|
On September 05 2013 12:36 LeLoup wrote:
Watch the game and then please explain to me how the invoker is the one that lost the game for you
again, if you post like this you clearly havent read my post. The POINT isnt about winning or losing.+ Show Spoiler +(in fact, i think the invoker played decently and should be avg skill level wise, the one who lost that game for us was rubick with the lack of supportive play)
1/ The matchmaking doesnt function correctly mathematically. It is matching new players with veterans.
2/ The definition of 'good match' that they are working with are generally wrong. With the skill differences too wide, players will not be able to understand each others actions and therefore result in no coordination. The match will be 1 sided regardless of you being on winning or losing side.
All of my complains has never been about why i lose a game. As you can see, the very first game that in my quote were me winning vs Luminous team and i still complain on how that was a shit match.
On September 05 2013 15:44 ReignSupreme. wrote:
Could you check the history of "pgg" and see if his games are him stacking with anyone else on your team previously?
That could 100% account for him being matched with you, think about it
"pgg" and person b stack together for 3 matches, win all 3 and "pgg"'s mmr is boosted substantially because of person b's existing mmr leading to them vs'ing high mmr players straight off the bat.
edit: the emulator sums it up pretty well, likely better worded than I did but you get the idea
I didnt check his match details but since he has only 2 'friends' attached to his account, assuming at least 1 of those is his main account, I skipped the party MMR boost checking. But you are completely correct that the possibility is there.
On September 05 2013 13:44 WindWolf wrote:
As a new player (~ 40 played games) I don't feel that this has been a problem for me.
E: Meaning that I haven't run into players who has way more games played than me
Thats because you dont understand the full impact of the problem. Imagine yourself being a noob FPS player and jump into Quake right now, or noob broodwar player trying to ladder on iccup, with no functioning matchmaking, you WILL get matched with veterans who will crush you down brutally and repeatedly. Having such games exist from the first place will discourage new players from playing+learning the game as well as demoralizing the veterans enjoyment for the game plus nullify the match result.
Given this is a minor case of valve matchmaking, things are not that terrible YET but a fix is definitely required.
On September 05 2013 16:44 run.at.me wrote:
You have over 4000 games played, and your complaining about one. These types of mmr mixes are few and far between.
While I agree the mm needs work, the example you have provided is a 'outlier' which we can ignore...
i agree with the possibility of the match being 'outlier'. But I get 3 games like such in the time frame of less than 48 hours, I dont think it could be ignored. Remember that dota2 games have the avg length of 35-45 mins. That combines with the match finding time make it 1 hour. So if the matchmaking essentially telling 10 people to waste 1 hour each of their life because they are too new or too old for it to function properly, I think there are fixes need to be done.
|
I have no idea how you can say this:
1/ The matchmaking doesnt function correctly mathematically. It is matching new players with veterans.
There's no "math" in your posts that show the matchmaking doesn't work. The system is not supposed to separate new players from veterans or match people by hours played. It's supposed to match people by skill level. Obviously there is some correlation between the two, but if in that game the Invoker wasn't an worse player than his team members, there's no reason why he shouldn't be there.
There is an issue if the system matches people of vastly diferent skill levels, no matter the amount of games they played. There is no issue if the system matches people with a vastly diferent amount of games played, if they have similar skill.
That's why how the Invoker played is very relevant, and why good players quickly moving to the top is a positive, not negative, effect. The issue is when the system does that to not so good players.
You say the issue is that the system is matching completelly new players with veterans, yet gives an example of a guy that is clearly not a new player. He knew how to play the game.
|
How do you know the enemy players weren't a five stack or something? There's nothing in place that prevents players who are good at the game from partying up with total noobs, and the matchmaking may be trying to balance the teams by approximating this disparity on the other side.
