357 replies to this topic

[General Solo]

I think such things are best looked at only when MM brings the skill gap between teams in games with in acceptable margins. Cannot balance on a non level playing field.

Such margins are number of people not playing the game due to match maker. Fullstop.

Thats your leak.

Edit: A lot of effort was than less efficent due to the fact that match maker was not balanced first and other balances done afterward.

An analogy is A lerm boat in skilled hands is a totally different thing than to ,now I am sorry for the vanackula to a lerm boat in non dominant hands.

Likewise the ability to counter lerms with skilled piloting is a totally different thing than to ,now again I am sorry for the vanackula to the ability to counter lerms with non dominant piloting. No AMS or ECM
Terrian masking only.

So if your going to balance fairly you have to judge lerms skilled vs skilled and non dominant vs non dominant. to judge the weapon not the weapon and MM.

TLDR: You have to balance MM first to balance other things more efficient.

Edited by OZHomerOZ, 18 June 2020 - 01:43 AM.

[Nearly Dead]

Tamerlin, on 17 June 2020 - 02:27 PM, said:

Does MWO allow for one team to have a group and not require the other team to have a like-size group? If so, it shouldn't.

Supposedly there aren't enough premade groups at any given time to match them up against each other so some matches you will drop with, or opposing, premade groups of 4, 3 or 2. I have seen one match a few weeks ago where a unit of 4 dropped in alpha with two other unit members on the same team and one poor guy who got stuck on the red team. Sync dropping, I assume.

Aside from splitting the queues again the next best thing they could do would be to only drop a 4 man group if there is another to balance it. The Holy Drop Time precludes it.

I think PGI bought Karl Kasarda's tee shirts for the dev team. The one that has the "We considered fairness, but decided against it" logo.

[Sitzheizung]

Jay Z, on 17 June 2020 - 07:00 AM, said:

Each and every solution presented will cause the entire playerbase to continually spread out until T3 is less populated and pilots settle towards T1 and T5.

Sjorpha, on 17 June 2020 - 07:27 AM, said:

I appreciate that you are trying to solve the divergence issue, but it seems to me that a psr formula that needs periodical adjustment of this kind has some fundamental issues.

Addressing this divergence issue...

Disclaimer: given numbers are fictional picked for ease explanation.

TLDR:

PSR ranking is capped at some "0" Tier 5 skill at bottom and "5000" Tier 1 at top. the worst and best players there cannot drop or rise any further and will push other players a bit more towards the middle of the ranking ladder, therefore Tier 3 will not be empty i belive.
Conclusion: periodically resets or adjusts are not necessary.

Oh, i`m bored, lets spend my time on some unnecessary wall of text:

lets say we have 2 palyers sitting on each edge of the ranking ladder:

- Player "doublelefthanded" (unfortunately he has no mouse, no controller, no pad, but only a keyboard to play)
- Player "Over9000" (top of the notch, he winns every game and always has best match score)

so Mr. Over9000 is on Rank 1, PSR 5000, and he is the only one who ever gets there, no one can compare to him.
He just did win a game, top matchscore, and is rewarded with +25 PSR points, which will be added to his account, bringing him up to 50??.... what? still 5000???
Yes, still 5000. there is a cap. which means that the game he did, not adds up to a Zero Sum for the Ranking. well, the game itself does, but not if you addup the points to the ranks...
Lets say there was a total of +100 PSR and a total of -100 PSR in the game which is the requested Zero Sum. it turns out to be a +75 and a -100 to the ranking. and dont get me wrong, this is actually good. because it leads to a push of slightly lower ranked players downwards the ladder. and this push will go through to the ones that will stay in Tier 3, just for that reason..

the same thing occures for our Mr doublelefthanded. He is at 0 PSR, last of the last, inhabiting the Shadowrealm of the PSR ranking ladder. in his case the points will be +100 and -75 to the ranking, because below 0 is not possible.
He is pushing slightly better players a bit upwards, and this going through until hitting Tier 3...

i am no scientist, engineer or game developer, so please correct me if i am wrong and my conclusions are incorrect.

Edited by Keldomet, 18 June 2020 - 03:02 AM.

