30 replies to this topic

[Evex]

Ever since I've heard of MW:O I've had a theory about how the game would go about creating matches. This is what I've come up with so far based on what we know about the game. The queue will sort through players based on affiliation. This means priority is given in order to Houses, Merc Corps, and finally Lone Wolfs. The queue then searches through the waiting players sorting them out by either BV or tonnage.

For example the match making program would sort through players of a house first, and assign them to one team. When the requirement isn't meet for any house players either by BV or tonnage. Then the program moves on to merc corps repeating the process, if no one in a merc corps doesn't meet the criteria then the program looks at the lone wolfs. This is done until you have two teams of what ever the size of the battle calls for. In short this makes sense when you consider how things would work in mechwarrior/battletech.

[Kanatta Jing]

Paul Inouye, on 14 May 2012 - 12:19 PM, said:

Yes, we have deemed that all your friends are not worth playing with and will hence hook you up with people our match making system determines as "Cool dudes".

Spoiler

Yes, you can create groups with friends and join a match.

Yes I was being sarcastic.

Yes, I still think your friends are not worth playing with.

You will be able to set up lances of your friends to play with.

Probably by all joining the same Merc or House Company.

[Blind Baku]

I think there should be an attractive lance that has to choose between three bachelor lances that ar hiding behind curtains...

[BenEEeees VAT GROWN BACON]

Match making will be based on the ramblings of an old hag reading a star chart. I know this to be true.

[Aedensin]

BenEEeee's VAT GROWN: BACON, on 15 May 2012 - 04:42 AM, said:

Match making will be based on the ramblings of an old hag reading a star chart. I know this to be true.

Made me think of the elders from 300 lol

[Woodstock]

As a serious answer to the OP

I quiet like your idea. My only addition would be that the system would group 'Lances/companies' together in its calculations.

So It would count 3 people from a merc corp that have 'Formed Lance' as effectively one entity.

So long as the value of each Entity is calculated by BV and that the sides DON'T have to be the same number of players, I think this system would work smoothly and quickly.

My only thought would be is how this would work when we consider the persistent meta game that will run over everything with contracts for worlds etc etc.

[Snotling]

Woodstock, on 15 May 2012 - 04:54 AM, said:

As a serious answer to the OP

I quiet like your idea. My only addition would be that the system would group 'Lances/companies' together in its calculations.

So It would count 3 people from a merc corp that have 'Formed Lance' as effectively one entity.

So long as the value of each Entity is calculated by BV and that the sides DON'T have to be the same number of players, I think this system would work smoothly and quickly.

My only thought would be is how this would work when we consider the persistent meta game that will run over everything with contracts for worlds etc etc.

Hmm, i like the idea. asymetrical matches wold make for a lot of variety.

But determining the BV of the Mechs/Players will be very hard. (I think nobody really thinks you could just take the BV from tabeltop. It just doesnt apply to a PC Game)

Also, BV of peolpe joining in groups should be higher, as they will be more effective most of the time.

Edited by Snotling, 15 May 2012 - 05:16 AM.

[Woodstock]

Snotling, on 15 May 2012 - 05:14 AM, said:

Hmm, i like the idea. asymetrical matches wold make for a lot of variety.

But determining the BV of the Mechs/Players will be very hard. (I think nobody really thinks you could just take the BV from tabeltop. It just doesnt apply to a PC Game)

Also, BV of peolpe joining in groups should be higher, as they will be more effective most of the time.

I think the BV from TT would work as an excellent starting point and once it is coded into the system then it would take a fraction of a second to calculate. Something that would be handled by the mechlab when you customise your mech. The BV should be listed on that page somewhere for every change you make.

Your 'lances' BV would also be listed in your 'Lance/company window'

But I think Asymetrical teams would be an excellent feature. How often do battles in the books etc have even numbers of opponents!

As far as Group BV being higher ... that all depends on the quality of voice comms the game provides ... if we are all linked via voice comms as part of the game ...the advantage of a group would be less.

Edited by Woodstock, 15 May 2012 - 05:24 AM.

[Redshift2k5]

People operating as a tight-knit team will always have an advantage over pure random groups. The solution is to try to pitch similar numbers of similarly sized groups on each team.

[Woodstock]

I play regularly with 3 other people. And just because we know each other and how we fight is a long way from describing us as a 'Tight nit' team. More a doddering arthritic dad's army!

A group does not = more effective. It does = potentially more effective.

But I concede your point.

[Redshift2k5]

Potentially more effective is much more accurate, I agree.

[MaddMaxx]

Redshift2k5, on 15 May 2012 - 06:15 AM, said:

Potentially more effective is much more accurate, I agree.

Thus how would you code "potential" into any Matchmaker and feel as if it would be "correct" at all?

I also like the idea but cannot see, in the early days for example, how any code will truly be able to determine how "good" a preset Lance/Group will play. What if that same Lance changes one member but the Lances Mech selection remains intact. How in hell does any system know, who this new guy is?

And if anyone says it will work out with "time played" then that is fine. Anyone who complains that they are getting stomped the first 2-3 months will have to stay silent, as obviously not enough time has passed to level the filed via MM.

I am not arguing against the thought, just the use of the apparent term BV + "Skill". One is measurable, the other, unless using Ladders, not so much....

Edited by MaddMaxx, 15 May 2012 - 06:36 AM.

