1 reply to this topic

[Moonlight Grimoire]

The Issue:

Currently PSR floats a player in one of two directions, up, or down. Either a player gradually floats to T5 or T1 creating an apparent bi-modal distribution of players. This is not healthy as T3 gets squished down as does T4 and T2 and forced to play with players that make it hard to actually allow the system to gauge their abilities.

Step One:

To address this issue there are a few steps to be taken, the first is to alter how PSR increase and decrease is awarded at the end of a match as well as dealing with the current stale situation of PSR. To address this PSR must be reset with all players set to the middle of T3, this ideally would be done every 3-6 months.

The second part is addressing allotment of PSR increase. I am not privy to the inner working of PSR, but, from the outside it looks a lot like an XP bar instead of a ranking of where one is within a Tier which is, bad. As of the past six months or so it is less of an XP bar and more of "how long have you played MWO" bar. This is bad as it means someone with a complete lack of skills but has mashed their face against the wall enough can somehow manage to get into Tier 1 and ruin matches for others with their lack of ability.

So how to address this?

Winning Team:
Top 4: +2 PSR
Middle 4: +1 PSR
Bottom 4: 0 PSR

Losing Team:
Top 4: 0 PSR
Middle 4: -1 PSR
Bottom 4: -2 PSR

Their ranking is based off of match score done, this makes it so the system is dynamic instead of static, so those players who did try their best in a bad matchup don't get washed out on PSR for placing top of their team, and not getting an arbitrary amount of Match Score.

This system also has the bonus of self sorting, routinely good players will slowly rise as they will have more gains than losses from +1, +2, +0, -1, -0, +1, +2, -0 netting them a PSR change of +5 over the course of 8 matches instead of the old system where they could have gotten 700 match score in each match and gotten +2, +2, +2, +1, +2, +2, +1 for a change of +12, even though they lost in 3 out of 8 matches. This goes to show how PSR as it works now seems to just be an XP bar and not at all reflective to how well a pilot actually does based on the situation of the match itself.

This would be the hopefully quick and easy to implement fix to PSR, each tier being divided into Z-scores, with T3 being the largest of Z-Score +/-1 for all of Standard Deviation 1, T2 being Z-Score 2, T1 being Z-Score 3+, T4 being Z-Score -2, T5 being Z-Score -3 (and all other beyond that), this will encompass all players and T5 and T1 should be hard to fall into, T1 will be truly a prestigious thing as it means you are 2.2% of the player base, or add in a T0 which is Z-Score -4 which is .1% just for extra chest pounding.

Longer Term Solutions:

1. So while the first proposal will fix a great majority of issues with PSR and tiers, there is a core issue with Match Score and it's calculation. This is that Damage, Kills, KMDD, and all that are put together in a sense that makes players want to play as ineffective as possible. What do I mean by this? It is more effective to blow off a side torso for two component destruction's and a guaranteed KMDD when you continue doing more damage and maybe kill the mech you crippled, creating ineffective kills and rewarding people for padding damage stats instead of just coring a mech out. In this Match Score should instead HEAVILY weigh a calculation of how much damage you did to the mech to kill it against the mech's overall health. This means something like (MS Kill)*(Mech HP)/(Damage done to mech), this will mean those who vomit damage out everywhere aren't going to get those huge match scores, instead it will be those who know how to back stab, finish off a crippled mech, do clean kills. Doing this with component destruction would be ideal as well as it rewards players for crippling a mech when seeing a mech with a damaged part exposed to them and taking the moment to slice it off instead of sitting there slowly blowing off a leg, both arms, a side torso then the CT for max stat padding.

