|
On December 08 2014 14:44 Goumindong wrote:
LOL how do you not understand that there exists no other way to find diamonds in the rough because no such measurement exists or can exist?
Literally you're asking "I want to know how good people are at winning" and I am sitting here saying "this is the stat which shows how good people are at winning, its literally the value for their winningness weighted against competition"
basing your valuation of something using the same readily available information that everyone else is also using to assign a value to that thing does nothing to help you find the quality hidden beneath the surface of these "diamonds in the rough" you keep talking about.
congrats on figuring out that winning players win more though, i appreciate the groundbreaking revelation. however, it seems clear that you do not in fact understand what i was asking, so i think everyone would benefit if you would refrain from further attempts at paraphrasing or explaining my intent.
|
the KDA chart shown isn't really surprising considering how it's pretty easy to get back to even once your team starts carrying you if you happen to win while having a bad game
|
You cannot make an accurate indicator of skill in any competitive setting as far as I'm concerned. The amount of variables required to accurately depict skill is unquantifiable, and anything assigned a number value is up to debate because we are inherently biased.
|
On December 08 2014 15:16 chalice wrote:Show nested quote +On December 08 2014 14:44 Goumindong wrote:
LOL how do you not understand that there exists no other way to find diamonds in the rough because no such measurement exists or can exist?
Literally you're asking "I want to know how good people are at winning" and I am sitting here saying "this is the stat which shows how good people are at winning, its literally the value for their winningness weighted against competition" basing your valuation of something using the same readily available information that everyone else is also using to assign a value to that thing does nothing to help you find the quality hidden beneath the surface of these "diamonds in the rough" you keep talking about. congrats on figuring out that winning players win more though, i appreciate the groundbreaking revelation. however, it seems clear that you do not in fact understand what i was asking, so i think everyone would benefit if you would refrain from further attempts at paraphrasing or explaining my intent.
On December 08 2014 10:14 chalice wrote:
the point isn't to find some hidden Diamond 5 gem who will take the LCS by storm, but rather to evaluate the play of those near the top of the ladder in comparison with one another.
like when you're putting together maknoon's NA team, you don't think there is any way that with enough research you might be able to find a challenger or master tier jungler with the potential to be better than nintendude instead of just sorting by elo and name recognition?
Those are your words, so that is what i was going for. When i say "diamond in the rough" its just a short hand for "finding a good player who isn't readily apparent already" it does not mean they have to be diamond 5 or diamond 1. It means any player and any relative ELO.
And I gave you a good method. Figure out what type of jungler your team needs, and go look at the best junglers by mmr who play those types of champions. MMR is literally "the stat for that"
If you want to talk about two players relative strengths and weaknesses then i can't help you. Because then we're back to those things like positioning et al, that we did not have a way to measure well, that are relative, that do not have easy breakdowns into situations which we can assign a score, like we can in basketball because we are dealing with points and independent events which have direct causation to the winning of the game, or that don't happen in a structured enough environment to make a difference(I.E. if your positioning goals change from game to game, determining where you ought to be and your relative position to that is much harder).
I know its not a groundbreaking revelation. But there isn't really any shortcut here.
|
If I was looking for this diamond, beyond pure MMR, I'd look at speed of victory (at least pre S5).
|
United States47024 Posts
On December 08 2014 13:05 chalice wrote:
you say this like it isn't basically the foundation of every scientific research work done ever. all i'm trying to do is see if anyone has any interesting hypotheses about what makes one player more skilled/win more games than another that might be worth testing.
For any work you do in science to be remotely acknowledged you have to spend years, if not decades studying the fundamentals and the work of others to build a baseline understanding that allows you to properly develop models to investigate. Any reasonable statistical model has to be based firmly upon fundamental knowledge that came before it (even if said knowledge is wrong, and requires the quantitative analysis to disprove it), much of which predates the use of such models in the first place.
It's actually the same with sports statistics as well. Most of the "good" models used in sports are also the product of professionals who spent years with the game either directly as part of a team or from a position where they followed the game in depth in a professional capacity (e.g. sports writers). And meaningful statistics could only be generated after a qualitative understanding of the game had developed after decades of play--even if that qualitative understanding led to some misconceptions that had to later be disproved. For the most part, you need a standardized, basic understanding of how the game "works" in order to develop meaningful models.
What people are doing now with LoL statistics are the equivalent of someone who's completed high school biology claiming that they can contribute meaningfully to the forefront of healthcare research. Or a high school physics student claiming that they can posit some groundbreaking insight in the field of particle physics.
