|
this place is amazing, its like "reddits community" but without downvotes and the moderation here is worse than playdota cuz u get "warned" for high quality posting.
On October 04 2014 00:42 ahswtini wrote:Show nested quote +On October 04 2014 00:24 wavebave wrote:On October 03 2014 23:17 Torte de Lini wrote:On October 03 2014 21:55 wavebave wrote:
how isnt it a viable solution? a 3k mmr player spamming zeus ulti whenever its off cd gives them 5k straight away
if valve decided to look through statistics of how many tbd players actually picked zeus it'd probably be the most picked hero with a rate of 80% among tbds because you're removing choice from players. It's something you need to uphold and removing Zeus will choose have TBD abusers switch to another hero that is a near second. lmfao ok give me another hero that doesn't require skills but can get to at least one third of zeus' average hero dmg at the end of a match fucking god this forum is like playdota's dota chat section except ppl try to be serious here lolz Except you're the one who's brainless. The problem here is that getting high hero damage inflates your calibration rating. Not that zeus is the easiest hero to get a high hero damage with. Your solution is to attack the symptom of the problem, not the problem itself. Worse still you do it with a nuke and not with a precision strike. That's great, you still haven't fixed the root of the problem so the abusers will just move onto the next best hero for the job.
ur like 5k posts in and ur most likely a tbd zeus player by the way u "argue". im asking u to show me a hero that can be played by 3ks and 4ks while still reaching 5k+ that can give at least ONE THIRD of zeus total hero dmg of an average match and u cant,
do u have to be a scientist to understand people pick zeus to inflate their rating by additional 1k+ points because its the easiest hero in regards to hero damage? HOLY FUCK you are a fucking troll
fuck it im done but these high quality bait posts just suck me right in
ive betrayed u playdota
User was banned for this post.
|
I don't know if there's an easy solution to this.
Suppose I'm a 6k player and I start a new account. Valve has no idea what my actual MMR might be so the reasonable thing to do would be guess that I'm average (~2.5k?) and work from there. If we ONLY use win/lose as the metric for calibration as some people suggest, by the time I'm done with the 40 unranked games, I'll only be at 3.5k even if I win all of them. After winning my 10 calibration matches, I'll get placed at 3.75k. In order to get up to 5k,I have to play through at least 100 games. This would suck for me and it would suck for my opponents since they're facing a 6k player and basically have no chance of winning.
Valve obviously thought that they need some metric to determine how I'm performing relative to everyone else in my games so they could bump my MMR up faster than 25 per game. Maybe they chose a formula like this:
(K + A - D)/(K_avg + A_avg - D_avg) + 0.5 * HD/HD_avg + 0.5 * HH/HH_avg + 0.3 * TD/TD_avg...
Just pulled random numbers out of my ass but guess which part of that formula is easiest to abuse. People are saying that Valve needs to address the root of the problem, which is that HD weights heavily into calibration. However, I think doing so will do more harm than good as stated above. It might be better just to do band-aid fixes that address specific abuses, like making Zeus ult and other globals not count towards HD or something.
|
On October 04 2014 04:14 B1nary wrote:
I don't know if there's an easy solution to this.
Suppose I'm a 6k player and I start a new account. Valve has no idea what my actual MMR might be so the reasonable thing to do would be guess that I'm average (~2.5k?) and work from there. If we ONLY use win/lose as the metric for calibration as some people suggest, by the time I'm done with the 40 unranked games, I'll only be at 3.5k even if I win all of them. After winning my 10 calibration matches, I'll get placed at 3.75k. In order to get up to 5k,I have to play through at least 100 games. This would suck for me and it would suck for my opponents since they're facing a 6k player and basically have no chance of winning.
Valve obviously thought that they need some metric to determine how I'm performing relative to everyone else in my games so they could bump my MMR up faster than 25 per game. Maybe they chose a formula like this:
(K + A - D)/(K_avg + A_avg - D_avg) + 0.5 * HD/HD_avg + 0.5 * HH/HH_avg + 0.3 * TD/TD_avg...
Just pulled random numbers out of my ass but guess which part of that formula is easiest to abuse. People are saying that Valve needs to address the root of the problem, which is that HD weights heavily into calibration. However, I think doing so will do more harm than good as stated above. It might be better just to do band-aid fixes that address specific abuses, like making Zeus ult and other globals not count towards HD or something.
If you're a 6k player, why would you start a new account
|
28090 Posts
On October 04 2014 05:18 evanthebouncy! wrote:Show nested quote +On October 04 2014 04:14 B1nary wrote:
I don't know if there's an easy solution to this.
