43 replies to this topic

[Agent 0 Fortune]

Elo per chassis not weight class, possibly per variant.
Elo modifiers for skill unlocks
Elo boundaries, limit the min and max Elo boundary in a match, if match cannot create, then drop to 8v8 as opposed to grabbing players outside the boundary.

Obviously this will only impact Elo Dependant queues, but considering 85% of the playerbase is PUGs, seems like a reasonable consideration.
Using Elo to offset bad mechs will work in tandem with quirks, so development will need to determine if every mech must be CW competitive level or if mechs will fall in the Elo bracket as determined by pilot skill and chassis.

Edited by Agent 0 Fortune, 03 April 2015 - 10:16 PM.

[Xetelian]

I would vote for all this.

My HGN piloting may be roughly the same as my DWF but the fire power difference is tremendous and I have ran standard and XL...XL is not a good time but it nascars so much better.

Don't go as far as per variant but seperate my AWS Pretty Baby drops from my DWF drops might satisfy some balance.

[Virgil Greyson]

I agree on the 'per chassis' part at least. I am not as good in my other mediums as I am in my HBK for instance, would be nice if the game took that into account.

[A Large Infant]

the matchmaker needs to take into account combat value of the equipment brought, period.

if you choose to bring a 19kph, armorless, weaponless stalker you should have hypothetically guaranteed a spot for someone else on your team, who is the strongest player in MWO using the strongest equipment, to compensate

Edited by Meeso Thorny, 04 April 2015 - 12:34 AM.

[Elizander]

Agent 0 Fortune, on 03 April 2015 - 10:06 PM, said:

Elo per chassis not weight class, possibly per variant.
Elo modifiers for skill unlocks
Elo boundaries, limit the min and max Elo boundary in a match, if match cannot create, then drop to 8v8 as opposed to grabbing players outside the boundary.

Obviously this will only impact Elo Dependant queues, but considering 85% of the playerbase is PUGs, seems like a reasonable consideration.
Using Elo to offset bad mechs will work in tandem with quirks, so development will need to determine if every mech must be CW competitive level or if mechs will fall in the Elo bracket as determined by pilot skill and chassis.

Chassis/Variant with under 25 games played: -ELO Modifier
Chassis/Variant without DHS upgrade: --ELO Modifier
Trial Mech: --ELO Modifier
Chassis/Variant without Basic Skills filled out: ---ELO Modifier
Chassis/Variant without Elite skills filled out: --ELO Modifier
Chassis/Variant without modules: -ELO Modifier
Unmodified Stock Mech: ---ELO modifier.



Subtle P2W version:

Hero Mech: ---ELO

Edited by Elizander, 04 April 2015 - 12:51 AM.

[XX Sulla XX]

It is hard to quantify Elo from game to game. Some games I am awake and in a TW and trying. Some games I am in a TW build trying something out and have not had any sleep and just do not care that much. Makes it a bit hard for the match maker.

But yes by chassis would help or even varient. Varients vary at least as much as chassis does in how viable the mech is.

[A Large Infant]

I confess that when I saw the thread title I assumed the topic would be how to play matches that are more fun:

Start a game then quit out of it in all owned mechs except the one you wish to play
Assuredly lose a greater amount of them than you would have had you been playing
Take Elo hit
???
Enjoy clubbing baby seals in favorite mech

[cleghorn6]

I don't understand. What problem are you trying to fix?

[Sjorpha]

Meeso Thorny, on 04 April 2015 - 12:31 AM, said:

the matchmaker needs to take into account combat value of the equipment brought, period.

I don't think that's possible, there is no way the MM could intelligently evaluate how good a build is.

It would also create another layer of exploits, where you learn how to fool the MM into underestimating your build. (For example dropping in a SHS dual gauss Jager etc.)

ELO per chassis? Maybe a good idea, but a good or bad player is still a good/bad player even in a newly bought mech. I think if you implement ELO per chassis you should start them at that players average ELO minus something rather than starting at the default value.

Stock mode with separate ELO would be sweet if the playerbase was larger, but I don't think it can handle additional splits right now.

[Aethon]

Or, make it simple and effective:

Remove Elo, match by weight or BV.

[A Large Infant]

Sjorpha, on 04 April 2015 - 06:06 AM, said:

I don't think that's possible, there is no way the MM could intelligently evaluate how good a build is.