[VileFalcon]

There is a lot to take in on this thread and sadly I don't have time to read all of it.
I realise that there is a move to implement one of the Core 1 options. I would like to add however that my preference would be to add emphasis to team work. This would be a vote for one of the Core 2 options.

[Jay Z]

Keldomet, on 18 June 2020 - 02:49 AM, said:

https://mwomercs.com...71#entry6338671

Here ya go mate. Have a look there.

[RRAMIREZ]

I would have backed models promoting win over "individual perf" (core 2*), if the queues weren't merged.
With that in the background, I really can't figure out how it's gonna go. I just know that I'll continue to drop solo.
So I won't vote and see if games will be more enjoyable (for me) or not.

Edited by RRAMIREZ, 18 June 2020 - 06:04 AM.

[Xiphias]

Tarl Cabot, on 17 June 2020 - 03:51 PM, said:

Isn't that making an assumption that the player would almost always be facing players who are hitting 300+ MS? He had peaked for about 1500 games, exceeding 200 -231 MS for several months (likely tier 3 ) and is likely in Tier 1 now, where he is facing a much tougher crowd.

No, as long as the player is playing against players who are on average scoring higher than they are they will move down. Take the extreme case for simplicity (e.g. top player moves up, bottom player moves down).

Let's say the community AMS is 250 and a player has an AMS of 249. Now that player doesn't always score 249, sometimes they move up and sometimes they move down. Let's say out of 100 matches, this player scores in the middle and doesn't move 97 times. Occasionally, they score really well and end up in the top 1 out of 100 times. However, sometimes they do poorly and end up in the bottom, this happens 2 out of 100 times. So over the course of 100 matches the trend is moving down by 1. If they play 1000 matches they will move down by 10 etc. Eventually they get to the bottom.

If a player with AMS of 251 plays against a bunch of players with AMS of 249, eventually that player will move up and all those other players will move down.

If T3 is filled with 249 and 251 AMS players, half those players will move up, half those players move down.

Think of it this way. If T3 is filled with players with an AMS score of exactly 250 and those players are completely even (they move up/down equally and stay where they are). Now a player with an AMS of 249 gets into a match with 23 players with an AMS of 250. They might go up some, but they will on average go down move than they go up (otherwise their AMS would be higher than 250). Similarly, a player with an AMS of 251 will on average go up. What that means is that T4 is now filled with all the 249 AMS players who got pushed out of T3, and T2 is filled with all the 251 AMS players.

Now a 248 AMS player moves from T5 to T4 and gets matched against 23 players with an AMS of 249. This player on average scores the lowest, so they move back down into T5. Anyone below them is locked into T5. The opposite happens in T2. Players with AMS of 252 on average move up when matched against players with AMS of 251, so they move up into T1 and anyone with a higher AMS stays in T1.

The result is that any system that rewards based on matchscore ranking ends up diverging and stratifying the player base into two main groups. Obviously, in reality the movements aren't quite as clean, take longer, end end up with more noise in the middle, but you still end up with two primary groups every time once enough matches are played.

[Xiphias]

Jay Z, on 17 June 2020 - 04:38 PM, said:

Every single option will of course cause uncontrolled divergence as there is no reference to global or stored stats.

As far as I understand it, we are limited to individual matches with Matchscore as an input and Player PSR Shift as an output.

I think a W/L system (not a WLR based MM) will only results in a 50% divergence (based on the sims that I've run). Basically, anyone with a WLR < 1 will diverge into T5. There's not really a way around that. However, the players with a WLR > 1 will end up separating out between the remaining tiers.

So you end up with the upper half of the community pretty well separated by skill, but the bottom half is the same as it is now. I'm not sure if this is an improvement over what we have right now, but I think it's better than a new system that diverges into two large buckets (matchscore based ranking).

[Horseman]

Keldomet, on 18 June 2020 - 02:49 AM, said:

PSR ranking is capped at some "0" Tier 5 skill at bottom and "5000" Tier 1 at top. the worst and best players there cannot drop or rise any further and will push other players a bit more towards the middle of the ranking ladder, therefore Tier 3 will not be empty i belive.
Conclusion: periodically resets or adjusts are not necessary.

On the contrary, the fact it's capped means the system is bound to lose accuracy over time and that periodic readjusts are necessary to keep things on a reasonable scale. Otherwise you will again eventually end up with a ton of players who are on one or the other extreme of the scale while not being equivalent to one another

Edited by Horseman, 18 June 2020 - 07:14 AM.

