139 replies to this topic

[Xavori]

MischiefSC, on 26 November 2017 - 02:42 PM, said:

As has been pointed out in detail repeatedly this is absolutely false. Here is a link to the actual stats and data analysis Nightbird did showing- unequivocally- that win/loss is the best indicator of ability to win matches.

There is absolutely, demonstratively and unquestionably no better metric for predicting win/loss than your win/loss. Any other data you use will only reduce accuracy.

In 12 v 12 or 100 v 100 or 1 v 1 there is, mathematically, no other useful or reliable metric for identifying a players impact on winning. Why this particular piece of misinformation keeps coming up I don't know but it's absolutely false.

https://mwomercs.com...ost__p__5952780

So. Many. Wrong. Assumptions.

What Nightbird showed was not that there is no better metric, what he showed is that we have no better metric available to us. I'd bet there are many better metrics. For example, I'd suspect that people who play mostly in units do better than people who pug drop. I'd suspect that people who run nothing but top comp build mechs do better than people who play a variety of mechs. Hell, I'd suspect that people who stay the hell away from PGI's events do better at W/L than people who try to game the system to get the shinies as quickly as possible.

I don't want to derail this thread (too much), but in a highly interdependent team game like MWO, win loss is a terrible metric since it hides what actually leads to success. Sadly, you are correct in that it's the closest thing we have to one right now tho.

[mistlynx4life]

Sorry I mispoke. I wasn't talking about a player's chances of success or something, merely damage as a contributing factor within any given match. All blind matches are easier if you're banking on at least a little damage. That's all. It's rarely not useful, unlike running away and avoiding combat or something. Sorry. My bad.

[MischiefSC]

Xavori, on 26 November 2017 - 03:10 PM, said:

So. Many. Wrong. Assumptions.

What Nightbird showed was not that there is no better metric, what he showed is that we have no better metric available to us. I'd bet there are many better metrics. For example, I'd suspect that people who play mostly in units do better than people who pug drop. I'd suspect that people who run nothing but top comp build mechs do better than people who play a variety of mechs. Hell, I'd suspect that people who stay the hell away from PGI's events do better at W/L than people who try to game the system to get the shinies as quickly as possible.

I don't want to derail this thread (too much), but in a highly interdependent team game like MWO, win loss is a terrible metric since it hides what actually leads to success. Sadly, you are correct in that it's the closest thing we have to one right now tho.

No. In the other thread we went over this as well. In fact it ended with someone posting you the link to Microsoft's TrueSkill system which, again, uses win/loss in team v team at it's core and is the most comprehensive matchmaker ever made.

There is no derailing to happen - this is basic, fundamental statistics and analytics. Basic. Team v team, 1 v 1, doesn't matter. 1 v 1, 1 billion v 1 billion, there is absolutely no metric of any sort in any circumstance better for calculating a customers impact on winning than their w/l. If you want to assert that's not the case I need you to provide the algorithm and formula showing so. Then you need to explain to Microsoft that they clearly don't understand math and Trueskill and the huge investment they made is worthless. Then you need to get all those pesky textbooks updated.

This argument is literally flat earth. Opinions on it, how it "feels", means **** all nothing. Statistical analysis and how to solve for a value in an equation is already a thing. There is no argument to have.

[Xavori]

MischiefSC, on 26 November 2017 - 03:35 PM, said:

No. In the other thread we went over this as well. In fact it ended with someone posting you the link to Microsoft's TrueSkill system which, again, uses win/loss in team v team at it's core and is the most comprehensive matchmaker ever made.

There is no derailing to happen - this is basic, fundamental statistics and analytics. Basic. Team v team, 1 v 1, doesn't matter. 1 v 1, 1 billion v 1 billion, there is absolutely no metric of any sort in any circumstance better for calculating a customers impact on winning than their w/l. If you want to assert that's not the case I need you to provide the algorithm and formula showing so. Then you need to explain to Microsoft that they clearly don't understand math and Trueskill and the huge investment they made is worthless. Then you need to get all those pesky textbooks updated.

