DUST Search - Sticky:[Feedback] Player rank and stats

shade emry3 wrote:

Regis Blackbird wrote:

CCP Rattati wrote:

Currently lifetime, but as Sponk says, a more complex system could have a decay methodology. Your Mu is persistent and updated every time you complete a battle. It is only set at 25 upon your very first battle.

Then this confuses me even more. "Most" (I guess) Dust players are veterans which means their Mu value would have had plenty of time converging to a reasonable value. Then, how come we have the loopsided matches we have today?

Personally I think it's good to get as far away as possible from Win/Loss ratio to determine player rank, since it does not really work with the current Dust mechanics/mentality. In my experience, the motivation to win is not as high as the motivation to preserve ISK. Since the ISK awards are not that different between winning and loosing, players tend to "give up" using high end gear, and/or switch to redline sniping if they see they are getting stomped, giving further disadvantage to the entire team.

I don't think a loss should give less ISK than today, but there should be further benefits of winning even in public matches. If we get that, I think players will push more for the win.

i would have to concur on this :), its a very good view. how would you propose that the winning side gets more? the better question is, do they get more in the form of salvage, mu maybe?


Author	Thread Statistics \| Show CCP posts - 23 post(s)
Regis Blackbird DUST University Ivy League 375	Posted - 2014.08.20 06:35:00 - [1] - Quote CCP Rattati wrote: We actively calculate each players rank based on the result of every match, and we lovingly call that rank "Mu". Before the first battle, the player is given a rank of 25. That rank is updated after the battle, based on the weighted average of the persons duration in the battle and whether he won or lost. Simplified (the calculation is considerably more complex) NewMu=OldMu+"Player seconds in Battle/Battle Duration" * Win/Loss So if you lose, your Mu goes down, and if you win, your Mu goes up. This is a modified version of elo ranking, used in many competitive sports such as chess, major league team sports and esports. Eventually your Mu will converge and stabilize around your "true skill", which is where you will win and lose equally against either players with the same Mu. However, since this is a team game, convergence will happen slower and you may find that even having the best game of your life, will not influence the match enough to secure a win. I will, however, demonstrate that it works very well to predict player skill. The underlying problem is that after one match, everyone is very close to 25, and 2 battles, even 10 battles in, only the very best (and worst) have begun to be different from the pack. That's why it's imperative to find a proxy for Mu for the first battles, which is what comes next in our findings. It unclear to me if the Mu score (current and future implementation) is calculated over the lifetime of the character or per play session? Perhaps you can clarify? I.e Will I start with Mu of 25 every time I log-in to Dust (before my first battle of the day?)
Regis Blackbird DUST University Ivy League 376	Posted - 2014.08.20 08:18:00 - [2] - Quote CCP Rattati wrote: Currently lifetime, but as Sponk says, a more complex system could have a decay methodology. Your Mu is persistent and updated every time you complete a battle. It is only set at 25 upon your very first battle. Then this confuses me even more. "Most" (I guess) Dust players are veterans which means their Mu value would have had plenty of time converging to a reasonable value. Then, how come we have the loopsided matches we have today? Personally I think it's good to get as far away as possible from Win/Loss ratio to determine player rank, since it does not really work with the current Dust mechanics/mentality. In my experience, the motivation to win is not as high as the motivation to preserve ISK. Since the ISK awards are not that different between winning and loosing, players tend to "give up" using high end gear, and/or switch to redline sniping if they see they are getting stomped, giving further disadvantage to the entire team. I don't think a loss should give less ISK than today, but there should be further benefits of winning even in public matches. If we get that, I think players will push more for the win.
Regis Blackbird DUST University Ivy League 376	Posted - 2014.08.20 08:31:00 - [3] - Quote shade emry3 wrote: Regis Blackbird wrote: CCP Rattati wrote: Currently lifetime, but as Sponk says, a more complex system could have a decay methodology. Your Mu is persistent and updated every time you complete a battle. It is only set at 25 upon your very first battle. Then this confuses me even more. "Most" (I guess) Dust players are veterans which means their Mu value would have had plenty of time converging to a reasonable value. Then, how come we have the loopsided matches we have today? Personally I think it's good to get as far away as possible from Win/Loss ratio to determine player rank, since it does not really work with the current Dust mechanics/mentality. In my experience, the motivation to win is not as high as the motivation to preserve ISK. Since the ISK awards are not that different between winning and loosing, players tend to "give up" using high end gear, and/or switch to redline sniping if they see they are getting stomped, giving further disadvantage to the entire team. I don't