Loading...
  OR  Zero-K Name:    Password:   

FFA - ELO and and things related.

7 posts, 820 views
Post comment
Filter:    Player:  
sort

2 years ago
I recall some years ago we had a FFA ladder, it was quickly removed and we reverted to casual\MM ladders....

I only bring this up because there are a few arguments for FFA ladder recently, a main one being the disparity of it with teams and MM...

I question MM and would argue if you play, you play... So ideally. it's teams, 1v1 or ffa... does the MM button request become a requisite?

Idk, thought this may worth revisiting.
+0 / -0
quote:
I recall some years ago we had a FFA ladder, it was quickly removed and we reverted to casual\MM ladders....

There was never a proper FFA ladder - but there was a thread in which DeinFreund ran rating calculations for FFA (among other interesting queries like nemesis players, synergies and antisynergies, etc) - all based on a scraped dataset.

It was not "removed", it's just that then Dein stopped doing it. Cuz, you know. Lots of manual work.

quote:
I only bring this up because there are a few arguments for FFA ladder recently, a main one being the disparity of it with teams and MM...

IIRC the previous discussion more or less converged on that making all FFA unranked would be a quick and easy thing to do. But it's touching infra, so that's unfortunately slow.

quote:
Idk, thought this may worth revisiting.

The optimal way to revisit this would likely be with a pull request. :P
+1 / -0

2 years ago
I didn't follow the discussion much, but two things I picked up are the request for FFA ladders and the request for FFA not to have ratings at all. Adding FFA ladders would mean adding an even less actively played ladder than the 1v1 ladder, which would thus be even less accurate and dilute our ladder spots. The other argument was that ratings distort FFA games because players base their strategic decision on the ratings of other players. I didn't quite understand this argument since it applies as long as any rating exists, not just FFA rating (1v1 and team ratings correlate strongly with FFA).
+1 / -0
I would express the second argument more along the lines of "some players may be motivated to throw or collude in FFA games to tank or otherwise manipulate casual rating".
+1 / -0

2 years ago
Which is just as easy in private rated team/1v1 games?
+1 / -0

2 years ago
For a single person trying to tank their rating, or multiple people manipulating their rankings in public rooms, deliberately losing 1v1 or team games is more obvious than deliberately losing FFAs, and the admins would likely do something about it eventually (this has happened once or twice). I think it would be fairly easy to throw in FFA games while doing nothing that was actionable.

A group of people manipulating their ratings in ranked private games is zero-sum. For one person or team to win the other always loses. A group of people colluding in FFA can probably increase the average winrate of everybody in the group at the expense of everybody else in the game.
+0 / -0


2 years ago
quote:
Which is just as easy in private rated team/1v1 games?

The point isn't how difficult it is to manipulate your rating, the point is who has their time wasted by people doing it. The same problem would exist if FFA did not affect ELO but people thought it did - to the point that people would ruin games trying to manipulate it.

It fundamentally does not matter how accurate ratings are. Accurate ratings are not an end goal. The goal is to generate good games. If measuring FFAs for minor boosts in accuracy has a noticeable cost for FFA quality, then the measurement is not worth it. This is true even if "quality" is just the experience of people who lose ELO from FFA, without any attempts at manipulation involved.
+1 / -0