On December 08 2014 13:38 Goumindong wrote:
The thing is, we know what makes players better, mechanics, decision making, positioning etc. What we don't have is a way of measuring those values in a way that means anything. This is partially because we don't track everything and partially because the structure of the game presents statistical issues which are hard to overcome.
"We" the League community as a whole might know these things. But "we" referring to the individuals who are poring over these statistics and developing shitty models don't.
If a reasonable model can be developed, it's FAR more likely to be done so by a coach/writer/analyst/ex-player who has knowledge of how the game works near the limit of what the entire League community collectively does and thinks about these kinds of things every day as part of their professional obligation. Not a Silver 2 struggler who can't make heads or tails of an LCS game and thinks of these things on a whim in his spare time.
|
|
|
what a fantastic rebuttal. i never thought of it that way
|
On December 08 2014 17:06 TheYango wrote:Show nested quote +On December 08 2014 13:38 Goumindong wrote:
The thing is, we know what makes players better, mechanics, decision making, positioning etc. What we don't have is a way of measuring those values in a way that means anything. This is partially because we don't track everything and partially because the structure of the game presents statistical issues which are hard to overcome.
"We" the League community as a whole might know these things. But "we" referring to the individuals who are poring over these statistics and developing shitty models don't. If a reasonable model can be developed, it's FAR more likely to be done so by a coach/writer/analyst/ex-player who has knowledge of how the game works near the limit of what the entire League community collectively does and thinks about these kinds of things every day as part of their professional obligation. Not a Silver 2 struggler who can't make heads or tails of an LCS game and thinks of these things on a whim in his spare time.
Yes. What i meant to convey was that those things which we qualify players as by watching their play, we still don't really have a way to quantify. We say doublelift has amazing mechanics but besides "he pushes the right buttons in the right order very fast, we think better than other people" that is all it means.
For instance it makes a lot of sense to say that its good to be between your opponents and their base (if you're on their side of the map and you've got a minion wave and can push a tower and they can't kill you as you do so) but how do you quantify that good positioning as a stat?
And then many times we don't even have a good understanding of what these "good play concepts" mean besides "they got good results". So, essentially, we don't have an equivalent to "being a good free throw shooter" because we barely even know what it means to be the equivalent of a good free throw shooter, let alone have can figure out a way to measure it. Certainly not in solo queue, certainly not on a systematic basis, we don't even record the type of data we might possibly need in order to do that and we don't even know what we would be looking for if we did.
|
![[image loading]](http://i.imgur.com/3pHpQNi.jpg)
Translation: Legendary Skin, Limited Edition in Transfer Windows, Miner Kakao Miner Rookie
Rumors said that Kakao&Rookie may need to play Chinese amateur league(not LSPL)
Chinese communities jokes that they are going mining
Unlike ex-Samsung players and Save,Fly etc, who joined big organizations, they choose to join an team even without a spot.
Rumors is that the boss/agent come from nowhere cheated them that he would buy a spot from LPL and form a team with Cool Uzi Gogoing, but obviously it is not happening.
iG want to buy Kakao, but the agent insist that rookie and kakao need to sell together, and the Deadline of transfer is 10th Dec, we don't know what will happen yet.
|
On December 08 2014 17:06 TheYango wrote:Show nested quote +On December 08 2014 13:05 chalice wrote:
you say this like it isn't basically the foundation of every scientific research work done ever. all i'm trying to do is see if anyone has any interesting hypotheses about what makes one player more skilled/win more games than another that might be worth testing.
For any work you do in science to be remotely acknowledged you have to spend years, if not decades studying the fundamentals and the work of others to build a baseline understanding that allows you to properly develop models to investigate. Any reasonable statistical model has to be based firmly upon fundamental knowledge that came before it (even if said knowledge is wrong, and requires the quantitative analysis to disprove it), much of which predates the use of such models in the first place. It's actually the same with sports statistics as well. Most of the "good" models used in sports are also the product of professionals who spent years with the game either directly as part of a team or from a position where they followed the game in depth in a professional capacity (e.g. sports writers). And meaningful statistics could only be generated after a qualitative understanding of the game had developed after decades of play--even if that qualitative understanding led to some misconceptions that had to later be disproved. For the most part, you need a standardized, basic understanding of how the game "works" in order to develop meaningful models. What people are doing now with LoL statistics are the equivalent of someone who's completed high school biology claiming that they can contribute meaningfully to the forefront of healthcare research. Or a high school physics student claiming that they can posit some groundbreaking insight in the field of particle physics. Show nested quote +On December 08 2014 13:38 Goumindong wrote:
The thing is, we know what makes players better, mechanics, decision making, positioning etc. What we don't have is a way of measuring those values in a way that means anything. This is partially because we don't track everything and partially because the structure of the game presents statistical issues which are hard to overcome.
