61 replies to this topic

[Maddermax]

Just a thought, but why not use Average kills per match as a main statistic rather than Kills/Death Ratio? In other shooters with respawns, KDR makes more sense as it denotes how often you have to respawn to get your kills. In a game like MWO though it makes much less sense, as it doesn’t matter if your KDR is 5 or 15 if you’re getting it by not engaging early and then hiding at the end of a match, rather than pushing for victory.

If Average Kills per Match were used instead, it actually draws a line between those who are getting KDR through actually bringing enemies down often, and those who do it by simply avoiding the fight, picking off DCs and wounded stragglers, and then hiding.

what do you think?

[Prototelis]

Divide number of matches played by total number of kills.

[Maddermax]

Prototelis, on 03 April 2019 - 09:15 PM, said:

Divide number of matches played by total number of kills.

Exactly, used as a headline stat.

[Curccu]

And KMDD could be included also....

[Khobai]

kills per match is a meaningless statistic

furthermore it encourages negative behaviors like kill stealing since the game stupidly doesnt make provisions for percentile kills (i.e. if you kill 95% of a mech you get credited with 0.95 kills even if someone else does the last 5% of the damage and steals the kill).

right now the only statistic that actually matters is win/loss ratio

Edited by Khobai, 03 April 2019 - 09:30 PM.

[Snowhawk]

I like to see that that the matchmaker uses average matchscore instead of the outdated tier-System....

[Maddermax]

Curccu, on 03 April 2019 - 09:25 PM, said:

And KMDD could be included also....

I'd like to see KMDDs be recorded as part of the stats too. This thread is more about replacing KDR as a headline stat, rather than recording new information.

Khobai, on 03 April 2019 - 09:27 PM, said:

kills per match is a meaningless statistic

furthermore it encourages negative behaviors like kill stealing since the game stupidly doesnt make provisions for percentile kills (i.e. if you kill 95% of a mech you get credited with 0.95 kills even if someone else does the last 5% of the damage and steals the kill).

right now the only statistic that actually matters is win/loss ratio

Win/Loss is the most important statistic, but also one of the most random at smaller sample sizes. An exactly average player who plays 100 games in a month will have between a 0.66 and 1.5 win/loss ratio (95% CI). The more matches you play, the more accurate the stat, but for most players it's fairly inaccurate on a month to month basis.

While Kills/Match isn't perfect, it's a lot better than KDR, which doesn't just prioritize killing, but also survival, even at the expense of your team. A person who is hiding and trying to jump out and grab kills will probably still have a lower Kills/match than someone who's going out and getting 95% damage on mechs regularly. Because currently KDR means that 2 kills/1 death in 10 matches looks the same as 20 kills over 10 matches if you died every time. The difference to the team is *massive* though, and that's what we should encourage in terms of stats.

[Prototelis]

Maddermax, on 03 April 2019 - 09:20 PM, said:

Exactly, used as a headline stat.

Exactly. Its irrelevant because it can be derived from the published statistics...

[Tiewolf]

First focusing on stats encourage negative player behaviour like hiding, killstealing, share no amor, seal clubbing, e.g..

Second they say next to nothing!
What mechs are you playing? Using meta only and spamming consumables? At which times are you playing EU or US? QP or GQ? With whom are you dropping in GQ? And so many variables more to influence your stats beyond your skills.


Use only tournament data (maybe compQ) for stats and get rid of the rest!
The ones that like to measure their epeen should do it in an tournament environment where everybody tries to play at the possible limits of the game. All other stats are useless anyway if you don`t know the environment in which they are achieved.

Edited by Tiewolf, 03 April 2019 - 11:31 PM.

[Maddermax]

Prototelis, on 03 April 2019 - 11:17 PM, said:

Exactly. Its irrelevant because it can be derived from the published statistics...

So can KDR, so it's an odd point to make. Replace KDR with average kills per match as a headline figure is what we're talking about.

