Greyhart, on 21 February 2019 - 01:52 AM, said:
For balancing in match maker barring some estimate of player skill the use of weight is the problem.
I am at a loss as to why they don't allocate a battle value to every mech and competent so that the deployed mech has a value based on what is in it and what the mech is. 100 ton mech with 1 small laser is not the same as a fully equiped mech.
Then balance the teams on battle value and player skill.
Player skill is always going to be an ify criteria and therefore a robust method of valuing the mech is needed.
Battle value can then be adjusted on a regular basis and used for FP rather than tonnage.
Now we have to go back into which game mode you are referring to, as well as player base.
If we are in QP game mode, it will try to pair corresponding weights weights on each side. If one side has three assaults, the MM will attempt to find three assaults for the other team as well (within the same PSR tier bracket).
If we are in GP, you are restricted by tonnage limited the more players that are within your group at the time, then try to pair you with other players from an average of your PSR ranks, but here GP is very loose on it's match making when related to PSR. This is typically because it's got fewer resources to pull from.
In FP there is no MM, in part due to limited resources to make teams by, as well as it's intended "hard mode". You are suppose to drop in FP in groups. Drop decks are limited by a set tonnage cap, so as to prevent too much armor/health from being on the other side as well as to encourage multi-role/tonnage spread from a single player.
I will remind that a lot of players have requested (demanded) that MM balance via tonnage. This was as requested as Skirmish game mode (and just like Skirmish, people complained about it as soon as it was implemented. In Skirmish, people complained that PGI should have done something to prevent players from hiding at the end of the match to "drag the match on needlessly". In this case here, we got tonnage balance, people don't like the restrictions (depending upon queue mode), and complain and now want a different system.
The ultimate problem with a battle value is the same problem that it had with BT, how do you rate each individual component? BV systems make it very hard to alter a gear's performance on an individual piece, as because as soon as you adjust something it's old value is incorrect and you now need to assign a new value. Then you need to consider some multiplier for specific weapon combos that are powerful together but not when apart, and some multiplier when weapons are boated in mass quantity... not to mention that the point of balancing gear is to have every piece of gear valid and equal (but different) to every other piece of gear.
BV also wouldn't consider player skill abilities within specific types of mechs and mech classes. BV will rank almost every light mech as very low numbers, and most pilots struggle using light mechs. However, an ace in a light mech can wreck face. In my specific case, I tend to perform well with medium class mechs over other classes (or mechs that perform similar to a medium mech), and benefit more from JJs compared to what I've seen of other players. An Assault mech may have more BV (because of number of gear it can take), but my performance would be (on average) worse in any assault mech compared to a much lower BV valued medium mech.
Trust me when I say I would love a BV type system and I do see it's potential strengths, but it also would have a lot of weaknesses and would be a lot of work to actually create and maintain, and it's actual value/impact on the game itself may be more limited than either of us may believe.
Bud Crue, on 21 February 2019 - 02:13 AM, said:
PGI may have tried in the past, but they are no longer tying at all these days, and that according to Paul is intentional. I don’t recall which NGNG twitch-cast-episode-thingy it was but a while back when complaint was made about the tier system and the resultant horrid MM system, Paul said (I’m paraphrasing here): “The system is working as intended. It is designed to keep truly new players away from the more experienced population and nothing more.” So, no, they are not trying to make a system that provides “proper match balance”, they are not trying at all.
The remark was that PGI never gave it a try to balance via any form of reasonable MM. As mentioned, PGI has obviously tried to create a reasonable MM system, several times.No matter what system may be adopted, people will complain about it and say it's not good. I see it in every game with forums that I play. I know of one game (Monster Hunter) where they announce this great patch of fixes, and people complain because "this exact resolution that you don't have to play the game on is still buggy, fix it". I mean, they could correct every other bug in the game, and people will still find a way to complain about something, even if it's completely trivial and easily bypassed.
I might also mention, PGI has been working on trying to balance gear, as they already have a MM system. It may not be perfect (I think it's better than Elo, but still too dependent on winning rather than just match performance, with not enough ways to go down... unless you of course lose a match...), but it is in fact a MM system and it does it's job of separating the different skill/experience levels. The hardest tier to play in though is T3, as they get dropped with anyone.
Sadly, just about any MM system put into place is most likely still going to reward selfish game play (personal glory) rather than teamwork... but player skill is probably one of the hardest things to place into a formula to use for any equation. You can't just look at someone and know how well they are going to play. You can't even look at end of match scores to necessarily get that information (because sometimes you take a hit for the team to get that victory). We humans can be rather unpredictable things, and any numbers crunched on us can be ruined by many things. (Played poorly because I'm in pain/sleepy/distracted/failing hardware/sick/etc. I played well because I got into a zen moment/got a new computer/replaced my mouse/upgraded my internet/figured out how this weapons work/had really good luck and was in the perfect spot this match/etc.)
I think this may be a case of "If it ain't broken, don't fix it". PSR may not be perfect, but for the moment it's not broken either. It seems better than Elo use to be, and it does try to divide players into different groups based on match performance. If anything, my only large critique on the system is it's reliance on win/loss modifiers, where a perfectly average match performance (250-300ish) can still see a reduction of PSR on a loss, but be an plus on a win. Or that an abysmal match of less than 100 can be an equal result on a win, and a reduction on a loss... I think if they adjusted this facet of PSR MM that it could really improve as an overall system.
7
