Pages: 1 2 [3] :: one page |
|
Author |
Thread Statistics | Show CCP posts - 3 post(s) |
ScI-Iurk
Molon Labe. General Tso's Alliance
62
|
Posted - 2015.03.09 18:10:00 -
[61] - Quote
I don't think there is anything wrong with the matchmaking as is. If Rattati says the numbers are working it is probably functioning as well as can be. If there is still problems in the match other solutions have to be found. Personally I have the feeling that a system which gives you a penalty for quitting combined with a system to fold a match might solve this issue. Example the matchmaker makes two even teams and the two best players on one team quit leaving the tree best players on the opposing team. It becomes a stompfest. two people quit the match and get fined by their hiring company for say 1mil isk (or more if need be). The 6 out of the 12 remaining players choose to fold but keep playing till its done. They get sp and maybe no or a smaller payout than normal. The stompers get normal reward. This cuts short stomping matches while preventing them from occurring on the other side. It would even be possible to up the reward for the losing side while the match is going to make sure people keep trying.
If these options are implemented properly it could change the dynamic of losing matches and also make it less crappy to end up in a match which wasn't going great for your side. |
Vesta Opalus
T.H.I.R.D R.O.C.K RISE of LEGION
484
|
Posted - 2015.03.09 19:22:00 -
[62] - Quote
CCP Rattati wrote:These are all answered in my previous blogs/threads.
Step 1) All squads and individuals are converted to units. Each unit is the sum of Mu of all the individuals in the squad.
Step 2) Multiplier is applied to all units
Step 3) The highest value unit is placed randomly on Team A or Team B, let's say Team A for simplicity.
Step 4) Units are placed on the opposite Team, Team B, until the sum on Team B exceeds Team A.
Step 5) Units are placed on Team A until the sum of Mu exceeds Team B or either team has 16 player. Then the rest go on the other team.
Step 6) Repeat until all units have been placed.
This method guarantees that the two best squads are always on opposite sides.
It also guarantees that the second best and third best squads are on the opposite side of the best squad.
If there is a single 6 man squad and no other squads, it gets placed on one Team and ALL the best individuals on the opposite side, due to the multiplier, usually 8-10 of the best solo players fight the 6 man squad.
Finally, this method has never been shown to have not worked as intended. Players concern is usually because the second best squad isn't all in the same corporation so it looks like randoms.
I had a few matches in a row last night where I was the only person on my team who went positive in KDR, I went like 11-1, 8-2, etc, everyone else on the team was completely stomped.
Could the current matchmaking system result in a circumstance where SQUAD 1 (the stompers on the other team) is #1 in MU, then one person on the other team is #2 in MU (in this example, that would be me), then a bunch of low MU players would fill out the rest of the slots?
I think the biggest remaining problem in matchmaking is that the matchmaking thinks its ok to put players with 1 MU up against people with 1000 MU (values are examples based on what I just pulled out of my butt).
All those low MU players just need to be removed from the equation and put into their own game. They shouldnt even be matched up in games where people like me or that opposing squad exist. They just cant compete. Even if they are matched with a high MU squad on the same team, they are still just food for the opposing high MU players. |
Dust 0001
Spyders Inc.
1
|
Posted - 2015.03.09 20:54:00 -
[63] - Quote
CCP Rattati sucks at his job. Get good Rattati. Get good. My 12 year old brother can code better than half your company. Eat sh*t maggot.
I am Big Burns.
|
Morbeus V Kain
The Naughty Ninjas Smart Deploy
0
|
Posted - 2015.03.09 21:13:00 -
[64] - Quote
Dust 0001 wrote:CCP Rattati sucks at his job. Get good Rattati. Get good. My 12 year old brother can code better than half your company. Eat sh*t maggot.
^^ Someone got banned. |
Nothing Certain
Bioshock Rejects
1570
|
Posted - 2015.03.09 21:16:00 -
[65] - Quote
CCP Rattati wrote:Mossellia Delt wrote:This doesn't seem to work at all though. When I'm playing its stacked, badly.
Team A will have a 6 man squad, plus a 3 man squad with a high MU While team B will have everyone solo with 2 good players and 14 blues who can't get more then 200 WPs. I have checked hundreds of battles, and demonstrated in a recent thread a sample of such a deep dive. How many have you checked, and where is your data? How do you determine who is in squads and who not?
I believe you when you say it is working, but on the other hand, it definitely isn't. Matches are terribly one sided and there is no way to balance a 6 man squad of good players if there isn't six even better players on the other side. I know 4 man squads and metalevels are being considered, just understand the huge amount of frustration this is causing people.
I'm kind of dense but it seems to me that making 1 man squads is detrimental as well.
Because, that's why.
|
Dragonmeballs
Better Hide R Die
83
|
Posted - 2015.03.10 19:17:00 -
[66] - Quote
A few thoughts on the comments made by Rattati and others so bear with me as this contains math and I beg your forgiveness for it length:
Team building as explained by Rattati is based on "initial conditions" before the match starts. Each merc or squad is assigned a value. The sum of each side is supposed to be similar. This should bring a balanced match.