[Maddermax]

Tiewolf, on 03 April 2019 - 11:17 PM, said:

First focusing on stats encourage negative player behaviour like hiding, killstealing, share no amor, seal clubbing, e.g..

Second they say next to nothing!
What mechs are you playing? Using meta only and spamming consumables? At which times are you playing EU or US? QP or GQ? With whom are you dropping in GQ? And so many variables more to influence your stats beyond your skills.


Use only tournament data for stats and get rid of the rest!
The ones that like to measure their epeen should do it in an tournament environment where everybody tries to play at the possible limits of the game. All other stats are useless anyway if you don`t know the environment in which they are achieved.

You may be right, and yet stats is brought up all the time, on the forums and in game. Some people like to know their stats, have something to work towards or as an indicator that they're doing well or at least improving. It's unhealthy when stats drive bad behavior though, and we see that in game at the moment, with people fleeing from combat or taking passive roles to pad their own stats at the expense of team chances.

What I'm saying is that if stats are to be had, at least make them more meaningful and less likely to drive bad behaviors.

[Anjian]

Maddermax, on 03 April 2019 - 09:09 PM, said:

Just a thought, but why not use Average kills per match as a main statistic rather than Kills/Death Ratio? In other shooters with respawns, KDR makes more sense as it denotes how often you have to respawn to get your kills. In a game like MWO though it makes much less sense, as it doesn’t matter if your KDR is 5 or 15 if you’re getting it by not engaging early and then hiding at the end of a match, rather than pushing for victory.

If Average Kills per Match were used instead, it actually draws a line between those who are getting KDR through actually bringing enemies down often, and those who do it by simply avoiding the fight, picking off DCs and wounded stragglers, and then hiding.

what do you think?

Kill to match ratio is better on Faction Play modes, because people will knowingly eject from damaged mechs for new ones in preparation for a new wave.

Average Kills per Match ratio is also what War Robots uses, because of that game's format of objective based game modes with respawns. Again, its for the same reason: people will knowingly eject from heavily damaged mechs for fresh ones off the drop deck, and its part of match strategy. The game also exhibits average damage per match which is the most important stat featured. These averages are all based on the last 50 games the player has played.


Not sure if KPM is the best stat for MWO's QP modes, because many kills are killing blows, but I like to see damage per match ratios instead. Having an average cut off like the last 50 or 100 games can help you determine your performance curve over a period.

Edited by Anjian, 03 April 2019 - 11:58 PM.

[Nesutizale]

Khobai, on 03 April 2019 - 09:27 PM, said:

kills per match is a meaningless statistic

furthermore it encourages negative behaviors like kill stealing since the game stupidly doesnt make provisions for percentile kills (i.e. if you kill 95% of a mech you get credited with 0.95 kills even if someone else does the last 5% of the damage and steals the kill).

right now the only statistic that actually matters is win/loss ratio

We allready have kill most damage dealt so somehow its allready counted who does the most damage for a mech. It should be possible to use that data to get the numbers you discribe.

Basicly both statistics should be fine, if you have enough matchers both values W/L K/M should balance out over time.
W/L shows more how well you play with the team as good teamwork should result in more wins for the team.
K/M is more a reflection of your personal skills or better said how often you manage to get the kill.

So the question is more, what is more of interest to you. Do you want to know if someone is more likely a good teamplayer, what I would prefere since, except for Solaris, I allways have people around me or do you want to see if you have more kills on average then the guy next to you...kinda bragging rights.

[PhoenixFire55]

Maddermax, on 03 April 2019 - 09:09 PM, said:

Just a thought, but why not use Average kills per match as a main statistic rather than Kills/Death Ratio?

And since when is KDR "a main statistic"? ... W/L is all that matters.

[Maddermax]

PhoenixFire55, on 03 April 2019 - 11:53 PM, said:

And since when is KDR "a main statistic"? ... W/L is all that matters.

