39 replies to this topic

[Wakashan]

Whether I am playing with a group of people I know, or if I am joining PUGs, I lose almost every game (I actually win more often in PUGs). My win-lose ratio is 16:48. Is such a high win-lose ratio normal? Back when I played beta, I was with a group of people that won 7 out of 10 games. Are all of the really good players on the clan side? I was told yesterday that a lot of people from Jade Falcon hack, but I don't see any evidence for that assertion.

[Euphoric1RW]

Hey Wakashan, whatz up! When pugging alone I find myself scoring much higher as well, like 300-400 more damage per match. I think that is because if you have been playing for a while you are probably a decent player, so in pugging you will have a higher score compared to every one else, and the competition for damage is less,

If you join a couple guys you may drop against a 8 to 12 man group and the team you pair up with will be scrambling for the damage and kills so it is a little more difficult to get the higher damage and kills unless you are really good. In addition the competition is much much higher once you group up even with one person. If you can find a group of pilots that are on when you are and can play with a bigger group (closer to 12 man team the better)and get organized you will win much much more.

As far as hacking, PGI unveiled a bunch of hackers and they were not from any specific faction but a varity including IS, so I would not listen to anyone telling you Clan is better or IS is better, there is a pretty good balance over all, just find a mech you like and join a group, figure out your play style and see you on the battle field.

Edited by Merroc Sandybanks of Frogmorton, 15 July 2015 - 07:34 AM.

[mailin]

If you only have 64 matches since the reset, the MM may still be trying to assign you an Elo. For example I have a win:loss of 1.15 with 6200+ matches in since the reset. Interestingly my w:l was 1.12 with almost exactly 12000 matches in before the reset. Otherwise, no, it's not normal to have such a disparity because the MM tries to set it up so that your w:l is nearly constant at 1.0, or at least that's what PGI would have us believe.

Just curious, but what was your w:l before the reset?

[SnagaDance]

Seems like you haven't returned for too long. You obviously left before the the reset and the changes since then have been MASSIVE. What with things like new maps, Clan mechs, weapon changes, quirks, HSR improvements, hit detection improvements etc. etc.

it's basically a new game since you've left. Not only are you likely to be rusty, but many of your dormant ingrained skills are actually wrong as well!

I'm sure you'll be able to get back into the groove though, it will just take some time.

[Revis Volek]

I have seen no evidence that ever supported MM tries to get you a 1:1 w/l ratio. If someone can show me it i will take it into corroboration. But i have yet to see this information...

Welcome to the world of MWO, it gonna be rough for a while until you get into the proper ELO pool, and if you had played before and you ELO was high then it may still be lingering there until you settle into place.

Clan Jade Falcon has no more hackers then the rest of the games factions and whoever told you that is lying through their teeth or just bad. Dont listen to either of them....Keep doing you and the wins will come.

Feel fee to find us in TS StranaMechty for some fun casual matches if you ever want.

IP Stranamechty.info
pw: StranaMechty

Edited by DarthRevis, 15 July 2015 - 10:29 AM.

[IraqiWalker]

DarthRevis, on 15 July 2015 - 10:29 AM, said:

I have seen no evidence that ever supported MM tries to get you a 1:1 w/l ratio. If someone can show me it i will take it into corroboration. But i have yet to see this information...

Welcome to the world of MWO, it gonna be rough for a while until you get into the proper ELO pool, and if you had played before and you ELO was high then it may still be lingering there until you settle into place.

Clan Jade Falcon has no more hackers then the rest of the games factions and whoever told you that is lying through their teeth or just bad. Dont listen to either of them....Keep doing you and the wins will come.

Feel fee to find us in TS StranaMechty for some fun casual matches if you ever want.

IP Stranamechty.info
pw: StranaMechty

Elo technically tries to set you to a 1:1 ratio. However, it's trying to do the same for the other 23 players in the match, so in effect, it's not really getting it done.

[Revis Volek]

IraqiWalker, on 15 July 2015 - 10:56 AM, said:

Elo technically tries to set you to a 1:1 ratio. However, it's trying to do the same for the other 23 players in the match, so in effect, it's not really getting it done.

