Dump Kleier Rating System

[sadistic]

Lets face it, it doesn't work. A player can win 15 straight and gain only 1 or 2 points! You can beat the highest rated player and not even gain a point. Many players here complain about it, so why not dump it altogether, or in the least, make the WSC the premier ranking page when we check rankings? There is a lot more movement with the WSC system and it is more exciting and accurate. With WSC, I noticed that when I lose a lot, my ranking sinks and conversely, when I win a lot it rises up as it should. This is what players want, not a dead ranking system.

[apollo]

I can't believe I am saying this, but I wholeheartedly agree with Sadistic! Like he said, you could win 20 games in a row against good competition, and gain NOTHING! At least with the WSC ranking, people's ranking actually changes once in a while.



Apollo

[sadistic]

Spion, you have a great site and you deserve a lot of credit. The only complaint that is common about gravon is in regards to the Kleier rating system. I think that you should make the WSC rating system the primary, if not sole, rating sytem. The Kleier system does not work and the WSC promotes movement and more interest in the ranks.

[skywalker]

sadistic may be right with his opinion but what i like about the kleier system you only see the active players on gravon and the players above 20 games
for me its annoying to see guys in the ranking who made 2 games on gravon in 2004
a nice solution would be a connection of both rating systems
_________________
"All you have to decide is what to do with the time that is given to you" Mithrandir

[spion]

Hello,

the Kleier System is not what you both assume it to be.

It is not a system awarding points to players winning games. It is a system to estimate playing strengths.

If you win 20 games in a row against players with a lesser rating than yours, this wont affect your "points" at all. If you win some games against players within +-200 rating points, your rating also wont change much, especially if you already played tons of games.

Why?

Winning 20 games in a row against weaker players does not show any improvements in your playing strength. Of course. According to your actual playing strength (that already is higher than your 20 opponents's strength), you have been expected to win the games. So if you did it, it's no surprise and does only proof your playing strength was estimated correctly.

Winning some games against players within +-200 rating points is nearly the same: You are expected to win some and to loose some because you have nearly the same strength than your opponents.

In order to improve your rating you would have to beat players with more than +200 rating points 20 times in a row

A single win against a high rated player is considered to be luck. Thats why one win does not affect your rating.

And last but not least, your rating wont change much after you played tons of games.
Thats because winning some games in a row is considered to having a lucky day compared to all other games.

If any only if you consequentially win games against players with a higher rating, your rating will slowly grow, because that will show a strength improvement.

It may be frustrating to see your real playing strength but thats it

The WSC system is much more dynamically. It forgets about games in the past very quickly so your rating shows something like an "actual playing strength". If you got a lucky month, your "monthly" playing strength will be higher than in other months.

The Kleier system may have some drawbacks but it remains what it was meant for: it estimates playing strengths.

Best regards
spion

[sadistic]

In other words, Kleier does not work. This is your site, Spion, so you can do as you wish with it. Most players here consider it a "dead" ranking system that is not accurate and up to date.
I have beaten several high ranked players and it did not do much to my rank at all. I did not realize that I had to beat leto 20 times in a row to move up. At times I was one of the hottest players on the site and my rank did not change. The good thing about the WSC rating is that it shows us who is playing well at the present time. There are times when players, including myself, play a bunch of games and don't take them seriously. When I do this, I lose a lot of games. However, when I am up against a better player, I take it more seriously. The games I lose because of recklessness is not a true indicator of my ability. I am sure that this is true of other players as well.
I personally feel that anybody in the top 100 is capable of beating any other player with some regularity. Anybody above this rank knows how to play and the game becomes a matter of luck and good instincts. Where Kleier fails is in motivating players who seek challenge in improving their respective ranking.

[antares]

Gentlemen & -women,

I bet there are 1,000 good reasons, why Kleier's Rating System is - scientifically speaking - a good standard of estimating a players long-period playing-strength.

But it is boring like something else!

Mr. Spion - tear down this wall! Good suggestion of Skywalker...

gg, Ant

PS: Sadi, 2nd useable post ever within a short while, respect! Do I have to start taking you for serious???
_________________
Ease on Perth!

[psychicwarrior]

I would like to call for a new rating system as the current Kleier system is simply inadequate. It doesn’t reflect accurately a players level and is strongly skewed towards those who have played fewer games.

Gravon is a great site, and I believe needs a great rating system to go with it.

