SE4 Rating System

[geoschmo]

I still prefer a method where you caclulate the points lost by averaging the Ratings and splitting points between the winners. It's not as complicated as it sounds when you get down to it.

Five player game:

A B C D E

E eliminated first

E loses points as if losing to one player of a rating equal to the average of the four players beating him.

A, B, C, D gain that number of points split between the four of them. Two ways to do this. Equal shares, or some sort of ratio based on their relative Ratings. The first is easier, but the second isn't that complicated.

Ultimate winner A gains x+.5x+.333x+.25x=2.08x points.

D loses about .75x points.

C loses about .4x points.

B gains about .1x points

One advantage of this is that nobody gets hammered with a 15X drop in score for getting toasted early in a game. The most you could ever lose would be points equal to losing a 1 v 1 game.

Another advantage is the ultimate winner doesn't get jacked up with a 15X jump in scores. The most the ultimate winner could gain is approximatly 3 times the score gained from winning a 1 x 1 game. The actual amount will approach 3x the larger the game gets, but never reach it.

Players aren't incentivised to hang on in a hopeless cause as Primitive pointed out.

Players will be more relaxed and able to play to win, instead of playing defensively for rank and trying not to lose.

Players with good Ratings are going to be ganged up on, that's a given. But this way they aren't unfairly penalized in their Ratings. Otherwise they will probably tend to avoid large games, causing a lowering of the quality of the competition in those games.

And with this method you can also calculate scores as the players are eliminated. Which I agree is important.

[ March 09, 2004, 15:31: Message edited by: geoschmo ]

[tesco samoa]

i like geo's idea.

[parabolize]

Originally posted by tesco samoa:
i like geo's idea.

I do too

[Slynky]

I think Geo's idea has a lot of merits, too. I'd prefer to do it with the "straight-line" method (the easier method he referred to), after all, not sure there is any really detailed need to be so careful to adjusting points in a ratio'd manner (I mean, why make sure we give the highest rated player a bigger proportion of the points when HE may be the next player out). And, I've always liked the idea of computing it as a player is beaten (assuming I get a report... ).

One of the good points, I think, is one of the points I worry about, though. And that is the skew that prevents the lowest finishing player getting points subtracted any worse than a 1 x 1 game while allowing the winner to gain points equal to 2 "1 x 1" games (or even more). While this is appealing to those who fear a person may hang around hoping that someone else will go out a turn before them, this could also lead to people overwhelming multi-player games. After all, you have a chance to gain an equivalent of winning 2 or 3 "1 x 1" games but NEVER having the fear of a loss being any worse than a single "1 x 1" game. In other words, you risk as many points in a "1 x 1" game as ANY multiplayer game yet the fruits of victory could yeild much more in point gain.

I'm not against it, mind you, I'm just listing my concern. If we get a lot of people who prefer this way to rate multiplayer games, this is what we will use.

[geoschmo]

That's a valid point, but I see it as a positive. More people and more games being rated.

While your downside risk is minimized, your chances of getting the 3x jump isn't all that great. That would require a 20 person game, and you'd have to be first out of 20. Most of the people would fall in the range in the middle, not getting much plus or minus.

[Slynky]

Originally posted by geoschmo:
That's a valid point, but I see it as a positive. More people and more games being rated.

While your downside risk is minimized, your chances of getting the 3x jump isn't all that great. That would require a 20 person game, and you'd have to be first out of 20. Most of the people would fall in the range in the middle, not getting much plus or minus.

Shall we wait for a few more readers and comments or put it into action?

[primitive]

One more vote for Geo

Winning a Huge game is a major feat, and should be richly rewarded.

[Asmala]

Originally posted by primitive:
One more vote for Geo

Winning a Huge game is a major feat, and should be richly rewarded.

But doesn't you get only three times more winning 20 players game than 1 vs 1 game?

[Parasite]

Originally posted by Slynky:

In other words, you risk as many points in a "1 x 1" game as ANY multiplayer game yet the fruits of victory could yield much more in point gain.

Note that (I believe) Multiplayer games take longer to play in real time as do two player games. Each turn is done at the slowest (for that turn) persons time. It also takes greater effort with diplomacy, ect. I would be agreeable to Geo's idea.

[Parasite]

My evil twin brother is asking what if someone had a small low maintance ship hiding in a red nebula. Therefore he would not be "Killed", yet he has clearly lost. Are there rules/gentlemans agreements on this that I have missed?