Jump to content

[Idea] Apply Elo Algorithm To Mech Variants


No replies to this topic

#1 Star Captain Obvious Kerensky

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 500 posts

Posted 27 February 2013 - 06:47 AM

Lets face it, not all mech variants are equal. The Raven 3L is superior to the Raven 4x and 2x. While its possible to do well in the "other" ravens, you are putting yourself at a disadvantage. If you're in a Commando or Spider, you're at an even greater disadvantage to the Raven 3L.

Similarly, The Atlas DDC is better than the Atlas D. The mechs are very similar hardpoint wise, except one has ECM.

Obviously a skilled pilot can excel in these less-popular variants. I am not here to debate which variant is better than the other. Instead, I am suggesting that PGI use the ELO algorithm to determine if a variant is better than others. We will get most of the benefits of Battle Value, but without having to try and balance each piece of equipment, the players will do that for you.

How to apply the ELO system to the mechs:
  • Record the Wins/Loss rate of each variant, just like is currently performed with players and assign an ELO score.
  • For each weight class (Light, Medium, Heavy, Assault) determine the average ELO value.
  • Calculate the variation from the average for each variant. This is the Variant Multiplier
  • When the Matchmaker sets up a match. Apply the Variant Multiplier to the ELO score for each pilot
  • When the match ends, repeat step #1.
Trial mechs do not count towards this ELO score, since players are only able to pick from 4 mechs out of dozens, and are unable to configure them. Trial mechs would artifically inflate/deflate ELO scores of whatever variant happened to be a trial mech for the week.

Here is an example:

Lets say that for the Heavy weight class, the Average Variant ELO score is 1000. And the Catapult variants have the following ELO scores. Note: These are made up values that are oversimplified, and show far more distribution in variant quality than we would see in practice.

A1 - 1500
K2 - 1250
C1 - 900
C4 - 800

This would result in Variant Multipliers as follows:

A1 - 1.5
K2 - 1.25
C1 - .9
C4 - .8


Now I decide to pilot the C1, and my ELO score is 2000. When I launch into a match, the Matchmaker multiplies my ELO score by the Variant Multiplier: .9, for a total ELO score of 1800. Meanwhile, someone else launches with an A1, and their ELO score is 1200, The matchmaker applies the Variant Multiplier and comes to a combined ELO score of 1800. Since our modified ELO scores are the same, we are put in the same match.

This system takes variant quality into account, and will automatically adjust as more game features are added and code changes are made.

Example: The "I-Win" Missile System is introduced, able to kill any mech on 1 shot. However to "balance" this weapon system, only the Catapult C1 can mount it. Players flock to using the Catapult C1, and the variant's ELO score skyrockets. When the I-Win missile is removed, the Catapult C1's ELO score drops.

This system could be further enhanced by adding an additional Weight Class multiplier (If weight-class balancing is never reintroduced).

With such a system in place, players will have an incentive to use mechs other than the exceptionally powerful builds, and will not be punished for grinding out the basic skills on new mechs.





2 user(s) are reading this topic

0 members, 2 guests, 0 anonymous users