This argument is literally flat earth. Opinions on it, how it "feels", means **** all nothing. Statistical analysis and how to solve for a value in an equation is already a thing. There is no argument to have.

1. TrueSkill has nothing to do with interdependent team based skill
2. Microsoft wastes tons of money on failed projects all the time, so not really a good example anyway.

By your logic, Barry Sanders sucked as a running back because his W/L record was below 500 and he never won a Superbowl.

In MWO, a player with a good W/L record might be a terrible player who just limits to playing with his good teammates. Not sure why you seem to have a problem understanding that...

[MischiefSC]

Xavori, on 26 November 2017 - 04:32 PM, said:

1. TrueSkill has nothing to do with interdependent team based skill
2. Microsoft wastes tons of money on failed projects all the time, so not really a good example anyway.

By your logic, Barry Sanders sucked as a running back because his W/L record was below 500 and he never won a Superbowl.

In MWO, a player with a good W/L record might be a terrible player who just limits to playing with his good teammates. Not sure why you seem to have a problem understanding that...

That's exactly what TrueSkill. It creates a projection and confidence in prediction metric foe cross platform sol and team v team performance. Your assumption that you understand analytics and statistics better than the team that developed TrueSkill (which is regarded worldwide as the best matchmaker ever designed, bordering on overkill) will need some support.

Please provide the math to support your position and why all current predictive analytical models for resolving for a factor in random sampling are false.
Equation please.

As said a million times before, group and solo queue needs separated.

Edited by MischiefSC, 26 November 2017 - 04:53 PM.

[Xavori]

MischiefSC, on 26 November 2017 - 04:52 PM, said:

That's exactly what TrueSkill. It creates a projection and confidence in prediction metric foe cross platform sol and team v team performance. Your assumption that you understand analytics and statistics better than the team that developed TrueSkill (which is regarded worldwide as the best matchmaker ever designed, bordering on overkill) will need some support.

Please provide the math to support your position and why all current predictive analytical models for resolving for a factor in random sampling are false.
Equation please.

As said a million times before, group and solo queue needs separated.

*sigh*

I provide actual, real world examples that prove, beyond any shadow of a doubt, that your assumptions about W/L are wrong, and you ask me for math. I don't need math when I have reality.

The reality is that in every competitive team sport on Earth, you do not evaluate players based on the team's win/loss record. Want to know if an MLB pitcher is any good? You look at ERA, hitting percentage, walks per inning, HR's allowed, etc. Want to know if a QB is any good? You look at completion percentage, yards per attempt, TD/Int ration, and so on. Want to know if a goalkeeper is any good? You look at goals allowed per minute played, saves, etc. You know what you don't look at in any of these cases? W/L. Because W/L evaluates a team, not the individual, and even then, W/L is only part of the answer.

So no, I cannot give you any math because MWO doesn't give me any statistics that are meaningful. I can, and did, manipulate my damage numbers (because event). I can boost my K/D ratios just by playing heavies more than lights, and not ever playing mechs that I haven't skilled or that I know aren't top tier. I can utterly, beyond any shadow of a doubt, make a mockery of every leaderboard statistic we have available which means you can look at my stats and know exactly nothing about my skill as a mech pilot in MWO. And I'm not a unique snowflake nor anything special. There are lots of good MWO players just like me.

[Mystere]

Khobai, on 26 November 2017 - 02:10 PM, said:

if its 1v1 I completely agree winning is the best metric.

but if its 12v12 its not the best metric. because you can literally do nothing and still win a game. you can just play with a group of really good players and let them carry you for example. or you can do extremely well and get the most kills and damage and still lose a game for reasons beyond your control. its not a good measure of individual prowess.

match score should be the metric used for increasing or decreasing player ranking. matchscore just needs a few adjustments so its not as easy to manipulate. but its entirely doable.

and since winning adds to your match score, winning is still a factor. just not the only factor.

