EPT[G] Comprehensive SC2 League and Ladder Guide - Page 41
| Forum Index > SC2 General |
|
IMScientist
Italy125 Posts
| ||
|
oZe
Sweden492 Posts
| ||
Excalibur_Z
United States12275 Posts
On August 23 2011 02:05 oZe wrote: Sry if this has been talked about but don't have time to read 41 pages. Is there an MMR for FFA? Good question. I don't play FFA or have any interest in FFA or know how popular FFA is, so I have no idea. Maybe it uses your 1v1 since it's basically a 1v1v1v1? Maybe it just grabs the first 4 players? How long does it take to find a FFA game? | ||
|
Brainling
United States660 Posts
I am not even sure what you would do entirely, but I would imagine it would have to be some sort of heuristic based on how you are losing your games. If a bunch of your losses are insta-quits or AFK snow jobs, the heuristic would need to kick in and say "hey, this person may be smurfing". Is this even something Blizzard cares to combat? | ||
|
Excalibur_Z
United States12275 Posts
On August 23 2011 06:32 Brainling wrote: This may have been discussed already, but I didn't see it (and the OP definitely doesn't touch on it): Smurfing. Is there anything in the system to try and combat smurfing? I am not even sure what you would do entirely, but I would imagine it would have to be some sort of heuristic based on how you are losing your games. If a bunch of your losses are insta-quits or AFK snow jobs, the heuristic would need to kick in and say "hey, this person may be smurfing". Is this even something Blizzard cares to combat? There is no protection against smurfing, and in fact, some players do abuse the system in this way. Blizzard's response and their approach to this is a little more manual: http://us.battle.net/sc2/en/forum/topic/2743728019 They've presented no information about any kind of automated system that flags players with a series of immediate losses. Maybe that exists, but we can't be sure until we start hearing reports of people who have had action taken against their account. | ||
|
aksfjh
United States4853 Posts
On August 23 2011 02:46 Excalibur_Z wrote: Good question. I don't play FFA or have any interest in FFA or know how popular FFA is, so I have no idea. Maybe it uses your 1v1 since it's basically a 1v1v1v1? Maybe it just grabs the first 4 players? How long does it take to find a FFA game? I think I remember reading somewhere that your FFA stats are tracked. They just don't keep a ladder for it. | ||
|
Vronti
United States111 Posts
| ||
|
Bamurdead4
3 Posts
| ||
|
Shield
Bulgaria4824 Posts
On August 23 2011 02:46 Excalibur_Z wrote: Good question. I don't play FFA or have any interest in FFA or know how popular FFA is, so I have no idea. Maybe it uses your 1v1 since it's basically a 1v1v1v1? Maybe it just grabs the first 4 players? How long does it take to find a FFA game? As far as I know, the more you win FFA games, the better players you will play against. However, I don't know if the system bases this only on FFA games. If I'm wrong, feel free to correct me.  | ||
|
LordOfDabu
United States394 Posts
| ||
|
ShaPinna
Germany1 Post
Could you tell Dario to translate all o' that to german, tho? ...  | ||
|
J_D
United States102 Posts
Master 3.7% Diamond 9.7% Platinum 18.5% Gold 21% Silver 22.7% Bronze 24.3% So I'm wondering if this is just a bug with sc2ranks, or does blizzard not want to strictly enforce the percentages and instead have them be a loose guideline, or maybe have set MMR thresholds for some leagues? | ||
|
LordOfDabu
United States394 Posts
| ||
|
aksfjh
United States4853 Posts
On August 31 2011 00:21 LordOfDabu wrote: The discrepancy is most likely that not every player on sc2ranks meets Blizzard's definition of active. While this is the primary reason, there can be other reasons for discrepancy as well. If we take a look at a picture in the OP, we can see that the population distribution is estimated to look like a bell-curve: http://imgur.com/5w92j.jpg The majority of the population and points lie within the middle, and slopes equally in each direction. However, there are statistical anomalies that occur when there is a clear advantage to getting better after playing more games. This skews the bell-curve to look more like a wave, where the majority of the points are in the right/high side, and a majority of the population is on the left/low side. Since Blizzard does everything by a static MMR threshold, where X MMR "points" SHOULD be a cutoff of Y% of the population, it actually marks a much smaller portion of the population. Sorry if this is too technical or doesn't make sense, I just felt like explaining it in full. | ||
