SE4 Rating System

[Grandpa Kim]

Are you actually trying to say that winning a 20 person game is 19 times harder to do then winning a 2 person game?

Yes, Geo, that is exactly what I'm saying. And vice versa. Are you saying that if you have a 50% chance of winning a one on one game, you have a better than 5% chance of winning a 20 player game?

Why should I get the same amount of points in a large game simply for outLasting Primitve, when we never met and I had no involvment in beating him at all,

Simply because you did outLast him! Either diplomatically or militarily or in some other way you were better. I feel you do deserve the full measure of points.

On the other hand why should the guy who finished second, beating out 18! count 'em, 18 other players get so much less than the winner? The power struggle at the end may well have been a close thing. The final result may have been decided by luck.

Unfortunatley we don't know all the factors. Was it a no-contest and the winner really should get all the points while all the others lose points, or was it anybody's game right to the end and even the first man out should have received some points for his outstanding effort? The fact is we just don't know and, after all, the rating system is all about winning and losing not about how hard it was.

Now apology time. I was incorrect in saying there is inflation in Geoschmo's system. I have no excuse for missing this fact and apologize, especially to Geoschmo as well as any others I may have offended.

FIDE: Federation Internationale de Echecs (any of our French speaking contributors feel free to correct my spelling)

Or in english, The International Federation of Chess.

FIDE uses the rating system upon which our own SEIV rating system is based. This system is so good you can accurately compare yourself to someone long dead... as long as you both have (had) a rating. Ratings range from about 2900 for a world champion to about 500 for a tyro who continually loses. The average is somewhere around 1500. A typical Master would be around 2200 points, a Grandmaster 2500.

Some years ago, if memory serves, the whole system (or was it just Canada?) had an adjustment because of deflation! This happens because people stop playing but while playing they got better and gained rating points. When they quit or died, those points were lost to the general pool ergo the deflation.

Now, back to Geo's system. Not inflation, not deflation, but stagflation. Geo's idea tends to keep Ratings bunched close together. (I for one am disappointed how closely bunched the Ratings still are. I fully expected someone to have achieved 1500 points by now.) It also makes playing rated multiplayer less attractive. No matter how well or poorly you do, there won't be much change. Why bother? After playing for a year and winning, you find you gained 18 points! Well whoopdedoo! Think I'll stick with KOTH!

I still can't see any good reason to fiddle with the Ratings in multiplayer.

[narf poit chez BOOM]

Some years ago, if memory serves, the whole system (or was it just Canada?) had an adjustment because of deflation! This happens because people stop playing but while playing they got better and gained rating points. When they quit or died, those points were lost to the general pool ergo the deflation.

i havn't really been keeping up with this debat, which may be why i can't wrap my head around that.

[geoschmo]

Originally posted by Grandpa Kim:

quote:

---

Are you actually trying to say that winning a 20 person game is 19 times harder to do then winning a 2 person game?

Yes, Geo, that is exactly what I'm saying. And vice versa. Are you saying that if you have a 50% chance of winning a one on one game, you have a better than 5% chance of winning a 20 player game?

---

Uh, huh? I don't understand the point you are trying to make here. It's probably me though.

The point I was trying to make though was in a 1v1 game it's easy to figure out the points. He lost, you were 100% responsible for him losing, so you get 100% of the spoils. If you have 20 players and one loses it's not so easy to say who is the most responsible. We could try to come up with some system to cover that, but it would be hopelessly complicated and very subjective. So the fairest thing in my mind is just divide the points up. Over the course of the game, the better player should Last longer and end up with more points anyway.

quote:

---

Why should I get the same amount of points in a large game simply for outLasting Primitve, when we never met and I had no involvment in beating him at all,

Simply because you did outLast him! Either diplomatically or militarily or in some other way you were better. I feel you do deserve the full measure of points.

---

Even if you never met him? Even if you never met anyone? You could be stuck in a corner and not meet anyone before 2 or three players get knocked out of the game. That's fine. It's the luck of the draw and all. But you haven't exactly earned anything. Certainly not three victories.

On the other hand why should the guy who finished second, beating out 18! count 'em, 18 other players get so much less than the winner? The power struggle at the end may well have been a close thing. The final result may have been decided by luck.

I don't think there is a huge difference between 1st and 2nd in what I am suggesting. In a 20 player game first would get close to 3x points. 2nd would get around 2x points. The farter down you go the lower the points get until you end up at -1x for the Last player. It's not a linear progression of course, The largest gap is between 1st and 2nd. But it's not that out of whack I dont't think.