A player who consistently wins more than loses must be doing something right -- even if it's just being the team mascot.

[Mystere]

MischiefSC, on 26 November 2017 - 03:01 PM, said:

However that's all pointless if there's no risk of getting attacked.

The point is that you can do that in scouting currently - while the other team tries to kill you. So up rewards for running objectives however realize that if the mode becomes more about chasing squirrels and avoiding combat then most the people currently playing will stop.

History shows that players did not stop. Instead, they threatened to stop playing -- by loud whining and incessant crying on the forums -- unless PGI did something to stop people from going for objectives instead of just running to map center to only play "Rock 'em, Sockl ;em Robots". Why else were turrets added in Assault, generators given invincible shields in Invasion, cap times lengthened in both Assault and Conquest, and rewards for doing anything other than killing made into a pittance?

[Mystere]

Xavori, on 26 November 2017 - 05:08 PM, said:

*sigh*

I provide actual, real world examples that prove, beyond any shadow of a doubt, that your assumptions about W/L are wrong, and you ask me for math. I don't need math when I have reality.

The reality is that in every competitive team sport on Earth, you do not evaluate players based on the team's win/loss record. Want to know if an MLB pitcher is any good? You look at ERA, hitting percentage, walks per inning, HR's allowed, etc. Want to know if a QB is any good? You look at completion percentage, yards per attempt, TD/Int ration, and so on. Want to know if a goalkeeper is any good? You look at goals allowed per minute played, saves, etc. You know what you don't look at in any of these cases? W/L. Because W/L evaluates a team, not the individual, and even then, W/L is only part of the answer.

So no, I cannot give you any math because MWO doesn't give me any statistics that are meaningful. I can, and did, manipulate my damage numbers (because event). I can boost my K/D ratios just by playing heavies more than lights, and not ever playing mechs that I haven't skilled or that I know aren't top tier. I can utterly, beyond any shadow of a doubt, make a mockery of every leaderboard statistic we have available which means you can look at my stats and know exactly nothing about my skill as a mech pilot in MWO. And I'm not a unique snowflake nor anything special. There are lots of good MWO players just like me.

You are mistaking measuring a player's pitching ability with measuring the same player's ability to win. You hinted it yourself, Barry Sanders sucked at winning even if he was a great running back.

How do you measure command ability, teamwork and coordination, distraction, baiting the enemy to draw a few away from the main fight, being a vanguard, sacrificing one's self for the win?

[Xavori]

Mystere, on 26 November 2017 - 05:42 PM, said:

You are mistaking measuring a player's pitching ability with measuring the same player's ability to win. You hinted it yourself, Barry Sanders sucked at winning even if he was a great running back.

How do you measure command ability, teamwork and coordination, distraction, baiting the enemy to draw a few away from the main fight, being a vanguard, sacrificing one's self for the win?

You're kidding, right? Barry Sanders sucked at winning?

No. Barry Sanders is the only reason the Lions had any success in the 10 years he played. Short of switching teams (can you imagine him behind the Dallas Cowboys o-line at that time?), he was doing everything he could to help his team. In addition to his amazeballs rushing ability, he was also a good receiver and could pick up the blitz. You literally could not ask anything more of him (except not to retire when he was still in peak physical condition, but given what we know of the toll that football takes on players, who could possibly fault him)

Barry Sanders was only 1/22nd of his starting lineup, and only 1/53rd of the players on the team. He was also not a coach or GM which makes his ability to influence the entirety of a team's success limited. Now, he did everything he could with that limited ability to skew things, but the Lions still lost more than they won. That's not on him at all.

In MWO, players don't even have that much control. We have a broken matchmaker using a completely flawed stat in a much too small playerbase creating teams that have huge impact on a player's stats. And yet you keep insisting that there is some meaning to be had in those stats...

