is better more consistently (i.e. the one with more skill), and you get lopsided W/L rates for better players, which is how it should be.
Were starcraft games to be played with the amount like tennis points in a match (oh gosh that would be exhausting) then it is my belief that you would start to see more dominant players like what you see in tennis. However that's pretty much impossible, so the best players simply have to play as if everything were riding on one point, one game.
Hope that helps! Sorry if it was super long or boring :/
|
On March 15 2011 15:03 kamicom wrote:
Again, to the ppl who said it's due to MMR, you have to realize a really good player should still rip through the matches made by the system and eventually make it out on top. It's just the fact that noone does that puzzled me.
You are comparing ladder to tournament play.
Let's see. You give the example of federer who has 80 percent win record against everyone in his wikipedia.But in ladder because of the matchmaking, Federer would only play against the big 3 ( nadal, murray,djokovic).
Their head to heads are
Federer-nadal 8 -14
Federer -murray 6-8
Federer-djokovic 13-8
total win-loss = 27 - 30 ( about 50 percent). Stats are from very quick googling so it might not be 100 percent accurate but i believe the point stands.
This is what is happeniing in the ladder. The MMR system makes the very good players only play against other very good players. That is why you dont see 70percent stats in ladder.
On the other hand, on tournament plays it is possible. MVP has 78 percent.
|
|
|
I think tennis is a particularly poor analogue to starcraft. One mistake in tennis only costs you 1 point. One mistake in starcraft can cost you the match. Tennis has so many chances for the better player to show he is better.
When federer or nadal makes a mistake, it may cost him a point, even a game, but there are a usually dozen more games before the match is over. His ability to consistantly outplay his opponent is rewarded. When IMmvp makes a single mistake in a BO3 in the GSL, it costs him the match. Two mistakes knock him out of the tournament.
I think a better analogy would be soccer. In soccer one mistake costs you the game, and the outcomes can be more unpredictable.
|
@ OP:
This is my answer for your question:
In computergames, you are very much limited to some point: e.g. You cannot mine faster than anybody else... you can play the exact same BO as MC and you will be exactly even the first few minutes until parameters like micro/game sense/decision making etc. start to make a difference
but in a sport like tennis, there are much more parameters on the one hand, and there is much more room for improvement for any of these parameters on the other hand
from the very beginning of any match, the skill gap in any parameter will kick in:
i don't know much about tennis but i think there are many factors how you approach to the ball, how you stand, how you hit the ball, how you aim etc... you can outplay your opponent in every single second unlike in a computer game
and even if you are unlucky or you make mistakes... that is not that bad because you just lost one point but there are many points to be won in tennis so the better player will most likely win in the long run of one match
also i think mindgames play a much bigger role in sports
|
Idra had a 70-80 win percentage on his NA account during beta.
|
because tennis games aren't best of 1/3/5/7 games, they are best of 3/5 sets of first to 6/7 games
if starcraft matches were the same, flash would be your federer.
|
I do enjoy all of the analogies being drawn to tennis in here though ^^. It warms my heart. Any college or former college players? :D
|
I'd be curious to see how dominant Federers numbers are if you calculate his Set win/loss ratio instead of his Match win/loss ratio. A Tennis volley might be comparable to a SC2 battle, and a Tennis Set comparable to a SC2 game, and finally a Tennis Match comparable to a SC2 BO5 or BO7.
So, I'd like to see Federers Set win/loss ratio, OR top SC2 players BO5 win/loss ratios
|
Also, players have thousands of games within months. Tennis players don't play that much, so they're less likely to lose due to little mistakes, exaustion, stress, etc.
//tx
|
Tournament vs ladder play, two completely different formats can partly explain the difference in win ratios. Add the fact that luck plays a bigger role in sc2 (happens more frequently and have a bigger impact on the outcome), and considering luck is evenly distributed over a large enough amount of games it's not hard to understand why the win ratios are much closer to 50% in sc2.