|
oZe
Sweden492 Posts
On August 23 2011 06:32 Brainling wrote: This may have been discussed already, but I didn't see it (and the OP definitely doesn't touch on it): Smurfing. Is there anything in the system to try and combat smurfing? I am not even sure what you would do entirely, but I would imagine it would have to be some sort of heuristic based on how you are losing your games. If a bunch of your losses are insta-quits or AFK snow jobs, the heuristic would need to kick in and say "hey, this person may be smurfing". Is this even something Blizzard cares to combat? Blizzard don't do anything against people win trading atleast in FFA. I was in a game where one player instaquit when he didn't see his win trading friend. Another player who pressumably plays alot of FFA's. Says he usually goes after that guy and have reported him but blizzard have done nothing. Later I play another FFA same guy is there again, doesn't instaquit. So I say something like you gonna team up on me now. He says "just quit" soon there after he reapers me, then his friend comes with DT's. After the game he even asks me if I want to trade wins with him. Was this guy incase you interested. http://sc2ranks.com/eu/896023/YuGi/a/4325378 | ||
|
slim_z
21 Posts
| ||
|
Not_That
287 Posts
Before I go into theorycrafting mode, am I missing something obvious here? I tried searching but I couldn't find the answer, so if it's been covered already I apologize. | ||
|
Vorenius
Denmark1979 Posts
For Random Teams, it is likely that the MMRs of all members are averaged first, then an opposing team with a similar MMR (based upon a degree of uncertainty) is located. But does that actually make any sense? Can you quantify skill like that? Is a 2v2 team with a bronze/master gonna be as good as a team with a gold/plat (or whatever the "middle" MMR between bronze and master is)? I'm also sure the majority of games I play consist mainly of player from the same league, only with the occasional one league above or below. | ||
|
aksfjh
United States4853 Posts
On September 03 2011 05:40 Not_That wrote: I'm having a hard time figuring out some of the match histories data when 2 high GMs play against each other. It looks like the winner usually gets a decent amount of points, but still below 12, while the loser still loses a substantial amount of points (well over 12). I thought that if both of their points are fairly close to representing their MMR, then when two players meet their combined points won / lost should be about 0, but this seems to me like it's regularly not the case for GMs. Before I go into theorycrafting mode, am I missing something obvious here? I tried searching but I couldn't find the answer, so if it's been covered already I apologize. There's theories surrounding a supposed MMR or adjusted ladder point soft cap. Basically, when certain players get so so high in the system, losing points becomes easier and gaining harder, even against even opponents. If it's a MMR cap, getting points higher than that cap would be discouraged by the whole "points converge on MMR" system. If it's a point cap, they just apply a different converging formula after the cap. So, for an even win you'd get a number less than 12, and for an even loss you'd lose some number more than 12. Both are feasible for very complex reasons. | ||
|
Excalibur_Z
United States12275 Posts
On September 03 2011 10:25 aksfjh wrote: There's theories surrounding a supposed MMR or adjusted ladder point soft cap. Basically, when certain players get so so high in the system, losing points becomes easier and gaining harder, even against even opponents. If it's a MMR cap, getting points higher than that cap would be discouraged by the whole "points converge on MMR" system. If it's a point cap, they just apply a different converging formula after the cap. So, for an even win you'd get a number less than 12, and for an even loss you'd lose some number more than 12. Both are feasible for very complex reasons. Yeah I think this is correct. Also, some high-level player said they spoke with David Kim who said that MMR caps out. Since it's either one or the other, I'd lean toward an MMR cap being the more likely possibility based on this secondhand confirmation. | ||
| ||
