41 replies to this topic

[wolfhnd]

This disaster reminds me that good intentions don't always lead to good results. It's a lot like trying to fit professional hockey players into an amateur game, it ain't going to be pretty and someone is going to get hurt. If you want the amateurs to play with the pros you will have to come up with some pretty extreme variation of the game rules and the current base capture mode is an example of how unsatisfying that is.

[Hauser]

SgtKinCaiD, on 27 March 2013 - 08:41 AM, said:

Hauser, you are stuck in it because the matchmaking make sure you get a 50/50 ratio of WL.

You're confusing cause and effect.

When matched against equals you'll win 50% of your games. They wouldn't be your equals otherwise. Now if you win when the matchmaker doesn't expect it your ELO goes up and you get matched against better players (or rather, more equal since you are apparently better too), hence start losing more, thus levelling out at 50%.

It's because your ELO changes that you should stay around 50% wins.


----


@focuspark, FromHell2k,

you wanna report that as a bug in the matchmaker. I don't think it has anything to do with ELO. It's just wrong.

[7 Nation Army]

As far as 8 mans dropping against pug teams is concerned as mentioned abovr, that was a bug a few weeks ago and isn;t a problem anymore, although if anyone here was around during closed beta that how it used to be.

I messaged the support staff today about the issue, I directed them to this forum thread and this was their reply:

"Hello 7 Nation Army,
The matchmaker isn't complete. Currently it is using ELO, but it's not doing any weight class balancing, which is pretty messy in some matches (8 Assaults vs 8 mediums/lights?) Add that to group launches and we end up with some potentially very unbalanced matches. It's definitely no fun when that happens, and this is why the matchmaker is still a work in progress as part of the open beta testing. This is the kind of thing we'd be unable to test properly in internal/controlled conditions.

Long story short: we know, the devs know, this is not the "finished" matchmaker yet. I can't say much about what's in the future other than re-introducing weight class balancing in some shape or form."


So atleast they know its an issue. I can imagine that programming this thing is a nightmare what with all the variables.

If what he said is true im looking forward to weight matching again, that alone would make a huge difference in weight balance. At least give me a friendly noob atlas to be my meat sheild right!

Edited by 7 Nation Army, 27 March 2013 - 07:24 PM.

[Space Odin]

The real problem with the matchmaking system right now is that not enough people play MWO. The matchmaker is designed to include weightclass balance and incorporate skill level. If the system was super strict about making the teams exactly equal in both skill level and weight class you'd be searching for a match for hours before actually getting to play one.

As for the OPs complaints about the system... Games were insanely one sided before ELO was implemented, nothings changed in that regard whatsoever. The only thing that's any different is there's no weight class balance anymore. Really thats the only change I've noticed playing about 5 hours a day since the ELO patch, other than playing with the same 1 or 2 guys all the time.

[focuspark]

Space Odin, on 27 March 2013 - 09:52 PM, said:

The real problem with the matchmaking system right now is that not enough people play MWO. The matchmaker is designed to include weightclass balance and incorporate skill level. If the system was super strict about making the teams exactly equal in both skill level and weight class you'd be searching for a match for hours before actually getting to play one.

As for the OPs complaints about the system... Games were insanely one sided before ELO was implemented, nothings changed in that regard whatsoever. The only thing that's any different is there's no weight class balance anymore. Really thats the only change I've noticed playing about 5 hours a day since the ELO patch, other than playing with the same 1 or 2 guys all the time.

My WP7 app has 2000+ unique downloads... given WP7's popularity there's no way MW:O doesn't have enough people. The Sunkworks desktop app has nearly as many as well. Last estimates I'd seen for MW:O were are 450K users. Hell the forums have more than that many accounts.

