GA Algorithm - Worse Than Before??

"Hortus;c-1898958" wrote:

"leef;c-1898846" wrote:

If, however, MM algorithm is based purely on actual player performance, this allows you don't care about it and develop roster purely around building better squads, counters, etc., as it should be imo.

There's something logically wrong with this paragraph. It's impossible for matchmaking to be purely based on actual player performance, but also allows you to build better squads, counters, etc etc.

It's definitely possible and actually exists in many games. All that you need is some form of permanent PvP rating like Elo or equivalents. Of course you will need some starting point for initial setup, and for this GP is as good as any other sane parameter which allows to approximate power. After initial setup the system will begin to balance itself, and the more games people will play, the less their rating will be affected by anything but actual PvP performance. The only question is "how many games we need before people will get more or less stable rating" but it requires simulation and testing.

Now let's see how different manipulations with roster will affect your matchmaking.

1. Let's assume you are not improving your roster for some time. People who do it will become more powerful but their rating still the same. You begin to lose more often and your rating will fall, as it should be.
2. Let's assume you want to raise some toon to G13 just for collection. If you have limited resources and do it instead of raising some useful toon - see before. People who do something useful for PvP will get advantage, as it should be. But if you just buy this toon and gear - you are not punished but neither get advantage. Again, as it should be.
3. Let's assume you are direct your resources into improving your roster specifically for PvP. Your are pursuing meta, developing counter squads, etc. If you do this better than other players - you will climb higher. If you do it worse - see before, you will fall. Again, as it should be.
4. Let's assume you want to try sandbagging. Well, it just completely impossible. If you will make your roster weaker in any way you still will face same players but with worse roster - it may or may not affect you in bad way but you won't get advantage.

GAC system with climbing into higher leagues may, in theory, produce something similar in long run... IF there were more leagues, movement between leagues in both ways instead climbing only, and no reset of player position with new championship. Sadly currently it's nothing like this.

Okay, it seems like i misunderstood what you meant by "player performance".
What you're proposing is basically just a ladder system in which your roster is the main determining factor. You may be able to compensate a little by being a better player, but not by a lot. Better player as in better at anything non-roster related like strategy, with which team you attack, battle skill etc. Basically the things a lot of players think should be the sole determining factor in winning GA; player skill.
As a side effect it will also result in a winrate of 50% across the board except for at the very top and very bottom.