Let’s look at other rating systems. Take for example the model used in Chess, which is obviously universal and could only achieve that status by having a valid rating system. In Chess, when a player wins against an equally rated opponent, his rating goes up 15 points. Against a vastly superior opponent, a max of 25 points and against a vastly inferior opponent, a minimum of 10 pts. This accomplishes 2 things, gives the player immediate feedback on his/her rating and accurately reflects their most recent level of play.

For example, if a player played 20 games, and lost the 1st ten and won the 2nd ten, their rating would be much higher than if the won and lost every other game and accurately so as if a player wins ten games in a row.. they are clearly playing in superior form and the rating should reflect that,

In any case a victory or defeat shows an immediate and predictable change in rating.

Take the example in the Kleier system which is based on % of wins ..and thus the more games you play, the lower each game counts in a rating, thus it inaccurately reflects a players most recent play and in effect penalizes players who play many games.

Take the most recent example extrapolated of say 1000 games. If a player, under the Kleier systetm loses all of his 1st 500 games, then proceeds to improve and win the next 500 in a row, the Kleier system shows his % of wins to be 50% and the player would be rated somewhere in the “middle” of player rankings. In chess, however, he would be rated amongst the top as he should be, winning 500 games in a row.

This the Kleier system in effect penalizes player who play often, as their wins mean less… and doesn’t accurately reflect the level of skill of a player but simply a statistical qualitative evaluation that is almost meaningless after a number of games have been played.

It is no wonder that in the Gravon Stratego rankings, most of the top players have played far fewer games.

In my case, specifically, having played over 1600 games, If I win 100 games in a row, against higher ranked players, my rating only improves marginally and I still would be somewhere in the middle of the pack. Contrast that with Chess, and my ranking would be amongst the top players as it should be for winning so often.

Thus I ask those in charge at the great site of Gravon to implement a rating system that reflects and responds to both the levels and improvements of its players.

Thank you.

Edfx

[sadistic]

That was very well written, edlottofx. Keep up the good work.

[esquire]

Is Sadistic a member of edlottofx's fan club now....Rumors are spreading wild.

[spion]

Hi,

edfx: you are wrong, Kleiers rating is NOT based on win%.

Wheter your rating is going to change or not depends on several facts.
If you win some games in a row against a player who's rating is lower than yours, your rating may not increase.
Winning some games in a row against a weak player does not indicate that you are a strong player, does it?

Winning many games in a row against an even player or even a better player will increase your rating.

You are right, after having played 1000 games, your rating will not vary much any longer. Why that?
Kleier tries to eleminate all "daily" interference. If you had a bad day and not enough sleep, you may loose a couple of games. If you have a good day, you may win some.
Perhaps you even got a good week or a bad week. Winning some games in a good week does not show you've become a better player, at least not if the next week is a bad one...

I guess what all of you request is a dynamic rating which reflects your "actual" playing strengths over the last days or weeks..
I'm thinking about a system that does that. But it will take time..

But the Kleier system will not be removed at all because it is the system used for the isf's real life world ranking list.

One thing should be clear:

I'll not add as many rating systems as needed until everyone is no 1 in some system

Best regards
spion

[-deleted594]

Hi all,
Kleier is like the BCS college rankings. Your ranking is based on strength of opponents just like on gravon. If you're number 1 then you've played some tough games. If your in the latter part of the standings then your games have been cakewalks. Just accept it, it is what it is.

[sadistic]

[samuel]

So in fact more people look at the WSC rankings list to get an idea of where they are? I always thought Klier was the premier one but I generally do better at WSC so I'd be happy to follow that one more if others do too! I've noticed my ranking is pretty much stagnant now in the mid-40s, even when I didnt play for a few weeks I went up a couple of places rather than down, and the number hardly changes with results. But on WSC I've been as high as 12th and and down in the low 30s, but I noticed lot of in-active names on that list too so I didn't think it was the better one to look at! Im still pretty new to the rankings game though. My other suggestion is for newbies, to actually explain how the systems work on their pages. I had to email spion to find out and it was still hard to follow for a while.

[acerimmer]

Kleier is a better indication of skill level Samuel.

Many of these guys just moan because they can never get high on kleier.

WSC is less accurate but anyone can be high on wsc if they have some good wins. WSC is more accurate at showing a players recent form rather than their actual skill level.