It would also create another layer of exploits, where you learn how to fool the MM into underestimating your build. (For example dropping in a SHS dual gauss Jager etc.)

ELO per chassis? Maybe a good idea, but a good or bad player is still a good/bad player even in a newly bought mech. I think if you implement ELO per chassis you should start them at that players average ELO minus something rather than starting at the default value.

Stock mode with separate ELO would be sweet if the playerbase was larger, but I don't think it can handle additional splits right now.

Just because the system is imperfect, does not mean it is ineffective or even an improvement over the current system. Player skill * combat value of equipment is the only way

[A Large Infant]

Additionally, a system similar to Elo can apply to equipment. How often was a certain build involved in a win. more details to follow.......


http://mwomercs.com/...e-value-system/

http://mwomercs.com/...ttle-value-dbv/

[Davers]

Having a different Elo score per chassis would eliminate the need to track individual components. Plus tracking components might lead to weird situations, like if the medium laser was the most 'expensive' weapon because it gets more kills than anything else...

[Ted Wayz]

ELO doesn't account for discos because the cause of disco cannot always be determined. A ladder system would as the chronic discos would eventually fall to the bottom tier.

A single disco happens in 50% of the fights and determines 90% of the outcome.

[A Large Infant]

A different Elo score per chassis leads to the currently dominant 6 large laser stalker being swapped for a 19kph, armorless, weaponless stalker with no penalty.

[Alistair Winter]

Kind of hard to get an accurate Elo value based on <50 matches, so if it was a separate value per chassis then you would see absolutely all kinds of players, from 3-year veterans to completely new guys, in the same matches very often. People who love to buy and master all the mechs in the game would spend pretty most of their time playing with a newly reset Elo.

Per chassis would be better. Although you could probably take (Variant Elo rank + Chassis Elo rank + Weight class Elo rank) / 3. In this case, elite veterans who start out with a new chassis would still not be at the same level as completely new players. And by the time you get to your second, third or fourth variant for that chassis, the game has a pretty good idea where you're at with that mech.

It would be nice if the matchmaker didn't treat my Locust as if it were as dangerous as my Raven or Firestarter, I'll give you that.

[Molossian Dog]

Forget all the other factors and calculations. Match players by numbers of games played.

Not by totals of the respective teams, mind you.
10 noobs plus two vets is not the same as 12 average players.
This is the black magic the ELO-Matchmaker does with other stats. And it sucks. Exhaust fumes.

Just try to keep all pilots with 0-100 games played together.
Those pilots with 101-500 games played together.
Etc

No need to hang an albatross of a complete noob team around the vet´s neck because he won "too much" lately.
No need to throw the noob to the wolves, because your team needed an ELO sink.

Whoever told PGI that an enforced win/loss ratio of 1,1 for all is
-achievable
-desirable
-fair
-fun
- or profitable
should be beaten up in some dark side alley. ELO hell is hell. Hence the name. For noobs and vets alike.

.
..
...

Edited by Molossian Dog, 04 April 2015 - 05:38 PM.

[Deathlike]

Just for random offtopicness, did you know "ELO" is a manufacturer of monitors?

At this point, I don't worry too much about it. Just know that if you succeed a lot more than you fail, your Elo is well reflected.

The most skewing aspects is solo vs group, as they are literally one and the same. When you succeed more in the group queue, your ELO would be very skewed in the solo queue (unless, you are that elite).

Just fair warning on that.

[Roadkill]

Meeso Thorny, on 04 April 2015 - 02:35 PM, said:

Player skill * combat value of equipment is the only way

Close, but not quite.

Each player can have an Elo rating for each piece of equipment when used on every variant. Yes, that's a crazy number of Elo ratings for each player.

Your formula assumes that the combat value of a particular weapon is the same for everyone, which is obviously not true. I totally suck with the Gauss Rifle because I simply cannot get used to the charge mechanic. Lasers are another example that's pretty obvious - some people are very good at holding lasers on target for their entire burn time, while others use them to spray & pray.

Your formula would probably be better than what we have, but it wouldn't be as good as it could be. And if we're going to start changing code, we might as well go whole hog.

[A Large Infant]

Alistair Winter, on 04 April 2015 - 04:30 PM, said:

Kind of hard to get an accurate Elo value based on <50 matches, so if it was a separate value per chassis then you would see absolutely all kinds of players, from 3-year veterans to completely new guys, in the same matches very often.

Does this not already occur?