233 replies to this topic

[Vasces Diablo]

Larger gap in skill level to accommodate weight matching concerns me a bit. This is a game where a good light pilot cans destroy a bad heavy pilot.

That said I'm glad that the tolerance is something that can be adjusted quickly and without patching.

[NinetyProof]

Deathlike, on 12 April 2013 - 12:21 PM, said:

I'm bored...

I wished ELO was based on the mech VARIANT, which would make comparisons to ELO from a Raven 2X or 4X DIFFERENT from a Raven 3L. They should be different, but they are valued the same, which, they are clearly not by design or balance.

Variant doesn't work either ... cause you can make the same variant work in different fashions (sniper, brawler, boat). You might be great at sniping, and terrible at brawling, and visa versa. So even tracking ELO by variant doesn't solve the issue.

To solve the issue you almost have to classify variant builds, then attach a weighting system to the build, then track the ELO per weighted variant ... for what? Should not really make much of a difference.

Why? ELO simply tries to predict if you will win or lose, and then sets how much you *move* based upon it, and it all works out in the end.

[Utilyan]

I better up my two small laser atlas to four small lasers!

[NinetyProof]

Vasces Diablo, on 12 April 2013 - 12:33 PM, said:

Larger gap in skill level to accommodate weight matching concerns me a bit. This is a game where a good light pilot cans destroy a bad heavy pilot.

That said I'm glad that the tolerance is something that can be adjusted quickly and without patching.

Wait ... a "good" light pilot would have a "good" elo ... and a "bad" heavy pilot would have a "bad" elo ... so your example should never really happen ... or, at worse, not happen much.

[One Medic Army]

Just going to point out that strict overall tonnage equivalency wouldn't work either.

Think which pair of mechs is generally more useful:
Dragon+Trebuchet (60+50=110)
Raven3L+Cataphract (35+70=105)

The second pair, despite being 5 tons less, is MUCH more effective in general.

Or in other words overall tonnage matchmaking means Mediums are actively gimping your team, since the other side gets 15 extra tons for Heavies/Assaults plus counters you with a 35ton light.

[Deathlike]

Wintersdark, on 12 April 2013 - 12:30 PM, said:

No, absolutely not.

Because most players don't play a lot of matches - I seem to remember... Garth? saying 6 matches/day was the average. At 6 matches per day, your Elo score will take months to settle unless you always play with the same mechs. Even assuming it was done by mech type not variant, it would still take ages, and change extremely slowly.

It's a balance between adjustment speed and accuracy. Weight class is really the best way to do it IMHO.

(Edit: to calculate Elo, that is)

Let's say you've grinded out the Ravens. This is fine. However, let's say you want to grind out a Spider. The Spider has a completely different playstyle than the Raven (particularly the 3L). So, if your best mech happens to be the 3L, your ELO when you are grinding for the Spider is LITERALLY the 3L... and that value will take a beating. It would take some time for "correct" that ELO and if you continue to grind with the 3L after grinding out the Spiders, you might actually start off in "easy land" (ELO having taken a beating). It's something you have to think about how one mech does not even translate to another. The 3L is not comparable to the 2X, and trying to equate these two as the same, is crazy... because under very limited circumstances where the 2X would actually be better than the 3L (ECM being the biggest factor).

NinetyProof, on 12 April 2013 - 12:35 PM, said:

Variant doesn't work either ... cause you can make the same variant work in different fashions (sniper, brawler, boat). You might be great at sniping, and terrible at brawling, and visa versa. So even tracking ELO by variant doesn't solve the issue.

To solve the issue you almost have to classify variant builds, then attach a weighting system to the build, then track the ELO per weighted variant ... for what? Should not really make much of a difference.

Why? ELO simply tries to predict if you will win or lose, and then sets how much you *move* based upon it, and it all works out in the end.

Do people actually buy duplicate mechs for this purpose? I will admit that if you have two different builds for the same variant, the system would not be accurate by any stretch. However, I'd think most players have one build for one variants and stick to that... and try to see if other variants can do the same job. Think Stalkers and Cataphracts, where most of the stuff you can do between the mechs are pretty much the same overall. I get the strong feeling that this is a less used "behavior" for most players, so I'm not dismissing your point outright, but it probably wouldn't be the common case.

Also, it may be difficult to "classify" a build.. you would have to institute some sort of arbitrary system and frankly it wouldn't be able to cover all possible cases.

Edited by Deathlike, 12 April 2013 - 12:50 PM.

[ohtochooseaname]

It took 2 months for them to figure out this problem? You know, people would have found it easily if they'd just have published ELO's. They'd see that their score went up too much/etc due to screen shots. We're beta testers: the more information they disclose the more we can debug.

[sarkun]

Skill based matching should be more important than weight class based one - I'd rather have a skilled teammate in a medium mech than a noob in an Atlas.

[NinetyProof]

Wintersdark, on 12 April 2013 - 12:27 PM, said:

With tonnage matching (+ keeping equal numbers, but ignoring actual classes) there's so much more value in bringing low-end mechs than there is currently.

What the heck are you talking about? You mean bringing say the worst spider chassis vs the best raven chassis?

Well Duh ...