Hauser, on 27 March 2013 - 02:54 PM, said:

@focuspark, FromHell2k,

you wanna report that as a bug in the matchmaker. I don't think it has anything to do with ELO. It's just wrong.

done

[Space Odin]

focuspark, on 27 March 2013 - 10:14 PM, said:

My WP7 app has 2000+ unique downloads... given WP7's popularity there's no way MW:O doesn't have enough people. The Sunkworks desktop app has nearly as many as well. Last estimates I'd seen for MW:O were are 450K users. Hell the forums have more than that many accounts.


done

I have no idea what apps youre talking about. Link please? 450k registered accounts is an extremely small number. Consider how many players are actually playing at a time.

[Lyteros]

The Problem I see:
The whole ELO only consists of win / loss.

And those statistics will be created by the WHOLE team, not taking personal achivements of each pilot into account. So your personal influence on what happens with your ELO is 1/16 (since you're one single pilot of the team and the whole other team wants your head)

Since your teams (when PUG) are randomly drawn, the team based ELO rating is just as much as a gamble as the last system.

Spoiler

So the whole Pilot rating is actually a team-rating. If you PUG and drop often enough with bad players, they will ruin your stats and you will be far underrated, if you have no idea what you are doing but drop with good players you will be far overrated. You will be dragged along the majority.

Also the system will not at all take into account what mechs / fitting people have used and how good they play. Think about a Trial commando against a 8+ million custom-ecm-deathmachine. The stock Atlas K in trial against a custom frontline brawler atlas.

Imho it was an improvement, but after all it is not enough to change the problems.

The system will not at all take into account what mechs / fitting people have used and how good they play themself.

I think the system will not solve the mess we have with the matchmaking. The proposed ELO misses a lot of available and easy to use information, which can vastly increase the effectivity of the ELO calculation and the match making.

Some simple examples:

Spoiler

Teamgames:
Rating is taken into account, but NOT the mechs and equipment. So the ECM (or insert any upcoming I-WIN button here) madness will continue. The more FOTM the more likely the team wins with equal ELO.

Also a team full of lights will just be able to outcap a mixed mech team (in all currently available game modes) even if the other team had the better ELO rating.

PUG:
Got a bad luck drawing your teams? Getting all the overrated people in your team while the opposing team gets the underrated pro's - getting once again 8:2 roflstomped.

The "Mech setup problem" will also apply to PUG.

Proposed Changes / additions:

Spoiler

• Tonnage matching teams with fallback to class matching when no equal teams can be drawn, but stopping at 10% weight difference and rather stopping the process without a game, instead of creating a vastly unbalanced one.

• Matching Trialmechs to Trialmechs, newbies to newbies, 4mans to 4mans

• Looking into the chosen mechs of each team, trying to match comparable setups based on tech in the mechs and especially I-WIN modules like currently ECM. Also keeping in mind mech DPS / heat efficiency (0,6 heat efficiency is generally not a good setup...). This is not about dropping only equal fittet mechs, its about trying to close the gap a little. (if possible a BV system)
  
  
  
  
  Imagine a stock commando against a commando that has DHS, ENDO, FF, ECM, XL and a complete weapon swap. Something like the BV calculation of classic TT for each mech and trying to get comparable with it... Has not to be perfect, but if at least somewhat decent, it will improve the matchmaking a lot. The calculated value can already be put up in the mechlab so the matchmaker has very little to calculate (less server traffic and match setup time, instead it's spread on the mechlab, where people have all the time and calculation power of their own machine).

• Taking the statistics of the whole match into account, that are displayed on the endscreen and incorporating them into the ELO calculation, so at least some of your own actions modify your ELO, instead of beeing completely dependand on the team. (Like: Damage done, hit%, spotting, rate of your remaining armor on death vs maximum armor to see how good the damage scattering and piloting was)

• The lowest third of the winner team and the highest third should be checked for particularly high / low score. If you carry your team (or pull it down) you should not get the full effect of the teams result - you should either have a lower loss of ELO or a lower gain of ELO depending if you carry or pull. Lets say like... 1 or 2 people are having more then 40% of the total generated match score, or someone has less then 10% of the total score.