The reason why bw managed to produce players with a higher ratio is simply due to the fact that it's more skillbased and thus luck has smaller impact on the outcome.
|
On March 15 2011 15:18 moose162 wrote:Show nested quote +On March 15 2011 14:14 lbmaian wrote:
1) Intense multitasking will inevitably lead to mistakes and large mental strain - you don't need to multitask when playing tennis
I'm all for the proliferation of e-sports but this needs to be clarified. I'm not going to make any judgments but perhaps some people on the forum aren't familiar with what runs through your head as an athlete. I say this because anyone who has played any sports would balk at this statement. More specifically, in tennis you need to be taking in consideration your opponent's positioning, be aware of where you are on the court, constantly on the move, etc. Of course these things become second nature with tons of experience, as is the case in SC. Sure Idra multi tasks really well but I'd imagine he's rarely in a novel position where his actions aren't like clockwork.
doesn't really matter because there is still way more things to think about in a SC match vs. any sport. there are more variables in a SC game than in tennis that can cost the game.
and like you said with experience, things become second nature, with SC2 strategies are always changing and their are many different strategies out there, with something like tennis their are not as many ways to play it compared to the many ways of SC2 which means you can become accustom to the limited way to play tennis.
And by limited i mean that it doesn't change much and most likely will not change very much from how it is now
|
TLDR: Have a think about why MVP's ladder win percentage is lower than his GSL win percentage. How many games did he win when he won GSL? How many did he lose when he went out in the RO32?
I suspected that players competing in ladders, round robins, and tournaments would have intrinsically different win ratios, so I did an EXPERIMENT!
Consider 4 players with Elo ratings of 0, 50, 100, and 150 and pretend (heroically) that this is our population of pro gamers.
In a ladder where you only ever play people within 50 points of you, win percentages are as follows:
Mr 150: 57%
Messrs 100, 50: 50%
Mr 0: 43%
In a round robin tournament (everyone plays everyone else) win percentages are as follows:
Mr 150: 64%
Mr 100: 55%
Mr 50: 45%
Mr 0: 36%
In a seeded BO1 tournament where Mr 150 plays Mr 0 in the first round:
Mr 150: 66%
Mr 100: 54%
Mr 50: 43%
Mr 0: 32%
Note that in the case of the tournament, the percentages are not symmetrical! This is because good players tend to play more games (because they qualify for the finals) and tend to win them, but poor Mr 0 usually only plays 1 game per tournament and usually loses it!
Because of this, I suspect that if I expanded my population to 8 players and ran my seeded tournament again, or if I made it BO3, Mr 150's win percentage would be higher, but I don't have time just now. You get the idea, though.
|
On March 15 2011 18:25 tetranoir wrote:
Because the more you win the higher ur ELO gets and the higher your ELO gets the harder your opponents become.
Also, percent win ratio doesn't matter, its (wins - loses) that really determine skill
solid post first post, using bold and underline to highlight something that has already been said countless times in the same thread and then coming up with this ratio bullshit
win/loss contains the same info as win/matches, you can convert them into each other (matches = wins+losses)
|
I only read the first page of this so apologies if this has already been said, but the answer seemed obvious to me. The reason tennis is this way, is essentially because of sample size. Roger federer may have a 80-90% match win rate, but his rally win rate is nowhere near that. A game of starcraft is more like a rally than a match. If every time you laddered you had to play the guy first to 50 wins to get the points, i bet thered be very different win %'s. Its the same principle a why pros hate BO1 in tournaments. Imagine if in tennis you just played 1 game. Think roger federer would still have a 80-90% win rate?
|
Maybe MVP wasn't playing to win on ladder, but rather to practice certain things, or try/refine out new builds.
|
So many reasons, I'll keep examples tennis.
-For starters, blizzard ladder is not a good reference since players often use this to experiment with builds, it's like a tennis player having a hit up with another pro.
-I believe federer has a win rate of around 80% from around 2004-2009, but this statistic is a very specific point where his dominance came to reign. I'm sure flash had a period of dominance very similar but maybe not so long (2 years or so).
-Plays many more lower rank players. Grand slams are top 100 players + 28 wild cards compare this to GSL which is just top 32 players. In effect top players can get up to 2 free wins per tournament. Obviously upsets still occur on occasion (though not to a player like federer.)
