Jump to content

Final Analysis Of Matchmaker

Balance Gameplay Social

40 replies to this topic

#1 MechWarrior319348

    Member

  • PipPipPipPipPipPipPip
  • Big Daddy
  • Big Daddy
  • 997 posts
  • LocationInside a straightjacket

Posted 21 April 2016 - 05:37 AM

Ok, I have literally spent almost an entire 24 hours playing this damn game.... I am exhausted. I made a post earlier this month here about this subject, and I would be an *** not follow up on it. So I done gone and went through the entire matchmaker cycle first hand. (Yes I was wrong in that thread)

I really needed to see what it looked like in game. I'm working off eastern U.S. time. I played on both North American servers, and European servers all damn day. Google can convert times for you.

=====================================================================


12:00 AM - 4:00 AM(Excellent) 5:00 AM to 5:00PM(Big problem) 6:00 PM to 11 PM (decent)

All the tiers are grouped well. / Tier 1-3 are mashed together. / Open to opinion for quality.

----------(hard to make a bum graph in a forum).

12:00 AM - 4:00 AM(Excellent) - All the tiers are grouped well.


5:00 AM to 5:00PM(Big problem) - Tier 1-3 are mashed together.


6:00 PM to 11 PM (decent) - Open to opinion for quality.

=====================================================================


The middle time is where we have our issue. Now they told us that tier 1-3 are supposed to be grouped together. But the programmers did not finish the feature. Management needs to deal out some spankings.

1. The matchmaker draws the correct number of players from the stream to be organized.

2. The matchmaker Does not balance out the number of players of each tier; to be placed on each team.

I mean, you would expect it to look like this:

Team A - 2 tier one players, 5 tier two players, 5 tier three players.

Team B - 2 tier one players, 5 tier two players, 5 tier three players.


But it does not go that far. It pulls these players from the stream, but doesn't organize them. They are distributed randomly among the two teams.


Actually... There, that's the problem^^^^^^^^


If they tell us the specifics of how tier 1-3 are grouped, then maybe we can clean it up a little bit. They dont seem to realize that they aren't in this programing fight by themselves. We can help.

Edited by Gigliowanananacom, 21 April 2016 - 07:38 PM.


#2 Coolant

    Member

  • PipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 3,079 posts
  • Facebook: Link
  • LocationCalifornia

Posted 21 April 2016 - 06:17 AM

Logically fallacy...your play time does not make something true or false. You could play a year straight (if that was physically possible) and it still wouldn't make something true or false. Astronauts that have been to the moon and back could say that the moon is made of cheese and despite their experience in space, doesn't make it true.

Quote

I swear this is the issue, I went through the effort to find it.


The amount of effort spent does not make something true or false. I could say,"I found a coin and it had Homer Simpson's face on it! I spend all day searching for it!" - but, doesn't make it true.

Edited by Coolant, 21 April 2016 - 06:19 AM.


#3 Almond Brown

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Philanthropist
  • Philanthropist
  • 5,851 posts

Posted 21 April 2016 - 06:22 AM

Well Finally! A "Final Analysis". Thank You good sir. It was a long wait but FINALLY! ;)

#4 MechWarrior319348

    Member

  • PipPipPipPipPipPipPip
  • Big Daddy
  • Big Daddy
  • 997 posts
  • LocationInside a straightjacket

Posted 21 April 2016 - 06:29 AM

View PostCoolant, on 21 April 2016 - 06:17 AM, said:

Logically fallacy...your play time does not make something true or false. You could play a year straight (if that was physically possible) and it still wouldn't make something true or false. Astronauts that have been to the moon and back could say that the moon is made of cheese and despite their experience in space, doesn't make it true.



The amount of effort spent does not make something true or false. I could say,"I found a coin and it had Homer Simpson's face on it! I spend all day searching for it!" - but, doesn't make it true.

Ugh... disgusting post. This is a video game, not a conspiracy theory. The moon being made of cheese? how does this compare to something like that? No thanks inspector gadget, this isn't a court room where you are trying to prove if god exists in your television.

Jesus....... fall back onto bureaucracy?

Edited by Gigliowanananacom, 21 April 2016 - 06:30 AM.


#5 MechWarrior319348

    Member

  • PipPipPipPipPipPipPip
  • Big Daddy
  • Big Daddy
  • 997 posts
  • LocationInside a straightjacket

Posted 21 April 2016 - 06:39 AM

View PostAlmond Brown, on 21 April 2016 - 06:22 AM, said:

Well Finally! A "Final Analysis". Thank You good sir. It was a long wait but FINALLY! Posted Image

Thanks! I was thinking of you the entire time! I had a picture of you in my cockpit next to the joystick!

#6 Almond Brown

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Philanthropist
  • Philanthropist
  • 5,851 posts

Posted 21 April 2016 - 09:08 AM

View PostGigliowanananacom, on 21 April 2016 - 06:39 AM, said:

Thanks! I was thinking of you the entire time! I had a picture of you in my cockpit next to the joystick!