[Hayashi]

ELO seems to be working as intended, largely, if you run more or less the same chassis. I've noticed disruptions when I use vastly different mechs - the ELO I had with a Commando (probably the weakest mech in the game) led to a ridiculous killrate of 14 vs 2 the moment I played with a 3L (probably the strongest light in the game) for a couple of games and a 3.5 win-loss rate. But otherwise, the win/loss rate tends to even out over time as you keep playing a certain chassis.

The problem with unbalanced matches is not a problem with ELO per se. There are actually three problems here:

1. People disconnecting or not getting added to a team. This should be solved by booting the DC and allowing the MM to patch in extras at the last minute.

2. The class matching was removed. This isn't a ELO problem, but a problem with removing any form of equipment matching entirely.

3. The weapons and equipment are hideously unbalanced at present.

All an ELO MM can do is ensure your win loss rate is ~50%, evening odds between matches, but it does not seem to do much for evening odds within matches.

Some setups are just significantly stronger than others (a COM-2D streakboat will always win a COM-3A pilot of equal skill, etc), and in other cases some are specifically weak in situations where others are strong.

For instance, a team with 8 Splatapults will be SOL if they drop on Alpine against a team of 8 Gausscats.

The effect of having unbalanced equipment is that your chance of winning an opponent of a certain loadout depends greatly on that loadout's effectiveness against yours, instead of skill. This makes kills/wins feel like 'it will happen anyway' and deaths/losses feel like there was no chance to win anyway. Which unfortunately, may be largely true in some cases.

Can you think of a game where the odds of winning are 50%, but every single game tends to be hideously unbalanced?

Yes, it's scissors-paper-stone.

[focuspark]

Space Odin, on 27 March 2013 - 10:28 PM, said:

I have no idea what apps youre talking about. Link please? 450k registered accounts is an extremely small number. Consider how many players are actually playing at a time.

Skunkworks: MechLab Utility for Desktop
Skunkworks for Windows Phone

[RoboPatton]

I've been on teams where I'm just yelling at my screen and cursing everyone. Of course, I keep it to myself, but Fail teams gonna fail.

Would be nice if there were a ranking system for the pilots that would factor in. Just thinking of the algorithms needed to pull that off makes me dizzy. Would be a nice feature though.

[SgtKinCaiD]

Space Odin, on 27 March 2013 - 09:52 PM, said:

The real problem with the matchmaking system right now is that not enough people play MWO. The matchmaker is designed to include weightclass balance and incorporate skill level. If the system was super strict about making the teams exactly equal in both skill level and weight class you'd be searching for a match for hours before actually getting to play one.

That's exactly the problem they corrected shortly after the ELO ratings system was introduced : long waiting time.
And they take a shortcurt instead of resolving the problem correctly : mixing high ELO and low ELO players to have an average ELO ratings that is easier to match with.
They should have put instead a solo only PUG queue which would have resolved many problems like PUGStomping, grinding XP and cbills, testing new mech builds, etc.
Now they are stuck with the ELO ratings system, spending time on it to make it works, time that will not be spent on CW or other features.

[ohtochooseaname]

Since the new stat taking was implemented, I have over 400 games played, and my win/loss ratio is over 1.7. That is over a 63% win rate in lights and mediums with a significant number of games played. I solo pug 100%. How does that happen, i wonder?

[Hauser]

ohtochooseaname, on 28 March 2013 - 02:00 AM, said:

Since the new stat taking was implemented, I have over 400 games played, and my win/loss ratio is over 1.7. That is over a 63% win rate in lights and mediums with a significant number of games played. I solo pug 100%. How does that happen, i wonder?

Two options:

1. You might be in the top of the elo population at the time you play.
2. You are not yet at your proper elo level.

Do you win allot by capping in lights?

SgtKinCaiD, on 28 March 2013 - 01:02 AM, said:

And they take a shortcurt instead of resolving the problem correctly : mixing high ELO and low ELO players to have an average ELO ratings that is easier to match with.

