|
East Gorteau22261 Posts
On April 25 2014 04:13 woopr wrote: What's the point of having a rating system if it means nothing?
How many kebabs does it take to fill a submarine?
The answer is I don't know, why do you ask?
|
On April 25 2014 04:16 Zealously wrote:Show nested quote +On April 25 2014 04:13 woopr wrote: What's the point of having a rating system if it means nothing? How many kebabs does it take to fill a submarine? The answer is I don't know, why do you ask?
We should ask the frenchies, their currency is kebab afterall.
|
On April 25 2014 04:13 woopr wrote: What's the point of having a rating system if it means nothing? You can say the same about anything that uses probability and statistical models. It does mean something, you just have to be careful applying it.
In this case. The ratings and predictions will be fine as long as the players are in the same "region". We don't have a system that works globally, yet.
|
So the blind romance between aligulac and ForGG remains...
|
On April 25 2014 03:33 Die4Ever wrote: I noticed this in the FAQ
"Rating gap cap. The idea here is to prevent players from "farming" much lower rated players. It is possible to artificially inflate a player's rating if he never plays other players close to himself in skill. The ideas usually consist of ignoring matches where players are farther away in rating than a given threshold. However, this can be seen as unfair to the lower rated players, who will have their wins against good players discarded."
Couldn't you implement this as "ignore the match if the weaker player loses", as if saying "it went as expected" so no change is needed? I'm not sure how helpful this would be though, just a random thought.
Also does a Bo5 win count more than a Bo3 win? I forget if you count by games or matches. I think a 4-3 win should count as better for the winner than a 2-1 win would, even though the differential is the same. The length of the match matters, but not nearly as much as the win percentage. So for example, 3-0 in a Bo5 is better than a 2-0 in Bo3, which is better than a 3-1 in Bo5, which is better than a 2-1 in Bo3, which is better than a 3-2 in Bo5, which is better than 2-3, 1-2, 1-3, 0-2, 0-3, etc.
Disclaimer: Math done while drunk, I can't guarantee correctness of details, only of intent.
On April 25 2014 03:57 herMan wrote:Show nested quote +On April 25 2014 01:03 TheBB wrote:ForGGeddon was a gimmick that was going on past summer/fall when he was threatening to make world #1 after moving to Europe. It was interesting then because there weren't very many Koreans who were competing abroad regularly. Nowadays they are much more common, and this has lead to a bit of an inflation in the top (> 2000 points is no longer uncommon). It's not so surprising at this point that he or someone like him would manage eventually. Yes, we're very aware of the shortcomings in the rating system. It's easy to come up with ideas but it's difficult to implement them in a way that doesn't create funny business in some places. I could elaborate but I've frankly had my fill of debating rating systems on TL. ForGG moved to Europe in 2012 so this gimmick thing emerged long after it. Too bad he didn't match the hype with success. Sorry yes, I underestimated the amount of time we've been alive.
|
On April 25 2014 01:27 vult wrote:Well let's see if he ever wins anything
He's the only oGs Terran who has remained relevant for so long!!!
|
Canada16217 Posts
On April 25 2014 08:04 Gamegene wrote:He's the only oGs Terran who has remained relevant for so long!!! SuperNova?
|
Pretty surprised given the low significance of any of his victories. I suppose he just has that much consistency against EU. I'd like for Aligulac to better adjust for this, because I think we all know that ForGG is NOT as good as many of the Code S pros (Maru, Zest, herO, even Polt)
|
On April 25 2014 08:05 NovemberstOrm wrote:Show nested quote +On April 25 2014 08:04 Gamegene wrote:On April 25 2014 01:27 vult wrote:Well let's see if he ever wins anything He's the only oGs Terran who has remained relevant for so long!!! SuperNova?
DAMN IT. lol
|
On April 25 2014 04:24 Prillan wrote:Show nested quote +On April 25 2014 04:13 woopr wrote: What's the point of having a rating system if it means nothing? You can say the same about anything that uses probability and statistical models. It does mean something, you just have to be careful applying it. In this case. The ratings and predictions will be fine as long as the players are in the same "region". We don't have a system that works globally, yet.
