Game Limit

[franz]

the amount of games shouldn't matter so much, as it's been shown people have qualified with the minimum 80 games when the people around had played 200+. the most important thing to focus on is winning vs quality opponents instead of focusing on other players having played so many games.

[Aerox]

the amount of games shouldn't matter so much, as it's been shown people have qualified with the minimum 80 games when the people around had played 200+. the most important thing to focus on is winning vs quality opponents instead of focusing on other players having played so many games.

It's been shown shouldn't be a valid reason for anything.


I'd like to hear how you think activity is not a clear handicap with this current system.

[franz]

there are ways to tinker the formula, if MI wants to change the amount of progression that ratings exchange between wins, he can do that (I remember something like a 'K Factor' in the formula that does this). If I remember right, this would make increasing one's seasonal rating move closer to their overall rating (given they win the appropriate amount of games vs quality opponents) --> and in turn, players who simply play a ton of games would just plateau at their lower overall rating (unless that actually manage to improve their skill that season, which is good, and they would deserve their playoff spot should they get one).

I haven't heard from MI yet if he's thought of this or ever experimented with it, but this is just another idea that might appeal to those intimidated by all the high ranked players with 200+ games. still, I argue that getting a high ranking is just as possible right now with the current system because it definitely still rewards players for beating quality opponents. just do that and play 80 games (the current minimum) and you should be up there.

[DarkOne]

No, K-factor won't have any effect on the mentioned problem - it'll just raise the rating, both overall and seasonal proportionally. Well, it might help somewhat for one season, but the seasonal rating will just catch up and you'd have to change your K-factor every couple of seasons and will give a massive rating inflation.

Rating calculations usually involve a logarithm. It determines your chance of winning (or rather, how many points you'd statistically win in a game against player B).
For example, in chess, your projected winning percentage against someone with 400 points lower than you is 1.0 (ie, you win all the games)
The best way to suppress noob bashing and activity playing too big a role (while maintaining the current system) is just to make the rating range (of 400 in the example) smaller: that way, you won't score any points against someone with 300 points less, rather than against someone with 400 points less.
You'll also gain your seasonal points faster, because the upper limit is also smaller (you win the same amount of points against someone rated 300 points as you would in the old system against somebody with 400 points higher).

But there's a trade-off if this happens: ratings will fluctuate more than they used to and ratings will become less reliable than they are now.

Setting a roof on your total amount of games also has a trade-off, though: avoiding.

[Aerox]

. just do that and play 80 games (the current minimum) and you should be up there.

Yes I guess I would. But it's inside the realm of possibility that I don't, due to the nature of the system.

Right?

[avirex]

good point ropa

[franz]

No, K-factor won't have any effect on the mentioned problem - it'll just raise the rating, both overall and seasonal proportionally. Well, it might help somewhat for one season, but the seasonal rating will just catch up and you'd have to change your K-factor every couple of seasons and will give a massive rating inflation.

Rating calculations usually involve a logarithm. It determines your chance of winning (or rather, how many points you'd statistically win in a game against player B).
For example, in chess, your projected winning percentage against someone with 400 points lower than you is 1.0 (ie, you win all the games)
The best way to suppress noob bashing and activity playing too big a role (while maintaining the current system) is just to make the rating range (of 400 in the example) smaller: that way, you won't score any points against someone with 300 points less, rather than against someone with 400 points less.
You'll also gain your seasonal points faster, because the upper limit is also smaller (you win the same amount of points against someone rated 300 points as you would in the old system against somebody with 400 points higher).

But there's a trade-off if this happens: ratings will fluctuate more than they used to and ratings will become less reliable than they are now.

Setting a roof on your total amount of games also has a trade-off, though: avoiding.

we're probably thinking of two different things, I'd honestly have to look at the formula again to refresh my memory, but that's up to MI if he wants to share.  I'm just going off of my memory of this from a few years back when I helped Veg with XTC's rating system, and I remember something very specific to what I described. still, all this may not even be necessary, but if so, at least our ideas are out there.

[Phanton]

No, K-factor won't have any effect on the mentioned problem - it'll just raise the rating, both overall and seasonal proportionally. Well, it might help somewhat for one season, but the seasonal rating will just catch up and you'd have to change your K-factor every couple of seasons and will give a massive rating inflation.

Rating calculations usually involve a logarithm. It determines your chance of winning (or rather, how many points you'd statistically win in a game against player B).
For example, in chess, your projected winning percentage against someone with 400 points lower than you is 1.0 (ie, you win all the games)
The best way to suppress noob bashing and activity playing too big a role (while maintaining the current system) is just to make the rating range (of 400 in the example) smaller: that way, you won't score any points against someone with 300 points less, rather than against someone with 400 points less.
You'll also gain your seasonal points faster, because the upper limit is also smaller (you win the same amount of points against someone rated 300 points as you would in the old system against somebody with 400 points higher).

But there's a trade-off if this happens: ratings will fluctuate more than they used to and ratings will become less reliable than they are now.

Setting a roof on your total amount of games also has a trade-off, though: avoiding.

we're probably thinking of two different things, I'd honestly have to look at the formula again to refresh my memory, but that's up to MI if he wants to share.  I'm just going off of my memory of this from a few years back when I helped Veg with XTC's rating system, and I remember something very specific to what I described. still, all this may not even be necessary, but if so, at least our ideas are out there.

sure