1 reply to this topic

[Deathlike]

I understand this system needs to be tweaked but no word has come as to progress. So, I'm kinda asking for an update, since you have collected two weeks of data that should be worth a look (especially after the tourney).

I do have a few questions...

1) I understand that the MM attempts to speedily match ELOs with people.. and it's quick... possibly too quick. I think we're on the other extreme of the v1 of ELO MM deployment (too slow/impossible to get a match). Shouldn't part of this be tightened up a little? People that I speak to are still annoyed that they still find mismatches (newer players) in their game... which sounds like the range is too wide/broad. How much more time is needed to produce the data you need and/or revising of the ranges?

2) What exactly is the progress to match weights (to a lesser extent, tonnage) in the current revision of the MM? This is a more common issue that does plague a number of matches... Is that even considered in part of changing the MM process?

3) MM needs to consider the mech variant (not just chassis) when it does ELO score adjusting. Right now, a Raven 3L is far superior to a 2X or 4X (there are issues with ECM and streaks, but that's not the discussion). Even if you fix the problems that ails this particular mech chassis, some mech variants will simply work better for a person over a different light mech (like a Commando 2D for instance). So, inevitably it is hard to properly differentiate ELO from a mech that does one thing well, than another that does well at something else altogether. Differentiating by weight class is not an accurate, but also weak measure of trying to analyze ELO. Is there a shot that ELO will treat mech variants more than just weight classes?

4) With premades (this is not a discussion about premades vs solo PUGs) there should be a different kind of ELO that is generated from them. The average of solo-PUG ELOs is simply not enough, but even worse that it is not accurate. The success of a group changes the dynamics greatly among drops. I understand the demand for a separate queue for these games, but IMO these are not necessary. Rather, they should be populated with other premades of close/equal strength (ideally, of the same sizes). This might make MM more difficult, because premades can be of various sizes, but inevitably a very successful group of players has a greater team effect than individual players that probably/can/will work as a team, just on the natural side of things. This has to be thought out long term, and I don't expect this to be resolved quickly or soon... but you have to address this problem at some point. What do you think?

5) Part of the MM's problem is with not filling out the entire match with 8v8.. I get the feeling that the MM doesn't have enough time to make a proper decision. I had requested in a different thread that players should be allowed to vote (usually the team with the downed players) to allow MM to get people to replace those that have been DCed or simply empty spots in this process. It would only be a 15-30 second extension, probably a 5-10 second vote.. and it is unlikely to be abused (you get only one vote in this anyways). IMO, this is a must-have feature if you are to see less complaining about potential lopsided matches. I know that this would be revisited if/when 12v12 is deployed, but this is a very important step to getting a more complete and competitive system going. Does this sound fair?

6) Last, but not least, MM should have some special considerations about the mechs chosen on particular maps. Alpine is unfortunately decided by the number of light mechs... and it is vital to have speed available in these matches. Although Alpine itself needs balancing (especially for Conquest), the MM has to consider whether a mech is fast and should reasonably attempt to evenly match faster/light mechs on each side. This is not necessary for smaller maps, but it critical for the balance of bigger maps. I fear that there's an actual possibility that a 12v12 match would still have "assault scouts".. which is sub-optimal in trying to make these matches competitive and fair. Is this reasonable?

So.. outside of complaining about what mechs are OP and/or why something is unbalanced (because, they have very little to do with ELO or MM), the ongoing ELO and MM changes needs to be addressed at some point.

[Deathlike]

Bump