This is not based on any facts I know, just some random speculation, so don't take it too seriously unless someone happens to confirm it :D
|
On September 06 2013 01:45 SKC wrote:I have no idea how you can say this: Show nested quote +1/ The matchmaking doesnt function correctly mathematically. It is matching new players with veterans. There's no "math" in your posts that show the matchmaking doesn't work. The system is not supposed to separate new players from veterans or match people by hours played. It's supposed to match people by skill level. Obviously there is some correlation between the two, but if in that game the Invoker wasn't an worse player than his team members, there's no reason why he shouldn't be there. There is an issue if the system matches people of vastly diferent skill levels, no matter the amount of games they played. There is no issue if the system matches people with a vastly diferent amount of games played, if they have similar skill. That's why how the Invoker played is very relevant, and why good players quickly moving to the top is a positive, not negative, effect. The issue is when the system does that to not so good players. You say the issue is that the system is matching completelly new players with veterans, yet gives an example of a guy that is clearly not a new player. He knew how to play the game.
Feel free to check out the 2 games that i have in the quote for example. What i mean by math was the invoker guy had exactly 50% winrate with a tiny sample size and being matched into very high game. Yes his mmr could be skewered up by his very high MMR friend but isnt that the flaw of the system? In the game he plays with me, he was solo queue and there were no one there to skewer him up. The fact that he lose repeatedly after reach 50% point out that his predicted MMR was inaccurate when he got matched with me and the system is no punishing him for previously partying with his friend.
Look at the first game where i beat luminous for example: very high MMR, the RIKI had 25 last hit in a 30 mins game. Imagine you being a new player and got matched with all the veterans who expected you to do your job like a pro. Its demoralizing for both him and the rest of the players in that game including me. That riki is a clear example of a false smurf detection which result a failure in match making generating good result.
|
On September 06 2013 02:17 NB wrote:Show nested quote +On September 06 2013 01:45 SKC wrote:I have no idea how you can say this: 1/ The matchmaking doesnt function correctly mathematically. It is matching new players with veterans. There's no "math" in your posts that show the matchmaking doesn't work. The system is not supposed to separate new players from veterans or match people by hours played. It's supposed to match people by skill level. Obviously there is some correlation between the two, but if in that game the Invoker wasn't an worse player than his team members, there's no reason why he shouldn't be there. There is an issue if the system matches people of vastly diferent skill levels, no matter the amount of games they played. There is no issue if the system matches people with a vastly diferent amount of games played, if they have similar skill. That's why how the Invoker played is very relevant, and why good players quickly moving to the top is a positive, not negative, effect. The issue is when the system does that to not so good players. You say the issue is that the system is matching completelly new players with veterans, yet gives an example of a guy that is clearly not a new player. He knew how to play the game. Feel free to check out the 2 games that i have in the quote for example. What i mean by math was the invoker guy had exactly 50% winrate with a tiny sample size and being matched into very high game. Yes his mmr could be skewered up by his very high MMR friend but isnt that the flaw of the system? In the game he plays with me, he was solo queue and there were no one there to skewer him up. The fact that he lose repeatedly after reach 50% point out that his predicted MMR was inaccurate when he got matched with me and the system is no punishing him for previously partying with his friend. Look at the first game where i beat luminous for example: very high MMR, the RIKI had 25 last hit in a 30 mins game. Imagine you being a new player and got matched with all the veterans who expected you to do your job like a pro. Its demoralizing for both him and the rest of the players in that game including me. That riki is a clear example of a false smurf detection which result a failure in match making generating good result.
The issue with the other examples is that they are anonymous, so we can't really see how they got there of if they were stacking. They were still more valid complaints than the Invoker guy.
The point is that you made a whole blog post just to talk about this guy that clearly (as you said so yourself) has the skill to play at that level. So you made a whole blog about something that is working the way it's supposed to. We can't know exactly how he got there, but there is no reason why him being there is an issue. Perhaps if you chose a better example things would be diferent, but then there could be other explanations on how they got there. We don't know, since you didn't.
You can't make a blog about how worried you are at the experience of new players facing veterans and then as your prime example show a match that clearly has no new players. It hurts your argument more than anything.
|
As I said in reply to your QQ post too, this isn't rare, uncommon or even unreasonable to happend. I've had one of the people in my stack changing account, and then still continue to play in very high bracket in his very first game.
