633 replies to this topic

[Sug]

Roadkill, on 23 January 2014 - 01:08 PM, said:

As I explained above, you should only need 25-30 matches in each weight class to reach reasonably accurate Elo ratings.

Why do you think that? If your Elo can only change significantly if you beat a higher Elo team or lose to a lower one and you're matched against people close to your Elo range and your WLR is 1.00 than your Elo should stay close to where it started.

The only way to really change your score is to group with people and stomp some pugs.

[RichAC]

People should be ranked by match score, period. ELO by win/loss only is not for individuals on constantly changing teams. I am in total shock right now.

It should be based on how well you do on that team. Your ranking at the end of the match, should be what your ranking calculates as. What you see on the scoreboard at the end of a match, from top to bottom, should be how you rank in that match compared to your other teamates. I can't believe what i'm hearing today.

One thing PGI has going for it that quakelive did not. Is that they have a match maker. Not a tiering system, and they don't have dedicated servers. SO it would be harder for people to sandbag their bracket since they don't pick their servers or know what their ranking or bracket even is.

But PGI definitely needs a new formula, this is the most ridiculous thing I ever heard.


This must be the twilight zone.

Edited by RichAC, 23 January 2014 - 03:23 PM.

[Deathlike]

RichAC, on 23 January 2014 - 03:19 PM, said:

People should be ranked by match score, period. ELO by win/loss only is not for individuals on constantly changing teams. I am in total shock right now.

Match score is heavily weighted towards fighting, so it's technically less effective when it comes to capping on Assault and worse on Conquest.

For instance, capping in an 0-0 Assault match is different than 11-10, or even 1-11. Match score isn't always adequate or perfect at describing a match.

The thing is, you can't easily have one number describing everything. You can only get a close approximation... and even then it won't be perfect.

Edited by Deathlike, 23 January 2014 - 03:25 PM.

[RichAC]

Deathlike, on 23 January 2014 - 03:23 PM, said:

Match score is heavily weighted towards fighting, so it's technically less effective when it comes to capping on Assault and worse on Conquest.

The thing is, you can't easily have one number describing everything. You can only get a close approximation... and even then it won't be perfect.

Thats all people play for. K/D and dmg done. Noone cares about winning on a random team, so they are always going to complain regardless.

But I did add they need to bring back the points for capping individual bases in conquest. I dont' know why they took that away in the first place.

And yes, its back to the same old story again. Not only adjust the points for capping in conquest....adjust the points for caps in assault. The whole team already gets the assist, maybe some more points for the guys doing the capping.

But this has defintely been the problem I have had with this game from the very beginning. None cares about winning.

Add ELO to the huge list of why noone gives a {Scrap} about winning in this game... Just another incentive for people to lose and just play for dmg and cbills instead.
We have skirmish now, It shoudlnt' even be an issue now for people who just want to fight.

Edited by RichAC, 23 January 2014 - 03:31 PM.

[Doctor Proctor]

Deathlike, on 23 January 2014 - 03:23 PM, said:

Match score is heavily weighted towards fighting, so it's technically less effective when it comes to capping on Assault and worse on Conquest.

The thing is, you can't easily have one number describing everything. You can only get a close approximation... and even then it won't be perfect.

Worse, match score is heavily weighted on damage. I've had more kills and assists than players and received a lower match score due to have lower damage. That could mean that I was either much more accurate and not spreading damage around as much, or it might mean that I was finishing off already damaged mechs whereas the other player was killing them relatively solo. It's hard to say and probably varies on a case by case basis.

Basically though, any single metric will always give you a rather narrow range of measurement. Elo seems to ignore a lot of influencing factors that could affect the viability of any particular single match, thus leading to a potential for a lot of blowouts that end up evening out to an approximate 1:1 w/l ratio in the end. Ranking via match score only would mostly distribute the rankings by what chassis/weight class people most commonly drove. Assaults will tend to move up faster, Lights will tend to move up slower. Players that work together as a team, such a dedicated LRM boat that always groups with his dedicated spotter buddy, would basically win or lose based on their respective performances in an Elo system while being on opposite ends of the spectrum in a match score system.

This is why I was saying earlier that some sort of hybrid that takes both Elo as well as some individual metrics (match score being one of them possibly) into account might do a better of ranking people. Then something like "Oh look, Doc does much better on Crimson Straits than on other maps" might translate into a higher MM rating for me on that map, thus leading to closer matches.