[Redshift2k5]

differentiating between pre-made teams vs all randoms is, in my opinion, more important than differentiating between pro teams and mediocre teams. People queueing as a group of 4, 5 ,6, or a dozen, should be matched against similarly sized teams. Multiple small groups or Ungrouped players would fill in the rest of the team, against with a bias towards similar size of clusters facing off across the battlefield.

worst case scenario, you pit good pre-made cluster vs a bad pre-made cluster, which won't be any worse than the same skilled premade cluster vs random ungrouped players.

I don't mean to assign a value to premade groups or premade teams, assigning a value would then mean you have to balance that value possibly with weird numbers of players, or saying 'we get more BV worth of mechs because you guys have a bigger team'.

Just a selection bias to put similar sized groups together. This would be especially noticeable for groups queueing as a full team of 12, who should ideally be pitted against other teams of 12 first, then if none are available, 9 or 10 and some randoms, or two 6's, etc.

[AlanEsh]

Snotling, on 15 May 2012 - 05:14 AM, said:

Also, BV of peolpe joining in groups should be higher, as they will be more effective most of the time.

No NO NO.
You can't code for player skill or group organization.
a ) it's never going to be "accurate"
<-- (die smiley!) b ) you're penalizing players grouping up as friends.

Just don't do this, please.

Edited by Angelicon, 15 May 2012 - 07:08 AM.

[Snotling]

In the beginning everybody will have the same tec anyways, so just match everybody vs a similar sized mech (+-5 tonns) + the group thing redshift explaned. I bet this wohl word fairly well. Then tweak it from there.

Edited by Snotling, 15 May 2012 - 07:36 AM.

[Aedensin]

Snotling, on 15 May 2012 - 07:18 AM, said:

In the beginning everybody will have the same tec anyways, so just match everybody vs a similar sized mech (+-5 tonns) + the broup thing redshift explaned. I bet this wohl word fairly well. Then tweak it from there.

The problem with that is it takes the light/medium scouts out of the heavier matches, that's not very fun.

[Snotling]

Aedensin, on 15 May 2012 - 07:26 AM, said:

The problem with that is it takes the light/medium scouts out of the heavier matches, that's not very fun.

why? lets say your team takes 3 30ton mech, 5 50 ton mechs, 3 70 ton mechs and 1 atlas.
so your opponent team gets 3 mechs 25-35 tonns, 5 mechs 45-55 tonns, 3 mechs 65-75 tonns and one 95-100 ton mech.

as you will likely just get trown in a team in pub matches (if everybody could join a server, then choose his mech, then change it around a dozend of times......we wold se no playtime at all), it should work fine

Edited by Snotling, 15 May 2012 - 07:37 AM.

[Aedensin]

Snotling, on 15 May 2012 - 07:36 AM, said:

lets say your team takes 3 30ton mech, 5 50 ton mechs, 3 70 ton mechs and 1 atlas.
so your opponent team gets 3 mechs 25-35 tonns, 5 mechs 45-55 tonns, 3 mechs 65-75 tonns and one 95-100 ton mech.

That sounds more reasonable, the "(+-5 tons)" thing you mentioned made it sound like you would see matches of nothing but heavies or nothing but lights.

[Sassori]

Snotling, on 15 May 2012 - 07:36 AM, said:

why? lets say your team takes 3 30ton mech, 5 50 ton mechs, 3 70 ton mechs and 1 atlas.
so your opponent team gets 3 mechs 25-35 tonns, 5 mechs 45-55 tonns, 3 mechs 65-75 tonns and one 95-100 ton mech.

as you will likely just get trown in a team in pub matches (if everybody could join a server, then choose his mech, then change it around a dozend of times......we wold se no playtime at all), it should work fine

And then nobody can play Cicada's or Dragons or Awesome's as they're not within the tonnage limits.

You're better off grouping by class and saying 3 Lights, 5 Mediums, 3 Heavy, and 1 Assault. Just my two cents.

[Banditman]

You certainly can factor "potential" into your matchmaking, you simply have to come up with a way to measure it. In this day and age, frankly, it's not that hard.

Any game worth it's salt will be tracking and maintaining a myriad of stats about each match played. Believe me when I tell you, game developers these days KNOW what happens in their games. It's especially true when the developer is the sole host for all matches.

I would certainly hope that when the game launches, I'll be able to look at my pilot profile and see interesting stats like my win / loss record, number of kills of any given type of mech, what mechs I pilot most often, etc. In general, I should be able to find out pretty much anything about myself. What that means is that it's being tracked, and once it's tracked, it can be used.

I'm fairly confident that given all of this data a good developer can come up with a formula for combat effectiveness for each pilot. Present it to us much like a baseball batting average. It's certainly not going to tell you who would win any given engagement, but it would be a good general indicator of what to expect.

You might even see this extended to player run units, for instance only applied when a full lance of players from team "whatever" drops together. There is so much potential here it's silly not to think far, far ahead.

Edit: As far as tonnage limits go, frankly, that should be completely on the shoulders of the developers to make it work. They are telling us that they want battles to be diverse . . . great! So do most of us! However, it is encumbent upon them to make that happen though incentive and disincentive leveraged by the game, not from some silly, arbitrary drop weight limit.

Edited by Banditman, 15 May 2012 - 10:06 AM.