Take for example the typical squad of stompers we see 5-9 pm EST every freaking day it seems: The squad would look like this: A,B,C,D,E where A,B,C are the shooters and D,E are logistical/support.
Individual MUs (assume 1-10): A,B,C=8; D,E=7: This squad's Mu = 38.
This seems fair but as you know the key to winning is communication/coordination among the squad.
Rattati explained there is a multiplier (Step 2) say 1.2. Now this squad's Mu is 45.6. In other words you going to need at least 9 random solo players on the other side with an average Mu of 5 just to match this squad initially.
Would a 5 v 9 match work, probably not--I'll expand on this later.
Fill out the rest of the teams. Add 11 more to the stomp team and 7 more to the other team selecting Mu values so the sum of each side is similar.
Now here is the problem: As soon as the match starts the 11 added to the stomp team realize they are with a stomp squad. They forgo the cheap fits and bring out ADV or better. They run with the stompers and now a herd is formed. This herd is difficult to turn back for a team of randoms. As soon as the team of randoms is on their heals the momentum has shifted and a stomp unfolds. (We all know that tactically there are ways to mitigate a stomp but lets save that for another thread)
No amount of matchmaking math can predict the result of herd mentality. The herd mentality is not confined to the winning team. It occurs on the loosing side as well. Nobody in their right mind will keep spamming PRO suits in attempt to turn back a stomp that is fully underway. That is why an OMS that is 78-60 at 12 minutes unravels into a 75-24 in the span of the next 3 minutes.
Therefore the mid match conditions (mid game team Mu) are very different than the initial conditions. And here lies the problem and therefore the reason stomps have become so prevalent. There is no easy way to predict how each of the herds will alter the team Mu values.
I said no easy way but I believe there is a way to deal with this problem before the next match. I submit the multiplier for the squad Mu of should be a variable.
A fair question to ask is how is CCP supposed to adjust this multiplier (Step 2)? Simple. These squads don't stomp in a one and done style. They persist...for hours. What if the the squad Mu multiplier were to increase until a suitable outcome is found? (I define a suitable outcome as a match that ends with a 8-10 clone difference between the teams--these are the matches we all like to be part of btw)
The rub here is the multiplier delta for a given squad would be based on their performance over the course of hours. Difficult to do I admit but I don't believe it to be impossible.
Don't like the adjusting the multiplier then I see nothing wrong (as a means to compensate) by shorting the stomp team a randoms as a means to balance the match. I also see nothing wrong with making a stomp squad wait until there is a suitable squad for them to take on. Note: I do not propose this for Dom or Skirm--only Ambush. I'm not that well versed in those game modes so I won't recommend changes at this time.
More than a few times in the last few weeks I have seen two hours go by without a win. Is it the same stomp squad? NO! Some nights there are three or more but they never seem to find each other when I'm in the match. However, when they do it is tempting to grab popcorn and watch how things unfold when the balancing equation works as intended.
To be clear Rattati I am not criticizing the work you have done to date on this. I don't envy your task. I just submit your math needs to add one more dimension.
Blueberry!....Make yourself useful and shoot the blurry thing running this way!
|
Soldier Sorajord
Subsonic Synthesis RISE of LEGION
187
|
Posted - 2015.03.10 19:44:00 -
[67] - Quote
Funny enough, You can actually get destroyed for being a squad. I'm not sure how the system works, but The MU I guess likes putting me and the people I run with (who run typically std/adv gear) In with people who Pub stomp 24/7. 11/10 IGN Matchmaking. *Slow clap*
Because Matchmaking is totally effective when You see someone go 15/0 every single game they play.
I'm one of the blueberries that Isn't useless, cause I always have a rep tool, and USE IT. B)
|
Soldner VonKuechle
SAM-MIK General Tso's Alliance
1693
|
Posted - 2015.03.10 19:49:00 -
[68] - Quote
Soldier Sorajord wrote:Funny enough, You can actually get destroyed for being a squad. I'm not sure how the system works, but The MU I guess likes putting me and the people I run with (who run typically std/adv gear) In with people who Pub stomp 24/7. 11/10 IGN Matchmaking. *Slow clap*
Because Matchmaking is totally effective when You see someone go 15/0 every single game they play.
Mu =team balancer =/= match making. There is no match making, only team balancing.
BigBurns is exiled,and so with that, the forum's jimmies will no longer be rustled.
|
Maken Tosch
DUST University Ivy League
11126
|
Posted - 2015.03.11 00:08:00 -
[69] - Quote
Mossellia Delt wrote:CCP Rattati wrote:Mossellia Delt wrote:This doesn't seem to work at all though. When I'm playing its stacked, badly.
Team A will have a 6 man squad, plus a 3 man squad with a high MU While team B will have everyone solo with 2 good players and 14 blues who can't get more then 200 WPs. I have checked hundreds of battles, and demonstrated in a recent thread a sample of such a deep dive. How many have you checked, and where is your data? How do you determine who is in squads and who not? Well, I am guessing the 6 and 3 who are of the same corps are in squads, and I could see using the neocom that my team didn't have a squad. And that my team had 12 NPC corp guys who looked fresh out of the academy
Those are very poor indicators of determination. Let me explain.