RichAC, on 23 January 2014 - 03:28 PM, said:

We have skirmish now, It shoudlnt' even be an issue now for people who just want to fight.

BTW, just as an aside, I actually saw someone complaining in a game about a week ago because the team I was on ended up capping. I was like "This was literally why they created Skirmish mode. Why are you even playing Assault if all you want to do is fight without having to worry about capping?"

WTF people?!?!

[RichAC]

Doctor Proctor, on 23 January 2014 - 03:31 PM, said:

Worse, match score is heavily weighted on damage. I've had more kills and assists than players and received a lower match score due to have lower damage. That could mean that I was either much more accurate and not spreading damage around as much, or it might mean that I was finishing off already damaged mechs whereas the other player was killing them relatively solo. It's hard to say and probably varies on a case by case basis.

Basically though, any single metric will always give you a rather narrow range of measurement. Elo seems to ignore a lot of influencing factors that could affect the viability of any particular single match, thus leading to a potential for a lot of blowouts that end up evening out to an approximate 1:1 w/l ratio in the end. Ranking via match score only would mostly distribute the rankings by what chassis/weight class people most commonly drove. Assaults will tend to move up faster, Lights will tend to move up slower. Players that work together as a team, such a dedicated LRM boat that always groups with his dedicated spotter buddy, would basically win or lose based on their respective performances in an Elo system while being on opposite ends of the spectrum in a match score system.

This is why I was saying earlier that some sort of hybrid that takes both Elo as well as some individual metrics (match score being one of them possibly) into account might do a better of ranking people. Then something like "Oh look, Doc does much better on Crimson Straits than on other maps" might translate into a higher MM rating for me on that map, thus leading to closer matches.

What kills me, is they give no incentives for winning a match, and then they base the ELO off of it? hahaha

Although somewhat understandable when you are on a team with random players.

But I mean from day one i coudlnt' believe the game was all about damage done. Which as you say has more to do with your weight class and loadout.

You get as much cbills for losing a match as you do winning! You get an obscene amount of cbills for dmg done over kills,kill assists, cap assaults, cap defends, spots assists, etc combined! Now its coming back down to that root problem with this game once again imo.

Your absolutely right dude. Match score + W/L would be ok with me but not W/L alone. But again they definitely have to give more points for people doing things besides damage. Especially capping especially since We have skirmish now, whats the problem?

But its still a fact thats what people play for. Winning means nothing to them. so they are always going to complain.

These same people who are always crying about cap warriors are probably the same ones complaining about weight limits now and ELO or matchmakers, which are problems they brought on themselves from the beginning.. its Ironic.

Basing the it off of match score, will at least please most of the community and make more sense, since people are only playing for match score, even if it does not nescessarily mean they are more likely to help their team win.

Hey guys dont' worry, it will even out in the end

Edited by RichAC, 23 January 2014 - 03:45 PM.

[RussianWolf]

On the match maker balance I said I'll pull a bit of data. Nothing huge in the 5 games I was able to play last night

2L 3M 4H 3A 740 vs 3L 3M 2H 4A 745
1L 4M 4H 3A 805 vs 1L 4M 2H 5A 855
3L 2M 4H 3A 725 vs 3L 3M 3H 3A 705
1L 3M 6H 2A 760 vs 1L 4M 3H 4A 780
3L 2M 4H 3A 765 vs 0L 4M 8H 0A 725

So 50 tons was the largest difference, but you can see there were some odd groups especially the last one with the team with no Lights or Assaults. The teams on the right were the losers.

[Sarsaparilla Kid]

IceCase88, on 23 January 2014 - 01:25 PM, said:

Who cares whether your ELO is high or low? It's just a game and being the best at it will have exactly zero effect on your daily life. When the game goes away, and it will one day, what then?

No one knows what their ELO score is or how it is computed. Is it based off wins/losses, damage, KDR, is it per mech, total of all your mechs, an average of all stats, a ratio there of, per capita of the total sum? Speculating is just dumb. Dr. Seuss hour is over.

This was billed as a "thinking person's game"...pardon us while we think out loud...

[MischiefSC]

Abivard, on 23 January 2014 - 03:46 AM, said:

The above is Intelligent, well-educated, ignorance.

Nice ad hominem, textbook even, but I'd love some math to back up your position or contest mine. Roland put forward some very intelligent arguments with points to back them up and with a larger playerbase I'd agree with him.

So, do you actually have any valid data to back up your position? If you're going to call me ignorant please identify what I'm ignorant about or where I'm wrong?