Suppose I'm a 6k player and I start a new account. Valve has no idea what my actual MMR might be so the reasonable thing to do would be guess that I'm average (~2.5k?) and work from there. If we ONLY use win/lose as the metric for calibration as some people suggest, by the time I'm done with the 40 unranked games, I'll only be at 3.5k even if I win all of them. After winning my 10 calibration matches, I'll get placed at 3.75k. In order to get up to 5k,I have to play through at least 100 games. This would suck for me and it would suck for my opponents since they're facing a 6k player and basically have no chance of winning.
Valve obviously thought that they need some metric to determine how I'm performing relative to everyone else in my games so they could bump my MMR up faster than 25 per game. Maybe they chose a formula like this:
(K + A - D)/(K_avg + A_avg - D_avg) + 0.5 * HD/HD_avg + 0.5 * HH/HH_avg + 0.3 * TD/TD_avg...
Just pulled random numbers out of my ass but guess which part of that formula is easiest to abuse. People are saying that Valve needs to address the root of the problem, which is that HD weights heavily into calibration. However, I think doing so will do more harm than good as stated above. It might be better just to do band-aid fixes that address specific abuses, like making Zeus ult and other globals not count towards HD or something. If you're a 6k player, why would you start a new account
Maybe you want to play some more ranked dota without the possibility of losing your 6k 
|
Or you're just Emulator and you're going for highest number of smurf accounts that one person can possibly have.
Honestly I don't even know, is there an original Emulator anymore? Is Emu even real?
|
Northern Ireland22208 Posts
Emu is like voldemort, shattered into multiple pieces
|
On October 04 2014 05:45 TheEmulator wrote:Show nested quote +On October 04 2014 05:18 evanthebouncy! wrote:On October 04 2014 04:14 B1nary wrote:
I don't know if there's an easy solution to this.
Suppose I'm a 6k player and I start a new account. Valve has no idea what my actual MMR might be so the reasonable thing to do would be guess that I'm average (~2.5k?) and work from there. If we ONLY use win/lose as the metric for calibration as some people suggest, by the time I'm done with the 40 unranked games, I'll only be at 3.5k even if I win all of them. After winning my 10 calibration matches, I'll get placed at 3.75k. In order to get up to 5k,I have to play through at least 100 games. This would suck for me and it would suck for my opponents since they're facing a 6k player and basically have no chance of winning.
Valve obviously thought that they need some metric to determine how I'm performing relative to everyone else in my games so they could bump my MMR up faster than 25 per game. Maybe they chose a formula like this:
(K + A - D)/(K_avg + A_avg - D_avg) + 0.5 * HD/HD_avg + 0.5 * HH/HH_avg + 0.3 * TD/TD_avg...
Just pulled random numbers out of my ass but guess which part of that formula is easiest to abuse. People are saying that Valve needs to address the root of the problem, which is that HD weights heavily into calibration. However, I think doing so will do more harm than good as stated above. It might be better just to do band-aid fixes that address specific abuses, like making Zeus ult and other globals not count towards HD or something. If you're a 6k player, why would you start a new account Maybe you want to play some more ranked dota without the possibility of losing your 6k
Because you zues abused 
|
On October 04 2014 06:25 ahswtini wrote:
Emu is like voldemort, shattered into multiple pieces
But isnt Voldemort powerful?
|
that is ridiculous, but im not surprised. I too, started a 2nd account (im almost exclusive support) and despite a 60% winrate, it put me at 3800. I had no idea how that was possible, even lower than my main. Now I understand why, even though I definitely had the biggest impact in most games with my support play, I never dealt the most damage. Oh well, least I got to play with Tobi a good 10 times or so because of it.
I managed to take my main from 3900 to 4800 within a few months lately when I got more serious, but since 6.82 where carries basically autofarm from the return gold, I dropped down to 4500. I'm taking an extended break from dota because its just a bullshit system, and unless you pick mid everygame, your ability to climb is hindered greatly. I did my best and somehow climbed 1000 pts almost to 5k with support, but this patch ruins that even more and you basically need to pick a support with lategame semi-carry capabilities (Mirana, Alch, Sven, etc) if you want to win in this patch.