No offense was taken, no appology is neccesary. Were simply having a lively debate on a subject we disagree on.

I am not totally sure the Chess system correlates directly to SEIV. You don't have 20 player games in chess after all. If the Ratings system were limited to 1 on 1 games, and we were talking about a 20 player round robin SE4 tourney then I would agree 100% with using that system. But in my mind there is a significant difference between a 20 player tournament of 1 v 1 games and a single 20 player game of SE4.

I can see your point about the lack of movement in points with my suggestion. And I did consider one factor that makes it not so bad. Since 20 player games take so long to complete, and players get knocked out one at a time over the course of the game, the winner isn't going to get a one-time 300+ point jump in their standings. It will happen gradually over the course of the game 16 or so points at a time. So in this regard it's not so bad.

But frankly human nature is what it is. You can't deny it or try to fight it. No matter how ideal we'd like to think people act, there will be a tendancy for players once they get near the top of the Ratings to "do the math" and have second thoughts about whether joining a 20 player game is worth the risk of getting ganged up on early and taking a 300+ point drop in their Ratings. Do you honestly see no problem at all with this?

I just can't think of an acceptable solution to what I see as two separate problems. It's not like we can give the winner credit for 19 victories and only give the loser credit for one loss, can we?

Wait, can we?

It sounds kind of top heavy, but what if we do that. As each person loses they get a loss equal to the average of the points of the players remaining in the game, and each player gets a win against them. So the ultimate winner gets 19 wins. The 2nd place guy gets 18 wins and one loss. 3rd place gets 16 wins, and still just one loss. 4th place, 15 wins, still one loss. et cetera.

What effect would such a top heavy system have on the Ratings. Would it cause serious point inflation?

Geoschmo

[ March 12, 2004, 13:25: Message edited by: geoschmo ]

[narf poit chez BOOM]

Originally posted by Slynky:

quote:

---

Originally posted by narf poit chez BOOM:
i havn't really been keeping up with this debat, which may be why i can't wrap my head around that.

Try, maybe, wrapping your ears instead [/QB]

---

[ March 12, 2004, 23:04: Message edited by: narf poit chez BOOM ]

[Slynky]

Originally posted by narf poit chez BOOM:

quote:

---

Originally posted by Slynky:

quote:

---

Originally posted by narf poit chez BOOM:
i havn't really been keeping up with this debat, which may be why i can't wrap my head around that.

Try, maybe, wrapping your ears instead

---

[/QB]

---

Well, they looked floppy enough...LOL!

[Slynky]

Originally posted by narf poit chez BOOM:

quote:

---

Some years ago, if memory serves, the whole system (or was it just Canada?) had an adjustment because of deflation! This happens because people stop playing but while playing they got better and gained rating points. When they quit or died, those points were lost to the general pool ergo the deflation.

i havn't really been keeping up with this debat, which may be why i can't wrap my head around that.

---

Try, maybe, wrapping your ears instead

[narf poit chez BOOM]

Well, they looked floppy enough...LOL!

*waggles his ears, twitches his wiskers and scratches an ear with a hind leg*

[ March 13, 2004, 01:34: Message edited by: narf poit chez BOOM ]

[Grandpa Kim]

Originally posted by Grandpa Kim:

quote: Are you actually trying to say that winning a 20 person game is 19 times harder to do then winning a 2 person game?

Yes, Geo, that is exactly what I'm saying. And vice versa. Are you saying that if you have a 50% chance of winning a one on one game, you have a better than 5% chance of winning a 20 player game?


Uh, huh? I don't understand the point you are trying to make here. It's probably me though.

I can be obscure at times. The point I was making is that it is proportionately more difficult to win against 19 opponents than against one and the winner should receive full credit for this. I think he should get credit for 19 victories not a mere 3- victories. The rewards do not match to performance.
There is precedent though for your idea. In business it is common to add a fancy title and a slew of new duties but only a 10% raise in pay (or sometimes none). Not fair? Of course not, but it happens all the time.

The point I was trying to make though was in a 1v1 game it's easy to figure out the points. He lost, you were 100% responsible for him losing, so you get 100% of the spoils. If you have 20 players and one loses it's not so easy to say who is the most responsible. We could try to come up with some system to cover that, but it would be hopelessly complicated and very subjective. So the fairest thing in my mind is just divide the points up. Over the course of the game, the better player should Last longer and end up with more points anyway.

Here I agree with everything but the solution. In my mind since it is impractical to determine who really is most deserving (or least deserving) we shouldn't even try, but rather, just let the chips fall where they may.