[Ghogiel]

Sug, on 23 January 2014 - 03:14 PM, said:

Why do you think that? If your Elo can only change significantly if you beat a higher Elo team or lose to a lower one and you're matched against people close to your Elo range and your WLR is 1.00 than your Elo should stay close to where it started.

The only way to really change your score is to group with people and stomp some pugs.

Because the MM is bad due to player count. The MM now theoretically allows players of any elo rating to be put into the same match as other players.

Paul in his infinite wisdom addressed a nonsense point in his recent command chair. He said there were a few players who sent him screens of games that were stomps and he worked it out that the averaged team Elo rating between the teams is quite closely matched actually.

Well that is one thing yeah, the thing I and I am sure everyone else gives a **** about and the real reason why they even subbmitted those screens to him was to demonstrate low rated players AND high rated players on the same team.

It just makes the games play whacky.

Elo is actually not a problem at all. It is still tracking your own ability to influence a matches outcome. It's the MM allowing for more goofy matches by dropping 4 high elo rated players in amongst scrubs and the pitting them with mediocre players and saying the team Elo is matched ggclose.

All the players who want match scores or some **** in there now.. High Elo players can just premade up and get a much improved average match score and improve their W/L since the loosening up of the limits.

[MischiefSC]

For everyone talking about adding a criteria other than win/loss for player ranking -

Short version - you can't if you want the system to be accurate.

Long version -

Any and every other factor save win/loss can be gamed. Some builds do more damage, some do more kills. Sometimes it's the organizational skills someone brings to the match that gives their presence value. Sometimes it's the guy who draws fire, not the guy who fires, that decides a win. The guy who stops a cap and keeps 3 lights occupied long enough for the rest of your team to exploit their numerical advantage for the win, etc. etc. etc.

Win/loss can't be 'gamed'. You either helped win or you didn't. That's why, end of the day, it's the only valid metric for player ranking.

Now, you want an e-peen measuring system you could have some combo of damage/kills/assists/average weapon accuracy/caps and the like for a composite score of some sort but again, a kill-stealing punk who just drops every match that looks like his team is losing but exploits every match where he can steal kills with his 2PPC Cicada from a team that's winning without him can easily inflate performance. Drop in a premade in an LRM stalker with two tag-carrying Ravens and you'll do tons of damage, especially after the new NARC update. You may have a minimal impact on your teams wins and losses but you'll blanket a lot of hurt.

Everything you do contributes to win/loss. Everything, good or bad, loadout, mech choice, positioning, situational awareness, teamwork, chat use, VOIP or not, accuracy, damage, kills, assists, tag, scouting, *everything* contributes to your teams total odds of winning.

Win/loss. Every other game system (TrueSkill as a premier example) only uses win/loss as well. It may gather tons of data about how you won and lost, who with and who against, but the metric that places you in its 50 levels of ranking is win/loss.

[Abivard]

Elo Was created as a means to rate chess players.

It works fine for that, it works fine for any one on one sport, or even team sports where the teams are consistent. Once you get outside of what it was designed for, it stops working.

[RichAC]

Abivard, on 23 January 2014 - 03:55 PM, said:

Elo Was created as a means to rate chess players.

It works fine for that, it works fine for any one on one sport, or even team sports where the teams are consistent. Once you get outside of what it was designed for, it stops working.

Should be common sense, but apparenlty we have some geniuses here who think otherwise...lol

They should also just get rid of the scoreboard at the end of matches, since it has absolutely nothing to do with a players rank or skill rating and is totally meaningless.

[RichAC]

MischiefSC, on 23 January 2014 - 03:54 PM, said:

For everyone talking about adding a criteria other than win/loss for player ranking -

Short version - you can't if you want the system to be accurate.

Long version -

Any and every other factor save win/loss can be gamed. Some builds do more damage, some do more kills. Sometimes it's the organizational skills someone brings to the match that gives their presence value. Sometimes it's the guy who draws fire, not the guy who fires, that decides a win. The guy who stops a cap and keeps 3 lights occupied long enough for the rest of your team to exploit their numerical advantage for the win, etc. etc. etc.

Win/loss can't be 'gamed'. You either helped win or you didn't. That's why, end of the day, it's the only valid metric for player ranking.