[Nightbird]

Horseman, on 18 June 2020 - 07:09 AM, said:

On the contrary, the periodic readjusts are necessary to keep things on a reasonable scale. otherwise you will again eventually end up with a ton of players who are on one or the other extreme and the system will lose accuracy over time!

On the contrary, the math says otherwise. If you're referring to Xiphas graph, the nice linear relationship before hitting the caps is an illusion, and it will take a lot of work to make it look right.

If there was an accuracy graph, it starts at 0%, and slowly moves up to 5%. Resetting it moves it down to 2.5% again.

[DevinMace]

I think groups are only an issue when it is the top % of players As someone who is in the bottom tiers when I see a group I expect to win. Only a few groups are fairly strong that have beaten my solo sides consistently, Atleast that I have noticed.

[wasder unguided]

I would vote 2a, with 2b my second choice. Qualified as follows:

A roadmap which would attract more players to the game, which could make all game modes viable (Quick Play, Faction Play, Competitive & Solaris) and allow us to cater to all types of players... well I'm in favour of anything that leads to that.

I guess that ship has sailed though. So what are we doing here? What is the objective? If it is to prevent stomps... well is that even desirable? I'm a bang average player and my win/loss ration from the very beginning is around 50/50 (I enjoy the stomps I win and can handle the stomps I lose). Do we know how many players have a seriously negative win/loss ratio? Is it those we are trying to help?

Sorry I don't normally participate in the forums so if this is just ******** I apologise.

[wasder unguided]

DevinMace, on 18 June 2020 - 09:07 AM, said:

I think groups are only an issue when it is the top % of players As someone who is in the bottom tiers when I see a group I expect to win. Only a few groups are fairly strong that have beaten my solo sides consistently, Atleast that I have noticed.

As someone who plays in a group I can confirm we routinely get a good thrashing. Lose 50% of the time with a good proportion of that are stomps as mentioned above. We don't coordinate with each other let alone the rest of the team

P.S. I'm tier 3 and happy to remain there for all time, I have fun.

P.P.S I blame my ping from India mostly.

[DevinMace]

A lot of new teams, with new cadets having joined them as well.

[Blechreiz]

What people seem to forget is that, for the more or less average player, the likelyhood of them managing to achieve a high or low matchscore will change depending on their current tier. Meaning, the lower the tier they're in ,the higher the likelihood of achieving a high matchscore and vice versa. Once they reach the "appropriate" for their personal skill-level, their matchscore should be, more or less, in equilibrium. So, without them changing anything, they will remain in the same tier.

This is something that was quite noticeable when I started a 2nd account: With each tier I advanced, my average matchscore dropped on average by about 50 points. Assuming that trend continued, i would, under the current system, end up with the same average matchscore as I got on the main account.

Hence, imo, it's unlikely that we will end up in a situation where there are only people in T1 and T5.

[Xiphias]

Blechreiz, on 18 June 2020 - 12:30 PM, said:

What people seem to forget is that, for the more or less average player, the likelyhood of them managing to achieve a high or low matchscore will change depending on their current tier. Meaning, the lower the tier they're in ,the higher the likelihood of achieving a high matchscore and vice versa. Once they reach the "appropriate" for their personal skill-level, their matchscore should be, more or less, in equilibrium. So, without them changing anything, they will remain in the same tier.

This is something that was quite noticeable when I started a 2nd account: With each tier I advanced, my average matchscore dropped on average by about 50 points. Assuming that trend continued, i would, under the current system, end up with the same average matchscore as I got on the main account.

Hence, imo, it's unlikely that we will end up in a situation where there are only people in T1 and T5.

It's not a player's actual matchscore that matters though. It's a player's matchscore compared to other players. Yes, a T5 player will get less matchscore (on average) playing against a bunch of T1s than if they played against a bunch of T4s, but they will still (on average) score lower than the T4s and as a result will move down. It doesn't matter if you get 249 vs 251 pr 349 vs 351, the end result is the same.

If all players were being compared to global AMS values for the community and went up/down based on that you would be correct that players would reach tier equilibrium. With a relative matchscore comparison in matches, players who are below the average skill for the game will eventually get shifted down and players above it will eventually get shifted up. If everyone increases no one does in a relative matchscore system.