"We" the League community as a whole might know these things. But "we" referring to the individuals who are poring over these statistics and developing shitty models don't. If a reasonable model can be developed, it's FAR more likely to be done so by a coach/writer/analyst/ex-player who has knowledge of how the game works near the limit of what the entire League community collectively does and thinks about these kinds of things every day as part of their professional obligation. Not a Silver 2 struggler who can't make heads or tails of an LCS game and thinks of these things on a whim in his spare time.
lol, you tried so hard to sound like you know what you're talking about and be vague in your insults, but the salt still shines through.
|
He's only saying that we're all shitters, whatever some people's ego may lead them to believe here.
It just so happens that said people's egos when not kept in check shit up GD even more than it already is and bystanders wonder why this thread isn't talking about League of Legends but about people mixing passive-agressiveness and mediocre trash talk about maths and their own fields and opinions on statistics.
|
On December 08 2014 19:27 Alaric wrote:
He's only saying that we're all shitters, whatever some people's ego may lead them to believe here.
It just so happens that said people's egos when not kept in check shit up GD even more than it already is and bystanders wonder why this thread isn't talking about League of Legends but about people mixing passive-agressiveness and mediocre trash talk about maths and their own fields and opinions on statistics.
It's too bad whenever people get called out on their nonsense without being named John Galt they get warned. =\
|
On December 08 2014 19:27 Alaric wrote:
He's only saying that we're all shitters, whatever some people's ego may lead them to believe here.
It just so happens that said people's egos when not kept in check shit up GD even more than it already is and bystanders wonder why this thread isn't talking about League of Legends but about people mixing passive-agressiveness and mediocre trash talk about maths and their own fields and opinions on statistics.
not that you're one to give advice about shitting up GD, but i do apologize for responding to goumindong's posts, i really should know better.
|
On December 08 2014 17:06 TheYango wrote:Show nested quote +On December 08 2014 13:05 chalice wrote:
you say this like it isn't basically the foundation of every scientific research work done ever. all i'm trying to do is see if anyone has any interesting hypotheses about what makes one player more skilled/win more games than another that might be worth testing.
For any work you do in science to be remotely acknowledged you have to spend years, if not decades studying the fundamentals and the work of others to build a baseline understanding that allows you to properly develop models to investigate. Any reasonable statistical model has to be based firmly upon fundamental knowledge that came before it (even if said knowledge is wrong, and requires the quantitative analysis to disprove it), much of which predates the use of such models in the first place. It's actually the same with sports statistics as well. Most of the "good" models used in sports are also the product of professionals who spent years with the game either directly as part of a team or from a position where they followed the game in depth in a professional capacity (e.g. sports writers). And meaningful statistics could only be generated after a qualitative understanding of the game had developed after decades of play--even if that qualitative understanding led to some misconceptions that had to later be disproved. For the most part, you need a standardized, basic understanding of how the game "works" in order to develop meaningful models. What people are doing now with LoL statistics are the equivalent of someone who's completed high school biology claiming that they can contribute meaningfully to the forefront of healthcare research. Or a high school physics student claiming that they can posit some groundbreaking insight in the field of particle physics.
I don't get it. You are saying that we already know how to construct great statistics and the people currently trying to come up with statistics are creating bad ones? I thought the whole point was that the field of LoL statistics is still very underdeveloped? In which case, of course any stats we come up with will suck for a long time, but that's not a reason to give up. That's like deciding to become an artist, drawing one picture, saying "well this sucks" and giving up. You have to go through the shitty to get to the good, that's how progress is made. It's an iterative process and we are still at the beginning.
Also, while MMR is a great measure for soloqueue performance, certainly something else is needed for analysing players in pro teams. Since there we have many games with the same players winrate doesn't mean a whole lot (we've seen a number of bad players getting carried on good teams and good players getting dragged down by bad teams). When a manager is looking to poach a player for a specific position they will want stats to look at to help them make an objective decision.
|
The general public does take longer to accept certain scientific theories, often due to their lack of training in scientific method and general unfamiliarity with nee concepts. But if you are at least lightly trained on the subject matter it's not hard to see the merits of someone else's theory by yourself.