How good is it in predicting actual results? If it's an ELO based system and you have thousands of games it should be pretty good. On the other hand if the rating system has ForGG at the top of the heap I have serious concerns with how it was programmed and perhaps it's not so good. I think because of the limited foreign tournament opportunities for most of the Korean KESPA players, you should create a Korean database (based on GSL, GSTL, Proleague performance and performance in GOM sponsored events) and then a database that rates everyone based on all other tournaments.
|
On April 25 2014 04:16 Zealously wrote:Show nested quote +On April 25 2014 04:13 woopr wrote: What's the point of having a rating system if it means nothing? How many kebabs does it take to fill a submarine? The answer is I don't know, why do you ask?
with regards to woopr, it's because people like to put numbers on everything.
|
On April 25 2014 08:25 The_Darkness wrote:Show nested quote +On April 25 2014 04:24 Prillan wrote:On April 25 2014 04:13 woopr wrote: What's the point of having a rating system if it means nothing? You can say the same about anything that uses probability and statistical models. It does mean something, you just have to be careful applying it. In this case. The ratings and predictions will be fine as long as the players are in the same "region". We don't have a system that works globally, yet. How good is it in predicting actual results? If it's an ELO based system and you have thousands of games it should be pretty good. On the other hand if the rating system has ForGG at the top of the heap I have serious concerns with how it was programmed and perhaps it's not so good. I think because of the limited foreign tournament opportunities for most of the Korean KESPA players, you should create a Korean database (based on GSL, GSTL, Proleague performance and performance in GOM sponsored events) and then a database that rates everyone based on all other tournaments.
Maybe just limit it to tournaments that have WCS points (IEM, Dreamhack, Red Bull, WCS, etc.) and also major team leagues like Pro league, GSTL and Acer Team Story Cup.
|
On April 25 2014 00:10 Hider wrote: Aliguac really needs to come up with a smart way of quantifying the skill-difference of the korean and foreign scene. Otherwise their rankings are basically useless.
i think the only possible way to do this is to have 2 different ratings, one against korean, one against foreigner.
|
Actually there needs to be points decay on this system. MVP, MC, Nestea and a few otehr all have inflated scores due to WoL success.
|
Haha nice :D Aligulac isn't doing so great on that front :/
|
On April 25 2014 08:40 Orcasgt24 wrote: Actually there needs to be points decay on this system. MVP, MC, Nestea and a few otehr all have inflated scores due to WoL success.
For Nestea and in some parts Mvp, i'd agree with you, but MC's rating isn't inflated because of his WoL success, hes been successful for almost all of HOTS as well, hell, he just won WCS EU.
|
On April 25 2014 08:48 Josh_Video wrote:Show nested quote +On April 25 2014 08:40 Orcasgt24 wrote: Actually there needs to be points decay on this system. MVP, MC, Nestea and a few otehr all have inflated scores due to WoL success. For Nestea and in some parts Mvp, i'd agree with you, but MC's rating isn't inflated because of his WoL success, hes been successful for almost all of HOTS as well, hell, he just won WCS EU. For the same reason ForGG is so highly ranked. MC's win-rate vs Koreans is actually sub-50% in HotS.
|
On April 25 2014 04:16 Zealously wrote:Show nested quote +On April 25 2014 04:13 woopr wrote: What's the point of having a rating system if it means nothing? How many kebabs does it take to fill a submarine? The answer is I don't know, why do you ask?
To criticize the rating system?
|
i'll never believe forgg is better than polt for example. if you watch both their streams for like an hour it's pretty apparent Polt is better.
|
I really wonder if it wouldn't be possible to implement some type of regression where you looked at historial results from players whom have played most of their games against foreigners and players whom have played the majority of their games against Koreans. Now if the latter group outperforms compared to what you would expect given the "naive" Aliguac predictions, you could adjust the rankings for that.
|
|
|
|