EPTRanked matchmaking coming to Dota 2 - Page 115
| Forum Index > Closed |
Please keep the QQ to a minimum if you do not like this update. We are happy to hear your reasoning for not liking a ranked system, but no "OMG VOLVO WHY" posts. | ||
|
Plansix
United States60190 Posts
| ||
|
Sn0_Man
Tebellong44238 Posts
On January 22 2014 05:28 Plansix wrote: Who cares? As long as I have a player with 3.9k MMR on my team too, the game will work out. Interestingly, that isn't necessarily always the case. If I stack with say my brother (hasn't finished calibration but I'd put him around 2K mmr) and some of his dota friends at the same level, then chances are we end up vs a similar stack of something like 4x2k and 1x4K players. Maybe. However, I'm just bad at certain things/roles aka snowballing and such. So I pick a support (lets say lich), and proceed to crushhhhh the early game, roaming, winning all 3 lanes, etc. However, their good player is playing, lets say carry. And he efficiently rices their jungle and split pushes and whatever. It rapidly becomes very hard for me to win no matter how hard I crushed my lanes because my allies are so bad they can't even lasthit. I'm not mad at them, in fact I just assumed that this would happen. But I find it very hard to win when I'm seriously ahead of the rest of my team in MMR because I simply am not good at leveraging that advantage. If I choose to go mid then sure I'd crush mid but I only play mid at say a 3K level becuase I'm fucking awfull at it and that ends up meaning my team is still at a big disadvantage because I can't play my role at a good enough skill level to carry. So if you got me as your token "good" player you'd be very sad basically. Meanwhile I know a guy who's really not that amazing or well rounded but he plays a mean invoker and he'd probably just stomp the whole game if you had him instead of me regardless of his rating being lower. I mean, every player is different but when the MMRs diverge that much it gets pretty seriously wonky. | ||
|
Plansix
United States60190 Posts
| ||
|
Sn0_Man
Tebellong44238 Posts
I don't think theres a better way to do it, that isn't what I'm implying, but I have serious difficulty shouldering the amount of weight the MM system expects me to when I'm wayyyy above the level of my teammates. That is just my experience. I'm quite certain that valve has run the numbers and your average 4.2K mmr player should shoulder exactly as much as MM expects them to, I just can't in that situation. | ||
|
SkelA
Macedonia13069 Posts
On January 22 2014 05:39 Plansix wrote: As long as both sides had an equal chance of winning, then I dont see an issue. Non of that stuff is world ending. Heh lets take for example my last game. I get an 2 russian allies with sub 49% winratio, one of them went nyx treads bfury mid. Other team gets one guy with 58% winratio and 2 with 53% winratio. Game wasnt even close they completly wrecked us. You wanna know what was our chance for win ? It surely wasnt 50% , not even close to 10% . Thanks valve for making balanced teams. | ||
|
Sn0_Man
Tebellong44238 Posts
On January 22 2014 06:09 SkelA wrote: Heh lets take for example my last game. I get an 2 russian allies with sub 49% winratio, one of them went nyx treads bfury mid. Other team gets one guy with 58% winratio and 2 with 53% winratio. Game wasnt even close they completly wrecked us. You wanna know what was our chance for win ? It surely wasnt 50% , not even close to 10% . Thanks valve for making balanced teams. Dis aint the qq thread though this is for actual discussion. | ||
|
crms
United States11933 Posts
On January 22 2014 05:43 Sn0_Man wrote: I disagree that the chances of victory are equal though. Isn't the statistics of valves recording keeping showing that those match-ups are indeed 50/50 or extremely close? I'd have to imagine millions of games worth of data is more telling than your hunch or anecdote. | ||
|
sCCrooked
Korea (South)1306 Posts
On January 21 2014 19:16 Erasme wrote: You don't understand. A 5k player can play 1game a week, if they have the level, they will be placed around 5k. The 200games are the unranked games you have to play to reach lvl 13. In your imaginary world where 6k players get placed at 2k for some unknown reasons, the ranked matchmaking needs to change. Luckily, in reality, it doesn't exist. Wrong, you are the one who does not understand. Even worse is you do not understand that you do not understand. Your understanding of how the matchmaking system works is extremely poor and the proof is in your own words. When 5k friends invite me to 5k rating games and I perform better than my reciprocal role on the opposing team, its also proof. Your argument is also based on 200 games where you are assuming a completely new player is getting adjusted to the game. What you should really be giving for an example would be an account where a person new to DOTA played the 200 and then gave it to a much stronger player who tries to shoot it up to 5k+. So far, there is no account of anyone managing to do that, while there IS sufficient evidence for the opposing side. I feel sorry for people like you since they can't even tell you are taking the blue pill. | ||
|
Sn0_Man
Tebellong44238 Posts
On January 22 2014 06:15 crms wrote: Isn't the statistics of valves recording keeping showing that those match-ups are indeed 50/50 or extremely close? I'd have to imagine millions of games worth of data is more telling then your hunch or anecdote. lets boil this down. My chance of victory in a 1v1 matchup against 4.2k mmr players is NOWHERE NEAR 50/50 because I'm awful at 1v1. That doesn't mean that I'm misranked or that others are misranked, just that people are good at different things. If you actually read my post you'd see that I agree that the average 4.2K mmr player would play exactly well enough to balance it in that situation. I'm just not the average 4.2K mmr player. Nor is almost any other person exactly average. I'm positing that despite achieving "50/50 overall winrates", the problem of having wildly mismatched MMRs in a game isn't particularly solved, because players function strangely in that circumstance. Valve are doing as well as is reasonably possible though. I'm not calling for change, I'm just positing that both teams having one super high rated player isn't gonna guarantee a fair match. | ||
