EPT[Patch 4.4] CUDDLY INCOMING! ヽ(*・ω・)ﾉ - Page 99
| Forum Index > LoL General |
|
FinestHour
United States18466 Posts
| ||
|
Goumindong
United States3529 Posts
On March 28 2014 10:02 Sufficiency wrote: how does it matter for two values?I think you should posting too. I think I have lost about all the respect I had for you. Sigh. I mean your values are either zero or one. Either Leona is in a game or not. Either Leona and Ashe are in a game or not. It doesn't matter what values you assign to "Leona and Ashe" are in a game and "Leona and Ashe aren't in a game" it's the same math. Edit: it changes the coefficient but only in a way which requires you to modify your interpretation of the results. This is why logistic regressions tend to look at continuous independent variables rather than binary. | ||
AsmodeusXI
United States15536 Posts
On March 28 2014 10:02 Sufficiency wrote: I think you should posting too. I think I have lost about all the respect I had for you. Sigh. If this is going to be your attitude now, then you should stop posting. If you actually believe yourself to be correct, then resorting to condescension towards those arguing with you is not going to make your arguments hold any more weight. | ||
|
petered
United States1817 Posts
On March 28 2014 10:12 FinestHour wrote: shots firing everywhere more at 11 What is funny is that I stopped lurking in order to argue with TheYango. So many stupid people on the internet I could have argued with and I had to pick one who is most definitely smarter than me. | ||
|
canikizu
4860 Posts
| ||
|
petered
United States1817 Posts
On March 28 2014 10:35 canikizu wrote: Can somebody tldr what people are arguing about here? There is a lot to cover but let me try: Is Warwick a viable support? The answer is maybe, but only if his coefficient is high enough or something like that. | ||
TheYango
United States47024 Posts
On March 28 2014 10:04 petered wrote: So you need 100+ years of history to be able to understand a game through numbers? Yango you disappoint, that is nonsensical. The math behind data analysis is not that drastically different from one area of research to the next, making it possible to draw from the rich experience of other fields of research. The sheer volume of LoL games played and the numerous quantitative measures that can be drawn from each game make it a fantastic target for this type of analysis. You need a qualitative understanding of what your numbers mean and what you are trying to model with them for a model to be meaningful. A high school biology student who has not completed a semester of basic biology isn't going to take large scale DNA sequencing data and use it to build a meaningful model for anything. Now, I'll be honest, I didn't even look very much at what Sufficiency posted. I am just so confused when people flip their shit about someone trying to use numbers to better understand the game. Sure there are challenges and caveats with some studies, but that does not mean that he is going down the wrong path or that what he has presented is useless. Disagree with his study/conclusions fine, but this will inevitably be the future of understanding the game, since that is what is happening is just about every other field. The reason people "flip their shit" is because he hasn't demonstrated any understanding of the game. He has quite literally compiled statistics that someone who has never seen a game of LoL in their life could have compiled and tried to sell them off as meaningful. Relative winrates don't model any actual interactions that take place in the game, but by titling his post as "champion counters" he's implying that they do. | ||
|
Kupon3ss
時の回廊10066 Posts
If one takes a random subset of 100 colors and splash them randomly onto millions of paintings, then go through the paintings to calculate which colors overlap the most, is that still art? futhermore, if a champion counters every other champion + Show Spoiler + ![[image loading]](http://2.bp.blogspot.com/-c3_e2hbA2u8/UzJwWEzF5hI/AAAAAAAAAGg/s_nZ1b6v6VY/s1600/wukong.png) Is it better to blame riot for OP or bad stats And finally, if I ran 100 P-tests for each of 100 possible variables, what is the percentage of P-tests that would return has statistically significant assuming that everything is based on random chance? | ||
|
Sufficiency
Canada23833 Posts
On March 28 2014 10:04 petered wrote: So you need 100+ years of history to be able to understand a game through numbers? Yango you disappoint, that is nonsensical. The math behind data analysis is not that drastically different from one area of research to the next, making it possible to draw from the rich experience of other fields of research. The sheer volume of LoL games played and the numerous quantitative measures that can be drawn from each game make it a fantastic target for this type of analysis. Now, I'll be honest, I didn't even look very much at what Sufficiency posted. I am just so confused when people flip their shit about someone trying to use numbers to better understand the game. Sure there are challenges and caveats with some studies, but that does not mean that he is going down the wrong path or that what he has presented is useless. Disagree with his study/conclusions fine, but this will inevitably be the future of understanding the game, since that is what is happening is just about every other field. Thank you. Finally someone who actually has some ideas of what he is talking about. | ||