Warning: The following is not intended as a slight or insult but only to help me make a point.

The system you describe penalizes the good players while rewarding the poor players. To me this sounds like communism or a welfare state. You've probably guessed, I'm not a big fan of either.

Even if you never met him? Even if you never met anyone? You could be stuck in a corner and not meet anyone before 2 or three players get knocked out of the game. That's fine. It's the luck of the draw and all. But you haven't exactly earned anything. Certainly not three victories.

Luck? Geo, luck has played a part in every single KOTH game I've played. Minor in a couple, moderate in most and the determining factor in at least two. In my very first game against Gozguy, I utterly creamed him without breaking a sweat yet I consider him my equal. I had plenty of systems with planets on my side of the map while Goz's was littered with empty systems. He didn't stand a chance and I walked all over him. ...But I did receive full credit for the victory while Goz took his lumps with a smile. Should a panel judge the game to determine an appropriate point adjustment? I think we will all agree that's not wanted after all this ain't figure skating.

The very nature of SEIV makes this inequity inevitable and has led yourself and others to find ways to eliminate the element of luck-- a project I heartily endorse! But make no mistake, you will never eliminate luck without changing the essential character of the game.


The chess rating system works perfectly for any solo (as opposed to team) competition that is scored by win/draw/loss. It would work great for tennis but not so well for the 100 meter dash (the time clock is best here) and very poorly for hockey (where personnel changes taint the result). It works better for chess then SEIV because of the luck factor. In chess there is no luck. In SEIV luck tends to keep Ratings close by allowing a very poor player to win more games than he should.

But frankly human nature is what it is. You can't deny it or try to fight it. No matter how ideal we'd like to think people act, there will be a tendancy for players once they get near the top of the Ratings to "do the math" and have second thoughts about whether joining a 20 player game is worth the risk of getting ganged up on early and taking a 300+ point drop in their Ratings. Do you honestly see no problem at all with this?

Alas, I see no foolproof solution to this. For practical purposes, it is unlikely to find a 20 player game with 20 rated players. Even 10 would be unlikely. Also, as a lower rated player, do you really want to grab the tiger by the tail? How reliable are your cohorts? And where the heck did the game hide his empire? No, these are not solutions but I for one, would be up for the challenge... if I had that high rating to go along with the bravado.

Finally, your latest suggestion would be very inflationary. New players joining the system would have an impossibly steep hill to climb.

Geo, I went into such detail regarding the FIDE system to emphasize the large point spread. Currently SEIV has less than 200 points between best and worst. What that means is our best player against our worst in a 16 game match would have a 10/6 result. Hardly remarkable.

The system is young, the spread will grow. A personal, 300 point drop does not scare me. Does it really scare any of you?

[Slynky]

I see both points here and I must say, I fall inside them. That is to say, somewhere in the middle of what Grandpa Kim and Geo are discussing. Alas, I'm not sure of a solution.

On one hand, a multi-player game is like a tournement of several games yet it is unlike it. Of course, due to luck, SE4 is unlike a chess game. It becomes a bit more unlike a chess match in a multi-player game. More luck and more events to "upset" a good player. And, in multi-player games, it's the only place to add the dimension of diplomacy (not found in 1 x 1 games) and this is yet another side of a skilled player. (that AND team coordination).

If we made scoring like a series of 1 x 1 games for each to be rated, I fear there would be no one joining in them. After all, even the best diplomat would be at a disadantage if starting in the corner where a lack of empires to "chat" with would be a bit of an unfair start. (but, that is the "luck" discussion we have had several times...yet, who would want to chance the loss of a 100 points?) In contrast, 5 (for example) different games would "balance" out bad starting positions...assuming a person wasn't unlucky for 5 games in a row .

To make it a personal example, I'd much rather play (and get the rating from) 5 1 x 1 games than take a chance on being Last in a game with 5 rating players. While I recognize I might possible end in first place, I wouldn't want to wager a bad starting position on it.

Without looking back at the formula Geo has suggested, I would propose a "final adjustment" to it that goes like this: after all the points have been computed, we add (and subtract) 1 point for each person beaten (and lost to). A variation of this would be, instead, to add (and subtract) the number of points equal to the number of rated players in the game TIMES the number of players beaten (or lost to). So, in a 5-player game, the loser would get an adjustment of -4 or (considering the second suggestion) -20.

Just some food for thought.

[Grandpa Kim]

Quoting myself:

A personal, 300 point drop does not scare me. Does it really scare any of you?