And as far as those people doing the "get out of 3k MMR hell to 6k challenge" leveling up a low level account, I have yet to see anyone accomplish that without playing mid DP/ember/razor type heroes everygame. Its obviously not impossible, as I've almost done it up to 5k, but its WAY WAY WAY harder and would take much more than the 100 games it takes if you spam DP mid.
|
On October 04 2014 02:00 wavebave wrote:this place is amazing, its like "reddits community" but without downvotes and the moderation here is worse than playdota cuz u get "warned" for high quality posting. Show nested quote +On October 04 2014 00:42 ahswtini wrote:On October 04 2014 00:24 wavebave wrote:On October 03 2014 23:17 Torte de Lini wrote:On October 03 2014 21:55 wavebave wrote:
how isnt it a viable solution? a 3k mmr player spamming zeus ulti whenever its off cd gives them 5k straight away
if valve decided to look through statistics of how many tbd players actually picked zeus it'd probably be the most picked hero with a rate of 80% among tbds because you're removing choice from players. It's something you need to uphold and removing Zeus will choose have TBD abusers switch to another hero that is a near second. lmfao ok give me another hero that doesn't require skills but can get to at least one third of zeus' average hero dmg at the end of a match fucking god this forum is like playdota's dota chat section except ppl try to be serious here lolz Except you're the one who's brainless. The problem here is that getting high hero damage inflates your calibration rating. Not that zeus is the easiest hero to get a high hero damage with. Your solution is to attack the symptom of the problem, not the problem itself. Worse still you do it with a nuke and not with a precision strike. That's great, you still haven't fixed the root of the problem so the abusers will just move onto the next best hero for the job. ur like 5k posts in and ur most likely a tbd zeus player by the way u "argue". im asking u to show me a hero that can be played by 3ks and 4ks while still reaching 5k+ that can give at least ONE THIRD of zeus total hero dmg of an average match and u cant, do u have to be a scientist to understand people pick zeus to inflate their rating by additional 1k+ points because its the easiest hero in regards to hero damage? HOLY FUCK you are a fucking troll fuck it im done but these high quality bait posts just suck me right in ive betrayed u playdota User was banned for this post.
That took way longer than it should have.
|
United States12235 Posts
I'm sure Valve has metrics to back up their smurf detection system, and it seemed to work pretty well because a lot of players said that the placement/calibration accuracy was very good. However, the heroes in Dota are so varied in role and impact potential that calibration can't simply rely on a single factor to blanket them all.
If I were to guess, I would say that they probably parsed out average performance on a per-hero basis (and I'm using "performance" in the most nebulous sense, it could be just hero damage alone) and broke down those averages across various skill rating ranges. For example, we know that Zeus has a lot of hero damage potential because of his ability set. If in fact they were looking only at hero damage separately per hero, then they would know based on internal data that 3k-rated players do an average of 20k hero damage as Zeus in a 30-35min game, while 5k players do an average of 50k hero damage under the same conditions. By isolating variables, they can more directly compare data sets for the most accurate readings. That is, comparing total hero damage as Zeus from a 15 minute 3k game with a Crystal Maiden from a 75 minute 5k game wouldn't make sense because it's apples to oranges: the heroes are different, the game durations are different. How many matches have been registered so far? Tens of millions? Hundreds of millions? They have access to this level of detail.
So, let's go back to the example I listed before. On average, let's say a 5k Zeus in a 30-35 min game does 50k hero damage, while you under the same game conditions did 60k even though you just spammed your ult on cooldown. You must be better than 5k according to the governing metric. That could be okay in the short term, but what happens when there are so many other 60k-damage Zeuses doing this that the expected hero damage for a particular rating gets raised? What happens when the next hero trend emerges, like old-style Deso/Manta Venge (avg hero damage at 5k MMR is 10k and I got 40k, I'm 6k rating!)? There's probably more that goes into it than a single factor like hero damage alone, but people are always going to try to exploit for easy and fast rating if they know the conditions, and it's up to Valve to keep those metrics evolving and well-guarded.
|
I'm sorry for you, OP, but it seems like the only reason you are noticing this is because most people who are at 5k+ MMR are more or less properly ranked and you're noticing the ones that don't. In lower MMRs, even up close to 4k, there are still often raidcal differences in the skills of the players. There are more TBDs, there are more people who have ridden lucky streaks and more people who have exploited a single nuance in the game to rise above their general skill level.
Let's be honest, if there was a 3k player complaining that there was always one awful player in his games that decided which team won or lost, you would tell him that he should just play better and carry that guy and that his overall record would be his true skill. So....what should I tell you?
And as far as those people doing the "get out of 3k MMR hell to 6k challenge" leveling up a low level account, I have yet to see anyone accomplish that without playing mid DP/ember/razor type heroes everygame. Its obviously not impossible, as I've almost done it up to 5k, but its WAY WAY WAY harder and would take much more than the 100 games it takes if you spam DP mid.