[Mystere]

Xavori, on 26 November 2017 - 06:05 PM, said:

In MWO, players don't even have that much control. We have a broken matchmaker using a completely flawed stat in a much too small playerbase creating teams that have huge impact on a player's stats. And yet you keep insisting that there is some meaning to be had in those stats...

As people have been telling you, the only stat that actually has any meaning is wins.

How you deduced that all other stats have any meaning to me makes we wonder.

[MischiefSC]

Xavori, on 26 November 2017 - 05:08 PM, said:

*sigh*

I provide actual, real world examples that prove, beyond any shadow of a doubt, that your assumptions about W/L are wrong, and you ask me for math. I don't need math when I have reality.

The reality is that in every competitive team sport on Earth, you do not evaluate players based on the team's win/loss record. Want to know if an MLB pitcher is any good? You look at ERA, hitting percentage, walks per inning, HR's allowed, etc. Want to know if a QB is any good? You look at completion percentage, yards per attempt, TD/Int ration, and so on. Want to know if a goalkeeper is any good? You look at goals allowed per minute played, saves, etc. You know what you don't look at in any of these cases? W/L. Because W/L evaluates a team, not the individual, and even then, W/L is only part of the answer.

So no, I cannot give you any math because MWO doesn't give me any statistics that are meaningful. I can, and did, manipulate my damage numbers (because event). I can boost my K/D ratios just by playing heavies more than lights, and not ever playing mechs that I haven't skilled or that I know aren't top tier. I can utterly, beyond any shadow of a doubt, make a mockery of every leaderboard statistic we have available which means you can look at my stats and know exactly nothing about my skill as a mech pilot in MWO. And I'm not a unique snowflake nor anything special. There are lots of good MWO players just like me.

You're examples were not real world nor math. You've been provided the links to how analytics and predictive statistical analysis works. All you provided were anecdotes and assumptions.

Again. Because you seem unwilling to address this.

[b] The existence of statistics and analytics and predictive analysis exists to solve for factors like this situation. Anyone who works in the analytics field creates processes to solve for more obscure factors than your w/l in pug queue. You can not "game" win/loss because if you're the kind of person who takes bad mechs and loses matches intentionally then you are inherently lower in prediction for winning. Everything you do, inta finally or not, is reflected in your win/loss. This is directly then represented in the prediction of your future win/loss. It can not be "gamed" because as an average anything you do to game it is reflected in the average.

Also as Nightbird showed, the best data point by far for prediction is win/loss. You not understanding why that number is useful is irrelevant to how math works.

Again, there's no opinion here. No anecdotal reference to make. This is basic statistics. Again, google it. Inam strongly suspicious that it would be a waste of time for me to actually post foe you the predictive analysis process, the formulas and the mathematical principles that explain exactly why you're wrong.

[Xiphias]

Xavori, on 26 November 2017 - 05:08 PM, said:

I provide actual, real world examples that prove, beyond any shadow of a doubt, that your assumptions about W/L are wrong, and you ask me for math. I don't need math when I have reality.

You haven't proved anything. You've simply posted a poor example and completely misunderstood its meaning and application. Kind of like how your "super simple example" actually proved my point in that other thread and how you conveniently ignored it when I showed how it actually proved my point.

That's fine though. Continue to substitute your own "reality" instead of using actual math to back up your point like the numerous people who have tried to explain these concepts to you have done.

Quote

The reality is that in every competitive team sport on Earth, you do not evaluate players based on the team's win/loss record.

Because teams aren't randomly assigned in team sports, they stay constant. In a solo queue situation your team is random and your WLR is a direct reflection of your contribution to your team winning.

Quote

I can utterly, beyond any shadow of a doubt, make a mockery of every leaderboard statistic we have available which means you can look at my stats and know exactly nothing about my skill as a mech pilot in MWO. And I'm not a unique snowflake nor anything special.

Here's a good chance to prove your point then. Next leader board season why don't you farm your stats to the best of your ability for a week or two and then post the results? Needs to be a decent sample size, say 80 matches or so though.

