3 replies to this topic

[Kamenjar]

To my knowledge, we have no exact knowledge about how Elo is calculated, but there are dev posts that imply the following:
o Elo is solely dependent on team win/loss and not individual's contribution
o MM may pit a higher Elo team against a lower Elo team. If higher Elo team wins, there is little or no change. Otherwise, the lower Elo team gets a boost and higher Elo gets a slight drop.

Suggestion with "rank", where "score" is not the "match score" or Elo:

• ++rank/cbill/xp for hitting the most damaged components on an enemy mech. This value increases at long range.
  
• --rank when hitting a mech without obtaining target data for 3 seconds, when target data is available (AKA. hit R, so you don't blind-fire and hit stuff that don't matter)
  
• --rank when having no targeted mech for 1 second while doritos are avialble (AKA. hit friggin R, if nothing, memorize loadout and paperdoll status)
  
• +rank when revealing a hollow dorito to the team (AKA. hit R, so team knows about unaccounted for mech)
  
• -rank/cbill/xp for hitting armored part over hitting a heavily damaged exposed part, while exposed part is in line of sight. This penalty increases at close range.
  
• +rank +cbill depending on how your "match score" compares to the team's. The guys who pulled the most weight on the team should gain most "rank".
  
• -rank when 2 nearest friendlies are engaged, but you are not for 10 seconds. This penalty decreases with maximum speed, availability of jump jets and maximum range of your weapons.

Now rank all players on the team when the match completes and as opposed to giving everyone equal Elo gain or penalty, adjust the Elo gain/penalty per value of "rank"



Before everyone freaks out about "well what if this or that situation/exploit", I want to point out hat over long time, if implemented correctly, I think that things will even out if this list is fine-tuned. At least, it should be a step in the right direction, so I'd prefer suggestions for improvements over edge-case critique. I assume that there are other, more sophisticated ways of improving this, but may be hard to implement in the game - for example being simply out of position, or penalty when having full armor while team is damaged heavily and not presenting yourself as a target.

Goals of this ranking and c-bill gain system is to promote smart play and reward players who contribute to the team.

What are your thoughts?

Edited by Kamenjar, 27 February 2015 - 02:46 PM.

[Quick n Fast]

Kamenjar, on 27 February 2015 - 11:33 AM, said:

To my knowledge, we have no direct knowledge about how Elo is calculated

ELO = Win/Loss ratio per weight class

http://soundcloud.co...-heat-scale-elo (@ 38:30 mins in)

[EgoSlayer]

How MWO elo works post here:
http://mwomercs.com/...65#entry1626065

Only thing out of date I think is the K value (maximum change) which was changed from 50 to something else, I think 40. That information is somewhere in the Ask Karl Berg thread.

Edited by EgoSlayer, 27 February 2015 - 02:02 PM.

[Kamenjar]

Thanks to both of you for updating with references. I modified the post to correct my statement.

However, we still don't know exactly what "minor" modifications they may have made to the Elo system and how exactly does MatchMaker still very easily pit very experienced players with or against players that are completely new to the game so easily. In theory, If Elo system was perfect, our W/L ratio would be always 0.5, but mine is 0.55 - then why is it not?

The reason why I'm bringing this up is because as of late I've (including a lot of other guys in BWC) been getting some horrendous PUG matches, which should point out that there's something that PGI is not doing right, to the point of the game causing these rage-inducing matches where I'm barely getting the satisfaction out of this game. It could be me getting more frustrated with the game, and I should probably move on after 800 in-match hours.

Having Elo (or just some matchamker factor) be tuned in this way, in my opinion, would at least make it possible to at have quality matches where even if you lose, you enjoyed the game and you saw that your team did all they could, but the other team was just simply better. For all I care, I don't mind losing on regular basis, but at least I'll be able to play with a team that I can work with, rather than get frustrated on regular basis.

Don't get me wrong, I know that we were all new players at some point - but unless MWO is suffering from serious population problem, I don't see the reason to place new players in teams with some very good players. Good players won't be getting the backing of their team and new players won't be having too much fun when facing those good ones either.

Edited by Kamenjar, 27 February 2015 - 03:08 PM.