Only if the two players that are pitted against each other have similar ELO scores do the actually have a chance of 1:1. The more skewed the numbers are one way or the other the worse this gets.

and like you said, throw 23 other players into the mix and this is hardly the case...ever.

[Void Angel]

Actually, Elo does not try to get you a 1:1 win/loss - that's an oversimplification that gets bandied around - and is responsible for many of the wrong ideas about Elo that you see. What Elo tries to do is assign you a rating that is consistent with your level of skill; it then adjusts this rating (and your opponent's) based on the results of matches you're in. Over enough time for the rating system to adjust, the hypothetical Best Player in the Game, having the highest Elo, would win most of his matches, but still receive very slight adjustments in his Elo - and maintain his rating as an average. Of course this doesn't account for the matchmaker and the fact that it is a team game. Contrary to what some people will try to tell you, Elo will work even for random teams, but there are distortions because of the size of your teams and the learning curve of this particular game. The way in which the matchmaker works (and in which Elo is applied,) is a much larger subject than pure Elo. However - the point here is that neither Elo nor the matchmaker actually try to force your stats to a 1:1 win/loss.

[Void Angel]

Wakashan, on 15 July 2015 - 07:12 AM, said:

Whether I am playing with a group of people I know, or if I am joining PUGs, I lose almost every game (I actually win more often in PUGs). My win-lose ratio is 16:48. Is such a high win-lose ratio normal? Back when I played beta, I was with a group of people that won 7 out of 10 games. Are all of the really good players on the clan side? I was told yesterday that a lot of people from Jade Falcon hack, but I don't see any evidence for that assertion.

Actually, that's not a high win/loss ration - it's a high loss rate. =) Still, as others have told you, the game has changed drastically. Part of your difficulty lies in those changes, and is exacerbated by the fact that the Group Queue is a much nastier environment than PuGs; mistakes cost more there because of the superior coordination (and I've been told there's an Elo penalty to taking a 12-man as opposed to smaller groups; haven't confirmed, so do not quote me.) Keep working at it and focus on the social interactions with your group of friends - it'll turn out right.

Edited by Void Angel, 15 July 2015 - 11:16 AM.

[Wakashan]

Thanks for the comments. I did realize that I had the title backwards, but once the thread is started, it is too late.

Yesterday evening I started looking at some of the new aspects of the games (and mechs) since Beta. In Beta, I mainly piloted a Dragon 1N and an Awesome 5Q (I think). I sometimes piloted a Commando as well. My friends and I usually had a full three lances and we were on TS. The group of people I play with rarely have a full three lances whilst being on TS. This means that we usually end up working with PUGs. I remember when one of my friends yesterday told a Stalker NOT to go out into the water on a map. The goober went anyway and the rest is history.

Anyway, I think that is one of the primary issues I am facing. We are facing many well-coordinated teams and my friends seem to prefer talking to each other on TS instead of keeping the PUGs informed. Of course, we also have a problem with PUGs not listening at times, but I have seen a lot of really good players in PUGs.

I started playing with my King Crab (I named it BOSS. ), and my Hunchback (I named him Bullseye until I can get something better). I also picked up an Enforcer that I plan on outfitting with a bunch of U/AC5 autocannons. Things really started improving for me yesterday. I found that in Beta, because I had so many well-coordinated players on my team, I could go out in the open and wreak havoc. Now that I am in groups that are not quite as coordinated, I have to be really sneaky, especially with brawler mechs. I have to wait behind cover and then come out at the right time to lay down some serious pain. When I started to do this, I noticed that my teams started winning more. So, I think I was the weak link. I am stilling looking for that balance between me not going into the open too fast and joining the brawl before my fellow mechwarriors are torn up.

I did notice, however, that clan mech technology is vastly superior to innersphere technology. Of course, that is how it was early in the clan wars in the war. The quirks seem to help, but it does seem that the clans have a leg up on the IS so far as technology is concerned. That may explain why I lost every CW yesterday.