I've had very high within 10 games on a new account duo queuing, because the smurf detection is really good, and I honestly belong there over playing with normal players for the next 100 games.
Like I said there, hours ingame on that account, games played on that account, winrate, all of that is useless information, because it's so easy to just change steam account for a new account. It tells you nothing about the player, except how much he played on that specific account alone.
Lots of people complain about mmr in general too, but that's only because it's such a snowball game that can so easily go either way. Just look at the international finals. The two best teams stomped each other the 3 first games, it wasnt even close.
Personal skill, mood, daily form, luck/rng all factor in how people play, and it will make one game to the next vastly different.
Just look at my meepo stats, how vastly up/down they go from game to game. It can be anything from me being horrible, outlaning, heroes,tons of mud golems spawning early game to prevent my levels to skyrocket.
Dota has so many details that matters so much more than a few hours of gameplay, no matter how entitled you are to stay away from those "pesky 10 hours played scrubs".
On September 06 2013 02:17 NB wrote:Show nested quote +On September 06 2013 01:45 SKC wrote:I have no idea how you can say this: 1/ The matchmaking doesnt function correctly mathematically. It is matching new players with veterans. There's no "math" in your posts that show the matchmaking doesn't work. The system is not supposed to separate new players from veterans or match people by hours played. It's supposed to match people by skill level. Obviously there is some correlation between the two, but if in that game the Invoker wasn't an worse player than his team members, there's no reason why he shouldn't be there. There is an issue if the system matches people of vastly diferent skill levels, no matter the amount of games they played. There is no issue if the system matches people with a vastly diferent amount of games played, if they have similar skill. That's why how the Invoker played is very relevant, and why good players quickly moving to the top is a positive, not negative, effect. The issue is when the system does that to not so good players. You say the issue is that the system is matching completelly new players with veterans, yet gives an example of a guy that is clearly not a new player. He knew how to play the game. Feel free to check out the 2 games that i have in the quote for example. What i mean by math was the invoker guy had exactly 50% winrate with a tiny sample size and being matched into very high game. Yes his mmr could be skewered up by his very high MMR friend but isnt that the flaw of the system? In the game he plays with me, he was solo queue and there were no one there to skewer him up. The fact that he lose repeatedly after reach 50% point out that his predicted MMR was inaccurate when he got matched with me and the system is no punishing him for previously partying with his friend. Look at the first game where i beat luminous for example: very high MMR, the RIKI had 25 last hit in a 30 mins game. Imagine you being a new player and got matched with all the veterans who expected you to do your job like a pro. Its demoralizing for both him and the rest of the players in that game including me. That riki is a clear example of a false smurf detection which result a failure in match making generating good result.
Isn't that reasonable? No system can't possibly be perfect with a game like dota, so maybe he was wrong a game or two. Are you seriously going to complain about that?
|
On September 06 2013 01:15 NB wrote:Show nested quote +On September 05 2013 13:44 WindWolf wrote:
As a new player (~ 40 played games) I don't feel that this has been a problem for me.
E: Meaning that I haven't run into players who has way more games played than me Thats because you dont understand the full impact of the problem.