Wintersdark, on 12 April 2013 - 12:27 PM, said:

Without tonnage matching, you're just going back to the place where the fellow who brings a Dragon just give the opposing team a 3D or dual AC20 Jager, etc. Awesome vs. Atlas, etc.

Maybe ... but if all things are equal, then there should be the same *chance* to get the same number of *bad* chassis on each team.

Wintersdark, on 12 April 2013 - 12:27 PM, said:

With tonnage matching, things are so much more interesting.

No ... with tonnage matching your restricting the *off* tonnage players to purgatory. Your basically telling the awesome that there has to be another awesome for him to play against ... and if no other awesome is queuing, he is not going to get to play.

or catapult ... or spider.

Anyways ... going to be better then what we have now that is for sure. It might be good ... and frankly, I don't know if "mirror" teams is really such a great idea anyways.

Do you really *really* want to have no diversity in team match up? Do you really want to look at your team and know the other team is the EXACT opposite EVERY ... SINGLE ... TIME?

Nah, a little team diversity is good.

[Asakara]

The population's Elo stats went from a "Perfect Bell Curve" on April 8th to "Top-Heavy" by April 12?

[One Medic Army]

Thontor, on 12 April 2013 - 12:54 PM, said:

with HSR, lights, even the RVN-3L, are not that dangerous.

Oh I agree, and I leg them all the time.
Still I fear and target Lights before Mediums, since I know the mediums can't run away and they the mediums almost never present a higher threat.

[p00k]

"Of the remaining 26% of matches, about 7% are what we would call "horrendously bad" (i.e., they carry the weight difference of one or two full assault Mechs between teams)."
the other 19% are what we would call "super horrendously bad" (i.e. they carry the weight difference of one or two full dropships of mechs between teams).

[FrDrake]

Asakara, on 12 April 2013 - 12:54 PM, said:

The population's Elo stats went from a "Perfect Bell Curve" on April 8th to "Top-Heavy" by April 12?

If an entire set of data is inflated it's still a bell curve.

[Faolan65]

All I ask for is to know where I fall on the bell curve, I don't need to know my ELO score, I just want to know where I'm at on the curve....

[Wintersdark]

NinetyProof, on 12 April 2013 - 12:51 PM, said:

No ... with tonnage matching your restricting the *off* tonnage players to purgatory. Your basically telling the awesome that there has to be another awesome for him to play against ... and if no other awesome is queuing, he is not going to get to play.

or catapult ... or spider.

Anyways ... going to be better then what we have now that is for sure. It might be good ... and frankly, I don't know if "mirror" teams is really such a great idea anyways.

Do you really *really* want to have no diversity in team match up? Do you really want to look at your team and know the other team is the EXACT opposite EVERY ... SINGLE ... TIME?

Nah, a little team diversity is good.

Wait, what? No, silly monkey, tonnage matching just means the total tonnage for each team is equal(or as close to equal as possible given player pool constraints), not that each 70ton mech means the other team has a 70 ton mech.

[Deathlike]

p00k, on 12 April 2013 - 12:56 PM, said:

"Of the remaining 26% of matches, about 7% are what we would call "horrendously bad" (i.e., they carry the weight difference of one or two full assault Mechs between teams)."
the other 19% are what we would call "super horrendously bad" (i.e. they carry the weight difference of one or two full dropships of mechs between teams).

So you're saying, they've been hiding the Fatlases from the Biggest Loser?

[Sagamore]

I'm guessing that tonnage caps on teams will come in with some kind of dropship mode. The dropship can only hold a certain maximum weight, right?

[EvangelionUnit]

yeah, 4 more days i can play other games !

[Wintersdark]

Deathlike, on 12 April 2013 - 12:47 PM, said:

Let's say you've grinded out the Ravens. This is fine. However, let's say you want to grind out a Spider. The Spider has a completely different playstyle than the Raven (particularly the 3L). So, if your best mech happens to be the 3L, your ELO when you are grinding for the Spider is LITERALLY the 3L... and that value will take a beating. It would take some time for "correct" that ELO and if you continue to grind with the 3L after grinding out the Spiders, you might actually start off in "easy land" (ELO having taken a beating). It's something you have to think about how one mech does not even translate to another. The 3L is not comparable to the 2X, and trying to equate these two as the same, is crazy... because under very limited circumstances where the 2X would actually be better than the 3L (ECM being the biggest factor).

Your example doesn't work. Barring the (comparatively rare) players who play LOTS of games, you'd basically always have new user Elo for every single mech you played. You just don't get sufficient matches to sort out your Elo if it were different for each mech.

It takes a LOT of matches to compensate for calculating Elo by team success/failure.

Sagamore, on 12 April 2013 - 01:04 PM, said:

I'm guessing that tonnage caps on teams will come in with some kind of dropship mode. The dropship can only hold a certain maximum weight, right?

If that ever happens. I can dream

[ElLocoMarko]

First of all, sounds like good stuff.

Second, there seems to be some resistance against weight class matching versus tonnage matching.
I see this as two different flavors of chocolate. It's still chocolate.

Edited by ElLocoMarko, 12 April 2013 - 01:30 PM.