[Vlad Striker]

First train yourself to press R targeting key every time you see the enemy. This gives you mates tactical info.
Second - watch your flanks.
Third - shoot enemy UAV ASAP. UAV looks like little blue star in the air.

[Thorqemada]

I started with a ~0.3 W/L back in the Beta - its normal...

[mogs01gt]

Void Angel, on 15 July 2015 - 11:11 AM, said:

Actually, Elo does not try to get you a 1:1 win/loss - that's an oversimplification that gets bandied around - and is responsible for many of the wrong ideas about Elo that you see. What Elo tries to do is assign you a rating that is consistent with your level of skill; it then adjusts this rating (and your opponent's) based on the results of matches you're in. Over enough time for the rating system to adjust, the hypothetical Best Player in the Game, having the highest Elo, would win most of his matches, but still receive very slight adjustments in his Elo - and maintain his rating as an average. Of course this doesn't account for the matchmaker and the fact that it is a team game. Contrary to what some people will try to tell you, Elo will work even for random teams, but there are distortions because of the size of your teams and the learning curve of this particular game. The way in which the matchmaker works (and in which Elo is applied,) is a much larger subject than pure Elo. However - the point here is that neither Elo nor the matchmaker actually try to force your stats to a 1:1 win/loss.

That is a tad bit wrong when it comes to MWO's MM. The MM tries to even out the ELO levels per team which translates into MM trying to make every match 50:50, meaning both teams have an equal shot at winning. This is what causes the stomps you see often because ELO isnt always the true sign of someone's playing ability. Matching team ELO is part of the issue.

[Anunknownlurker]

As Elo was developed for 1v1 games (e.g. chess) it does not translate particularly well into team games such as MWO. You can try to match individual players with similar Elo ratings; you can even try to create two teams of players with similar Elos but there are so many variables in a relatively small sample group such as the currently active MWO community, that it can never attain an even win:loss ratio even if that is the aim. Still, it sort of works at the moment and will do until something better comes along

[Void Angel]

mogs01gt, on 16 July 2015 - 12:44 PM, said:

That is a tad bit wrong when it comes to MWO's MM. The MM tries to even out the ELO levels per team which translates into MM trying to make every match 50:50, meaning both teams have an equal shot at winning. This is what causes the stomps you see often because ELO isnt always the true sign of someone's playing ability. Matching team ELO is part of the issue.

Void Angel, on 15 July 2015 - 11:11 AM, said:

The way in which the matchmaker works (and in which Elo is applied,) is a much larger subject than pure Elo. However - the point here is that neither Elo nor the matchmaker actually try to force your stats to a 1:1 win/loss.

The reason you get lopsided match results in some matches is not due to any one factor - and is certainly not solely the result of having a matchmaker. The cumulative effect of 'mech attrition plays a much larger role, as does a lack of team cohesion. The way that Elo is applied to the matchmaker is not perfect, and is under review by PGI, but it's not causing a significant number of imbalanced matches. Analysis of the matchmaker by PGI (the only party with demographic data on it) indicates that it mostly does predict match results correctly. There are some problems, but the system is not really broken, nor is it responsible for the perceived imbalance in matches that is most often attributed to it when players start talking about "the stomps." So, while the way that Elo is applied could (and will be) better, simply using Elo isn't problematic.

And again, the matchmaker isn't trying to set your win/loss ratio to 1. It. Does. Not. Care. What it cares about is getting you matches, preferably without having to wait five minutes then re-queue if you're a top player. As a secondary goal, it tries to get you matches with players near your own Elo rating, but it does not in any case try to point you to a certain win-loss ratio. In a theoretically perfect matchmaker, given enough time and players who all reach their potentials: yes, the matchmaker would produce roughly 1:1 win-loss for everyone. But, and this difference is important, win-loss is a side effect of the matchmaker, not its goal.


PS: Elo is not an acronym.

[Void Angel]

Wakashan, on 16 July 2015 - 08:28 AM, said:

Anyway, I think that is one of the primary issues I am facing. We are facing many well-coordinated teams and my friends seem to prefer talking to each other on TS instead of keeping the PUGs informed. Of course, we also have a problem with PUGs not listening at times, but I have seen a lot of really good players in PUGs.