"Decades" is a gross exaggeration only for the untrained eyes.
|
As far as "choosing players" go, I feel that mechanics is really not that important. After all, most of the mechanics can be trained and as far as I can tell, most pro players are pretty comparable except a few exceptions. I think it's more important to choose motivated players who are easy to work with and coachable.
One additional argument I got is that wildturtle and pob are on top of NA solo Q ladder, but their performance in LCS is pretty meh.
|
On December 08 2014 17:06 TheYango wrote:Show nested quote +On December 08 2014 13:05 chalice wrote:
you say this like it isn't basically the foundation of every scientific research work done ever. all i'm trying to do is see if anyone has any interesting hypotheses about what makes one player more skilled/win more games than another that might be worth testing.
For any work you do in science to be remotely acknowledged you have to spend years, if not decades studying the fundamentals and the work of others to build a baseline understanding that allows you to properly develop models to investigate. Any reasonable statistical model has to be based firmly upon fundamental knowledge that came before it (even if said knowledge is wrong, and requires the quantitative analysis to disprove it), much of which predates the use of such models in the first place. It's actually the same with sports statistics as well. Most of the "good" models used in sports are also the product of professionals who spent years with the game either directly as part of a team or from a position where they followed the game in depth in a professional capacity (e.g. sports writers). And meaningful statistics could only be generated after a qualitative understanding of the game had developed after decades of play--even if that qualitative understanding led to some misconceptions that had to later be disproved. For the most part, you need a standardized, basic understanding of how the game "works" in order to develop meaningful models. What people are doing now with LoL statistics are the equivalent of someone who's completed high school biology claiming that they can contribute meaningfully to the forefront of healthcare research. Or a high school physics student claiming that they can posit some groundbreaking insight in the field of particle physics. Show nested quote +On December 08 2014 13:38 Goumindong wrote:
The thing is, we know what makes players better, mechanics, decision making, positioning etc. What we don't have is a way of measuring those values in a way that means anything. This is partially because we don't track everything and partially because the structure of the game presents statistical issues which are hard to overcome.
"We" the League community as a whole might know these things. But "we" referring to the individuals who are poring over these statistics and developing shitty models don't. If a reasonable model can be developed, it's FAR more likely to be done so by a coach/writer/analyst/ex-player who has knowledge of how the game works near the limit of what the entire League community collectively does and thinks about these kinds of things every day as part of their professional obligation. Not a Silver 2 struggler who can't make heads or tails of an LCS game and thinks of these things on a whim in his spare time.
Actually, Yango, Bill James (Mr. SABRmetrics in baseball) and Aaron Schatz (Creator of DVOA, and WARYAR, the two most widely used, football advanced statistics) got into stats because they couldn't break into "real" sports journalism. Developing a predictive model would be one of the best ways to break into the Esports scene.
|
On December 08 2014 23:44 cLutZ wrote:Show nested quote +On December 08 2014 17:06 TheYango wrote:On December 08 2014 13:05 chalice wrote:
you say this like it isn't basically the foundation of every scientific research work done ever. all i'm trying to do is see if anyone has any interesting hypotheses about what makes one player more skilled/win more games than another that might be worth testing.
For any work you do in science to be remotely acknowledged you have to spend years, if not decades studying the fundamentals and the work of others to build a baseline understanding that allows you to properly develop models to investigate. Any reasonable statistical model has to be based firmly upon fundamental knowledge that came before it (even if said knowledge is wrong, and requires the quantitative analysis to disprove it), much of which predates the use of such models in the first place. It's actually the same with sports statistics as well. Most of the "good" models used in sports are also the product of professionals who spent years with the game either directly as part of a team or from a position where they followed the game in depth in a professional capacity (e.g. sports writers). And meaningful statistics could only be generated after a qualitative understanding of the game had developed after decades of play--even if that qualitative understanding led to some misconceptions that had to later be disproved. For the most part, you need a standardized, basic understanding of how the game "works" in order to develop meaningful models. What people are doing now with LoL statistics are the equivalent of someone who's completed high school biology claiming that they can contribute meaningfully to the forefront of healthcare research. Or a high school physics student claiming that they can posit some groundbreaking insight in the field of particle physics. On December 08 2014 13:38 Goumindong wrote:
The thing is, we know what makes players better, mechanics, decision making, positioning etc. What we don't have is a way of measuring those values in a way that means anything. This is partially because we don't track everything and partially because the structure of the game presents statistical issues which are hard to overcome.