If it's really so easy you should have no problem maintaining a match score of at least 400 and a KDR of 3.0+. While you're at it go ahead and average 1 kill per match too. If you can keep that up over a decent sample size I'll concede your point and that you actually are a good player. I've done it for the last two seasons so it's certainly achievable.

Talk is cheap. If you can do it, prove it.

If not you're going to need to revise your quote:

Quote

There are lots of good bad MWO players just like me.

*In before Xavori starts making excuses for not doing it like, "I'm not willing to stoop to that level", "I don't need to prove anything", "I'd rather just have fun with the game" etc.

MischiefSC, on 26 November 2017 - 03:35 PM, said:

This argument is literally flat earth. Opinions on it, how it "feels", means **** all nothing. Statistical analysis and how to solve for a value in an equation is already a thing. There is no argument to have.

Thank you for this.

[Asym]

Mystere, on 26 November 2017 - 06:15 PM, said:

As people have been telling you, the only stat that actually has any meaning is wins.

How you deduced that all other stats have any meaning to me makes we wonder.

Are we really this far off topic?

Winning? Seriously? Winning is a mostly random event (i.e. the mascot example earlier); the straphanger on a comp team; the other team being that bad; winning dispite yourself; and, karma........in all it's forms...

Winning, orlosing for that matter, are the most unreliable numbers there are..... Winning contains skill. Skill is the metric we seek.

Sorry OP, we're WAY off topic and you are right, Scouting is a mess at the moment.

[arcana75]

I like Scouting. You just need a change in mindset. It's not Scouting. It's QuickPlay 4v4. It's a short, quick and tight experience for exactly what I want out of MWO: brawling.

When events rolls around, it evolves into damage farming when events centre around match score. I'm perfectly fine with that.

[MischiefSC]

Asym, on 26 November 2017 - 10:02 PM, said:

Are we really this far off topic?

Winning? Seriously? Winning is a mostly random event (i.e. the mascot example earlier); the straphanger on a comp team; the other team being that bad; winning dispite yourself; and, karma........in all it's forms...

Winning, orlosing for that matter, are the most unreliable numbers there are..... Winning contains skill. Skill is the metric we seek.

Sorry OP, we're WAY off topic and you are right, Scouting is a mess at the moment.

Your win/loss for QP is pretty accurate after 80 matches, solidly accurate after 100 matches.

Your stats are not random - they are a direct result of your gameplay. The random opportunities you get to swing a win or loss are the same for everyone - how well you exploit them are a direct representation of your skill at winning.

Hence it reflects accurately in the averages.

[Xavori]

Mystere, on 26 November 2017 - 06:15 PM, said:

As people have been telling you, the only stat that actually has any meaning is wins.

How you deduced that all other stats have any meaning to me makes we wonder.

No. Wins has no meaning. I'm always on great teams? A WINNER IS ME!

That's meaningless.

And no, I didn't deduce all other stats have meaning. I've kinda been pretty consistent in calling every leaderboard stat garbage. They don't give you any idea of whether or not a player is a good mech pilot or not. And I'm definitely including W/L in that.

MischiefSC, on 26 November 2017 - 11:51 PM, said:

Your win/loss for QP is pretty accurate after 80 matches, solidly accurate after 100 matches.

Your stats are not random - they are a direct result of your gameplay. The random opportunities you get to swing a win or loss are the same for everyone - how well you exploit them are a direct representation of your skill at winning.

Hence it reflects accurately in the averages.

No. My W/L after 100 matches is an utterly meaningless thing. It has zero relation to my skill as a pilot. Nada. Zilch. None.

You are absolutely the guy with a hammer who thinks the world is nothing but nails. I'm the guy with a zamboni who thinks the entire world is cartoon characters covered in chocolate sauce doing the macarena while singing out tech manuals as lyrics to old show tunes.

In chess, you can look at ELO and have an idea about the probability of the two players winning or losing. That's because chess only has individuals, and they're both playing with exactly the same pieces under the same set of conditions. The game itself isolates everything that is not the individual playing.