2. Changing how Match Score values are calculated per mode is also a change that is a must, if people are not getting rewards for an activity they will, generally, stop doing said activity. Less rewards for damage, kills, ect outside of skirmish is obvious, much larger rewards for playing the objective. Rewards for interrupting the cap in Domination through shooting or entering and staying in the cap after the enemy to prevent the enemy counter from going down. Increasing rewards for capturing objectives in conquest, denying objectives in conquest. Damaging energy cell carriers as well as disrupting energy cell captures in incursion. Match Score specifically for spotting the VIP, for taking damage while near the VIP, for flipping the towers, and for doing damage to the VIP instead of it being lumped in with total damage as well as VIP kill, VIP Leg destruction being a reward as well. Assault is just Skirmish with a way to end the match other than killing the enemy's last player, give rewards like conquest for capturing and disrupting capture, don't change the value for killing mechs because in my opinion Assault should replace Skirmish as the two have negligible differences.

This will mean in game modes outside of Skirmish and Assault those players who aren't always actively fighting and instead have to run back to emergency flip a cap, or have to pop damage into someone standing in the domination circle, or are taking out towers to prevent enemies from raining LRM's on them will get ample rewards instead of rewards that are equal to connecting to a match. This is more for rewarding players for what they are already doing and making it more rewarding to do what has to be done to win in a crippled mech or opening/end of match situations when there isn't much fighting left to be done instead of the current match score set up which means if you have to sit behind a building in the Grim Plexus Dom Cap to hold the timer for your team while they faff about so you don't lose you don't get nothing for your time making certain you did your part.

3. A multiplier per chassis, and for omni's a multiplier per pod based on hard points, hard point type, tonnage, and placement is also likely a valuable addition as a mech with only 4xERML pulling 400 damage playing their heart out under the current system will be penalized because they only did 400 damage and damage is king even with the nerfs to damage factoring into match score. So instead a multiplier per chassis based on how easily players are breaking a threshold of some amount of damage with a mech per match with the mech would be a good idea. This means a mech with low hard points doing game critical damage gets rewarded in match score just as much as some one with LRM's sand blasting away a mountain side and still racking up over 1k damage in their LRM80 King Crab. This helps also with rewards for playing mechs other than Heavy mechs and meta mechs as players know they aren't gimping themselves by playing a mech that isn't meta when it comes to their PSR and match score. This could also filter into cbills, but, I am uncertain, that would have to be investigated, but, it probably should, but, at a lesser factor.

Conclusion:

The quick change and PSR reset is a must, while I would greatly like all of these to be implimented the quick fix alone would do a lot of the general ground work of addressing many people's concerns with PSR. Each of the three longer term fixes are seperate, while I would love all three, any of them coming in is a good thing, having them come in one at a time would be likely the healthiest way to go about things for addressing how easily unbalanced match score still is by damage vomit in the game while clean play and play that prevents an early loss is unrewarded, and securing an early win is berated because you didn't let your teammates farm the enemy, so might as well at least make it reward players better and farming outside of farming modes not give as great match score.

[BTGbullseye]

No to your initial suggestion. You can have a full team of great players on both sides, and have 2/3 of the losing team have a worse rating despite being great players? Not a good representation of skill-based rating.

Better solution is to have a PSR for overall ability, one for chassis ability, and one for variant ability. For matchmaking purposes, use whichever is highest. You might be a PSR 5 overall, but a PSR 1 in a Hellbringer-Prime. In that situation you shouldn't be stuck against PSR 1 players when you're testing out your first Locust, likewise you shouldn't get stuck with PSR 5 players when obliterating in the Hellbringer.

Also, basing the PSR solely on winning is always going to be bad. Winning is not a good test of a PLAYER's skill, only of the TEAM's capability of winning. PSR rating should have win/loss as a significant part of its scoring system, but it shouldn't be the most significant. Map mode objective scores should ALWAYS come first, then kills, assists, win/loss, damage and other, in that order. Not playing to the objectives should massively hurt your score, but right now it doesn't. Right now the scoring system rewards playing every mode like a skirmish, which is rapidly starting to bore me.

Edited by BTGbullseye, 03 March 2018 - 08:59 PM.