-Federer became pro in 1998, and won his first grand slam in 2003 when he was 21. He probably started playing tennis when he was anywhere from 5-11, so he'd been playing for at least 10 years beforehand.
|
IdrA has something like a 75% win rate on the NA server with like 400 games played.
|
Assuming that OP isn't referring to ladder accounts (it doesn't sound like that's the point of the question): Doesn't the way that tournaments are structured (with highest seed playing against lowest) more or less guarantee that a good player will have a pretty high win percentage? For each tournament a player enters, he'll have at most one loss (the set in which he's knocked out), and as many wins as he manages to advance.
|
The more your win rate deviates from 50% the more skilled you are...at least on ladder. Average players have between 51-53% if they have played at least 1000 games already.
|
You can lose a game on one mistake.. or just based on build orders.
|
are you talking about on ladder or in tournaments? i mean its a reasonably balanced game, but the way the ladder works is that the system tries its hardest to keep you at a 50% win rate, and then im pretty sure it tries to spread your MU fairly evenely so i think that that could be part of the reason we see close to 50% across the board
|
there is a part of luck and postional imabalances on map depending on races.
Making it nearly impossible for a player to get a high win rate ratio.
The more the game get refined, and the map get positional balanced (no random close spots)
Basicaly, the more the game will be balanced, the more there are chances that high win rate exists.
As other said, the game is quite new and nobody has perfect solid plays on every match-up / maps right now. As the time progresses, I'm pretty sure the top players will have better win/loss ratio
|
On March 15 2011 21:04 CravenRaven wrote:
I only read the first page of this so apologies if this has already been said, but the answer seemed obvious to me. The reason tennis is this way, is essentially because of sample size. Roger federer may have a 80-90% match win rate, but his rally win rate is nowhere near that. A game of starcraft is more like a rally than a match. If every time you laddered you had to play the guy first to 50 wins to get the points, i bet thered be very different win %'s. Its the same principle a why pros hate BO1 in tournaments. Imagine if in tennis you just played 1 game. Think roger federer would still have a 80-90% win rate?
Read this. This guy is correct.
Laddering win percentages are entirely different than tournament win percentages and should not be used to determine skill or who is a better player.
That's why in competitive sports and e-sports (tournaments) it's rare to see a best of 1 in playoff brackets or championship matches.
The more games you play against your opponent, the less randomness affects the outcome overall/entire match; therefore, the higher chance of the 'true' better player coming out on top.
Let's say hypothetically if tennis matches were best of 21 sets instead of best of 5. I'm sure Federer's match winning percentage would be even higher. His rally win percentage would probably stay the same but those small percentages in each rally add up.
|
On March 15 2011 21:23 Hypatio wrote:
IdrA has something like a 75% win rate on the NA server with like 400 games played.
Yeah he's in my division lol. Id say that he would stay at a relatively high percentage but he doesn't have a lot of points currently so I'm not sure if he's playing top notch players.
|
In theory you should have a win % very higher undtill you get to rankings with equal skill where it will even out more.
This proces speedens up because of the 5 placement matches and the explosiveness of which you can escalate in the ladder.
|
because sc2 is so new noone has had a chance to show themselves to be "the greatest (so far). when we get our first flash or jaedong im sure they will have a 70-80% win ratio, atleast at lans.
another reason is that alot of pros try random stuff on the ladder, if it doesnt work out and they take a loss its not a big deal, but it does hurt their win ratios
in before this topic becomes about idra having a higher win percentage if he was protoss :D
|
Mh, close to 50? Top players usually get close to 60% on ladder?
Look at the tournament stats, last i checked tennis players didn't ladder and rarely tried new builds.
i'm pretty sure MC has something like a 80% win ratio for GSL.
edit: ok, its' 72% throughout, 80+ for pvp.
|
1.) For the most part BW pro-gamers continue playing long after everyone has caught up to them and they are past their prime, a lot of people comparing new gamers to old gamers forget this. Even  Flash will see his win rate fall quite a bit if he continues playing for 5 more years, goes to ACE, becomes a figurehead, etc. Guys like  Reach and Anytime used to enjoy dominant win rates when they were in the prime of their careers too, and guys like  sAviOr, Oov and NaDa used to be at or above 70% in their best match-ups for long stretches. Anytime is now below 50%, Reach is somewhere around there, and Oov, Nada, and Savior retired with a 57-60% win rate. EffOrt retired with a 60% win rate due to being his prime. For the best players it all comes down to playing long after you've lost your edge. If he's like his predecessors Flash will play for a long time after he is no longer able to qualify for a Starleague, or anchor a team in pro-league. Federer will most likely retire when he becomes unable to make quarterfinals appearances in slams anymore.