Profile>Leaderboards, look along the top row. There are three derivative stats - Win/Loss, Average score and KDR. So, since whenever it was put there, and people started playing weirdly to get better at that stat.

I'm sure some people would like to see stats removed entirely, but if they're there, may as well use a better one.

Edited by Maddermax, 04 April 2019 - 12:04 AM.

[PhoenixFire55]

Maddermax, on 04 April 2019 - 12:03 AM, said:

Profile>Leaderboards, look along the top row. There are three derivative stats - Win/Loss, Average score and KDR. So, since whenever it was put there, and people started playing weirdly to get better at that stat.

I'm sure some people would like to see stats removed entirely, but if they're there, may as well use a better one.

Trust me, nothing changed. The 0.1% of population that is hiding and farming random kills in the end were hiding and farming random kills in the end way before any stats were available for the public.

You should never look at any stat by itself, but rather the combination. When a player has high KDR, avg match score and most importantly W/L at the same time, it means he is good. You can't actively farm all three at the same time if you aren't simply good. In turn it means if just one of the three is decent, its just farming that particular stat.

Stats are necessery. But they should be easily accessible while in a match, i.e. a new player clicks on a guy who makes calls on VoIP, sees that his W/L is way above 1.0 and follows his calls. Same way if he sees its a sub 0.5 W/L player he can ignore him, etc.

Public stats don't change anything. Cancerous behavior was present in the game since closed beta, had a huge influx when game went open and remains at about the same % since then. E-peen and whatnot is likewise a part of online gaming, and was present in MWO way before stats were introduced.

[Vellron2005]

Maddermax, on 03 April 2019 - 09:09 PM, said:

Just a thought, but why not use Average kills per match as a main statistic rather than Kills/Death Ratio? In other shooters with respawns, KDR makes more sense as it denotes how often you have to respawn to get your kills. In a game like MWO though it makes much less sense, as it doesn’t matter if your KDR is 5 or 15 if you’re getting it by not engaging early and then hiding at the end of a match, rather than pushing for victory.

If Average Kills per Match were used instead, it actually draws a line between those who are getting KDR through actually bringing enemies down often, and those who do it by simply avoiding the fight, picking off DCs and wounded stragglers, and then hiding.

what do you think?

Both are useless mechanics, since neither is an accurate measure of skill.. because there's many other factors involved, luck, map, mech chassis, build, group you're stuck with, and many others..

It's just e-peen stroking..

I seriously pity anyone who thinks such stats are important..

[Nesutizale]

These stats might be usefull in Solaris where you can directly compare one person to another but as soon as you have lots of other peoples who influance the outcome of the game...they become more watered down.
You can have a good team, a bad team, map or mode voteing dosn't go or goes in your favor someone can block you and you die no matter how good you are....stuff like that.

[The6thMessenger]

Vellron2005, on 04 April 2019 - 12:22 AM, said:

Both are useless mechanics, since neither is an accurate measure of skill.. because there's many other factors involved, luck, map, mech chassis, build, group you're stuck with, and many others..

It's just e-peen stroking..

I seriously pity anyone who thinks such stats are important..

It's important, with context.

If you're just having fun, then don't bother.
If you want to better yourself, it's telling you whether you are doing something wrong or right.

If you want to engage in discussions about balance, it's quite important, because it's a look at the current overall balance.

[PhoenixFire55]

Nesutizale, on 04 April 2019 - 12:44 AM, said:

These stats might be usefull in Solaris where you can directly compare one person to another but as soon as you have lots of other peoples who influance the outcome of the game...they become more watered down.
You can have a good team, a bad team, map or mode voteing dosn't go or goes in your favor someone can block you and you die no matter how good you are....stuff like that.

All stuff like that averages itself out ... for every good team you'll have a bad team, for every block you had you'll have a player who saves your butt and so on ... After about 200-300 matches your stats represent your contribution and your contribution alone.