"We" the League community as a whole might know these things. But "we" referring to the individuals who are poring over these statistics and developing shitty models don't. If a reasonable model can be developed, it's FAR more likely to be done so by a coach/writer/analyst/ex-player who has knowledge of how the game works near the limit of what the entire League community collectively does and thinks about these kinds of things every day as part of their professional obligation. Not a Silver 2 struggler who can't make heads or tails of an LCS game and thinks of these things on a whim in his spare time. Actually, Yango, Bill James (Mr. SABRmetrics in baseball) and Aaron Schatz (Creator of DVOA, and WAR, the two most widely used, football advanced statistics) got into stats because they couldn't break into "real" sports journalism. Developing a predictive model would be one of the best ways to break into the Esports scene.
Remind me, how long professional baseball has been around? Baseball is also the easiest sports I can think of to make stats for since a majority of interactions are on a solitary or 1v1 basis.
|
On December 08 2014 23:51 Gahlo wrote:Show nested quote +On December 08 2014 23:44 cLutZ wrote:On December 08 2014 17:06 TheYango wrote:On December 08 2014 13:05 chalice wrote:
you say this like it isn't basically the foundation of every scientific research work done ever. all i'm trying to do is see if anyone has any interesting hypotheses about what makes one player more skilled/win more games than another that might be worth testing.
For any work you do in science to be remotely acknowledged you have to spend years, if not decades studying the fundamentals and the work of others to build a baseline understanding that allows you to properly develop models to investigate. Any reasonable statistical model has to be based firmly upon fundamental knowledge that came before it (even if said knowledge is wrong, and requires the quantitative analysis to disprove it), much of which predates the use of such models in the first place. It's actually the same with sports statistics as well. Most of the "good" models used in sports are also the product of professionals who spent years with the game either directly as part of a team or from a position where they followed the game in depth in a professional capacity (e.g. sports writers). And meaningful statistics could only be generated after a qualitative understanding of the game had developed after decades of play--even if that qualitative understanding led to some misconceptions that had to later be disproved. For the most part, you need a standardized, basic understanding of how the game "works" in order to develop meaningful models. What people are doing now with LoL statistics are the equivalent of someone who's completed high school biology claiming that they can contribute meaningfully to the forefront of healthcare research. Or a high school physics student claiming that they can posit some groundbreaking insight in the field of particle physics. On December 08 2014 13:38 Goumindong wrote:
The thing is, we know what makes players better, mechanics, decision making, positioning etc. What we don't have is a way of measuring those values in a way that means anything. This is partially because we don't track everything and partially because the structure of the game presents statistical issues which are hard to overcome.
"We" the League community as a whole might know these things. But "we" referring to the individuals who are poring over these statistics and developing shitty models don't. If a reasonable model can be developed, it's FAR more likely to be done so by a coach/writer/analyst/ex-player who has knowledge of how the game works near the limit of what the entire League community collectively does and thinks about these kinds of things every day as part of their professional obligation. Not a Silver 2 struggler who can't make heads or tails of an LCS game and thinks of these things on a whim in his spare time. Actually, Yango, Bill James (Mr. SABRmetrics in baseball) and Aaron Schatz (Creator of DVOA, and WAR, the two most widely used, football advanced statistics) got into stats because they couldn't break into "real" sports journalism. Developing a predictive model would be one of the best ways to break into the Esports scene. Remind me, how long professional baseball has been around? Baseball is also the easiest sports I can think of to make stats for since a majority of interactions are on a solitary or 1v1 basis.
The point I was making has nothing to do with the difficulty of doing it. Instead it was to demonstrate that entrenched journalists and public speakers are among the least likely to develop a working model. Peter King did not, Peter Gammons has not, ESPN developed "Total QBR", which is not a good advanced metric, its just an easy metric to understand (with almost no predictive power).
The other route is people developing proprietary systems. The Red Sox, Spurs, Oakland-A's, and some others have well developed proprietary advanced stats, but that helps us none here. SK Telecom might actually have a really sweet algorithm for finding SoloQ talent that will translate, but they will not share that with us, because it is part of the competitive advantage.
So no, Travis, Montecristo, and Richard Lewis will not be developing league stats, because that is not in their interest. They sell access, developing advanced stats could be counterproductive for them because it can turn off players/teams/etc that don't like the stats and how they reflect on them.
|
|
|
|
|
|
EPT
