EPTGeneral Discussion - Page 4429
| Forum Index > Dota 2 General |
IRC chatter should remain in IRC - http://webchat.quakenet.org/?channels=tl.dota2 Posts that relate to topics with their own thread, such as in News, Tournaments or Strategy should go in those threads. | ||
|
Erasme
Bahamas15899 Posts
| ||
|
crms
United States11933 Posts
On November 30 2013 16:16 Erasme wrote: Yes. lewl. anything for site traffic. | ||
TheEmulator
28103 Posts
| ||
|
crms
United States11933 Posts
On November 30 2013 16:26 TheEmulator wrote: Well I've already heard of multiple people who solo queue into page 1-5 regularly that have like 800-1200 ratings. This site is definitely broken. the author said it goes off of win rate, kda, xpm, gpm compared to other players of that same hero. however, it doesn't take into consideration the level of players you're currently playing. basically what that means is all your pub stomping noob smurfs will likely have a super high rating because your stats will be hyper inflated vs low level players. one of my smurfs is 'pro' level while my main was like barely gold (maybe silver? can't remember) even though I've regularly queued top 10 pages and it's been in 'very high' (valves top ~2.5%) since valve introduced the brackets, and subsequently removed them, or made them retarded to find I guess is more accurate. tl;dr this site is ass. also, the site rating system puts loda as alliance's best player and EGM as the 'worst' which should show it's clearly retarded. huehue. | ||
|
teapoted
United Kingdom24425 Posts
Like in looking at deaths it would need to also look at where you're dying, if you got anything out of it for you or your team, if it led to you losing. The death count at the end is only like 20% of the story. And the same goes for all the statistics. Like Loda has the highest average GPM in competitive Dota apart from lakelz who has significantly less matches. But how much does win percentage play into that. And how often is that number a sign of games already won or if it's actually contributing to the win. Black has very high average GPM, and it's arguable that it's due to issues in his playstyle, not because of great play. | ||
|
crms
United States11933 Posts
if valve added some sort of in-game ranking, you guys wanna discuss that? guys? HEHEHEHEHEHHHEHEHEHEHEHEHEHEHE (pls dont ban) | ||
|
teapoted
United Kingdom24425 Posts
But it's a very complicated problem both technically and in regards to community management. | ||
|
Staboteur
Canada1873 Posts
On November 30 2013 16:45 crms wrote: you know what would be cool way to stop all of these really bad sites from making really bad ratings? if valve added some sort of in-game ranking, you guys wanna discuss that? guys? HEHEHEHEHEHHHEHEHEHEHEHEHEHEHE (pls dont ban) Side one: I AM OFFENDED BY YOUR REQUEST FOR SOMETHING THAT I WOULD NEVER USE, AND ITS EXISTENCE WOULD BE DETRIMENTAL TO MY ENJOYMENT OF DOTA. Side two: I HAVE A HARD TIME RESPECTING THAT THIS WOULD BE DETRIMENTAL TO YOUR ENJOYMENT OF DOTA, AND THEREFORE CANNOT PRODUCTIVELY CONTINUE THIS DISCUSSION. BUT I'LL YELL AT YOU ANYWAYS BECAUSE I DON'T UNDERSTAND YOU. Ad infinitum. Best discussion ever. On November 30 2013 16:47 teapoted wrote: I think if they actually had a good way to quantify it where it isn't just a number that you can stroke yourself over and would actually be useful feedback then they would have implemented it. But it's a very complicated problem both technically and in regards to community management. Sc2 and game-we-don't-name use leagues as vague indicators of skill bracket, and it doesn't cause their communities to implode. The framework has already been shown to exist in Dota with the Vhigh / high / etc skill brackets, which gives me hope that Volvo has some plans regarding visible indicators of skill in the future. What's nice about leagues is that they're extremely vague. Hard-number indicators like this new thing are kinda ridiculous because they're aiming to be far more precise than possible. | ||
|
FinestHour
United States18466 Posts
| ||
|
synapse
China13814 Posts
On November 30 2013 16:36 crms wrote: the author said it goes off of win rate, kda, xpm, gpm compared to other players of that same hero. however, it doesn't take into consideration the level of players you're currently playing. basically what that means is all your pub stomping noob smurfs will likely have a super high rating because your stats will be hyper inflated vs low level players. one of my smurfs is 'pro' level while my main was like barely gold (maybe silver? can't remember) even though I've regularly queued top 10 pages and it's been in 'very high' (valves top ~2.5%) since valve introduced the brackets, and subsequently removed them, or made them retarded to find I guess is more accurate. tl;dr this site is ass. also, the site rating system puts loda as alliance's best player and EGM as the 'worst' which should show it's clearly retarded. huehue. haha the rating fucks me over then, ive basically queue'd with friends in very high after only learning for 2 days.... fed so much for the first month or so... | ||
|
Cool Cat
United States1644 Posts
On November 30 2013 16:59 FinestHour wrote: holy ufck listening to pucks voice in korean makes me almost want to play this game again korean voices own dawg | ||
|
teapoted
United Kingdom24425 Posts
On November 30 2013 16:59 FinestHour wrote: I just checked my dotabuff and it's actually been over a year since I played a match of Dota 2. I've even played Dota 1 in that time lol.holy ufck listening to pucks voice in korean makes me almost want to play this game again And shit I buy a lot of bracers. | ||
|
KlaCkoN
Sweden1661 Posts
On November 30 2013 16:41 teapoted wrote: An accurate skill rating would have to be significantly more complex than looking at the end game score screen. Like in looking at deaths it would need to also look at where you're dying, if you got anything out of it for you or your team, if it led to you losing. The death count at the end is only like 20% of the story. And the same goes for all the statistics. Like Loda has the highest average GPM in competitive Dota apart from lakelz who has significantly less matches. But how much does win percentage play into that. And how often is that number a sign of games already won or if it's actually contributing to the win. Black has very high average GPM, and it's arguable that it's due to issues in his playstyle, not because of great play. I don't think there is anything a fancy "skill" algorithm could do that a simple win/lose Elo system wouldn't do better. The better players you win against the higher your ranking. It would take longer to converge than it would for a 1v1 game but it would converge just the same. | ||
Firebolt145
Lalalaland34507 Posts
Player number 36931272 DSR is: 1025 This player has a Gold DSR score (top 30%). I'm awful at this game. | ||
|
teapoted
United Kingdom24425 Posts
On November 30 2013 18:04 Firebolt145 wrote: When Dotabuff had their rating thing I was top 10%, this rating thing I'm top 90% (whatever that means). I guess it doesn't take kindly to ward bitching bracer gamers.Player number 36931272 DSR is: 1025 This player has a Gold DSR score (top 30%). I'm awful at this game. | ||
|
Firebolt145
Lalalaland34507 Posts
DSR obviously more accurate guys | ||
|
Shaella
United States14828 Posts
I guess vhigh is shit tier now guys | ||
|
cecek
Czech Republic18921 Posts
| ||
|
Cool Cat
United States1644 Posts
| ||
|
TheEmulator
28103 Posts
On November 30 2013 19:11 Cool Cat wrote: i'm at 888 dsr naisu How are you not 2k+ Captain yomi. | ||
| ||