In MWO, you can't do that. You cannot say at the start of a match if a player is going to win or lose with any realistic probability. You have to start isolating other variables before you even begin to get hints. Which mech is said player running? Is said player on a premade team? Is the player on a solid internet connection? What time of day is the person playing at? What are the skill of his 11 teammates. Which mechs are his teammates running? And so on. And even then, your probabilities are going to still be in the area of 50/50. So, meaningless when it comes to pilot skill.

I do believe there are good pilots and bad pilots. What I don't believe is that you can look at anything on the current leaderboard and tell me who they are. All you can tell me is which pilots play on good teams more often than not.

[Asym]

MischiefSC, on 26 November 2017 - 11:51 PM, said:

Your win/loss for QP is pretty accurate after 80 matches, solidly accurate after 100 matches.

Your stats are not random - they are a direct result of your gameplay. The random opportunities you get to swing a win or loss are the same for everyone - how well you exploit them are a direct representation of your skill at winning.

Hence it reflects accurately in the averages.

Way off topic and another jab as if the leaderboard has any value... Xavori has it right and ,many on this forum throw "leaderboard grenades" to justify their beliefs: for good or ill. I'll not judge you.

Factually, my stats are a result of the environment first and foremost and then, skill. If I drop in FP on a pick-up-team and get farmed versus a Scouting match with peers.... I 've had 3 kill-900 damage scouting matches and <1000 dmg FP matches... I had ZERO control of the environment and skill made little or any difference.... Some players do control their environments by only playing where winning is a given (Ah, those teams that farm noobs we encounter each and everyday...)

But, let's see, according to the leaderboard (you can hear the choir music in the background), I've played 300+ games to your 100+ games...... Hmmmm? You need to get hopping to catch up with me !!! Geeze Marie Lad, only 100+ games...........(sigh...what is the world coming to?)

Scouting is still a mess OP and I lament having to talk about statistics vis-a'-vis positive enhancements we'd like to see....

[ANOM O MECH]

Xavori, on 26 November 2017 - 05:08 PM, said:

*sigh*

I provide actual, real world examples that prove, beyond any shadow of a doubt, that your assumptions about W/L are wrong, and you ask me for math. I don't need math when I have reality.

The reality is that in every competitive team sport on Earth, you do not evaluate players based on the team's win/loss record. Want to know if an MLB pitcher is any good? You look at ERA, hitting percentage, walks per inning, HR's allowed, etc. Want to know if a QB is any good? You look at completion percentage, yards per attempt, TD/Int ration, and so on. Want to know if a goalkeeper is any good? You look at goals allowed per minute played, saves, etc. You know what you don't look at in any of these cases? W/L. Because W/L evaluates a team, not the individual, and even then, W/L is only part of the answer.

So no, I cannot give you any math because MWO doesn't give me any statistics that are meaningful. I can, and did, manipulate my damage numbers (because event). I can boost my K/D ratios just by playing heavies more than lights, and not ever playing mechs that I haven't skilled or that I know aren't top tier. I can utterly, beyond any shadow of a doubt, make a mockery of every leaderboard statistic we have available which means you can look at my stats and know exactly nothing about my skill as a mech pilot in MWO. And I'm not a unique snowflake nor anything special. There are lots of good MWO players just like me.

You are absolutely not providing anything that proves your point.

You also can't articulate anything that even sounds remotely reasonable as to why the stats are meaningless. You have only provided an opinion which seems hinged around your theory that you are much better at the game than the stats portray. I have seen zero evidence that your position is anything other than your ego, unable to accept the numbers you have.

Argue and keep telling yourself that you're really good if only there was some magic stats showing how much sensor range you have added to your team....

Like it or not, match score, damage output, win/loss are absolutely important stats because killing mechs is the best way to consistently do well.

[arcana75]

So with all the data available, what's the best gauge of a person's .... shall we say collective wisdom or cumulative skill, in MWO?