Now, you want an e-peen measuring system you could have some combo of damage/kills/assists/average weapon accuracy/caps and the like for a composite score of some sort but again, a kill-stealing punk who just drops every match that looks like his team is losing but exploits every match where he can steal kills with his 2PPC Cicada from a team that's winning without him can easily inflate performance. Drop in a premade in an LRM stalker with two tag-carrying Ravens and you'll do tons of damage, especially after the new NARC update. You may have a minimal impact on your teams wins and losses but you'll blanket a lot of hurt.

Everything you do contributes to win/loss. Everything, good or bad, loadout, mech choice, positioning, situational awareness, teamwork, chat use, VOIP or not, accuracy, damage, kills, assists, tag, scouting, *everything* contributes to your teams total odds of winning.

Win/loss. Every other game system (TrueSkill as a premier example) only uses win/loss as well. It may gather tons of data about how you won and lost, who with and who against, but the metric that places you in its 50 levels of ranking is win/loss.

Yes we need a e-peen measuring system. Everything that gives you "points" should "count" Or else your gonna end up with a game that nobody plays. Its that simple. IF you have a problem with that, adjust the points, which I felt was a problem with the game since I started.

But What games use trueskill?

I don't know of any other fps random team game, that uses any skill rating system except quakelive. Let alone W/L only. So I applaud PGI for trying. They are still breaking new ground imo. Quakelives original rating system they paid another company for called GaimTheory, was designed by accountants, hedge fund stock brokers and casino owners.... Thats the one I would vote for.

Edited by RichAC, 23 January 2014 - 04:07 PM.

[Deathlike]

MischiefSC... you're not going to convince everyone with numbers, especially if they focus solely on "how they feel at the time".

I'm frustrated at the MM moreso than my ELO (I'm not even sure if I'm properly ranked in the system TBH). The thing is that people are not always observant of the people that play the... whether by the numbers, or how they go about playing the game (with whatever loadout they choose). It is often times impossible to quantify or qualify people when you see them in game, unless... you play with them often enough (either being part of premade, or just happening to bump into them more often than not).

It's kind of like the "black box" system used in college football to determine rankings. Frankly, it's probably mostly accurate, but on the other hand, you only play so few teams relatively speaking (because, there's a ton of them) and it's hard to make sense of it all in the grand scheme of things. The only actual arbitrary thing to this though is how it sets up the "BCS Championship game", because that is unlikely set up by a machine, but people that have money and influence. That's besides the point... but that's how ELO is being compared against... it's harder for some people to qualify or quantify how good/better/worse they are from the next person. It doesn't have to be perfect... in fact it simply IS NOT PERFECT EVER. It has to be "accurate enough" for people that see them and understand "hey, why is this guy successful" and "what do I have to do to be as successful as that guy". That is ultimately the more important thing that people don't always seem to grasp.

[Ghogiel]

Abivard, on 23 January 2014 - 03:55 PM, said:

Elo Was created as a means to rate chess players.

It works fine for that, it works fine for any one on one sport, or even team sports where the teams are consistent. Once you get outside of what it was designed for, it stops working.

Why does it stop working?

[MischiefSC]

Abivard, on 23 January 2014 - 03:55 PM, said:

Elo Was created as a means to rate chess players.

It works fine for that, it works fine for any one on one sport, or even team sports where the teams are consistent. Once you get outside of what it was designed for, it stops working.

How does it stop working. Please, enlighten me. League of Legends, association football (soccer for us Americans), college sports leagues for basketball, football and baseball, the MLB, not to mention individual and group ranking in pretty much every Korean esports league. Elo is the basis of pretty much every ranking system of every major and minor sport that does serious rankings.

Even LoL and WoW plus TrueSkill and Glicko use Elo as the core of their matchmaking - all use the win/loss metric and at their core use the Elo equation. They just drill down on your win/loss record to adjust for how your win/loss changes based on your teammates and the composition of the other team. All of them absolutely use your win loss in a team game to establish your ranking as an individual and without requiring you to play always with and against the same teams.

So please identify for me, so that I can achieve world renown and fame for disproving the fundamental mathematical basis of pretty much every major rating system.

Edited by MischiefSC, 23 January 2014 - 04:09 PM.

[RichAC]

MischiefSC, on 23 January 2014 - 04:07 PM, said:

How does it stop working. Please, enlighten me. League of Legends, association football (soccer for us Americans), college sports leagues for basketball, football and baseball, the MLB, not to mention individual and group ranking in pretty much every Korean esports league. Elo is the basis of pretty much every ranking system of every major and minor sport that does serious rankings.

