Cathy, on 29 October 2016 - 01:52 PM, said:
When is there going to be any sanity in what people are facing.
Not only are the gates all open so that there can be no mediums and no lights in a game these days, but the tonnage variation is disgusting.
Just finished a game with my alt.
Both sides had three assaults side A had Zeus, Victor, Awesome, side B had KDK-3 Atlas D-DC, Direwolf.
This kind of imbalance went all through the sides if it wasn't gross disparity between weight side B ended up with better mechs, Timberwolf against Orion. Only in one mech was there any parity and that was both sides shared a Rifleman.
It's getting rather stupid
I agree that the matchmaker could be better. Balancing tonnage a bit better would be great especially since assaults are not balanced.
However, I would like to point out that nothing has changed. The matchmaker has been the same as far as weight class matching goes for years now. It has never looked at tonnage ... only weight class. There have been many threads with comments and suggestions for improvements on this algorithm but it hasn't changed since PGI doesn't seem to understand how to make the matchmaker better and doesn't see any revenue potential in investing in seeing if they can come up with something a bit better.
Personally, I think that the situation could be somewhat improved in the quick play queue with a very small investment of resources. Once the matchmaker has 24 players in the required weight classes and in the proper pilot skill range ... it could run a very fast tonnage optimization pass to avoid situations with 3x100 on one team and 3x80 on the other. No matter what happens it would still wind up with 2x100 1x80 vs 1x100 2x80 ... but at least it would be more even and the matchmaker could use "pilot skill" as the deciding factor on which to move.
This kind of approach would not require writing a whole new matchmaker ... just a quick post-processor that could re-arrange the teams before passing the data off to the game server. The algorithm could be made to address more cases by looking at total tonnage on both sides as well as the total tonnage in each weight class. Finally, if they wanted to get even more complicated they could assign a rough scaling factor representing the average effectiveness of each chassis so that the more effective mechs would count as additional tonnage in the balancing. Anyway, since all of this would be a post-processor once the 24 players have been selected ... I think it would be fast, efficient and relatively effective at reducing the frequency of large tonnage/effectiveness imbalances in matches.