5 replies to this topic

[MadTulip]

Ok i thought i spend 5mins on this.

The idea is to install an algorythm ment to balance teams by using dynamic weights for all involved mechs and theire components. its similar to ELO in the working principle, but it rates builds over the whole player base instead of rating the individual player.

So lets call it dynamic battle value (DBV) instead of ELO for each component.

Basic idea, same as in ELO is that if you win with your build that build gets rated higher for everyone. If you loose it gets rated lower. this can be done on a chassis basis (to corse imo) or on a component basis (DBV for each medium laser no matter the chassis) or imo even better on a component relativ to a certain chassis basis (i.e. DBV for medium laser on HBK-4SP). you could further group the same components like.:

on HBK-4SP:
DBV for 1 medium laser = 1000
DBV for 2 medium laser = 2020
DBV for 3 medium laser = 3500

So if everyone wins with a HBK-4SP with 7 medium lasers that build gets rated higher and higher until it has to fight against only the best say Timberwolf builds. Then HBK-4SP will loose a bit more and it gets cheaper until the optimum DBV is reached.

games are started with both teams having about the same ELO and about the same DBV. you can even just multiply ELO and DBV and match that.

PROs:
- you have a high number of games each day -> the DBV balances in to the right value very fast.
- you do not need to spend any more time on balancing (read "messing things up") anything as it autobalances itself no matter how you mess it up.
- the "metagame" is reduced to something like a living market. Imba builds will just stop to exist 10000 games after they have been introduced (which is probably in the range of a few hours counting all games) they are already that expansive in DBV that they are not worth it anymore.
- you can read out the DBV to see where you have to balance and how much.

CONs:
- you will earn less money by putting out the new imba things for sale every other month solely based on the profit intention.

- dampen things a bit so the stuff doesnt oszillate all the time. i gues its already included in the elo algorythm.
- elo should be visible to everyone
- dbv should be visible for every component.
- current number of games and players should be visible (cowards ).

ok that was 10minutes. maybe employ someone who knows how to implement that. but i guess you know yourself but just dont want to cause of the point listed under contras.

Edited by MadTulip, 25 June 2014 - 10:58 AM.

[101011]

The problem with this is that a member of, say, House of Lords would do far better than I would in a 7 ML HBK.

[MadTulip]

101011, on 25 June 2014 - 11:02 AM, said:

The problem with this is that a member of, say, House of Lords would do far better than I would in a 7 ML HBK.

yea but he would in any chassis. lets say hes better by factor 1.5 then you are in every mech. his elo is higher by simply put 1.5 then also. this is why i said:
"multiply ELO and DBV and match that"

so that is not a problem for the statistics.

you could also think that he better player plays against other players who are as good as he is. on the higher level theire win chances are then even again and it comes down to the same problem of balancing the equipment used.

and even if an imba mech is only imba for people above a certain elo then the DBV values could be tied to the elo tiers. so that there is one dbv for a medium laser on a HBK-4SP for a player who has an ELO of 1500 to 1600.

Edited by MadTulip, 25 June 2014 - 11:15 AM.

[101011]

MadTulip, on 25 June 2014 - 11:13 AM, said:

yea but he would in any chassis. lets say hes better by factor 1.5 then you are in every mech. his elo is higher by simply put 1.5 then also. this is why i said:
"multiply ELO and DBV and match that"

so that is not a problem for the statistics.

Apologies, I missed that line.

Quote

and even if an imba mech is only imba for people above a certain elo then the DBV values could be tied to the elo tiers. so that there is one dbv for a medium laser on a HBK-4SP for a player who has an ELO of 1500 to 1600.

I am unsure as to how this would actually balance anything. PPC and AC/5's will still move at roughly the same speed, they will still apply all their damage to the same point, poptarts will still poptart, and the DS will still be king.

[MadTulip]

like in this example.:

team one has:
- JNR worth 1000 DBV
- JNR worth 1000 DBV
- JNR worth 1000 DBV
- JNR worth 1000 DBV
- DS poptart worth 21000DBV
= 25000DBV

team two has:
- ATL worth 5000DBV
- ATL worth 5000DBV
- ATL worth 5000DBV
- ATL worth 5000DBV
- ATL worth 5000DBV
= 25000 DBV

i think the first team would loose and the price of the jenners and dragon slayer will afterwards be slightly reduced. As the DS is just not ~4 times better then an ATL. its just an also unfair and expreme example, but there is a fair ratio and by just counting wins and losses this ratio will auto balance just like the price for every product automaticaly balances on a free market

the good point is that it doesnt matter anymore if something is extremly good as it will to the same ammount get also extremly expansive to bring it and both teams "spend" the same ammount of dbv. so if everything is worth exactly the same value as it is good then it effectively doesnt matter what you bring as your oponent gets the same "value" -> balanced.

Edited by MadTulip, 25 June 2014 - 11:54 AM.

[Roland]

I actually laid out a detailed description of a system that I think is what you're trying to get at, and actually implemented the algorithm and generated test results.

The link to the thread is in my signature (second link) if you are interested. Feel free to comment on it if you like.