Ohhh stop that you bad boy. You may end up with dual "joysticks" in your cockpit it you keep that sort of talk up... ;) ;)

#7 Dread Render

    Member

  • PipPipPipPipPipPipPip
  • Philanthropist
  • 847 posts
  • LocationSouth River NJ

Posted 21 April 2016 - 09:22 AM

View PostCoolant, on 21 April 2016 - 06:17 AM, said:

Logically fallacy...your play time does not make something true or false.

The amount of effort spent does not make something true or false.


Your post is pure punk...

The time he spent is just the Sample time.
The effort he put in to it was just the amount required to get the results he posted.
Go find some other forum to poop on why don't you...

Thanks OP for the info... My experience with the game tells me you are very close if not 100% correct.

Edited by Render, 21 April 2016 - 10:00 AM.


#8 occusoj

    Member

  • PipPipPipPipPipPip
  • Ace Of Spades
  • 452 posts

Posted 21 April 2016 - 09:53 AM

By large and all the time based variation in matches is something I can confirm. The more MM runs out of players, the worse it gets - not that surprising on its own.
What IS surprising though is how grossly mismatched some games are.

I recorded thousands of matches and the intra-day trend you mentioned is also found there. But then, there are stronger and slower trends as wells as significant, probably non-MM related, influences. Like HSR parameters and so on.

View PostCoolant, on 21 April 2016 - 06:17 AM, said:

Logically fallacy...

You werent by any chance in the group of people screaming "gamblers fallacy" whenever someone posted about ELO based MM?

#9 Wintersdark

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • 13,375 posts
  • Google+: Link
  • Twitter: Link
  • LocationCalgary, AB

Posted 21 April 2016 - 09:57 AM

Population is the problem. The coding is fine, and we know exactly what the algorithm is.

Nobody is "grouped together by tiers", that isn't how it works.

Players are added by simply having the right weight classes and being as close to the seed PSR rating as possible. The MM stretches to +/-1 tier after one minute, +/-2 after two minutes. And that's since the MM started building the match, NOT since you joined it.

#10 MechWarrior319348

    Member

  • PipPipPipPipPipPipPip
  • Big Daddy
  • Big Daddy
  • 997 posts
  • LocationInside a straightjacket

Posted 21 April 2016 - 10:05 AM

View PostWintersdark, on 21 April 2016 - 09:57 AM, said:

Population is the problem. The coding is fine, and we know exactly what the algorithm is.

Nobody is "grouped together by tiers", that isn't how it works.

Players are added by simply having the right weight classes and being as close to the seed PSR rating as possible. The MM stretches to +/-1 tier after one minute, +/-2 after two minutes. And that's since the MM started building the match, NOT since you joined it.

Hello there, can you direct me to where you got the info?

#11 MechWarrior319348

    Member

  • PipPipPipPipPipPipPip
  • Big Daddy
  • Big Daddy
  • 997 posts
  • LocationInside a straightjacket

Posted 21 April 2016 - 10:51 AM

View PostCoolant, on 21 April 2016 - 06:17 AM, said:

Logically fallacy...your play time does not make something true or false. You could play a year straight (if that was physically possible) and it still wouldn't make something true or false. Astronauts that have been to the moon and back could say that the moon is made of cheese and despite their experience in space, doesn't make it true.



The amount of effort spent does not make something true or false. I could say,"I found a coin and it had Homer Simpson's face on it! I spend all day searching for it!" - but, doesn't make it true.

I think he was trying to refer to the face on the moon. The pile of dirt that has a grumpy face on it. Became a big deal because the light from two different angles gave the same shape... or something like that. I'm sure cheese doesn't have anything to do with it though..... I think.

"I know cheese, but I'm not a cottage guy"

-Charlie Kelly

Edited by Gigliowanananacom, 21 April 2016 - 10:55 AM.


#12 Wintersdark

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • 13,375 posts
  • Google+: Link
  • Twitter: Link
  • LocationCalgary, AB

Posted 21 April 2016 - 11:39 AM

View PostGigliowanananacom, on 21 April 2016 - 10:05 AM, said:

Hello there, can you direct me to where you got the info?
Karl Berg's megathread. He wrote the Matchmaker, and was quite detailed in explaining it. (Off Topic section, Paging Karl Berg).

He's gone now, and while the system used Elo then and PSR now, they are (for the purposes of building matches) the same.

Russ has also spoken during town halls of this, and how the MM actually uses your PSR number (not tier) when seeking initially. It's just tier bounds that handle how the MM spreads open its net to find players over time.



#13 Kekkone

    Member

  • PipPipPipPipPip
  • Mercenary Rank 1
  • 144 posts

Posted 21 April 2016 - 11:53 AM

Is there an option to show the tiers of people in-game that i've missed?