It would be nice if some one did this with the idea of showing lower ranked players what things to work on or how to improve, but it seems that they just do it to say "lol 4k players so stupid, 3k players are mental deficients and 2k players are amoebas lol!"
It's not useful to show a 2800 player how a 5k player beats 3kers. He needs to know how to play like a 3k player.
|
On October 05 2014 02:41 Jerubaal wrote:
Let's be honest, if there was a 3k player complaining that there was always one awful player in his games that decided which team won or lost, you would tell him that he should just play better and carry that guy and that his overall record would be his true skill. So....what should I tell you?
Man, it's so sweet to see that turned around on someone. :D
|
On October 05 2014 02:10 Yergidy wrote:Show nested quote +On October 04 2014 02:00 wavebave wrote:this place is amazing, its like "reddits community" but without downvotes and the moderation here is worse than playdota cuz u get "warned" for high quality posting. On October 04 2014 00:42 ahswtini wrote:On October 04 2014 00:24 wavebave wrote:On October 03 2014 23:17 Torte de Lini wrote:On October 03 2014 21:55 wavebave wrote:
how isnt it a viable solution? a 3k mmr player spamming zeus ulti whenever its off cd gives them 5k straight away
if valve decided to look through statistics of how many tbd players actually picked zeus it'd probably be the most picked hero with a rate of 80% among tbds because you're removing choice from players. It's something you need to uphold and removing Zeus will choose have TBD abusers switch to another hero that is a near second. lmfao ok give me another hero that doesn't require skills but can get to at least one third of zeus' average hero dmg at the end of a match fucking god this forum is like playdota's dota chat section except ppl try to be serious here lolz Except you're the one who's brainless. The problem here is that getting high hero damage inflates your calibration rating. Not that zeus is the easiest hero to get a high hero damage with. Your solution is to attack the symptom of the problem, not the problem itself. Worse still you do it with a nuke and not with a precision strike. That's great, you still haven't fixed the root of the problem so the abusers will just move onto the next best hero for the job. ur like 5k posts in and ur most likely a tbd zeus player by the way u "argue". im asking u to show me a hero that can be played by 3ks and 4ks while still reaching 5k+ that can give at least ONE THIRD of zeus total hero dmg of an average match and u cant, do u have to be a scientist to understand people pick zeus to inflate their rating by additional 1k+ points because its the easiest hero in regards to hero damage? HOLY FUCK you are a fucking troll fuck it im done but these high quality bait posts just suck me right in ive betrayed u playdota User was banned for this post. That took way longer than it should have.
Not gonna lie, I was expecting the hammer to drop around the time he said "i just made this acc to flame OP" (sic).
Sorta glad it didn't though, this thread is gold.
|
Writing this blog was a big mistake, the number of abusers increased a lot since it appeared. It leads to some stupid 5o4... I understand it's a real problem for many players but it should have been posted on dev forums, not on a big traffic site.
|
On October 07 2014 19:16 nojok wrote:
Writing this blog was a big mistake, the number of abusers increased a lot since it appeared. It leads to some stupid 5o4... I understand it's a real problem for many players but it should have been posted on dev forums, not on a big traffic site.
You think the rise of tbd zeus is due to this blog in particular? This info is everywhere...
|
the only way u can rise is when u can actually 1v5 in ur current mmr despite having lots of feeders and retards in ur team. otherwise ure not good enough to rise, if you are depending on anyone but urself to win a game then ure thinking about it the wrong way. also everyone gets tbd players just as much as you do, so dont act like they are the ones that are stopping you from rising
|
Sanya12364 Posts
Pretty much, rising when you are below real MMR is hard unless you can Pub Star as the carry or the mid.
|
On October 02 2014 23:20 Ler wrote:Show nested quote +On October 02 2014 23:09 ahswtini wrote:
So do you get calibrated high even if you lose most of those games where you just zeus ult everytime it's off cooldown? Yes.
Wait is this still the case? O_O holy shit??
|
28090 Posts
On October 18 2014 15:27 Aerisky wrote:Show nested quote +On October 02 2014 23:20 Ler wrote:On October 02 2014 23:09 ahswtini wrote:
So do you get calibrated high even if you lose most of those games where you just zeus ult everytime it's off cooldown? Yes. Wait is this still the case? O_O holy shit??
Not recently. The other week it was patched.
|
|
|
|
|
|
EPT