|
Sn0_Man
Tebellong44238 Posts
On January 22 2014 06:20 sCCrooked wrote: Wrong, you are the one who does not understand. Even worse is you do not understand that you do not understand. Your understanding of how the matchmaking system works is extremely poor and the proof is in your own words. When 5k friends invite me to 5k rating games and I perform better than my reciprocal role on the opposing team, its also proof. I feel sorry for people like you since they can't even tell you are taking the blue pill. Let me assure you that you are both the one who is wrong and posting very poorly here. You cannot play ranked to see your mmr without having about 200 games played (level 13 account). Over the course of those 200 games you are capable of reaching ANY RATING since valve's uncertainty regarding your skill is quite high at that stage. Once you enter ranked the system in fact has a very good idea of your rating. However, 200 games isn't nearly enough to "learn" dota so you can of course improve. The only possible issue is if you magically go from "3K mmr skill" to "5K mmr skill" over the course of less than 40 games (the amount it would take to climb that much in terms of the number by ur name). In which case at 10 games a week it would take you less than a month to "achieve" your actual mmr based on skill. Which isn't very relevant. Of course, if you achieve your "ranked" rating, then play a lot of unranked and improve a ton, then go back to ranked, yes I suppose this could happen. however even if you improved an obscene amount, it still takes well under a month for any player to raise his MMR high enough to have fair games. | ||
|
makmeatt
2024 Posts
| ||
|
Sokrates
738 Posts
On January 22 2014 05:39 Plansix wrote: As long as both sides had an equal chance of winning, then I dont see an issue. Non of that stuff is world ending. The problem is that games snowball pretty fast if the variance in a team is high. If the best player that is much better than the rest gets fed then he will also wreck the best player of the other team. So there is a bit of luck involved who gets fed and who plays which role. It doesnt always happen but in some games this is true. http://dotabuff.com/matches/478699390 Look at this game, the russian storm named )))) and the bounty completely WRECKED the opposing team, me being at 4.3k mmr. They made 20-1 without me being involved. He had a bloodstone finished in 11mins. | ||
|
Makro
France16890 Posts
btw at what point u always have decent game in ranked ? 4000 ? 4500 ? | ||
|
IPA
United States3206 Posts
I'm a few games over .500 at 3450 mmr. Feels like a good place to be. Rarely am I the asshat costing my team the game and yet I feel like I'm learning constantly with a really pleasant level of challenge and moderate improvement. Dota <3 | ||
|
Doraemon
Australia14949 Posts
On January 22 2014 06:09 SkelA wrote: Heh lets take for example my last game. I get an 2 russian allies with sub 49% winratio, one of them went nyx treads bfury mid. Other team gets one guy with 58% winratio and 2 with 53% winratio. Game wasnt even close they completly wrecked us. i just don't understand why you spend so much time dotabuffing all the players in the game, did it make you feel better in the end? no. will it ever? no. just get on with it and play the damn game. | ||
|
Makro
France16890 Posts
On January 22 2014 07:18 IPA wrote: Completely anecdotal but I've noticed a distinct rise in the amount of good games, even when I'm on the losing side. We generally have supports, logical compositions, active warding/dewarding, reasonable builds and decent communication. Negatively, the tryhard aspect (which I actually love) comes with the price tag of some hypercompetitive dudes who lose their shit sometimes. I'm ok with that. I want to play with people who want to win. I'm a few games over .500 at 3450 mmr. Feels like a good place to be. Rarely am I the asshat costing my team the game and yet I feel like I'm learning constantly with a really pleasant level of challenge and moderate improvement. Dota <3 i envy you, maybe it's like this on us server because i'm actually 3328 and it's basicaly one very shitty game on two | ||
|
DrPandaPhD
5188 Posts
On January 22 2014 07:17 Makro wrote: it's hard to have an idea of the ranking of players because when you ask nicely at the end of the game no one wants to answer btw at what point u always have decent game in ranked ? 4000 ? 4500 ? you are never going to always have decent games. Stomps happen in every bracket, even at the TI3 finals. | ||
|
SeakayKu
United States128 Posts
especially the game we got stomp i could tell we are actually very close, but just 1 small mistake that is good enough for enemy team to snowball nobody blamed anyone, we drafted early game and did not win early game so result was clear by 15 min mark, it was like 30-5 game | ||
|
govie
9334 Posts
On January 22 2014 07:08 Sokrates wrote: The problem is that games snowball pretty fast if the variance in a team is high. If the best player that is much better than the rest gets fed then he will also wreck the best player of the other team. So there is a bit of luck involved who gets fed and who plays which role. It doesnt always happen but in some games this is true. http://dotabuff.com/matches/478699390 Look at this game, the russian storm named )))) and the bounty completely WRECKED the opposing team, me being at 4.3k mmr. They made 20-1 without me being involved. He had a bloodstone finished in 11mins. Yes, this could be true. But it should also be tweakable by valve because i have read somewhere that valve changed the matchmakingstuff recently, so that people find matches sooner. Singsing and others had to wait a long time to find matches sometimes. If true that dota now has more snowbally matches since the change, than valve could/should record that and tweak it so that there is more balanc/optimume between the number of snowball matches and waitingtime for a match. | ||
|
Wrath 2.1
Germany880 Posts
| ||
| ||