I did notice, however, that clan mech technology is vastly superior to innersphere technology. Of course, that is how it was early in the clan wars in the war. The quirks seem to help, but it does seem that the clans have a leg up on the IS so far as technology is concerned. That may explain why I lost every CW yesterday.

Use the in-game VoiP to talk to the team during matches, or just detail someone to be the PuG Whisperer. That will help the Puggles coordinate, but if push comes to shove and they won't listen, best results usually still come from following along with them and supporting what they decide to do.

As for brawling, stick with it - it's a hard role to play these days, because the game inadvertently trains people to play at long ranges, but it's worth the pain. Players are rewarded instantly for long-range combat, and punished instantly when they screw up or get outmaneuvered trying to get to close range - but the rewards for close-range combat are both deferred, and inconsistent (because you often have to take fire or wait to get close, and if your team pulls back when you engage, you can just die for nothing.) This makes players gravitate toward long-range combat and wetting themselves at the first hint of damage. Consequently, as you've discovered, timing your engagement with a brawler is something of an art.

Actually, though, the Clans do not have an advantage overall - the two tech bases are actually very balanced. However, in PuG matches (including CW,) the operant conditioning I outlined last paragraph pushes players into the Clans' strengths and away from the Inner Sphere's. The Clans have massive advantages in overall damage, range, and the mobility of nearly all their 'mechs. They also have superior XL engines and the ability to swap hardpoints from any and all variants on a per-component basis. They pay for these by having longer burn times, burst-fire autocannons, and substantial heat inefficiency at close range - as well as somewhat constrained 'mech construction and customization (many of their internals are hard-locked: they cannot change their engines or the location of many of their internal components; even stuff like Endo-steel is hard-locked to a specific location internally.) For this reason, the hide-and snipe mentality of so many Puggles causes the Clans to overperform in that environment - in organized matches, it's another story. Just bear in mind that most successful engagements against the Clans involve close range.

PS:My main King Crabs are named Pinchy and Cakes, while my Hunchbacks are Quasimodo, Igor, and Lurch. You are welcome.

[mogs01gt]

Void Angel, on 16 July 2015 - 02:26 PM, said:

And again, the matchmaker isn't trying to set your win/loss ratio to 1. It. Does. Not. Care. What it cares about is getting you matches, preferably without having to wait five minutes then re-queue if you're a top player. As a secondary goal, it tries to get you matches with players near your own Elo rating, but it does not in any case try to point you to a certain win-loss ratio. In a theoretically perfect matchmaker, given enough time and players who all reach their potentials: yes, the matchmaker would produce roughly 1:1 win-loss for everyone. But, and this difference is important, win-loss is a side effect of the matchmaker, not its goal.
PS: Elo is not an acronym.

You dont get it. MWO's MM is trying to make matches even by using the teams composed ELO. When any MM is attempting to make the came even or "fair" that is directly making it a 50:50 shot or 1:1. If MWO 's MM was going on individual ELO and trying to match players up that way, then you would be correct. We simply do not have the player base for that to function.

The reason why this causes stomps is when one team gets an extremely high elo player and that player then dies or has a bad match, it drastically effects the out come of what MM was attempting to do. A new player has the exact opposite effect, it essentially makes the match 11v12.

Edited by mogs01gt, 16 July 2015 - 05:47 PM.

[Karamarka]

My Theory on MWO elo.

2 Teams: Average ELO = 1500

Team 1 has a 3000 elo player
Team 1 has a 1 elo player

Team 2 has 2 1500 elo players

if the 3000 elo player is playing bad today, or in a bad mech - kiss the game good bye.

I have played thousands of matches, and easily can hit 600+ dmg avg. But today, i got in a game where 4 of my team died in first 45 seconds, and 1 person even overheated and killed himself.