[ESC 907]

IMHO, it seems to me that rating PSR solely on W/L would be a very macro solution. Certainly, it should cause players on-average to reach a point where they are playing with people that win & lose at an equal rate. Start losing a bunch, you go down and face easier opposition and begin winning. Start winning a bunch, you go up and eventually face tougher opposition and begin losing. Equilibrium should be reached eventually.
To rate scores and measure PSR on something other than W/L (like MS), the components for calculating the MS need to be figured out properly. And it seems a bit more of a micro-solution to me. Ideally, someone doing well in matches should go up until they are an average player in their matches, and someone doing poorly should go down until they are also an average player in those matches.
Either solution should work fine, so long as it is balanced properly.

[Nightbird]

ESC 907, on 18 June 2020 - 05:09 PM, said:

IMHO, it seems to me that rating PSR solely on W/L would be a very macro solution. Certainly, it should cause players on-average to reach a point where they are playing with people that win & lose at an equal rate. Start losing a bunch, you go down and face easier opposition and begin winning. Start winning a bunch, you go up and eventually face tougher opposition and begin losing. Equilibrium should be reached eventually.
To rate scores and measure PSR on something other than W/L (like MS), the components for calculating the MS need to be figured out properly. And it seems a bit more of a micro-solution to me. Ideally, someone doing well in matches should go up until they are an average player in their matches, and someone doing poorly should go down until they are also an average player in those matches.
Either solution should work fine, so long as it is balanced properly.

Good thinking, it's all about the strength of the convergence. Let's just call the force that keeps people in the right Tier 'gravity'. If what you want is the Earth's gravity - jump all you want and it immediately pulls you back into the right position, the gravity of the options in the OP is the international space station's. One jump and you're off to the abyss. The ISS has gravity for sure... just not enough.

Edited by Nightbird, 18 June 2020 - 05:23 PM.

[TheCaptainJZ]

I support 1A. With the right MS Kicker values, most of the time, almost all the winners will be at the top and all the losers will be at the bottom, but it allows for DC's or slackers, on the winning side to move down slightly while exceptional losers rise slightly. If you win, usually you'll gain a higher match score, naturally. But we can also boost actions that contribute to this. A win is an event that can be rewarded a MS kicker, but so are kills and assists. If it's a blowout match, there won't be many kills on the losing side or assists. If it's a very close match, everyone will be a lot closer together, in theory. But the winners will still be rewarded more because of the MS kicker for a win.
Just throwing random numbers out there, but someone on the losing side that had 5 kills and 1000 damage is a better player than someone on the winning side with no kills, 2 assists, and 200 damage and should move higher in PSR (depending on other MS kickers). When teams are random and uncoordinated, you're really just playing for yourself. Of course team play will net more wins and more score generally.
The only disadvantage to this is that in a match of all good players, half will will have to move down whereas in a match of all bad players, half will have to move up. But after a number of matches and playing against other players, this should wash out to something acceptable. If you are always a bad player, you will only move up a little at most or down a lot, until you run into players just as bad when you might start outranking them and moving back up a little or fluctuating within a tier. Same for good players--they will likely just trade back and forth at that higher tier. If they consistently under perform their peers, they'll move down in PSR until they start outperforming their peers at the lower level and move up again.

I don't have time to theorycraft numbers but let's say 1 pt for every 10 dmg. 500 dmg = 50 MS. I'd say a kill is worth about 200 dmg and 3 or 4 assists should equal a kill. Winning can be worth 2 kills.
So just ballpark figures:
Win-400 pts
Kill-200 pts
Assist-50 pts
Dmg-0.1 per dmg
A player who won with 500 damage and 2 kills and 3 assists would be 800 pts. Multiply that by some value, let's say 0.5 for easy math. MS would be 400. Compare that against the other players. (Today such a player might be 300-400 MS in the end.)
Same stats except on the losing team would be 400 pts or 200 MS. Multiply that by .5 for a total of 100. The loser would need to have made 2 more kills or 4 more assists to be equal.

Those numbers are way unbalanced because I simply made them up, but you can catch my drift.

[Valdorel]

Definitely voting 2B with Jay Z's modifications.