|
Sufficiency
Canada23833 Posts
On March 28 2014 10:39 TheYango wrote: You need a qualitative understanding of what your numbers mean and what you are trying to model with them for a model to be meaningful. A high school biology student who has not completed a semester of basic biology isn't going to take large scale DNA sequencing data and use it to build a meaningful model for anything. The reason people "flip their shit" is because he hasn't demonstrated any understanding of the game. He has quite literally compiled statistics that someone who has never seen a game of LoL in their life could have compiled and tried to sell them off as meaningful. Relative winrates don't model any actual interactions that take place in the game, but by titling his post as "champion counters" he's implying that they do. So you are saying I do not have a basic understanding of the game? I think I do, thank you very much. I think you do not have a basic understanding of how statistical analysis works, which brought you to this (IMO) very unfair criticism of my results. Then you try to justify your thoughts with this "we need time for qualitative analysis" kind of thing. In any case, let's not talk about this anymore. On March 28 2014 10:41 Kupon3ss wrote: The argument is more fundamental, If one takes a random subset of 100 colors and splash them randomly onto millions of paintings, then go through the paintings to calculate which colors overlap the most, is that still art? futhermore, if a champion counters every other champion + Show Spoiler + Is it better to blame riot for OP or bad stats And finally, if I ran 100 P-tests for each of 100 possible variables, what is the percentage of P-tests that would return has statistically significant assuming that everything is based on random chance? Yes, I remember that one. It's very interesting how it turned out. The short answer is that the table is filtered, so it does not list every single champion - only the ones which it thinks the effect is strong. | ||
|
TheYango
United States47024 Posts
This kind of analysis would actually have some more real applicability. Once you use the data to establish the validity of your statistics based on known outcomes (i.e. these are games where we can definitively say player A won the lane, and the statistic matches that), we can use this to address the gray area cases where qualitative analysis fails (e.g. champion A gets behind champion B on CS but typically gank assists better and can gank other lanes more easily--who "wins"?). A winrate table doesn't do anything. On March 28 2014 10:46 Sufficiency wrote: So you are saying I do not have a basic understanding of the game? I think I do, thank you very much. I think you do not have a basic understanding of how statistical analysis works, which brought you to this (IMO) very unfair criticism of my results. Then you try to justify your thoughts with this "we need time for qualitative analysis" kind of thing. In any case, let's not talk about this anymore. I'm not saying you don't have one. I'm saying that an relative winrate table doesn't utilize it whatsoever because at no point does your analysis ever derive anything from anything directly related to the game. This has been my point the whole time--"champion A wins game against champion B" is so far removed from any process that occurs within the game that it simply cannot be used to model any of those processes--which are where practical application of such analyses might occur. | ||
|
Kupon3ss
時の回廊10066 Posts
PS: Did you know that the P-value of a test that evaluates whether or not Blue Side counters Purple Side in ARAM is 0? | ||
|
Sufficiency
Canada23833 Posts
On March 28 2014 10:50 Kupon3ss wrote: But its a winrate table that combines the aggregate of everything from the equivalent of little league to semi-pro games with 0 rigor or segregation of results operating under a model in which each champion's winrate is extrapolate through all levels of play and possible matchups. How could it possibly be wrong, its got millions of datapoints and a minuscule P-value. PS: Did you know that the P-value of a test that evaluates whether or not Blue Side counters Purple Side in ARAM is 0? Yes. If you do it properly, you can clearly show there blue side has an advantage. This is in ARAM as well as on Summoners' Rift. | ||
|
Ketara
United States15065 Posts