TLDR: Matchmaker has no elo brackets for individual players. It just averages the teams

In games like DOTA 2. The min and max elo between 2 players is limited - and the average elo for both teams are equal. Therefore, the worst player on team 1 is close to the best player on team 1 in terms of skills.

mogs01gt, on 16 July 2015 - 05:45 PM, said:

You dont get it. MWO's MM is trying to make matches even by using the teams composed ELO. When any MM is attempting to make the came even or "fair" that is directly making it a 50:50 shot or 1:1. If MWO 's MM was going on individual ELO and trying to match players up that way, then you would be correct. We simply do not have the player base for that to function.

The reason why this causes stomps is when one team gets an extremely high elo player and that player then dies or has a bad match, it drastically effects the out come of what MM was attempting to do. A new player has the exact opposite effect, it essentially makes the match 11v12.

Exactly my theory. And it's disgusting losing because of 3 suiciders.

Edited by Karamarka, 16 July 2015 - 06:15 PM.

[Void Angel]

mogs01gt, on 16 July 2015 - 05:45 PM, said:

You dont get it. MWO's MM is trying to make matches even by using the teams composed ELO. When any MM is attempting to make the came even or "fair" that is directly making it a 50:50 shot or 1:1. If MWO 's MM was going on individual ELO and trying to match players up that way, then you would be correct. We simply do not have the player base for that to function.

The reason why this causes stomps is when one team gets an extremely high elo player and that player then dies or has a bad match, it drastically effects the out come of what MM was attempting to do. A new player has the exact opposite effect, it essentially makes the match 11v12.

No, I understand what you are saying quite well - you're just insisting on a faulty idea about what the matchmaker does. The matchmaker is not attempting to make your win-loss ratio even; that is a side effect, but it's not the matchmaker's goal. Its' goal is to get you a match in a reasonable time, with as much Elo fidelity as possible. The difference may seem like splitting hairs to you, but precision in thought is important. Your theory is certainly plausible - but you're trying to stretch it over a lot more than it can actually cover. You're making an unsupported claim - given the information you have it's not even falsifiable. The most you can get out of your reasoning is that your scenario can at times occur, but you're arguing much more from that inadequate premise.

When you say "this causes the stomps you see often [emphasis mine,]" you're not only passing off a subjective claim as to frequency, you're also claiming that this is the primary (or only) cause of one-sided matches - feel free to correct yourself at any time, but so far that's what you've said. This is simply not correct; I'm sure that this scenario happens, but you do not have the data required to make more than that claim. I know you don't have it, because PGI hasn't released it - yet they have published metrics which indicate that Elo fidelity is actually pretty good, but could be better; it was in an announcement on the forums somewhere a while back.

In any case in order to get to this "every stomp is the matchmaker trying to use Elo," claim, you have to know what the Elo numbers are for everyone in the match; you don't have that information. You need to be able to quantify every player's performance, including things like an Atlas pilot who doesn't do much damage, yet soaks up a round or two of alpha strikes at a critical moment; you don't have that information. You also have to be able to discount alternate causes, like the cumulative effect of focus fire; you don't have that information. As well, you need to be able to deal with confounding variables, such as where along the bell curve team Elos lie, or how well the weapons mix on your team matches the map; you don't have that information. As far as I can see, you have a hypothesis about what might occur, and are simply interpreting your and others' anecdotal experiences toward that conclusion.

This is not a good argument, and telling new players that your guess is the reason lopsided matches happen is nothing short of misinformation.

Edited by Void Angel, 16 July 2015 - 08:50 PM.

[Void Angel]

Karamarka, on 16 July 2015 - 06:12 PM, said:

TLDR: Matchmaker has no elo brackets for individual players. It just averages the teams

Well - and this is good news, really - your hypothesis is incorrect. The matchmaker uses a target Elo to build both teams in a match. This means that while the availability of players close to that Elo may be limited based on a variety of factors invisible to us, the times you're actually matched with such a huge disparity of players will be rare along most of the Elo curve. If you are getting a lot of Elo mismatches (and you need to actually measure a large sample of all your matches and collate that data to be sure; don't just rely on your impressions) it means you're at one end of the Elo distribution curve.