Without having any clear indicators on the roster or without first knowing ahead of time that someone you know is definitely in a squad, you can't tell beyond a shadow of a doubt that some of the players you have seen who happen to be from the same corp are in a squad. Take me as an example. I very seldom join a squad and usually only after I finish hosting a lecture on the weekends. I'm a solo player. A lone wolf. Yet every now and then I run into someone in my team who happens to be from Dust University. Yet we don't squad up and I hardly communicate unless I'm in a squad which, again, is a post-lecture thing for me on the weekends.
Also, the same is true on the opposite team filled only with randoms. They may seem random but they could be alts from veterans or they could already be their own squads without looking like they do. Again, take me as an example. After I finish a lecture, I tend to squad up with random players who may be coming from other corps and not necessarily from Dust University. So me and the other guys/girls might seem like nothing more than random players without any squad formed, but in reality we are coordinating.
So until we have visual indicators to show who is in a squad in the opposing team, we will never know for sure.
On Twitter: @HilmarVeigar #greenlightlegion #dust514 players are waiting.
|
TITAIN RUNNER
OUTCAST MERCS RISE of LEGION
1
|
Posted - 2015.03.11 09:38:00 -
[70] - Quote
CCP Rattati wrote:These are all answered in my previous blogs/threads.
Step 1) All squads and individuals are converted to units. Each unit is the sum of Mu of all the individuals in the squad.
Step 2) Multiplier is applied to all units
Step 3) The highest value unit is placed randomly on Team A or Team B, let's say Team A for simplicity.
Step 4) Units are placed on the opposite Team, Team B, until the sum on Team B exceeds Team A.
Step 5) Units are placed on Team A until the sum of Mu exceeds Team B or either team has 16 player. Then the rest go on the other team.
Step 6) Repeat until all units have been placed.
This method guarantees that the two best squads are always on opposite sides.
It also guarantees that the second best and third best squads are on the opposite side of the best squad.
If there is a single 6 man squad and no other squads, it gets placed on one Team and ALL the best individuals on the opposite side, due to the multiplier, usually 8-10 of the best solo players fight the 6 man squad.
Finally, this method has never been shown to have not worked as intended. Players concern is usually because the second best squad isn't all in the same corporation so it looks like randoms. then why was me any my squad mate in a dom 2 vs 4 an the 4 proceeded to proto stomp us into oblivion
theirs the big an the bad.....then theirs me
|
|
Raptor Princess
ALLOTEC INC
165
|
Posted - 2015.03.11 13:09:00 -
[71] - Quote
Maken Tosch wrote:Mossellia Delt wrote:CCP Rattati wrote:Mossellia Delt wrote:This doesn't seem to work at all though. When I'm playing its stacked, badly.
Team A will have a 6 man squad, plus a 3 man squad with a high MU While team B will have everyone solo with 2 good players and 14 blues who can't get more then 200 WPs. I have checked hundreds of battles, and demonstrated in a recent thread a sample of such a deep dive. How many have you checked, and where is your data? How do you determine who is in squads and who not? Well, I am guessing the 6 and 3 who are of the same corps are in squads, and I could see using the neocom that my team didn't have a squad. And that my team had 12 NPC corp guys who looked fresh out of the academy Those are very poor indicators of determination. Let me explain. Without having any clear indicators on the roster or without first knowing ahead of time that someone you know is definitely in a squad, you can't tell beyond a shadow of a doubt that some of the players you have seen who happen to be from the same corp are in a squad. Take me as an example. I very seldom join a squad and usually only after I finish hosting a lecture on the weekends. I'm a solo player. A lone wolf. Yet every now and then I run into someone in my team who happens to be from Dust University. Yet we don't squad up and I hardly communicate unless I'm in a squad which, again, is a post-lecture thing for me on the weekends. Also, the same is true on the opposite team filled only with randoms. They may seem random but they could be alts from veterans or they could already be their own squads without looking like they do. Again, take me as an example. After I finish a lecture, I tend to squad up with random players who may be coming from other corps and not necessarily from Dust University. So me and the other guys/girls might seem like nothing more than random players without any squad formed, but in reality we are coordinating. So until we have visual indicators to show who is in a squad in the opposing team, we will never know for sure.
If there were 6 players from a corp all running in a pack, it's safe to assume they're a squad (especially when only one is a squad leader). If it was dust university, I wouldn't assume they were in a squad because I've never seen a squad of them in the year I've been playing. Other corps however, regularly work together.
You also tend to bump into the same people over and over again, so know which randoms are in squads together.
|
Stormblade Green
KnightKiller's inc.
48
|
Posted - 2015.03.11 13:26:00 -
[72] - Quote
There is tend to and always...... I went 5 matches yesterday in a Dom. against the exact same people....... Why?
One might say... I'm very skilled... yet I'm his apprentice... So what does that say about my mentor?
|
|
|
|
Pages: 1 2 [3] :: one page |
First page | Previous page | Next page | Last page |