think a loss should give less ISK than today, but there should be further benefits of winning even in public matches. If we get that, I think players will push more for the win. i would have to concur on this :), its a very good view. how would you propose that the winning side gets more? the better question is, do they get more in the form of salvage, mu maybe? This is a tricky one. - More salvage is nice, but since we can't sell surplus salvage this might get more annoying then useful. - Increasing ISK rewards for the winning side might just fuel more proto-gear for the winning side. - (idea): Increasing the skill point multiplier at the end of game might work, but might also further separate veterans and noobs. However, it think this is a discussion for another thread. I am very exited about Rattati's proposal and want to see how this conversation progresses
Regis Blackbird DUST University Ivy League 376	Posted - 2014.08.20 09:40:00 - [4] - Quote Vell0cet wrote: While my intuition agrees with you, the data appears to indicate that win/loss is as good a metric as any other one. I'm surprised by that, but it's hard to argue against data. I am not so sure. The Mu distribution of players indicate the majority of the players are between 22,5 and 27,5, thus very close to the initial 25. Unless the sample distribution data is taken from only fairly new players, this means the majority of the Dust vets have not strayed far from the initial 25. Thus your Win / Loss ratio over (long) time is almost equal, or I would say almost random. Rattati's comparison were Win/Loss vs K/D, WP/s etc, which should correlate nicely (since usually the higher WP and Kill = Win). In any system where you want to distinguish or separate data (like the matchmaking system), you don't want to see such a major bulk of the players be so close to the initial value. The Mu should be more distributed over the entire graph to really distinguish people. I would like to see the Win / Loss ratio over entire lifetime of one or several sample characters to see how random it is preferably using older characters. If the system works, you should see a player with final Mu of 27 to have a initial period of primarily Win, which levels out and becomes roughly equal. If it don't work, you will see a very random pattern which overall will be close to 50% Win/Loss. Just my thoughts.
Regis Blackbird DUST University Ivy League 379	Posted - 2014.08.20 15:20:00 - [5] - Quote To come back to the proposed matchmaking mechanics using WP/s, I think it's a very good step in the right direction My only worry have already been voiced by others in this thread, which is that this system might give a bias to certain roles, such as the Logi. Since we ARE mercenaries, I wouldn't mind having ISK as a factor in matchmaking. The ability to go ISK positive is a powerful force in players minds, and if one can introduce a factor based on ISK destroyed vs ISK lost would help I think. As hinted before, I suspect the loopy matchmaking is due to the fact that most players seems to reside within the Mu 22,5 - 27,5, which is not a wide enough spread to reflect player skill (Disclamer: statement made without any knowledge how big pool of players CCP Rattati have used for his statistics). My point being; It might be dangerous to try and correlate any new mechanic to this value, as we know it does not work very good to begin with. I would be interested to see how the distribution (of WP/s) would look (recalculated based on the lifetime battles of all players in the statistics, if such data is available) If players are more distributed along the curve, I say we are on the right track.
Regis Blackbird DUST University Ivy League 381	Posted - 2014.08.20 16:23:00 - [6] - Quote CCP Rattati wrote: Sorry, but you are assuming that because matchmaking is not effective currently, then Mu must be inherently wrong. The data says otherwise, KDR and WP ratios verify that Mu correlates with skill, regardless of the matchmaker. Your other point, the distribution around 25, is just the fact that relatively new players outnumber the old players, that's just how the online business is, we have a very long tail, just like EVE, but only a small part of all the new blood retains for years. Thanks for the reply I didn't say Mu was wrong, I said I suspected the distribution might not be enough to make a real distinction between the player skill. I guess I expected a much wider distribution of skill in the player base, but it is next to impossible to make a judgement without the source data (which you have). Please don't take my comments the wrong way, as I (and I think we all) very much appreciate that you address this particular issue Keep on the good work!
Regis Blackbird DUST University Ivy League 381	Posted - 2014.08.20 16:28:00 - [7] - Quote SHANN da MAN wrote: The problem with your current Mu calculation is that you are using Win/Loss (a TEAM metric) to calculate an INDIVIDUAL player's Mu. Players have NO ability to control who is on their team (other than the 6 in their squad, and even those aren't the same all the time) and this affects their Win/Loss ratio which in turn affects their Mu. A better way to calculate Mu would be to use WP/Death (Individual Statistic) in place of Win/Loss (Team statistic). This would more accurately indicate a player's INDIVIDUAL effectiveness without being skewed by a TEAM based metric Well, if Rattati introduce WP/s (which is a individual parameter) to complement or even replace Mu, this would solve that issue, right?