Just a nitpick but the matchmaker doesn't try to average the elo level out. It starts with a specific level in mind and starts finding players for that level. As it has to wait longer and longer for suitable players to show up it starts to get less and less picky about elo level and weight class.

Elo level is only averaged when a group drops in. The matchmaker then uses the average elo level of the group for each player.

Edited by Hauser, 28 March 2013 - 04:00 AM.

[Sinthrow]

The Problem that I see with the current ELO system is that it is scored by Weight class not by varient.

It is easy for me to see that
spider 5d is not = to spider 5v yet I have the same rating in elo lol
so if in one game I take my ecm 2ml 1LL 5d into a match I bring more to that game..
next game I take my 1LL spider 5v I am not as much help to the team as before but I have the same ELO

This is the part of the business I say.. Please rethink what you do before you do it. Just because it is easier to set up or code doesn't make it right.

[ohtochooseaname]

Hauser, on 28 March 2013 - 03:54 AM, said:

Two options:

1. You might be in the top of the elo population at the time you play.
2. You are not yet at your proper elo level.

Do you win allot by capping in lights?




Just a nitpick but the matchmaker doesn't try to average the elo level out. It starts with a specific level in mind and starts finding players for that level. As it has to wait longer and longer for suitable players to show up it starts to get less and less picky about elo level and weight class.

Elo level is only averaged when a group drops in. The matchmaker then uses the average elo level of the group for each player.

I don't usually win by capping. in fact, if I lose, it is disproportionately by capping, I think. I think I actually have a better win rate at higher population times: when I play between midnight and 2 am PST, the games tend to be more one sided, so it is more like a 50-50 win rate.

[focuspark]

RoboPatton, on 28 March 2013 - 12:26 AM, said:

Would be nice if there were a ranking system for the pilots that would factor in. Just thinking of the algorithms needed to pull that off makes me dizzy. Would be a nice feature though.

Google (or Bing) "ELO".

Sinthrow, on 28 March 2013 - 06:52 AM, said:

The Problem that I see with the current ELO system is that it is scored by Weight class not by varient.

It is easy for me to see that
spider 5d is not = to spider 5v yet I have the same rating in elo lol

That's because the devs actually believe ECM is balanced.

[Mechteric]

focuspark, on 28 March 2013 - 08:46 AM, said:

That's because the devs actually believe ECM is balanced.

http://mwomercs.com/...-matter-of-ecm/

Quote

We are still analyzing data and examining how ECM impacts the game. We understand that some people love it, and some people hate it. That's exactly the polarizing type of content we want! However we believe it's not fully balanced yet, and are looking at ways to address community concerns, while maintaining the integrity of our vision for MWO.

[Davers]

The problems the OP is having is due to unbalanced teams, and the prevalence of high alpha builds. If your team's Atlas goes down in 2 volleys he isn't getting much chance to do damage himself. Sometimes you just turn the wrong corner and it's game over. Once you are down 2 or more mechs, especially if they didn't do too much before they died, it will be a very uphill battle to win.

[AnEnragedGamer]

I've actually quit playing because of the horrendous matchmaking.

After taking down two Atlas DDCs in a monumental battle with my C4 splatcat (hate all you like, but they rend assault mechs with good positioning), and then getting killed by the third on their team, and having the rest of my team killed by their Stalker. Because the rest of my team consisted of heavies and mediums.

Nobody can call a system that can match three DDCs against a team of heavy and mediums with no ECM, fair. I've had friends say they've been in matches against FOUR DDCs.

Until they fix it, I no play.

[focuspark]

CapperDeluxe, on 28 March 2013 - 12:06 PM, said:

http://mwomercs.com/...-matter-of-ecm/

Hardly helpful. They nerf'd the hell outta missiles because a feature most people didn't want had a bug but after months of complains about ECM they're thinking about maybe looking into the issue. Yeah, no. Devs don't have a clue about balance and have proven that they just don't give a damn about it as well.