2.) Most importantly. Win rate is measured differently in Tennis and Starcraft. Flash's 70% win rate measures his success in winning individual games of Starcraft. It makes sense to do it this way as pro-league makes up the majority of his games and is just a series of Bo1 between members of a team. However there is no pro-league equivalent in Tennis. Every "win" that is measured by Federer's win rate is essentially a best of 5 (for slams) or a best of 3 (for every other tournament) in which he defeated his opponent. Flash's win rate in best of x? When is the last time he lost one? To Effort almost a year ago? That's almost like winning a calendar slam in Tennis. Something that hasn't been done in half a century.
It has nothing to do with how the tournaments are structured in Tennis and BW. There are qualifiers in Slams too, and most top players (Federer included) don't play in more than a couple tournaments outside of slams and masters each year so its not like Federer is constantly newbie bashing his way into the later rounds and Flash isn't. Winning in straight sets over a washed up Gilles Simon in the opening rounds of a slam isn't much different than rolling over Ssak and Classic on your way to yet another MSL..... Oh wait. Have we decided if we're just going to pretend that didn't happen?
EDIT: Ever went back to link to players in TLPD and just TLPDized the whole post against your better judgment? Who the fuck is WhO
|
There isnt the same genetical cap in starcraft. Thats about it.
|
I dunno about overall win ratios, but if you look at people like MC, Nestea and MVP right now, they got 80+% in their respective mirror matchups, showing how good they are at their own races.
|
under 100 i bet. either that or you are REALLY good, which you claim not to be. I was at 160ish before I quit I think. But really, just stay one step of the metagame and put your opponents into new situations they've never been in.[/QUOTE]
I think one of the issues is that as the game evolves and the players get more experienced they have in a sense seen it all. Then it comes much more down to execution than trying to throw your opponent of with a play they've never seen before. But I totally agree that not following the mainstream trends of playstyle is a smart thing
The game is so complex that there is a lot more different paths and factors that leads to victory then say tennis. If a guy is gonna be as dominant as Fedrer it wont do with just great multitasking abilitys. He needs great mental strenght, stratigic thinking, creativity, understanding, adaptivity etc.. And seeing this game is a bit more advanced than hitting a ball back and forth over a fence you need super human abilitys to excel at all the parts of the game at once I think.
|
Some of the game is rock/paper/scissors since you can't see what the other guy is doing.
The game is played at a pace where concentration often backfires because you have to multitask so much. Games like tennis need intense concentration at all times, I'd imagine.
I've heard that in Warcraft 3 a couple of guys dominated. I'm betting because it's played at a pace (and designed) to need more concentration and focus and less multi-tasking. But I never played war3 seriously, so somebody else should correct me if I'm wrong.
Plus Blizzard can't nerf real life people, but they can nerf the game. And since there are three races, they pretty much need to nerf/patch things because they might be imbalanced. It's tricky. But if all races were the same in the game... and bloodlust was over powered, but everybody had it, then it'd be fine... whoever could use bloodlust the best wins.
|
Looking at the stats for last year on Federer (http://www.atpworldtour.com/Tennis/Players/Top-Players/Roger-Federer.aspx?t=mf) we have 89% of service games won, and 30% of return games won.
Assuming they serve roughly as often as they return (since it alternates) we have an average win% of 59.5%, which tallies up with the OP.
As many have commented above the nature of what constitutes a win in tennis vs SC2 makes up for a lot of the differences, and as in almost all sufficiently deep competitions the differences at the top are so small that to eek out even a small advantage will put someone 'way out in front' even if their win ratios are never in the 80 or 90% range.
|
|
|
the skill ceiling in tennis is a lot higher then SC hate to break it to you nerds
|
|
|
|
|
|
EPT