For example, I asked for Sufficiency to make a table with the Lux data for his winrates, and I put it in the Lux guide. In said guide I've already got rather detailed matchup analysis, both from my own experience and from talking to Dai, who is our resident D1 Lux player. Sufficiency's table doesn't really contradict anything we've said in the matchup section. But what it does do is add a little bit of reliability to what we've said. I didn't put it in the guide because I thought it was a definitive set of statistics on Lux matchups. If it came up with a matchup that looked wildly different to my own experience, I'd be pretty confident in calling the statistic model wrong. But since said statistics and our game experience more or less agree, I think it's good to include because they add weight to each other. I think that's worth an hour of Sufficiency's PC resources, and I think it's pretty cool that he decided to do it for me, and I don't know why he's getting such shit for it. | ||
|
TheYango
United States47024 Posts
On March 28 2014 11:06 Ketara wrote: But what it does do is add a little bit of reliability to what we've said. I didn't put it in the guide because I thought it was a definitive set of statistics on Lux matchups. If it came up with a matchup that looked wildly different to my own experience, I'd be pretty confident in calling the statistic model wrong. But since said statistics and our game experience more or less agree, I think it's good to include because they add weight to each other. So then where does the value come from If it holds zero weight when it differs from actual experience? | ||
|
cLutZ
United States19573 Posts
On March 28 2014 09:51 TheYango wrote: "Every facet of modern research" demands numbers because they are necessary for the precision and rigor required in their respective fields. Not to mention that they all went through hundreds of years of qualitative analysis and general theory before they reached the point where numbers could be practically applied. We're talking about a game that has only existed for less than 10 years. Even when applied to a game like baseball there was at least some qualitative understanding of the statistics being mined (and like 100+ years of baseball theory) before something like Moneyball could happen. Nobody has the qualitative understanding of the game necessary to draw proper conclusions from data like this and build meaningful models. The qualitative understanding of what a lot of numbers actually mean isn't there yet. The statistics we have are the analogues for stuff like batting averages that were proven to be useless. The complex aggregate statistics which are actually meaningful don't even exist yet--and many of those were developed through anecdotal impressions of their relevance before statistics showed them to be so. I dont really want to get into the Math Shitstorm, but I do know something about baseball stats. And what we know about baseball stats is that before people looked at the new, so called "advanced", stats they were wrong about like 1/2+ of the stats they thought were important for winning. So, just FYI, maybe avoid that analogy in the future. | ||
|
Shikyo
Finland33997 Posts
On March 28 2014 10:38 petered wrote: There is a lot to cover but let me try: Is Warwick a viable support? The answer is maybe, but only if his coefficient is high enough or something like that. I have actually played plenty of warwick support in the low-mid platinum range. In my opinion, he's a somewhat viable support. His bottom lane presence is so-so, but he camps the brush pretty well and he's very hard to push off the lane. After 6, you win practically every 2v2 and your gank assist is world class. In the lategame, your peel against a single target(bruiser diver) is world class as well. I have maxed his W first and practically pretended I'm a Nunu who doesn't give movespeed and I didn't exactly win lanes but they didn't go horribly either and in the lategame I believe that Warwick outscales most supports. His W is just incredibly powerful in teamfights. | ||
|
Sufficiency
Canada23833 Posts
On March 28 2014 11:11 TheYango wrote: So then where does the value come from If it holds zero weight when it differs from actual experience? I just solved this math problem and the answer I got was 42. I asked my friend, he told me he got 42 as well using a different method. Unfortunately his answer held zero weight. | ||
|
Requizen
United States33802 Posts
On March 28 2014 11:16 Sufficiency wrote: I just solved this math problem and the answer I got was 42. I asked my friend, he told me he got 42 as well using a different method. Unfortunately his answer held zero weight. I punch those numbers into my calculator and it makes a happy face. What are we doing? | ||
|
TheYango
United States47024 Posts
On March 28 2014 11:12 cLutZ wrote: I dont really want to get into the Math Shitstorm, but I do know something about baseball stats. And what we know about baseball stats is that before people looked at the new, so called "advanced", stats they were wrong about like 1/2+ of the stats they thought were important for winning. So, just FYI, maybe avoid that analogy in the future. That's exactly the point I was trying to get at, though and precisely WHY I used the baseball analogy. | ||
| ||