I do understand the problem; just not in this game. The match-making for the PC version of Super Street Fighter 4: Arcade Edition sucks to put it mildly. It has affected me both as a new player when I started and it happens now at times as well when I'm more experienced with the game
|
On September 06 2013 02:36 SKC wrote:Show nested quote +On September 06 2013 02:17 NB wrote:On September 06 2013 01:45 SKC wrote:I have no idea how you can say this: 1/ The matchmaking doesnt function correctly mathematically. It is matching new players with veterans. There's no "math" in your posts that show the matchmaking doesn't work. The system is not supposed to separate new players from veterans or match people by hours played. It's supposed to match people by skill level. Obviously there is some correlation between the two, but if in that game the Invoker wasn't an worse player than his team members, there's no reason why he shouldn't be there. There is an issue if the system matches people of vastly diferent skill levels, no matter the amount of games they played. There is no issue if the system matches people with a vastly diferent amount of games played, if they have similar skill. That's why how the Invoker played is very relevant, and why good players quickly moving to the top is a positive, not negative, effect. The issue is when the system does that to not so good players. You say the issue is that the system is matching completelly new players with veterans, yet gives an example of a guy that is clearly not a new player. He knew how to play the game. Feel free to check out the 2 games that i have in the quote for example. What i mean by math was the invoker guy had exactly 50% winrate with a tiny sample size and being matched into very high game. Yes his mmr could be skewered up by his very high MMR friend but isnt that the flaw of the system? In the game he plays with me, he was solo queue and there were no one there to skewer him up. The fact that he lose repeatedly after reach 50% point out that his predicted MMR was inaccurate when he got matched with me and the system is no punishing him for previously partying with his friend. Look at the first game where i beat luminous for example: very high MMR, the RIKI had 25 last hit in a 30 mins game. Imagine you being a new player and got matched with all the veterans who expected you to do your job like a pro. Its demoralizing for both him and the rest of the players in that game including me. That riki is a clear example of a false smurf detection which result a failure in match making generating good result. The issue with the other examples is that they are anonymous, so we can't really see how they got there of if they were stacking. They were still more valid complaints than the Invoker guy. The point is that you made a whole blog post just to talk about this guy that clearly (as you said so yourself) has the skill to play at that level. So you made a whole blog about something that is working the way it's supposed to. We can't know exactly how he got there, but there is no reason why him being there is an issue. Perhaps if you chose a better example things would be diferent, but then there could be other explanations on how they got there. We don't know, since you didn't. You can't make a blog about how worried you are at the experience of new players facing veterans and then as your prime example show a match that clearly has no new players. It hurts your argument more than anything.
I think i didnt make it clear. I think that because there are a smurf detection system in place, Valve indirectly, intentional or not, encouraging players to have smurf account. If you use any other existing matchmaking method, sc2 for example, you could see that the veteran ranking are rewarded with bonus pool and separate them from from the new ones which make smurfing useless because it will take them much longer to reach where they supposed to be.
Tbh, im not a design expert. If i were i would be doing this for money instead of writing a public blog and let everyone see it for free. But I see a flaw in the game i enjoy and I want it to be fixed. The blog didnt use that particular game as a PRIME example. This blog is simply just a follow up to the QQ post that previous posted and the game i posted should not be weighted any higher or lower than the other 2 games.
|
On September 06 2013 04:23 WindWolf wrote:Show nested quote +On September 06 2013 01:15 NB wrote:On September 05 2013 13:44 WindWolf wrote:
As a new player (~ 40 played games) I don't feel that this has been a problem for me.
E: Meaning that I haven't run into players who has way more games played than me Thats because you dont understand the full impact of the problem. I do understand the problem; just not in this game. The match-making for the PC version of Super Street Fighter 4: Arcade Edition sucks to put it mildly. It has affected me both as a new player when I started and it happens now at times as well when I'm more experienced with the game
again, i used the word IMPACT.
I think a lot of people understand the problem but not the consequence of having such problems.
On September 06 2013 03:02 kaztah wrote:
Isn't that reasonable? No system can't possibly be perfect with a game like dota, so maybe he was wrong a game or two. Are you seriously going to complain about that?
Again, you can see that in under 48 hours i get 3 games, 2 were consecutively, having such problem. And im pretty sure im not the only one as well as this is not the very first time such thing happen. The issue does have an impact on a group of player, no matter big or small, should still be resolved.
|
Isn't it possible to get into top MM as a beginner by simply being part of a 5 man stack where the other 4 are top tier players? There's no way the system can balance the teams if the stack itself has such huge disparity in skill level.
|
28090 Posts
On September 06 2013 04:40 Glacierz wrote:
Isn't it possible to get into top MM as a beginner by simply being part of a 5 man stack where the other 4 are top tier players? There's no way the system can balance the teams if the stack itself has such huge disparity in skill level.
Yeah. I 5 stacked with some friends once and one of them was on an account that had 0 games, and we still hit a page 3 game. Valve has mentioned that the mmr will skew a bit towards the higher rated players in a party, so that explains that.
|
|
|
|
|
|
EPT