#14 MechWarrior319348

    Member

  • PipPipPipPipPipPipPip
  • Big Daddy
  • Big Daddy
  • 997 posts
  • LocationInside a straightjacket

Posted 21 April 2016 - 12:42 PM

View PostWintersdark, on 21 April 2016 - 11:39 AM, said:

Karl Berg's megathread. He wrote the Matchmaker, and was quite detailed in explaining it. (Off Topic section, Paging Karl Berg).

He's gone now, and while the system used Elo then and PSR now, they are (for the purposes of building matches) the same.

Russ has also spoken during town halls of this, and how the MM actually uses your PSR number (not tier) when seeking initially. It's just tier bounds that handle how the MM spreads open its net to find players over time.

I see... yes..

Ok, I suppose that makes sense. My specific concern is what happens after matchmaker has picked out the players. The step after what you just mentioned.

- So at this point, the matchmaker has all 24 players picked out, but then how does it divide them between the two teams?

View PostKekkone, on 21 April 2016 - 11:53 AM, said:

Is there an option to show the tiers of people in-game that i've missed?

No, you have to openly ask in chat. When the match is winding down. Its tricky though, because asking people about their tier can become offensive quickly; other times they turn it into a joke.

#15 Death Proof

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 546 posts

Posted 21 April 2016 - 12:45 PM

View PostKekkone, on 21 April 2016 - 11:53 AM, said:

Is there an option to show the tiers of people in-game that i've missed?


View PostGigliowanananacom, on 21 April 2016 - 12:42 PM, said:

No, you have to openly ask in chat. When the match is winding down. Its tricky though, because asking people about their tier can become offensive quickly; other times they turn it into a joke.


You can also look up players on the forum. Most active players display their tier.

#16 TKSax

    Member

  • PipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 1,057 posts
  • LocationNorth Carolina

Posted 21 April 2016 - 12:47 PM

View PostGigliowanananacom, on 21 April 2016 - 12:42 PM, said:

I see... yes..

Ok, I suppose that makes sense. My specific concern is what happens after matchmaker has picked out the players. The step after what you just mentioned.

- So at this point, the matchmaker has all 24 players picked out, but then how does it divide them between the two teams?


No, you have to openly ask in chat. When the match is winding down. Its tricky though, because asking people about their tier can become offensive quickly; other times they turn it into a joke.




Which is the problem with any analysis you do. I have been asked in game and I always respond with same Response, Tier 5 :) Which is not my tier.




View PostDeath Proof, on 21 April 2016 - 12:45 PM, said:


You can also look up players on the forum. Most active players display their tier.


I don't and I am very active.

#17 MechWarrior319348

    Member

  • PipPipPipPipPipPipPip
  • Big Daddy
  • Big Daddy
  • 997 posts
  • LocationInside a straightjacket

Posted 21 April 2016 - 12:53 PM

View PostTKSax, on 21 April 2016 - 12:47 PM, said:




Which is the problem with any analysis you do. I have been asked in game and I always respond with same Response, Tier 5 Posted Image Which is not my tier.






I don't and I am very active.

I cant completely verify information with that data, yes you are correct. But the experience was supposed to give me a good Idea of what was going on.

Now I cant verify that I am right, because of what you say. That's why I need to get something official that shows how the players are divided between the two teams; after they have been pulled out of the pool.

Its one last lego piece, and I need it.

Edited by Gigliowanananacom, 21 April 2016 - 12:55 PM.


#18 MechWarrior319348

    Member

  • PipPipPipPipPipPipPip
  • Big Daddy
  • Big Daddy
  • 997 posts
  • LocationInside a straightjacket

Posted 21 April 2016 - 01:01 PM

View PostDeath Proof, on 21 April 2016 - 12:45 PM, said:




You can also look up players on the forum. Most active players display their tier.

I didn't want to downsize the game and look up every player 1 by 1 though........ sounds like a nightmare.

#19 Wintersdark

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • 13,375 posts
  • Google+: Link
  • Twitter: Link
  • LocationCalgary, AB

Posted 21 April 2016 - 01:03 PM

View PostGigliowanananacom, on 21 April 2016 - 12:42 PM, said:

I see... yes..

Ok, I suppose that makes sense. My specific concern is what happens after matchmaker has picked out the players. The step after what you just mentioned.

- So at this point, the matchmaker has all 24 players picked out, but then how does it divide them between the two teams?
It builds one team, then it builds the second. It starts building the first team before it has 24 players - it just picks appropriate players vis a vis PSR and weight class.

Quote

No, you have to openly ask in chat. When the match is winding down. Its tricky though, because asking people about their tier can become offensive quickly; other times they turn it into a joke.
This is pointless. People lie. I always do; I NEVER answer my current tier. Sometimes I make a joke of it, sometimes I just say I'm T4 for the lols. A look here shows my tier of course.

#20 Death Proof

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 546 posts

Posted 21 April 2016 - 01:04 PM

View PostTKSax, on 21 April 2016 - 12:47 PM, said:

I don't and I am very active.


I said most.

Those that don't are usually tier 4 or 5...with the odd higher tier player trying to be deceptive about their tier for some weird reason. Posted Image





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users