Even LoL and WoW plus TrueSkill and Glicko use Elo as the core of their matchmaking - all use the win/loss metric and at their core use the Elo equation. They just drill down on your win/loss record to adjust for how your win/loss changes based on your teammates and the composition of the other team. All of them absolutely use your win loss in a team game to establish your ranking as an individual and without requiring you to play always with and against the same teams.

So please identify for me, so that I can achieve world renown and fame for disproving the fundamental mathematical basis of pretty much every major rating system in use in the world.

LoL apparenlty just switched to trueskill, becaues they realized an ELO system is not designed for a random team game. But Ironically trueskill is still ELO based.

Glicko, is not a game, its another ELO/ chess like based rating system. http://en.wikipedia....o_rating_system

LoL will eventually come the same conclusion, that trueskill is no different then any other ELO system. What LoL has going for it, is that it has millions of players, so there is less of a skill gap felt. And there is no need for them to try and create another industry standard.

But you still have threads like these http://forums.na.lea...d.php?t=3817616 that beg the same question.

How do scouts rank professional athletes? Do you think they do it based only on their team wins and losses? Absolute utter nonsense! What determines how much a pro player gets paid? THEIR STATS!! But I expect nothing less from a nerdy industry that doesnt' even sell sports games anymore.

Its sad that bad decisions in quakelive killed that game, because their rating system was pioneering a whole new world once again in online mutliplayers. Just like they have in the past....But more then ID, we have to blame the gaming community that exists nowadays.

its common sense as to why you can't rate and individual based on their team performance and needs no explanation.

Edited by RichAC, 23 January 2014 - 04:24 PM.

[Roland]

Ghogiel, on 23 January 2014 - 01:13 PM, said:

It felt more like 200 matches per weight class for me to press through most of the scrub tier {Scrap}.

I personally believe 25-30 matches is a pretty wrong estimate. It's closer than however many million the other guy was saying lol.

Having played with you, I would estimate that you're better than the vast majority of players in MWO currently. For you, your rating is going to rise up significantly faster and stabalize at the "better than most people" level quickly.. because really, for outliers on the edge of the Elo rank, it's much EASIER for Elo to work..because you are exerting a stronger force on it.

That is, in many games you participate in, you are going to have a much larger influence over the outcome than most players do, thus the actual result of the match actually corresponds more directly to your skill than it does to other players.

MischiefSC, on 23 January 2014 - 04:07 PM, said:

How does it stop working. Please, enlighten me. League of Legends, association football (soccer for us Americans), college sports leagues for basketball, football and baseball, the MLB, not to mention individual and group ranking in pretty much every Korean esports league. Elo is the basis of pretty much every ranking system of every major and minor sport that does serious rankings.

In all of those applications, each team is viewed as a single virtual player. The TEAM is rated, rather than any individual players.

That's how Elo works in those cases, because they aren't trying to determine individual ratings for players based on the outcomes of team games.

[MischiefSC]

RichAC, on 23 January 2014 - 03:57 PM, said:

Should be common sense, but apparenlty we have some geniuses here who think otherwise...lol

They should also just get rid of the scoreboard at the end of matches, since it has absolutely nothing to do with a players rank or skill rating and is totally meaningless.

No, the problem is that 'common sense' in the context you're using it is actually 'in my uneducated opinion'.

The problem is that someones opinion of using win/loss has absolutely nothing to do, what so ever, with how effective it is. I entreat you to read the links I posted on statistics, statistical analysis and the actual principles of predictive analysis so you understand why your opinion is wrong.

Which it is. Saying Elo doesn't work for MW:O, or any other win/loss competition, is flat out wrong. Team or individual it doesn't matter. You could have 100 teammates that change every single game it it still works - it would just take more games.

Let me put it to you this way.

Suppose you took LeBron James and put him randomly in to replace one player each game in high-school basketball games. He would play each game with a different team and against a different team.

At the end of the season the teams he played with would be far, far more likely to have won because he was there.

Do the same thing with Payton Manning in high-school football games. Every team he played with would be far more likely to win, right?

You don't have to be world-shaking good to impact your teams performance, that just makes it easier to identify. Those guys could trend their winning in 20 games. For comparison sake they'd have a win/loss in that situation of like 5.0 or 6.0 instead of a 1.2 or the like.

Elo works. You're welcome to research the math behind why it works but trying to say it doesn't work is literally like saying math doesn't work because you're writing 1+1+1+1+1+1+1+1 = 8 instead of 4+4=8. The value of each factor (+1 eight times vs +4 twice) may be reduced but the result is the same.