Massive Assault
http://www.massiveassaultnetwork.com/forum/

New rating system in MAN 2!
http://www.massiveassaultnetwork.com/forum/viewtopic.php?f=37&t=10467		 Page 7 of 7
Author:  Maelstrom [ Tue May 01, 2007 7:53 am ]
Post subject: 
This topic is becoming pretty bloated :) Hard to really make any progress with so many threads of conversation going on in the same topic.

See this post on how to analytically determine the value for a map:
http://massiveassault.com/forum/viewtopic.php?t=12755

If I'm totally off base let me know. Several have mentioned ideas related to analytically determining the value of a map. Lets talk them out somewhere besides this topic :P

Anyways, on guderian27 and Shadow_D's comments, I totally agree that we can use some kind of measuring tool to determine what the value of a map should be, trying to get an end result with a similar spread to Storm's list.

But the inclusion of your opponent's skill is already handled by the glicko system pretty well, with the problem being that in Glicko right now, all maps are being treated equal, when they definately are not. Correct me if I'm wrong, but we're trying to determine a scaling factor for the points from games based on the map played, both for the glicko based skill rating and the experience rating.
Author:  guderian27 [ Sat May 05, 2007 11:57 am ]
Post subject: 
Maelstrom wrote:" Correct me if I'm wrong, but we're trying to determine a scaling factor for the points from games based on the map played, both for the glicko based skill rating and the experience rating."

Yes of course, we should try to give a value to the maps that is an objective value.
Then, in order to have a right assignation of points/victory, we should conciliate this value with a value deduced by glicko and experience rating (subjective value). Correct?
Author:  Captain X [ Fri May 11, 2007 4:50 am ]
Post subject: 
I liked Placids idea to calculate the luck-factor of a map out of the statistics of the games played on the map, that will always be more accurate, than any thought out/guessed numbers.

one easy idea to achieve this would be to take the ratio of wins from PL/FNU. As you can clearly notice, people win more often with PL, as its the first side to make the turn. this is especially true in the statistics of good players.

on the thesis that PL has some advantage, it means those maps where you have a fair balance of wins between PL/FNU, the PL advantage is minor to the skill and the match is decided primary by strategic skill. so they should count more. those maps where PL wins most of the time, the impact of the side you are playing on is much bigger than the impact of skill you have, so they are no good to determine the better player and should count less.

Even small differences of this ratio gives lots of informations, as we have thousands of games recorded. We need a developer though to give some proofs in numbers and statistics...
But think about it! the more you do, the more it makes sense.

the other part to determine the winning points should be the average length off matches on maps, or even better the average amount of troops build, as the more work you put into a game, the higher the reaward should be. and the the work you put in is directly connected to the size of armies you have do order around.
Author:  Nick_WN [ Tue May 15, 2007 9:55 am ]
Post subject: 
Massive Assault Network 2: New Demo Presents Dual Rating System

New Orleans, LA - May 15, 2007 - Wargaming.net, Inc. release the v249 Demo version of its turn-based strategy game Massive Assault Network 2 (MAN 2). It features the innovative dual rating system, which caters to both hardcore and casual players.

The Demo can be downloaded at:
http://www.massiveassaultnetwork.com/do ... n2demo.exe

The patch to v249 available at: http://www.massiveassaultnetwork.com/do ... 9patch.exe

With this new version players are able to select from both skill-based and experience-based systems, matching the most appropriate opponent for a battle. Thus the newbie players can be content with experience-based system, as for them the ranking reflects their growing experience, which grows with every battle, no matter whether they win or loss; while skill-based system can give the best fit to MAN 2 hardcore players, who are able to evaluate the real skill level now. In addition, the new Demo evolves the Mentoring system aiming to stimulate rookies to learn battle tactics from the strongest players of MAN 2 universe.
Author:  ravermeister [ Tue May 15, 2007 4:34 pm ]
Post subject: 
THANK YOU !!!!!

All is right in the universe once again...


Did all games count towards the skill based rank? or only the ones before the new (old) system was implemented....?
Author:  Nick_WN [ Wed May 16, 2007 10:02 am ]
Post subject: 
OK, all the games BEFORE this patch (even with trial players) are counted toward your Glicko rating.

Starting from this version:
1. Games with AI are not counted in Glicko
2. Games with trial players are not counted in Glicko
3. You need to check box "Rating game" when posting new challenge in order to get this game counted toward your Glicko rating.
Author:  Alliance [ Thu May 17, 2007 6:24 am ]
Post subject: 
Thanks for the patch!
I am begginer and where i can read about rules of this Glicko rating system? And that means "subscription points"? How i can use this points for prolongation of my lifetime membership? :)
Author:  Tiger [ Thu May 17, 2007 3:47 pm ]
Post subject: 
Alliance wrote:
Thanks for the patch!
I am begginer and where i can read about rules of this Glicko rating system? And that means "subscription points"? How i can use this points for prolongation of my lifetime membership? :)


As dar as I know it's points in the "Nicko"-rating. 5 points for each turn in the battle with trial player, if he will subscribe (you'll get points for all your turns in the battles with this player BEFORE he purchase subscription).
Author:  Tiger [ Fri May 18, 2007 4:07 pm ]
Post subject: 
These are new coefficients of maps in the Glicko Rating:

Micronesia 1,0
Butterfly 1,2
Bizarria 1,3
V6 Demo 1,6
V6 1,6
Station 'Thunder' aka Iron 1,7
San Fernando 1,7
Lion's Eye 1,7
Brimstone 1,7
Emerald 1,8
Antarcticus 2,0
Midgard 2,4
Crateus 2,6
Treasure Island 2,6
Eden 2,9
Beotorum 3,0
Noble Rust 3,1
Anubis 3,5
Inferno 3,6
Wasserland 3,7
Swamp 3,7
New Paradise 4,5
Twin Islands 4,7
Trinity 5,1
Sea Switzerland 5,9

Coefficients were based on empirical data.
Author:  storm440 [ Sat May 19, 2007 8:03 pm ]
Post subject: 
while i would make a few small changes in the placement of some of the countries, overall i think that you have come up with a good system. Am i correct in assuming that if you play a map rated "1" your score will be whatever is listed while if you play a map with a "5" rating you will receive 5 times the points listed?
Author:  storm440 [ Sat May 19, 2007 8:07 pm ]
Post subject: 
also, i am curious as to how twin islands ranked heigher than new paradise? Pretty much every post had trinity, sea swiss, wasserland and new paradise at the top of the list.
Author:  Shadow_D [ Mon May 21, 2007 8:13 am ]
Post subject: 
I shall agree with storm440

As a whole I am happy with system.
But it seems to me, as New Paradise Wasserland it is much more difficult than Twin Islands
Author:  Morn [ Sun May 27, 2007 3:35 pm ]
Post subject: 
IMHO Twin Islands is rated much too high. Actually I would swap Wasserland with Twin Islands.
Page 7 of 7 All times are UTC - 5 hours
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/